Gifts of Political Science

Introduction to Voting Behavior

In this episode of UCL Uncovering Politics, the discussion centers around the fundamental aspects of how voters make their decisions regarding which political party to support. The podcast seeks to explore the extent to which political parties' policy offers influence voter choices and the nature of voters' own policy preferences. Alan Renick introduces the topic and the guest, Ben Lauderdale, a professor of political science at UCL, who is one of the authors of a new book that presents a fresh perspective on this long-standing debate.

The Role of Elections in Democracies

Renick emphasizes the theoretical importance of elections in modern democracies, which are designed to enable voters to influence government policies. However, he raises the question of whether elections effectively fulfill this purpose in practice. Two contrasting views emerge in political science regarding voter behavior: one suggests that voters have clear preferences and understand the ideological spectrum of political parties, while the other argues that voters often lack meaningful preferences and that their choices are influenced more by party identification than by policy positions.

Three Broad Explanations for Voting Behavior

Lauderdale categorizes existing theories of voting behavior into three broad types:

• Identity-Based Voting: This perspective posits that voters make choices based on their identities, such as class, race, ethnicity, or religion. Voters align themselves with parties that they believe represent their interests or identities.
• Performance-Based Voting: This view suggests that voters evaluate candidates or parties based on their past performance. Voters consider how well a party has governed in the past, particularly regarding economic performance and competence.
• Policy-Based Voting: This approach focuses on voters' preferences regarding public policy and what parties promise to do in the future. Lauderdale notes that while the book emphasizes this perspective, it does not dismiss the significance of the other two types.

Debate on the Importance of Policy Preferences

Renick and Lauderdale discuss the ongoing debate about the role of policy preferences in voting behavior. The traditional view holds that voters have ideological preferences that guide their choices. In contrast, the "innocent voter" perspective suggests that many voters lack strong policy views and instead rely on identity or performance considerations. This perspective implies that voters may express policy opinions that are unstable and not the basis for their party choice.

The Idiosyncratic Voter Perspective

Lauderdale introduces the concept of the "idiosyncratic voter perspective," which posits that voters hold real views on specific issues that may not be ideologically organized. These views can be stable over time, but they do not necessarily align with broader ideological dimensions. This perspective aims to bridge the gap between the ideological and innocent voter views by acknowledging that voters can have meaningful preferences that are not always coherent across different issues.

Implications for Political Strategy

The discussion highlights the implications of these perspectives for political strategy. If voters primarily form their views ideologically, political parties may benefit from adopting moderate positions to appeal to a broader electorate. Conversely, if voters' preferences are idiosyncratic, parties should focus on specific issues that resonate with voters at the time of the election, rather than attempting to position themselves along a generalized ideological spectrum.

Importance of Understanding Voter Preferences

Lauderdale emphasizes the necessity of understanding both idiosyncratic and ideological preferences to gauge the health of democratic processes. Ignoring idiosyncratic preferences could lead to an underestimation of the connection between voter preferences and electoral outcomes. The authors argue that a clearer understanding of voters' policy preferences is essential for evaluating whether elections reflect the public's will.

Research Methodology

The core of the book's analysis is based on a large panel survey experiment conducted in the UK from early 2018 to early 2019. The survey involved three waves of interviews with the same participants, allowing researchers to track changes in preferences over time. The design aimed to mitigate attrition by intentionally planning for a certain level of participant loss and targeting specific demographics likely to drop out.

Survey Design and Findings

In each wave, participants were asked about their preferences on various policy issues and were presented with hypothetical candidates taking positions on these issues. The survey included 34 different policy areas, with respondents randomly assigned to answer questions about seven issues. This design allowed researchers to assess the stability of preferences over time and the impact of these preferences on voting behavior.

Stability of Preferences

One of the key findings from the research is that while voters may express opinions on a range of issues, the stability of those opinions varies significantly. Some issues, such as foreign aid, demonstrated strong ideological structuring, with stable preferences closely related to other political views. In contrast, issues like cannabis regulation exhibited idiosyncratic structuring, where voters held stable views that did not correlate with their opinions on other issues.

Conclusion on Voter Preferences

The overall conclusion drawn from the research is that voters possess a mix of ideological, idiosyncratic, and unstable preferences. The analysis indicates that instability is more prevalent than idiosyncrasy, which in turn is more common than ideological preferences. This nuanced understanding challenges the binary debate of ideology versus nothing, suggesting that all three types of preferences play a role in shaping public opinion and electoral outcomes.

Implications for Political Science and Electoral Strategies

Lauderdale calls for political scientists to consider how these different perspectives on voting behavior coexist and influence electoral strategies. Recognizing the complexity of voter preferences can inform campaign strategies and party positioning, allowing for a more effective engagement with the electorate. For practitioners, understanding the nuances of voter preferences is crucial for identifying potential constituencies and addressing their concerns effectively.

Generalizability of Findings

Finally, the discussion touches on the generalizability of the findings beyond the UK context. While the study was conducted during a unique political moment characterized by Brexit, Lauderdale argues that the dynamics observed may be applicable to other democracies. The emergence of new issues can lead to shifts in voting coalitions and preferences, reflecting the idiosyncratic nature of public opinion. The authors suggest that similar patterns may be observed in other long-standing democracies, albeit with variations based on local political contexts and party systems.

Conclusion

In conclusion, the podcast episode provides a comprehensive exploration of the complexities of voter preferences and their implications for political behavior. By introducing the idiosyncratic voter perspective, the discussion enriches the existing literature on voting behavior and offers valuable insights for both political scientists and practitioners. The findings underscore the importance of recognizing the multifaceted nature of voter preferences in understanding electoral dynamics and shaping effective political strategies.

In the discussion of voter behavior, the speaker identifies several key drivers that influence how individuals make their electoral choices. These include identity, past performance, and public policy preferences. Each of these factors plays a crucial role in shaping how voters align with political parties and candidates.

Firstly, the identity-based explanations suggest that voters often make decisions based on who they perceive themselves to be, which can include aspects such as class, race, ethnicity, and religion. This perspective posits that voters might feel a stronger connection to parties that resonate with their personal identities or social groups.

Secondly, the past performance of candidates or parties significantly influences voter decisions. Voters assess how effectively a party has governed in the past, considering metrics such as economic performance and overall governance quality. This retrospective evaluation can shape expectations about future performance, thus impacting voting behavior.

Lastly, the speaker emphasizes the importance of public policy preferences, where voters evaluate the promises made by parties regarding their future policy directions. There exists a debate about whether voters possess coherent policy preferences or if these preferences are formed after aligning with a political party.

Interestingly, the speaker notes that some voters might deviate from purely rational choices due to factors such as cognitive biases and emotional reactions. This can lead to decisions that are not solely based on logical assessments of policy but also on personal feelings and social affiliations.

• [02:58] "One of the authors of that book is our very own Ben Lordell, friend of the podcast and professor of political science here in the UCL department of political science."
• [01:01] "Voters understand differences between say left and right. They know where parties stand on the left-right spectrum."
• [05:00] "...all of which are clearly very plausible on their face bases for making a vote choice."

The speaker discusses the impact of voter information and the formats of election debates on the quality of voters' choices. A critical aspect highlighted is the accessibility and clarity of information available to voters, which can significantly influence their decision-making process. Voters who have a clear understanding of party platforms and candidates' positions are more likely to make informed choices.

Moreover, the speaker points out that the structure of debates and how candidates present their policies can either illuminate or obscure the differences between parties. For instance, if debates focus excessively on personality traits rather than substantive policy discussions, voters may become confused or misled about what each party stands for. This can diminish the overall quality of electoral choices as voters might base their decisions on superficial impressions rather than informed evaluations of policy positions.

The speaker also touches on the idea that misinformation plays a role in distorting voter perceptions. When voters are bombarded with conflicting information or sensationalized narratives, their ability to discern truth from falsehood is compromised, leading to choices that may not reflect their actual preferences or interests.

In summary, the quality of the electoral choices made by voters is deeply intertwined with the nature of the information they receive and the frameworks through which they engage with that information, emphasizing the need for clearer communication from political actors.

• [01:51] "...it suggests that the existing debate may have set the bar too high in terms of whether we say voters have meaningful preferences."
• [04:56] "...what parties say they will do in the future."
• [04:18] "This is a view where things like economic performance and other measures of government competence performance are relevant and it's really retrospective in orientation."

In the transcript, the speaker addresses the influence of moral convictions and economic circumstances on ideological disagreements and the emergence of far-right leanings among voters. The speaker points out that moral convictions often play a pivotal role in shaping voters' perspectives on various issues, which can lead to strong ideological divisions. For instance, issues such as immigration or social justice are often framed through a moral lens, influencing how individuals align politically.

Furthermore, the speaker highlights the impact of economic conditions on voter behavior, particularly in times of economic uncertainty or distress. Economic challenges can lead individuals to seek out far-right parties that promise simplistic solutions or scapegoating of marginalized groups. This can create a fertile ground for extremist ideologies to thrive, as voters may prioritize immediate, visceral responses to complex economic problems over nuanced policy discussions.

Additionally, the speaker suggests that ideological disagreements are often rooted in these moral and economic contexts, making it difficult to bridge divides when voters are anchored in their beliefs. As different groups may interpret moral and economic issues in vastly differing ways, this leads to polarization within the electorate, complicating the democratic process and consensus-building.

In conclusion, the interplay between moral convictions and economic circumstances is significant in understanding ideological disagreements, particularly regarding the rise of far-right sentiments, as these factors shape the political landscape and influence voter alignment.

• [06:15] "...that the existing debate may have set the bar too high in terms of whether we say voters have meaningful preferences."
• [09:36] "...people have limited time to devote to politics."
• [09:37] "...that voters are innocent of policy considerations when they make their choices."

Throughout the conversation, the speaker implicitly emphasizes the importance of being critical consumers of information in the face of misinformation and political promises. One of the recommendations is for citizens to engage more deeply with policy discussions rather than relying solely on party identities or emotional appeals. This involves taking the time to understand the complexities of policies and candidates' positions rather than making snap judgments based on surface-level perceptions.

Moreover, the speaker suggests that individuals should actively seek out diverse sources of information to gain a more comprehensive understanding of political issues. This could involve participating in community discussions, attending town hall meetings, or consuming a wide range of media to avoid echo chambers that reinforce existing biases.

Additionally, the speaker indicates that citizens should be aware of the nuances in political promises. Understanding that policies often have both pros and cons can empower voters to ask critical questions and demand accountability from their elected officials. This can lead to a more informed electorate that holds politicians accountable for their commitments.

Ultimately, the speaker's recommendations suggest a need for civic engagement that goes beyond mere participation in elections, advocating for informed decision-making based on a thorough understanding of policy implications and a commitment to questioning and evaluating political discourse critically.

• [11:12] "...we think they make their choices based on policy offers."
• [39:02] "...we see that voters are making choices based on the issues where they do have stable views."
• [50:49] "...you know, we think they make their choices based on policy offers."

The speaker explains how voters often judge governments retrospectively, particularly focusing on economic outcomes. This retrospective evaluation means that voters look back at how well a government has performed during its term, assessing factors such as economic growth, employment rates, and government spending. The implications of this approach suggest that voters hold governments accountable for their actions based on observable results, which can influence their voting behavior in subsequent elections.

Furthermore, the speaker highlights that this retrospective judgment does not solely rest on economic indicators but also encompasses societal well-being. Voters may consider how government policies have impacted social issues like education, healthcare, and public safety. This broad evaluation allows voters to form a comprehensive view of a government's effectiveness.

The speaker also points out that such retrospective judgments can lead to a form of fair accountability, where governments are rewarded or punished based on their performance. However, this raises questions about the fairness of this accountability. If voters are not fully informed about the complexities of governance or the reasons behind certain economic outcomes, their judgments may be based on incomplete information, leading to potentially skewed electoral results.

In conclusion, the way voters assess government performance retrospectively underscores the importance of clear communication from political leaders about policy outcomes and the complexities involved, ensuring that the electorate has the necessary information to make informed judgments at the ballot box.

• [06:56] "...you might be thinking about the kind of potential competence of rival governments as well."
• [09:03] "...we think they make their choices based on policy offers."
• [17:06] "...we need to understand the role of idiosyncratic as well as ideological preferences."

[00:04] Hello, this is UCL Uncovering Politics
[00:06] and this week we're looking at the
[00:08] fundamentals of how people decide which
[00:10] way to vote. To what extent do the
[00:13] policy offers of different political
[00:14] parties shape those choices and just
[00:17] what is the nature of voters own policy
[00:20] preferences?
[00:24] Hello, my name is Alan Renick and
[00:26] welcome to UCL Uncovering Politics, the
[00:29] podcast of the School of Public Policy
[00:31] and Department of Political Science at
[00:33] University College London. Modern
[00:35] democracies have elections at their
[00:38] core. Elections are the main mechanism
[00:40] through which we the voters can shape
[00:42] what governments do. At least in theory,
[00:45] they help to ensure that the policies
[00:48] governments pursue are in line with what
[00:50] people want. But do elections really
[00:53] achieve that effect in practice? One
[00:56] view in political science is that yes,
[00:58] they probably can. Voters understand
[01:01] differences between say left and right.
[01:03] They know where parties stand on the
[01:05] leftight spectrum. They broadly know
[01:07] where they stand themselves and they can
[01:10] choose which party to support on that
[01:12] basis. But another long-standing strand
[01:15] of research argues that no, this
[01:18] mechanism doesn't really work because
[01:20] for the most part, voters don't actually
[01:23] have clear meaningful preferences for
[01:25] how governments should act. And to the
[01:28] extent that they do have such
[01:29] preferences, these come after voters
[01:31] have chosen a party, not before. Voters
[01:34] have a sense of which party they feel
[01:36] closest to, and they tend to follow that
[01:38] party's policy positions. Well, it's a
[01:41] debate that has been raging for decades,
[01:44] but now a new book has been published
[01:46] that offers a third perspective that
[01:48] might just help to resolve things. It
[01:51] suggests that the existing debate may
[01:53] have set the bar too high in terms of
[01:57] whether we say voters have meaningful
[01:59] preferences. And once we measure things
[02:02] correctly, we start to get a much
[02:04] clearer picture.
[02:07] One of the authors of that book is our
[02:09] very own Ben Lordell, friend of the
[02:11] podcast and professor of political
[02:13] science here in the UCL department of
[02:15] political science. And I'm delighted
[02:17] that Ben joins me now to take a look
[02:19] through the research and the findings.
[02:22] Ben, welcome back to UCL Uncovering
[02:24] Politics. It's wonderful to have you on
[02:26] again. And let's start with what we
[02:30] already know or think we know about what
[02:32] shapes voters decisions about whom to
[02:35] vote for. I've started to offer some
[02:37] hints in the intro there, but you can do
[02:39] this much better than I can. So, what
[02:42] what different theories are out there
[02:44] about how voters choose?
[02:47] So there are a number of ways that you
[02:49] could describe the existing political
[02:51] science literature, but uh one that I
[02:53] find useful is to think about
[02:55] explanations for voting behavior in as
[02:59] coming in three broad types. So there
[03:02] are a range of theoretical arguments
[03:05] that relate to people making choices on
[03:08] the basis of identity. So who they are,
[03:11] who they understand themselves to be,
[03:13] who candidates are, who parties are uh
[03:17] composed of, things about, you know, who
[03:20] the people are, maybe as a proxy for
[03:23] their interests, maybe as a more sort of
[03:26] identity- based, you know, feeling of
[03:28] proximity. So there are a whole range of
[03:31] ideas related to this, lots and lots of
[03:33] research about people voting on this
[03:36] basis. So this might be voting on the
[03:38] basis of class identity say that kind of
[03:40] thing.
[03:40] >> Yeah. Could be class, could be race,
[03:42] ethnicity, it could be religion, could
[03:45] be lots of other dimensions might play
[03:47] into people's sense of themselves and
[03:49] who the parties are for. So that's one
[03:52] class of explanation. Another class of
[03:55] explanation is voting on the basis of
[03:57] past performance. So this is the idea
[04:00] that when you look at a candidate um or
[04:02] a party, what you're really evaluating
[04:05] is how have they done in the past. So
[04:07] for a candidate that might be your
[04:09] perception of you know their personal
[04:11] history, their competence for the
[04:12] position. For a party it might be how
[04:14] have they performed in government when
[04:16] they've been in government recently. Um
[04:18] and this is a view where things like
[04:21] economic performance and other measures
[04:24] of of government competence performance
[04:26] are relevant and it's it's really
[04:29] retrospective in orientation
[04:31] >> presumably you could also be thinking
[04:33] about the kind of potential competence
[04:34] of rival rival governments as well.
[04:36] >> Yeah. So there are different different
[04:38] versions of this kind of explanation
[04:40] some of which have more of a perspective
[04:42] element and some of which are are more
[04:44] retrospective in orientation. Lots of
[04:47] ideas here, lots of research over many
[04:49] decades. So that's sort of the second
[04:51] second grouping of of explanations for
[04:54] voter behavior. The one that our book
[04:56] focuses on is the third, which is really
[04:58] what you which is considerations of
[05:00] public policy, which you can think of as
[05:03] you know what parties say they will do
[05:05] in the future. So we've had who people
[05:08] are, who parties are, what people have
[05:10] done in the past, and this is about what
[05:13] they say they'll do in the future. um
[05:15] all of which are clearly very plausible
[05:17] on their face bases for making a vote
[05:20] choice. And so the fact that our book is
[05:22] focused on future plans on people's
[05:24] views about public policy is not because
[05:27] we think those other two things don't
[05:28] matter. It's simply because that's where
[05:30] our focus is.
[05:31] >> And I was suggesting in the intro there
[05:33] that there are basically two views on on
[05:36] the question of whether policy
[05:39] this kind of perspective voting on what
[05:41] governments say they will do, whether
[05:43] that matters. Was that an accurate
[05:44] picture I gave for you? Should we say
[05:46] more?
[05:46] >> It's a reasonably accurate picture,
[05:48] although there are certainly exceptions
[05:49] and work that doesn't fit tidily into
[05:52] into that sort of typology. But so it we
[05:55] do argue in the book that a lot of the
[05:57] literature that exists on how citizens
[06:00] engage with public policy
[06:02] >> takes a view that that either the
[06:05] important bit of it or sort of an enough
[06:07] of it is ideological in nature that it
[06:10] is fair to describe people as having
[06:13] generally left-wing views or generally
[06:14] right-wing views or generally liberal
[06:16] views or generally conservative views.
[06:18] different typo, you know, different
[06:20] dimensions are and different terminology
[06:22] um is used in different countries and
[06:25] contexts, but that those are useful
[06:28] descriptions and we can think about
[06:30] politics as occurring in those
[06:32] ideological terms. So, for example, that
[06:34] it is a reasonable way to talk about an
[06:37] election as saying uh the Labor Party
[06:40] did better in 1997 than previously
[06:42] because it moderated. And what does
[06:44] moderate mean? It means they came closer
[06:47] to the center of opinion on an
[06:49] ideological dimension. Um people have
[06:51] said similar things about the
[06:53] conservatives in 2010 and about uh labor
[06:56] again in 2024. So that's a sort of
[06:59] argument about from ideology that in
[07:03] some sense voters are via various
[07:06] mechanisms sort of getting a sense of
[07:08] where the party is in general relative
[07:11] to where they are and making choices
[07:14] accordingly. And from that perspective,
[07:16] we would expect to see that voters who
[07:19] held relatively left-wing positions on
[07:22] one issue would also tend to hold
[07:24] relatively left-wing positions on other
[07:26] issues, particularly sort of related
[07:28] issues. So that's the ideological
[07:30] perspective. Um the other perspective
[07:34] which you alluded to, we call the sort
[07:37] of innocent voter perspective. The
[07:39] perspective that voters are innocent of
[07:41] policy considerations when they make
[07:43] their choices. And this tends to align
[07:45] with those other two broader senses of
[07:48] where voter behavior comes from, that
[07:50] it's actually driven by identity and
[07:53] performance considerations rather than
[07:56] public policy considerations. And so if
[07:58] you take the view that people don't
[08:01] really have very strong views about most
[08:04] policy questions, perhaps understandably
[08:06] they haven't spent the time to determine
[08:08] which policies are sensible ones to
[08:11] pursue. Public policy is complicated.
[08:13] governance is complicated. People have
[08:14] limited time to to devote to politics.
[08:17] That the So from this perspective,
[08:20] if we see people express policy views,
[08:24] we might expect that there's not a lot
[08:26] to them. They might be unstable over
[08:28] time. They might not be so much the
[08:32] basis by which people make their
[08:33] political choices between parties, but
[08:35] actually the result of the choices that
[08:37] people have made. they've decided that
[08:39] they're a conservative voter or that
[08:42] they um support the Labor Party and they
[08:44] have learned from the party what
[08:46] positions they they should take on
[08:48] issues. And so from that perspective, we
[08:50] should see some combination of
[08:52] instability in people's views because
[08:55] they don't necessarily have a
[08:56] well-worked out view. you ask them one
[08:58] day what their view is, they might say
[09:00] one thing. You come back to them some
[09:01] months later, they might say something
[09:03] else um because they haven't really
[09:05] thought it through and they're coming up
[09:06] with an answer on the spot. Or if they
[09:08] do have a stable position, it's actually
[09:10] not because that was the basis for their
[09:12] party choice, but because they learned
[09:14] from their party what position to take.
[09:17] >> Yeah. So we have these two views of what
[09:21] role issue preferences might be playing
[09:24] here. one role simply saying well they
[09:25] don't play very much role at all because
[09:27] people don't really have views and then
[09:29] another view saying that yeah people do
[09:31] have policy views and these are
[09:32] important and furthermore those policy
[09:34] views as you've said are ideological so
[09:36] they form kind of coherent sets across
[09:38] different issues and then comes your
[09:41] book and you argue for a third account
[09:45] or you're exploring a third account here
[09:47] tell us
[09:48] >> that's right and it's important to note
[09:50] that our view is not we are not the
[09:52] first to express elements of this
[09:54] perspective, but what we call the
[09:57] idiosyncratic voter perspective. And to
[09:59] be clear, we mean idiosyncratic in an
[10:01] entirely non-projorative sense, but in
[10:03] the sense of the that people have
[10:06] particular views on particular issues.
[10:09] They are real views that those people
[10:11] hold, but they're not necessarily
[10:13] ideologically organized, but at the same
[10:15] time, they are real views. They are
[10:17] stable over time. So that is that
[10:19] perspective that this is an important
[10:21] part of understanding voters policy
[10:23] views and ultimately as we'll talk about
[10:26] later how they translate those into
[10:28] votes is the perspective that we're
[10:30] trying to um resuscitate within the the
[10:33] political science literature. There are
[10:36] there are various authors over over the
[10:38] years who have articulated
[10:40] aspects of this but it is definitely not
[10:42] the dominant perspective um or one of
[10:45] the two dominant perspectives within the
[10:47] field. So there's a difference between
[10:50] people having ideological views where
[10:52] they have views and those views are
[10:54] structured in a way that we can make
[10:57] sense of kind of in terms of standard
[11:00] left right and other dimensions of
[11:02] politics and then people can that's the
[11:05] ideological view and then the
[11:06] idiosyncratic view people have
[11:09] preferences about policy issues but
[11:11] across different policy issues it's
[11:12] harder for us to work out what the
[11:14] structure is and I guess I mean it could
[11:16] be that there is a structure that people
[11:18] do have a coherent set of views in their
[11:20] own mind and they've they've kind of
[11:22] worked things out, but it's just quite
[11:25] an unusual set of connections that
[11:27] they've made or an unusual combination
[11:29] of considerations. Or it could be that
[11:31] they just have a kind of fairly random
[11:33] set of ideas on different issues and
[11:35] they haven't really thought about the
[11:36] fact that maybe they're a bit
[11:37] inconsistent.
[11:38] >> Yeah. I mean, one reason why you might
[11:40] expect people to not all hold the same
[11:43] combinations of views, to have these
[11:45] sort of more idiosyncratic patterns of
[11:47] of views, is all the different ways that
[11:50] people might come to hold their
[11:52] particular set of views. They might have
[11:55] sort of worked out how they think
[11:57] different issues go together. Um, in
[12:00] which case they might have done so
[12:01] differently, in which case they wouldn't
[12:02] necessarily have the same combinations
[12:04] of views. But they also might simply
[12:07] have learned from other people and from
[12:09] different other people their positions
[12:12] and you know learning can take lots of
[12:15] forms. Um this can be sort of hearing
[12:18] other people's arguments hearing
[12:20] different evidence with rig that relates
[12:23] in their mind to to different issues.
[12:25] It's also the case that people have
[12:27] different interests, different lived
[12:28] experiences. These sometimes overlap,
[12:31] but it's not necessarily the case that
[12:33] they all sort of align such that they're
[12:35] the people who have the left-wing set of
[12:38] experiences and sources of information
[12:40] and, you know, socialization and then
[12:42] there are the people who have the
[12:43] right-wing versions of those, it's much
[12:46] more complicated than that. And in a
[12:47] world where it's much more complicated
[12:48] than that, you might reasonably expect
[12:50] that people will have their own peculiar
[12:53] combinations of views. And there might
[12:55] be some combinations that are more
[12:56] prevalent than others, but that's not
[12:58] going to be exhaustive of all the views
[13:01] that people have.
[13:02] >> Yeah. And when you're thinking about
[13:04] ideology, I mean, I'm jumping a little
[13:06] bit ahead in our conversation to how you
[13:08] measure things. Um, but you um you you
[13:12] don't have kind of a preset notion of
[13:15] what ideologies might exist. You you say
[13:18] that there is an ideological
[13:21] set of views. essentially of that is a
[13:24] set of views that roughly speaking is
[13:25] shared across lots and lots of people
[13:27] whereas an idiosyncratic idiosyncratic
[13:30] set of views is one that is quite
[13:32] unusual among all the many people that
[13:34] we can look at.
[13:35] >> That's exactly right. I mean there are
[13:37] there are more restrictive
[13:38] understandings of what ideology means
[13:41] than the one that we use which is a very
[13:43] minimalist understanding and one that is
[13:46] grounded not in an individual having
[13:49] arguments about what goes with what but
[13:51] rather about the patterns we see in the
[13:54] population of what goes with what in the
[13:57] data we've looked at and in almost every
[13:59] case looking across countries tend to
[14:01] resemble the combinations of views that
[14:04] are expressed by political parties.
[14:06] Before we get further into the detail of
[14:08] what you do in your analysis, um I guess
[14:12] I should ask why should we care about
[14:15] this difference between ideological
[14:17] preferences and idiosyncratic
[14:19] preferences. I guess some of our
[14:21] listeners might maybe be thinking okay
[14:23] that sounds maybe a little bit obstruuse
[14:24] that that that distinction. Why is this
[14:26] a distinction that matters?
[14:29] >> This is a very good question and the
[14:30] kind of question on which uh one is at
[14:33] risk of foundering. The answer I would
[14:35] give and the answer I think we give in
[14:37] the book is that one really important
[14:40] consequence of whether people are
[14:44] primarily forming their views
[14:46] ideologically or idiosyncratically is
[14:49] the downstream implications for how
[14:52] electoral competition goes. And in
[14:55] particular and these are very live
[14:57] debates in both the US and the UK right
[14:59] now. This relates to questions of how do
[15:03] you successfully position a party? How
[15:05] do you successfully run a campaign? In
[15:08] an ideological world, the answers tend
[15:10] to be some form of either moderation,
[15:14] looking for the center ground, trying to
[15:16] get as close to as many people as you
[15:17] can in some sort of simplified sense of
[15:21] left and right or liberal and
[15:22] conservative, or at least trying to make
[15:25] sure that you're not in a you're not in
[15:27] a crowded part of the space with too
[15:29] many other parties that are more or less
[15:31] interchangeable with you on that basis.
[15:33] If you however take the idiosyncratic
[15:37] perspective, you really want to a
[15:40] successful party is going to be one that
[15:43] takes popular positions on issues that
[15:45] people care about um specifically at the
[15:48] moment of of that election and that
[15:50] those could change over time. And it's
[15:52] and while that's related to this to
[15:55] being in the center ground in this sort
[15:57] of ideological sense, it can have
[15:59] somewhat different implications
[16:01] particularly to the extent that people's
[16:04] views aren't related to one another. And
[16:06] so you really need to be thinking about
[16:08] positioning on every issue on its own
[16:12] and not sort of thinking about sort of
[16:15] some synthesis of all of these into a
[16:17] sort of general centrist or moderate
[16:20] positioning. And so we think that some
[16:23] of the patterns that we see in political
[16:25] competition, candidates and parties
[16:28] being successful because they change
[16:30] their positions not in general across
[16:32] all issues but on very particular issues
[16:34] that have become salient are better
[16:37] explained by this perspective of really
[16:39] thinking about okay people have
[16:40] positions on individual issues. those
[16:42] aren't always very correlated to their
[16:44] positions on other issues. And so
[16:46] successful parties are ones that are
[16:48] able to find their way to positions that
[16:51] are relatively popular and are in
[16:53] particular popular on the specific
[16:55] issues that people care more about.
[16:57] >> So party strategists should certainly be
[16:59] listening to this conversation. And um
[17:02] and would I be right in thinking there's
[17:03] also an idea in the book that in order
[17:06] to kind of gauge the health of the
[17:09] democratic process, we need to uh
[17:12] understand the role of of idiosyncratic
[17:14] as well as ideological uh preferences.
[17:17] And that there's a danger that if we
[17:19] just think of preferences as ideological
[17:21] and we ignore the idiosyncratic
[17:23] preferences, then potentially we're kind
[17:25] of underounting the degree to which
[17:27] there is a connection between voters
[17:29] preferences and whom they elect and what
[17:33] governments do.
[17:34] >> Yeah. I mean, I think there's been a lot
[17:35] of concern in the political science
[17:37] literature expressed, again, coming very
[17:40] heavily from those who find the sort of
[17:42] innocent voter perspective most
[17:45] compelling, that there's just not much
[17:47] link between what comes out of the
[17:50] political system, who is elected and
[17:52] what policies they follow, and what the
[17:55] public actually wants. Um, because there
[17:58] isn't much to the what the public
[17:59] actually wants. And people have found
[18:02] this troubling. There are various
[18:03] articulations of this view um most
[18:06] prominently recently in a book by Aken
[18:07] and Bartell's that it's just a mistake
[18:10] to understand democratic politics as a
[18:13] process by which lots of people form
[18:16] views about issues they vote and then
[18:19] they get something resembling
[18:22] those views put into action. it to the
[18:25] extent that vote choices are ba made not
[18:28] on policy grounds we wouldn't
[18:31] necessarily expect that that vision to
[18:33] be realized
[18:35] but to tell the extent to which people
[18:37] are making choices on policy grounds we
[18:41] need to have a very clear idea of what
[18:42] they actually want in terms of policy
[18:44] and that's that's where this book comes
[18:46] in is trying to get a better handle on
[18:48] what policy preferences people actually
[18:51] have so that we can figure out if their
[18:53] voting can track that and does track
[18:56] that and the extent to which that's
[18:58] shaping shaping their choices.
[19:00] >> Great. So those are exactly the
[19:02] questions that you're exploring and we
[19:03] know why they're important. How do you
[19:05] go about doing this?
[19:08] >> Right. So at the core of this book is a
[19:12] large new data collection that we did.
[19:15] Um it's a large panel survey experiment
[19:18] that was conducted in the UK from early
[19:21] 2018 to early 2019. And we should just
[19:24] explain a panel experiment here is
[19:26] >> a panel a panel survey is one where you
[19:28] go back to the same people more than
[19:30] once. So we ask questions of people at
[19:32] one point in time and then we find them
[19:34] again later and we ask them in this case
[19:36] the same questions again. The experiment
[19:39] part means that there are elements of it
[19:41] that are randomized and and I'll come on
[19:43] to those later.
[19:44] >> So early 2018 to early 2019 we had three
[19:48] waves at six months six month intervals
[19:50] interviewing the same people. We started
[19:53] with 6,000 people and sort of haved that
[19:55] at each wave which was a sort of
[19:58] intentional design to to make
[19:59] >> I was really interested in that
[20:00] actually. I'd never seen that that that
[20:02] element of design before. Just explain
[20:04] the thinking there.
[20:04] >> I was not expecting this to be the thing
[20:06] that you would leap to, but um there's a
[20:09] problem in all panel surveys of survey
[20:11] attrition that if you try to go back and
[20:14] interview the same people again, you
[20:15] will not find all of them. And the worry
[20:18] is always that the people you can't find
[20:20] again are different than the people you
[20:22] were able to find again. And so one way
[20:25] to mitigate that which we followed here
[20:28] is to sort of build in an expectation
[20:31] that you are going to lose some people
[20:34] and then target in particular for rein
[20:39] those people who are in the categories
[20:41] of people who you're likely to lose. So
[20:43] people who in this case who are less
[20:45] interested in politics and some other
[20:47] characteristics and so basically by
[20:49] planning on a certain level of attrition
[20:53] and not just trying to grab every rein
[20:55] we can we're able to reinter people who
[20:59] remain relatively representative of the
[21:01] UK public at the second and third round.
[21:03] >> Great. Thank you.
[21:06] >> Right. So in each of those three waves
[21:09] we asked two kinds of questions. The
[21:11] first were questions about what policies
[21:14] respondents preferred. We had a pool of
[21:17] 34 different policy areas or policy
[21:21] issues. And on each of those, we wrote
[21:23] down five policy options that had some
[21:26] kind of logical order. So for example,
[21:28] on the question about migration, we had
[21:31] a range of migration targets for the
[21:34] government to aim for in terms of net
[21:36] migration to the UK. And those had a
[21:38] numerical ordering. And on an issue like
[21:42] cannabis regulation, it was a range from
[21:46] uh sort of full legalization to sort of
[21:49] criminalization with, you know, jail
[21:52] time for people who are found using
[21:56] cannabis. So we designed these five
[21:59] options to have a logical ordering and
[22:00] to cover a really wide range of views um
[22:03] including some views that are not being
[22:04] ad advocated by any political party in
[22:07] the UK at the moment.
[22:09] So each respondent didn't have to answer
[22:11] all 34 of those. That would be a bit
[22:13] much. We they each got a random seven
[22:15] out of the 34. And then the second kind
[22:18] of question that we gave people were
[22:20] candidate choice experiments. So these
[22:23] are hypothetical candidates taking
[22:26] positions on three of the seven issues
[22:28] that that particular respondent had
[22:30] expressed a view about in their issue
[22:32] questions. The possible positions for
[22:34] the candidates were exactly the same
[22:36] five positions the respondents had to
[22:38] choose from. And we recorded which of
[22:41] the two candidates the respondent
[22:43] thought they would vote for given those
[22:46] three positions for each of those two
[22:48] candidates knowing nothing else about
[22:50] about the candidates. So a distinctive
[22:52] aspect of this study is that we did what
[22:56] I just described and there are studies
[22:58] that have done that before. What studies
[23:00] have not done before is do what I just
[23:02] described and then come back and ask
[23:03] people the same questions later. And in
[23:08] particular, asking about the same
[23:10] hypothetical candidates in early 2018,
[23:13] mid 2018, and in early 2019 is not
[23:17] something that that anyone has had done
[23:19] before and gives us a new kind of
[23:22] evidence to think about whether people's
[23:25] policy views are sufficiently grounded
[23:27] that faced with the same hypothetical at
[23:30] at pretty distant points in time, they
[23:33] will make the same choices. So are there
[23:36] is there not just are their issue
[23:38] positions themselves stable which we can
[23:40] directly assess because we ask those at
[23:42] each of the three points in time but is
[23:45] their sort of v voting calculus also
[23:47] stable over time such that faced with
[23:49] the same kinds of choices as they often
[23:52] are in the real political system where
[23:54] candidates and parties don't change that
[23:56] much over time. Do we in fact see
[23:58] stability there as well?
[24:00] >> There's a lot there. Let's uh I mean
[24:02] that was really really clear and
[24:03] helpful. Thank you. But let's just break
[24:05] things down a little bit and focus
[24:06] firstly on your the questions about
[24:10] people's issue preferences.
[24:13] What exactly were you trying to find out
[24:15] here and what did you find out?
[24:18] >> Right. So we have asked each respondent
[24:23] their views on seven issues at three
[24:25] points in time and there are a few
[24:27] things that we look to derive from from
[24:30] that data. So one of the questions goes
[24:33] back to this point we started with about
[24:36] ideological versus idiosyncratic versus
[24:38] unstable opinion. So with that structure
[24:43] of data we can assess the extent to
[24:45] which people's issue positions are
[24:47] correlated with one another. Um and we
[24:50] have all the combinations of issue
[24:51] positions across across the full sample
[24:53] even though individuals haven't answered
[24:55] all the issue positions. We can look at
[24:59] the extent to which people's positions
[25:01] are not ideologically structured but
[25:04] stable over time. That is when they
[25:07] answer the first wave and they answer
[25:08] the second wave and they answer the
[25:10] third wave. They give a view on a
[25:12] particular issue that doesn't
[25:13] particularly line up with their views on
[25:15] other issues in the ways we might
[25:17] expect, but it's the same every time or
[25:19] it's very or their adjacent categories
[25:21] every time. So close to the same. And
[25:24] then we can also see just how much it
[25:26] changes. So we ask them their view on a
[25:29] particular issue in wave 1 and they say
[25:31] one thing and in wave two they say
[25:33] something else and in wave three they go
[25:35] back to the first one. So you're
[25:36] supposing here that if people's views
[25:40] change from wave to wave then that's
[25:43] probably an indicator that they don't
[25:45] really have a very strong view on on or
[25:47] or a clear view on this issue. Whereas
[25:50] if they hold the same view consistently
[25:52] over time then we think okay well that's
[25:53] a meaningful preference. they genuinely
[25:55] think what they're saying they think.
[25:56] >> Yeah. I mean, this is a this is an idea
[25:59] that has goes back sort of 60 years in
[26:02] public opinion research. But the thing
[26:05] that you can look at to distinguish
[26:09] change that might reflect sort of
[26:12] meaningful lasting change that is I
[26:15] actually just changed my view between
[26:17] when you interviewed me the first time
[26:18] and when you interviewed me the last
[26:19] time. and people sort of bouncing around
[26:23] giving different answers each time is
[26:25] whether in a three-wave panel like the
[26:28] one we have, there's a difference
[26:31] between how related wave 1 and three are
[26:35] from how related wave 1 and two are and
[26:38] two and three are. So the logic here is
[26:40] if it's the case that if I change from
[26:43] wave 1 to two, that's sort of the new
[26:45] basis for my position going into wave
[26:47] three, we would expect one and two to be
[26:50] more similar and wave two and three to
[26:52] be more similar than one and three are.
[26:54] One and three should be have bigger
[26:56] differences than wave than the the
[26:57] temporally adjacent waves
[26:59] >> just because there's more time for
[27:00] people to have changed their minds.
[27:01] >> Exactly. And what we see in this data
[27:04] and what everyone has seen in every data
[27:06] set where they've done this kind of
[27:07] analysis is that wave 1 and three look
[27:10] just about as similar as wave 1 and two
[27:13] and two and three which is another way
[27:15] of saying that while there may be real
[27:18] opinion change in the world I suspect
[27:20] there is um it's just much much smaller
[27:24] than the amount of noise variability
[27:27] that we get in survey data because
[27:29] people are answering quickly they
[27:32] haven't necessarily thought about the
[27:33] questions before and so you get a bunch
[27:35] of scattering around in their responses
[27:38] over time and that is sort of the
[27:41] dominant source of of noise um in panel
[27:45] panel data.
[27:46] >> Yeah. Okay. So, one thing you're
[27:48] measuring here is the degree to which
[27:50] people have stable preferences at all
[27:52] and then you're also looking at where
[27:54] they do have stable preferences, the
[27:55] degree to which those are, as you
[27:57] described before ideological or
[27:59] idiosyncratic. And one of the things we
[28:01] see is the extent to which people have
[28:04] ideological or idiosyncratic or unstable
[28:06] views. It varies across issues and it
[28:09] varies across people. Across issues, I
[28:11] can give you some examples. The issue on
[28:14] which opinion is most ideologically
[28:16] structured in our data is a question
[28:17] about foreign aid. So the level of
[28:19] foreign aid that that the UK should be
[28:21] providing. People have very stable views
[28:24] on this issue and it those views are
[28:26] very strongly related to their views on
[28:29] a several other issues including
[28:31] attitudes towards Brexit towards
[28:33] migration and towards the death penalty.
[28:36] All of these sort of align on what we
[28:38] refer to as a as a social liberal versus
[28:40] conservative dimension which we find in
[28:44] our data is distinct from an economic
[28:46] left-right dimension um which is also
[28:48] present in the data. Mhm.
[28:50] >> So that's the most ideologically
[28:51] structured issue in the data. The most
[28:54] idiosyncratically structured issue in
[28:56] the data is the legal regulation of
[28:59] cannabis. So people have views, stable
[29:03] views on this issue. You go back to them
[29:05] 6 months, 12 months later, they by and
[29:07] large give you very similar responses to
[29:09] what they gave you um in wave 1. But
[29:12] it's not particularly correlated with
[29:14] their views on other issues. Now, you
[29:16] might think cannabis regulation would
[29:18] also have a social, liberal,
[29:19] conservative element to it, and it has a
[29:22] little bit of that, but actually most of
[29:23] the variation on that issue doesn't seem
[29:26] to correlate with the views on people
[29:28] express on those other issues that are
[29:30] along that social, liberal, conservative
[29:32] dimension, but is just distinctive to
[29:34] their views on cannabis itself.
[29:37] >> And are there also issues where people
[29:39] just mostly have very unstable
[29:41] preferences?
[29:41] >> There are indeed. And and perhaps
[29:43] unsurprisingly, they're more they're the
[29:45] issues that you look at and you think,
[29:46] "Boy, that's an obscure issue, like, you
[29:48] know, rules around land development and
[29:50] subsidies for food production and bank
[29:53] insurance." We intentionally asked about
[29:55] a range of issues from ones that were
[29:57] very high-profile to ones that were very
[30:01] obscure precisely because we wanted to
[30:03] sort of be able to document that in
[30:04] fact, you know, it's not shocking that
[30:06] that the more obsc, you know, the more
[30:08] obscure ones, people tend to not have
[30:10] well-formed views, but it it's a way of
[30:12] reassuring ourselves that yeah, that is
[30:14] in fact what instability is capturing.
[30:16] And didn't you also have a question
[30:17] about the kind of trade-off in policy
[30:19] terms between inflation and
[30:20] unemployment? Whether governments should
[30:22] be targeting inflation more or targeting
[30:24] unemployment more and that that was
[30:27] quite high in terms of instability as
[30:28] well.
[30:29] >> That was also very unstable. Now you
[30:30] might think, oh maybe that's because in
[30:33] fact the right answer to that question
[30:35] changed over time. I'm sure there would
[30:37] be sort of macroeconomists who would
[30:40] have views on that question that would
[30:42] be responsive to to immediate
[30:44] circumstances, but mostly what we find
[30:46] is people are very unstable. Their
[30:49] responses don't really track with any of
[30:51] their other issue positions. It may be
[30:53] that they're responding to events, but
[30:55] we think it's probably unlikely. In
[30:57] fact, if I remember correctly, most
[30:59] responses on that were the middle
[31:00] category of we should care about them
[31:04] sort of similarly.
[31:05] >> Yeah. And and that's not reflecting a
[31:07] sense that people don't really care
[31:08] about inflation or unemployment. It's
[31:10] just it's a really hard question and and
[31:12] as you say, most people probably haven't
[31:15] thought through very much the necessity
[31:17] of a degree of trade-off between those
[31:19] as government policy priorities.
[31:21] >> I think that's exactly right.
[31:23] >> Yeah. Fantastic. So that's really
[31:24] interesting. So you've talked there
[31:26] about various specific issues and the
[31:28] way that we get quite different patterns
[31:29] on different issues. What's the overall
[31:32] picture? I guess the fundamental
[31:33] question that we want to answer here is
[31:36] to what degree do we have ideological
[31:38] preferences, idiosyncratic preferences,
[31:40] or just no stable preferences at all.
[31:42] Yeah. And and the answer is we've got
[31:44] some of each. And the way that we have
[31:48] mathematically decomposed things, we
[31:51] have a bit more instability than
[31:52] idiosyncrasy and a bit more idiosyncrasy
[31:55] than ideology.
[31:57] We're aware of and talk about in the
[31:59] book the fact that this depended in part
[32:01] on the issues we chose to ask about. Um
[32:04] if you filled a survey with a bunch of
[32:05] obscure issues um that no one had ever
[32:08] heard of, you could clearly get a higher
[32:10] number for instability. And if you
[32:12] picked only the most high-profile issues
[32:14] that are currently the subject of sort
[32:17] of political contestation, you could
[32:19] probably get the ideology number to go
[32:21] up. Then we have a comparison to a US
[32:23] data set which is sort of that latter
[32:26] category where indeed the ideology share
[32:28] that we estimate tends to be higher
[32:30] although that could be the US UK
[32:32] difference or it could be the particular
[32:34] set of issues very hard to to to be
[32:36] confident which one
[32:38] >> but the headline here is kind of this
[32:41] big debate in political science
[32:43] literature that as you say has been
[32:44] going on for 60 years has kind of got it
[32:46] wrong I mean in so far as people have
[32:47] said it's sort of ideology versus
[32:50] nothing in terms of preferences.
[32:52] Actually, you find the third category,
[32:54] idiosyncrat idiosyncratic preferences
[32:58] really matter.
[32:59] >> Yeah. I mean, our reading of the data is
[33:00] that all three matter and, you know, so,
[33:03] you know, in a sort of classic third way
[33:05] style, we have, you know, adopted
[33:07] elements of both of the existing
[33:09] perspectives, added a bit of our own and
[33:12] suggested that all three of these things
[33:13] are an important part of understanding
[33:15] public opinion. So that's the first part
[33:18] of the analysis looking at the nature of
[33:20] people's preferences. And then the
[33:21] second part looks at how these
[33:24] preferences then translate into actual
[33:27] voting behavior. You you've you've sort
[33:30] of said already how you go about doing
[33:32] that. Do you want to just remind us
[33:33] because it was a little while ago that
[33:35] we
[33:35] >> Right. So we've given people these
[33:37] questions where they choose between
[33:38] hypothetical candidates. the candidates
[33:40] are taking positions on three of the
[33:42] issues on which we know what the
[33:45] respondent said their own position was
[33:48] and we've asked them the same
[33:50] hypotheticals at three points in time so
[33:53] we can see if they give the same answers
[33:54] at three points in time. So there are a
[33:58] few things that we get from these data
[34:00] and I'll I'll try to sort of quickly
[34:02] outline what those are.
[34:04] So the first thing is we look at which
[34:07] issues people seem to care more about in
[34:09] the sense that they punish candidates
[34:11] who take positions that are different
[34:12] from their own positions. These turn out
[34:15] to be an interesting mixture of issues
[34:16] that are highly politicized like the EU
[34:19] withdrawal which was going on at the
[34:21] time of the survey and issues that were
[34:24] basically absent from politics at the
[34:25] time like the death penalty. It's worth
[34:28] noting that the UK is is currently
[34:31] treatybound um by membership in the
[34:33] European Convention on Human Rights to
[34:35] not use the death penalty and the death
[34:37] penalty was abolished in 1969 and none
[34:40] of the parties took prominent positions
[34:41] on this at the either the 2017 or the
[34:45] 2019 election. Um, but what we show is
[34:48] that there's actually a lot of
[34:50] disagreement about among the UK public
[34:52] about whether the death penalty should
[34:54] be used. And there was some previous
[34:56] survey evidence on that as well, not not
[34:58] entirely original to us. What we're able
[35:01] to show though is that if you put
[35:03] candidates in front of people who have
[35:05] ex contrasting views on on that issue,
[35:09] people put a lot of weight on it in
[35:11] their choice. So what we're measuring
[35:14] here is what we call choice importance.
[35:16] That is the importance to candidate
[35:18] choice um for issues that may or may not
[35:21] be currently the subject of politics. So
[35:24] the advantage of hypothetical candidates
[35:26] is that you can ask people about things
[35:27] that are not currently naturally
[35:29] occurring in the world and find out
[35:31] something about with some caveats about
[35:33] external validity what people how people
[35:36] might respond if their choice if the
[35:38] choices available to them were
[35:39] different. So that's one kind of thing
[35:41] that we're we do with these with these
[35:43] data. We can measure that choice
[35:46] importance for all of the 34 issues that
[35:48] we considered. And one of the things
[35:51] that we see very clearly in the data is
[35:54] how much weight people put on these
[35:56] issues on average across everyone is
[35:58] very highly correlated with opinion
[36:01] stability on those issues. So it's the
[36:03] same issues where people give the same
[36:05] answer for themselves at these three
[36:07] points in time where people also punish
[36:10] candidates who take positions that are
[36:13] different from them at one point in
[36:14] time. So this we think and we argue
[36:19] speaks to this idea which I alluded to
[36:21] at the beginning that it may be that
[36:24] people don't have well-formed views
[36:26] about every issue. They may only have
[36:29] well-formed views about a few issues.
[36:31] Those might be different for different
[36:32] people, but the tendency will be for
[36:34] people to put a lot of weight in their
[36:36] choices on the issues where they do have
[36:38] well-formed views. They're not just sort
[36:40] of checking their unstable view about an
[36:42] issue at one point in time and making a
[36:44] vote choice on the basis of that and
[36:45] then they have a totally different
[36:47] expressed pre view at a different point
[36:49] in time, but then they use that in their
[36:51] choice at that point in time. rather
[36:53] those that instability and issue
[36:55] position is going to be associated with
[36:57] people not actually using that issue in
[36:59] making their choices.
[37:01] >> So a crudder approach here might have
[37:03] just asked lots of questions about
[37:04] issues and found that people don't have
[37:06] strong opinions on most of those issues
[37:08] and concluded that therefore well they
[37:10] don't really have very many opinions
[37:11] about stuff but actually you're finding
[37:14] people do have opinions about some
[37:15] things and those are the opinions those
[37:17] are the issues that they're thinking
[37:18] about or at least that are shaping their
[37:20] vote choices.
[37:21] >> That's right. And we can see it in the
[37:23] aggregate that is when you look across
[37:25] everyone the issues where more people
[37:27] have stable views are the issues that
[37:29] people put more weight on in making
[37:31] their candidate choices. But it's also
[37:33] the case that looking within issues the
[37:35] people the specific people who had
[37:38] stable views across different waves of
[37:40] the survey survey tend to put more
[37:42] weight on those issues in their
[37:44] candidate choices than the people who
[37:46] didn't have stable views on those issues
[37:48] across the different waves of the
[37:49] survey. So, so we think that's a quite
[37:53] interesting kind of evidence that is
[37:55] gets us part of the way towards being
[37:57] able to show that
[38:01] even though many people don't have
[38:02] well-formed views, don't have stable
[38:04] views on issues, it does seem like those
[38:07] aren't sort of th there there's reason
[38:10] to think they aren't then sort of
[38:12] putting much weight on those and that
[38:13] but they are putting weight on the
[38:15] issues where they do have stable views.
[38:17] >> Yeah. Would you say something about the
[38:18] competence of voters? it leads to a more
[38:20] more optimistic view of the competence
[38:22] of voters than has come from some
[38:23] political scientists.
[38:24] >> I think that's right. I mean, it's a
[38:25] very sensible orientation for someone to
[38:28] take is if I'm choosing between two
[38:30] alternatives and I know some things that
[38:33] are relevant to that choice and then
[38:35] there are other things I could know but
[38:36] don't know that would be relevant to it,
[38:39] I'll make the choice based on the things
[38:40] I do know and that that seems a sensible
[38:43] way to approach the problem.
[38:44] >> Yeah. And then what about the difference
[38:46] between ideological preferences and
[38:48] idiosyncratic preferences? Do we see any
[38:50] difference in the aggregate in the level
[38:53] of importance of these in shaping voting
[38:54] behavior?
[38:55] >> The answer is that we don't. There's um
[38:58] what we find in the in the data and we
[39:00] look at this in a few different ways is
[39:02] that both forms of stable preference
[39:06] predict greater importance in candidate
[39:08] choices on those issues. And I'm not
[39:11] going to get into all the ways that we
[39:12] look at that here, but the basic idea is
[39:15] that what we, you know, that is in
[39:17] contrast to unstable opinion, which is
[39:22] doesn't predict people putting much
[39:24] importance on issues. So what we see is
[39:27] people put a lot of more weight on
[39:29] issues where their preferences are
[39:32] stable and it doesn't really matter
[39:34] whether they're that particular issue
[39:37] their views are aligned with sort of a
[39:40] more general ideology across multiple
[39:42] issues or whether it's just their view
[39:44] on that issue.
[39:46] Some listeners might be a bit skeptical
[39:49] about these kinds of experimental survey
[39:51] questions where you're asking people to
[39:53] choose between hypothetical candidates
[39:54] and they're, you know, they're often
[39:56] wildly unrealistic candidates that
[39:58] you're asking people choose to choose
[39:59] between and we might kind of wonder,
[40:01] well, does that really give us
[40:02] information about how people would
[40:03] actually vote in a proper election that
[40:05] is genuinely taking place?
[40:07] >> It's a fair question. We talk a bit
[40:09] about this in the book. There are
[40:11] obviously advantages to the experiments
[40:13] in which I've already alluded to. Um the
[40:15] disadvantages as you say are that we
[40:18] might worry that okay we've put people
[40:20] in an artificial context where they know
[40:22] nothing about these candidates except
[40:24] policy positions. So, we can't really
[40:26] say that we've proved that real people
[40:29] follow this logic because we've taken
[40:31] away the possibility of them making
[40:33] decisions based on other things like
[40:36] party identity, past performance of
[40:39] governments, etc. that they might be
[40:41] influenced by in a real election. Um,
[40:44] and that's absolutely right. And we talk
[40:46] about this and and one of the things we
[40:48] do is we talk about that what we're
[40:50] showing is sort of a possibility result
[40:52] that if people only had information
[40:56] about policy, what we're showing is that
[40:58] they can vote on the basis of that. They
[41:00] can vote in ways that are stable over
[41:03] time and they vote in ways that are sort
[41:06] of sensibly responsive to the positions
[41:09] that they hold, you know, more
[41:11] meaningfully and are not responsive to
[41:13] the issue positions that they've sort of
[41:15] seemed to have thrown out at, you know,
[41:17] a particular survey wave because they
[41:19] had to answer the question even though
[41:20] they hadn't thought about it before.
[41:24] >> So overall, we have the conclusion that
[41:27] idiosyncratic preferences matter. So do
[41:29] ideological preferences. So do I mean
[41:32] it's not the case that people have
[41:33] preferences in everything either. So so
[41:35] all three models say something. And all
[41:37] three to a degree predict voting
[41:40] behavior or or matter in our
[41:42] understanding of voting behavior. Um
[41:45] we're running out of time as we so often
[41:47] do in these conversations, but we should
[41:50] just think a little bit about the
[41:51] further implications of all of this. And
[41:55] one question in relation to that is
[41:58] do these findings travel beyond the
[42:00] particular context that you were looking
[42:01] at of the UK in 19 2018 2019 and
[42:06] listeners will not struggle to notice
[42:08] that that was a really weird period in
[42:11] British politics when Brexit was the top
[42:14] issue. So you know one could say we have
[42:17] this new issue relatively new issue on
[42:20] the agenda. It's kind of broken apart
[42:22] the voting coalitions of the major
[42:25] parties. Um it's a very complex issue.
[42:30] So you know people had strong views on
[42:32] this issue but they may have got to
[42:34] those views via via a range of different
[42:37] outcomes
[42:38] a range of different routes. Your
[42:41] analysis suggests that as as that would
[42:44] predict this was an issue that people
[42:45] did care about very much. They had
[42:47] stable preferences on it, but they were
[42:51] those preferences were partly
[42:52] ideological, but partly also
[42:54] idiosyncratic. So, both of those factors
[42:56] were very important. One could worry
[42:59] that you're just doing this study at a
[43:02] time when there was an issue that people
[43:05] cared about that was in significant part
[43:08] idiosyncratic.
[43:10] Um and that but that was quite unusual
[43:12] that normally politics goes along kind
[43:15] of leftright t terms or the social
[43:19] conservative socially liberal versus
[43:21] versus socially conservative dimension
[43:23] and that this was just a weird moment in
[43:25] British politics when a weird issue with
[43:28] idiosyncratic characteristics was
[43:31] important but that these findings might
[43:33] not travel beyond that context.
[43:36] So I think there there are a couple ways
[43:38] to answer that question about
[43:40] generalizability. On the sort of narrow
[43:42] point about Britain in 2018 2019, it is
[43:46] an interesting moment and I agree
[43:48] entirely with your characterization of
[43:51] the EU withdrawal issue being one where
[43:56] people did have idiosyncratic
[43:58] preferences. Preferences they had they
[44:00] held them. They held them strongly.
[44:02] There's lots of evidence of people
[44:04] switching their votes between the
[44:07] relevant elections in ways that bring
[44:10] those votes into alignment with their
[44:12] view about this issue. and that it's
[44:15] true that that doesn't happen all the
[44:17] time, but I think it is a common feature
[44:20] of electoral realignments and and the
[44:23] dynamics of elections and party systems
[44:25] over time in lots of places that part of
[44:28] what changes things over time is that
[44:32] new issues become important. um they
[44:35] become salient either because of
[44:36] external events or for for other reasons
[44:40] and that leads to changes in voting
[44:44] coalitions
[44:46] which is sort of what you would expect
[44:47] if you had this idiosyncratic preference
[44:50] perspective um is that there was the
[44:53] potential for that to happen. And so I I
[44:57] think you can see elements of this in
[44:59] other places on various time scales. We
[45:03] talk in the in the book a little bit
[45:05] about some US examples. The issue of
[45:08] abortion from the 1970s onwards is one
[45:11] where if you look initially views on
[45:14] abortion are almost unrelated to
[45:16] people's party preferences in at the
[45:19] point where Roie Wade is decided in 1972
[45:23] and over time people's views and to some
[45:27] extent this is people changing which
[45:28] party they support and to some extent
[45:30] it's people changing their views about
[45:32] abortion to bring them in line with the
[45:33] party they already supported you know
[45:36] reorient orient that issue and the
[45:38] party's coalitions. We've seen similar
[45:41] things with respect to views on
[45:42] immigration maybe starting in around the
[45:44] 1990s. Those are somewhat slower
[45:47] processes than the one than the one
[45:50] around Brexit. Although the one around
[45:52] Brexit really starts substantially
[45:54] earlier than the referendum. Um it's
[45:56] just that the referendum sort of creates
[45:58] an urgency and saliency to the issue
[46:01] that is perhaps particular to to to the
[46:04] UK at that moment in time. But there are
[46:06] issues that arise in lots of countries
[46:08] that sort of realign their political
[46:11] systems and and and I and we think that
[46:13] reflects the fact that there are these
[46:15] latent views out there in the public
[46:17] where people disagree where people have
[46:20] different views. They would feel
[46:21] strongly about it if it was presented to
[46:23] them as a relevant part of their
[46:25] political choice. And sometimes they're
[46:27] given that choice and sometimes they're
[46:29] not. and the issue just sort of lurks
[46:31] there in the background waiting for some
[46:33] political entrepreneur to to pick it up
[46:35] and run with it. So that's one point
[46:38] about generalizability. The other thing
[46:40] I would say is, you know, the UK is a
[46:42] particular country. It has a particular
[46:43] political history. We do nonetheless
[46:46] think that our analysis and sort of what
[46:50] we're saying about how citizens engage
[46:52] with politics and political issues and
[46:54] policy in general, we think that
[46:55] probably travels reasonably well to
[46:58] other long-standing democracies in North
[47:02] America, in Europe, in Australasia. Um,
[47:06] and part of the reason we think that is
[47:09] that there's lots of commonality between
[47:12] these countries in terms of the
[47:14] structures of their economies and the
[47:16] kinds of jobs and that people have. The
[47:19] political institutions vary in their
[47:21] details, but in the grand scheme of
[47:23] things, there's a lot of similarity,
[47:25] periodic elections, all of these things.
[47:28] The party systems are different and
[47:30] that's important. But all of these
[47:32] countries have experienced, you know,
[47:34] rising formal education levels,
[47:36] de-industrialization,
[47:38] perhaps increasing hostility to
[47:40] immigration and other sort of macro
[47:42] patterns that apply in lots of places.
[47:44] The one really important kind of
[47:46] variation we think that changes how
[47:49] these dynamics work in different places
[47:51] is this one related to party systems and
[47:53] sort of the menu of ideological
[47:56] groupings that are that are available to
[47:58] people to look to and to choose between.
[48:02] There are electoral systems that
[48:04] encourage two-party systems and there
[48:06] are electoral systems that encourage
[48:08] highly multi-party systems. And in this
[48:10] we think as well the UK is not a bad
[48:13] single case to look at because it sits
[48:15] sort of somewhere in between. It's
[48:17] neither a sort of two pure two-party
[48:19] system like the US nor is it a sort of
[48:22] highly multi-party system like say the
[48:25] Netherlands. It's got some elements of
[48:27] both and we can see you know not just
[48:29] one ideological dimension but two both
[48:32] in the parties and what they offer and
[48:33] also among the public and so you know it
[48:37] can speak to we think a wide range of
[48:39] cases.
[48:41] >> Finally then what are the lessons that
[48:44] you would like listeners to go away with
[48:46] from this analysis?
[48:49] That's probably the hardest question to
[48:52] answer. And you know, one of the things
[48:55] that we would like the political
[48:57] scientists among our listeners to think
[49:01] about is doing more to think about how
[49:07] these different perspectives on voting
[49:09] behavior coexist and how, you know, they
[49:12] all describe elements of what people are
[49:15] doing. We think the right perspective on
[49:17] public opinion is one that recognizes
[49:19] there is some ideology there is some
[49:21] idiosyncrasy and there's is some sort of
[49:23] innocence and instability as well and
[49:26] similarly we think you know zooming out
[49:28] to that sort of the different models of
[49:30] voting behavior that that I started with
[49:33] you know that it is true that to some
[49:35] extent people's vote choices are based
[49:37] on identity considerations
[49:39] considerations of you know who who
[49:41] people are who their who parties can you
[49:44] know base are and things like this. And
[49:47] also we think people, you know, make
[49:49] their choices in part based on sort of
[49:52] performance things like this. And also
[49:54] we think they make their choices based
[49:56] on policy offers. And we should do more
[50:00] to figure out how those all coexist and
[50:04] you know the extent to which there are
[50:05] some people doing one of those things
[50:07] and other people doing another or people
[50:09] doing interesting mixes of those and who
[50:11] those people are and when different
[50:12] things matter. you know, we think that
[50:15] sort of meta version of what we've done
[50:17] here would be really useful as a
[50:20] direction for the literature to go. So
[50:22] that's for the political scientists. Um,
[50:25] for the people who are interested in
[50:26] elections and and political behavior, I
[50:28] think it's important I I guess the thing
[50:30] I would say is it's important to take
[50:32] people's policy views seriously and to
[50:36] think about not just their views on the
[50:39] sort of the hot button issues, the ones
[50:42] that are currently the subject of
[50:43] debate, but also the various things
[50:45] people care about that are sort of
[50:47] lurking in the background where the
[50:49] parties aren't necessarily taking
[50:51] clearly opposing positions. But there
[50:53] might be a big constituency out there
[50:56] for some significant change as it turned
[50:59] out there was with Brexit. And I think
[51:04] finding those can be hard, but it's
[51:07] definitely possible. And we sort of show
[51:08] some of the toolkit here that you would
[51:10] use to go and figure out where those
[51:12] opportunities lie.
[51:15] >> Terrific, Ben. Thank you so much.
[51:17] There's so much in this book and we've
[51:19] uh covered a fair bit, but there's much
[51:20] much more. So uh it will be great if
[51:23] listeners appetites have been weted and
[51:25] uh they're able to go out and read the
[51:27] book as well. We have been discussing
[51:30] the book idiosyncratic issue opinion and
[51:32] political choice by Nick Viviian
[51:34] Benjamin E. Lauderdale and Chris
[51:37] Hanetti. It's published by Oxford
[51:39] University Press published at the end of
[51:41] 2025.
[51:42] It's available in print if you're
[51:44] willing to pay for it, but you can also
[51:46] read it online free of charge through
[51:49] the OUP website. And we will of course
[51:51] put the full notes uh the full details
[51:53] in the show notes for this episode. Next
[51:56] week we will be looking at the ethics of
[51:59] quitting social media. To make sure you
[52:01] don't miss out on that or other future
[52:03] episodes of UCL Uncovering Politics, all
[52:05] you need to do is subscribe. You can do
[52:07] so on Apple, Google Podcasts, or
[52:09] whatever podcast provider you use. And
[52:11] while you're there, we'd love it if you
[52:13] could take a moment of time to rate or
[52:15] review us, too. I'm Alan Renick. This
[52:18] episode was produced by Elellanena
[52:20] Kingwell Banham. Our theme music is
[52:22] written and performed by John Man. This
[52:24] has been UCL Uncovering Politics. Thank
[52:26] you for listening.

Introduction to the Module on Evidence Analysis for Policy Makers

In this episode of UCL Uncovering Politics, Alan Renwick introduces a discussion on how future policy makers are taught to utilize evidence effectively. The podcast aims to explore the core features of good evidence analysis that students need to grasp to make informed policy decisions. The episode features Dr. Julia de Romémont, who has developed a new module for master's courses at UCL's Department of Political Science and School of Public Policy. This module is designed to equip students with the skills to critically assess policy-relevant evidence, even if they do not delve deeply into complex statistical methods.

The Need for the Module

Dr. de Romémont explains that the module was created in response to the realization that while students are taught introductory quantitative methods, only a minority pursue advanced statistical techniques. The traditional approach often emphasizes coding and technical skills, which can be overwhelming for students. Consequently, many students disengage from the material. The new module focuses on building intuitions rather than delving into the mechanics of statistical methods. This approach aims to empower students to critically evaluate evidence without needing to understand complex mathematical formulas.

Applications of the Module in Real-World Scenarios

Dr. de Romémont highlights the importance of evidence-based policy making, particularly within the UK government, which has made significant strides in this area. Many students in public policy programs may find themselves in roles where they need to interpret and act upon evidence presented by data scientists and analysts. The module prepares them for these situations by instilling confidence in their ability to assess quantitative evidence critically. This is crucial because there is a common misconception that quantitative data is inherently objective, which can lead to misplaced trust in the findings presented.

Target Audience for the Module

The module is aimed not only at future civil servants but also at politicians, journalists, and commentators. Dr. de Romémont emphasizes that understanding evidence is vital for anyone involved in policy making or public discourse. The goal is to foster a generation of policy makers who can engage with evidence critically, leading to better-informed decisions that ultimately enhance public policy outcomes.

Core Skills and Intuitions Developed in the Module

The content of the module is structured around three types of inferences that students must learn to make: measurement inference, causal inference, and population inference.

• Measurement Inference: This involves understanding what the observed data signifies about the underlying concepts. For instance, survey questions about political attitudes must be critically examined to determine if they accurately reflect the respondents' views. Students learn to question the wording of survey questions and the implications of those questions on the data collected.
• Causal Inference: This section teaches students to differentiate between correlation and causation. Dr. de Romémont stresses that establishing causality requires careful consideration of alternative explanations and the comparability of groups being studied. The module introduces concepts like randomized control trials and the potential outcomes framework to help students grasp these ideas without getting bogged down in complex statistical methods.
• Population Inference: Here, students learn to generalize findings from a sample to a broader population. They explore the importance of representativeness in sampling and the implications of drawing conclusions based on limited data. The module encourages critical thinking about how well a sample reflects the larger population and the potential biases that may arise.

Engagement and Critical Thinking in the Classroom

The module employs a seminar format where students engage with real-world evidence reports, often produced by think tanks or research institutes. They are tasked with presenting these reports and critically analyzing their methodologies and findings. This hands-on approach fosters a collaborative learning environment where students can share insights and challenge each other's perspectives.

Dr. de Romémont notes that students often express surprise at how much they can critique evidence, even without a deep understanding of the statistical methods involved. This empowerment is a key outcome of the module, as students learn to focus on the quality of the data and the assumptions underlying the analyses they encounter.

Challenges and Limitations of Evidence-Based Policy Making

Throughout the module, students are made aware of the limitations of evidence-based policy making. Dr. de Romémont cautions against an over-reliance on quantifiable metrics, which can lead to overly cautious decision-making. She emphasizes that while evidence is crucial, it should not stifle creativity and boldness in policy development. Historical examples, such as the establishment of the NHS or the New Deal, illustrate that transformative policies often emerge from a willingness to take calculated risks, even in the absence of robust evidence.

Feedback and Student Reactions

Dr. de Romémont shares that student feedback has been overwhelmingly positive. Many students report feeling more confident in their ability to engage with evidence and critique research findings. The module's focus on intuitions rather than technicalities resonates with students who may have previously felt intimidated by quantitative methods. The collaborative atmosphere in seminars encourages students to voice their thoughts and engage in meaningful discussions about the evidence they encounter.

Conclusion: The Future of Evidence-Based Policy Making

In conclusion, the module designed by Dr. de Romémont represents a significant step toward equipping future policy makers with the necessary skills to navigate the complexities of evidence analysis. By focusing on critical thinking and intuitive understanding, the module prepares students to engage with evidence confidently and make informed decisions in their future careers. As the demand for evidence-based policy making continues to grow, the insights gained from this module will be invaluable for students as they enter the workforce.

Listeners are encouraged to consider enrolling in UCL's master's programs to benefit from this innovative module and the expertise of Dr. de Romémont and her colleagues. The podcast concludes with a preview of the next episode, which will delve into the tradition of political parades in Northern Ireland, inviting listeners to subscribe for future updates.

The transcript does not specifically delve into the main drivers of voter behavior, such as policy preferences, material interests, or identity. It focuses primarily on the need for policymakers to assess evidence critically rather than concentrating on voter behavior itself. The discussion is centered around the education of future policymakers on how to evaluate evidence effectively, which indirectly suggests that understanding these drivers is essential for creating informed policies that resonate with voters. It is implied that voters' decisions often stem from complex motivations, but the text does not elaborate on why voters might deviate from purely rational choices.

• [04:50] "And the danger often, especially when we talk about quantitative evidence, which this module focus on slightly more or quite a bit more because that’s more my personal expertise than qualitative methods, is that there’s a danger that like, oh, there’s numbers, it’s maths, it’s objective, and I can just believe what comes out of it."
• [06:05] "Exactly. And ultimately, the idea is that this would lead to better policy making in the future, that people are able to better assess also the state of the evidence because it’s one thing to only understand one study very well, but it’s another thing being presented with a whole like loads of different studies and try to make informed decisions on that basis."

The transcript does not explicitly discuss how voter information or election debate formats affect the quality of voters' choices. Instead, it emphasizes the role of education and understanding in helping future policymakers evaluate evidence. By focusing on how students can become better consumers of policy-relevant evidence, the speaker implies that the way information is presented and debated in elections is crucial for voters' comprehension and decision-making. The speaker highlights the importance of being able to critically engage with evidence, suggesting that a lack of understanding can lead to poor choices by voters.

• [05:11] "But there’s actually a lot of things that we need to question around that. And it’s basically giving students the confidence in the tools to know where to look without having to look at the mathematical formula in those."
• [06:02] "...it would be great if they did have the confidence to engage with this sort of evidence even if they don’t have the full grasp of the methods."

The transcript does not address the role of moral convictions or economic circumstances in driving ideological disagreement or far-right leanings. It primarily focuses on the education of students in evidence-based policymaking and the importance of understanding statistical methods and evidence assessment. While these elements are essential for informed political discourse, the speaker does not provide specific insights into how moral or economic factors influence voter ideology or behavior.

• [06:11] "...that people are able to better assess also the state of the evidence because it’s one thing to only understand one study very well, but it’s another thing being presented with a whole like loads of different studies and try to make informed decisions on that basis."
• [11:04] "...we’re always still making a leap of faith. We’re always still trying to just say something about something we can never directly observe."

The speaker emphasizes the importance of equipping citizens with the skills to critically evaluate evidence and political claims. While the transcript does not provide explicit recommendations for navigating misinformation, it suggests that understanding evidence is crucial for making informed decisions. The speaker states that it is essential for future policymakers to be able to assess the credibility of evidence without necessarily mastering the statistical methods used to produce that evidence. This aligns with the idea that citizens should cultivate a critical mindset towards the information presented to them, enabling them to evaluate political promises more effectively.

• [07:52] "...also think tanks, people who work in like regulation as well, so public service adjacent positions as well. I think that there’s a broad like a very broad population of people who this might be useful for."
• [11:02] "...understanding that any analysis makes a little bit of that leap of faith, I think also helps."

The transcript does not directly address how voters judge governments retrospectively based on economic outcomes. However, it implies that understanding evidence is vital for making informed judgments about policy effectiveness. The speaker mentions the importance of assessing various studies and their findings, indicating that citizens should be able to evaluate how past decisions have impacted current outcomes. This approach suggests that fair accountability in governance requires voters to have the tools to analyze the evidence presented to them, but specific details regarding retrospective judgment are not discussed.

• [06:13] "...it’s one thing to only understand one study very well, but it’s another thing being presented with a whole like loads of different studies and try to make informed decisions on that basis."
• [10:32] "...the fact that you’re doing inferences means that you’re trying to make claims about something that we never directly observe."

[00:05] Hello, this is UCL Uncovering Politics,
[00:07] and this week we're looking at how we
[00:09] teach future policy makers about using
[00:11] evidence. What core features of good
[00:14] evidence analysis do they need to know,
[00:17] and how do we get them across?
[00:24] Hello, my name is Alan Renwick, and
[00:26] welcome to UCL Uncovering Politics, the
[00:28] podcast of the School of Public Policy
[00:30] and Department of Political Science at
[00:32] University College London.
[00:34] Policy makers hopefully base their
[00:36] decisions on careful assessment of the
[00:38] available options, and at least part of
[00:41] that assessment is likely in most cases
[00:43] to involve systematic analysis of
[00:45] relevant evidence.
[00:47] But such analysis, at its best, often
[00:49] uses complex statistical methods, and
[00:52] few policy makers, whether they're
[00:54] politicians or officials, or indeed
[00:56] campaigners or commentators, have the
[00:58] capacity to learn those methods in
[01:01] depth. So, how can we ensure that policy
[01:03] making can be well grounded? Well, a
[01:06] module recently introduced on our
[01:08] master's courses here in the UCL
[01:10] Department of Political Science and
[01:11] School of Public Policy aims to square
[01:13] the circle. It teaches our students,
[01:15] many of whom will have great careers as
[01:17] policy makers in the future, how to be
[01:20] excellent consumers of policy relevant
[01:22] evidence.
[01:24] Even if they're not ready to apply
[01:26] methods of statistical analysis
[01:27] themselves, it helps students think
[01:29] through what conclusions they might be
[01:31] able to draw from any given piece of
[01:34] research. So, in a slightly different
[01:36] episode from the usual podcast, we're
[01:38] going to explore this module. What's it
[01:40] for? How does it work? And also, how
[01:43] much of the content can you, dear
[01:44] listener and I, lap up over the next
[01:47] half hour or so? To explore all of these
[01:49] things, we're joined by the module's
[01:51] deviser and convener, and my wonderful
[01:53] colleague, Dr. Julia de Romémont, who's
[01:56] lecturer teaching in quantitative
[01:58] research methods and political science
[02:00] here in our department. Welcome Julia to
[02:03] UCL Uncovering Politics. It's fantastic
[02:05] to have you on your first appearance on
[02:07] the show. And let's go straight into the
[02:10] heart of the matter. What's this module
[02:12] for? What was the need that you
[02:14] perceived for which you saw this module
[02:17] as the solution?
[02:19] >> So, this module was in response to the
[02:21] fact that at this point we teach our
[02:23] master's students an intro to
[02:25] quantitative methods, where we teach
[02:27] them both a little bit the methods but a
[02:28] lot of the coding as well. And what
[02:31] we've realized over the time is that
[02:33] only a minority of those students
[02:34] actually take to it and end up doing
[02:36] more advanced methods and actually
[02:37] applying it. And while obviously it
[02:40] makes sense to understand how the
[02:42] methods work to be able to then, you
[02:44] know, read evidence,
[02:46] often it doesn't cover enough ground for
[02:48] them to be actually able to do that
[02:50] effectively once they have a job where
[02:52] they're presented with a lot of
[02:53] evidence. And also a lot of the time is
[02:55] taken over by the coding and things that
[02:58] also some students feel very stressed
[03:00] about and therefore like don't engage as
[03:03] much as we would want to.
[03:05] And the discussions that me and
[03:07] colleagues had about this is just like
[03:09] how can we cover more ground and really
[03:11] focus on the intuitions rather than the
[03:13] mechanics of how these methods work.
[03:15] Because ultimately you don't need to
[03:18] understand the mechanics in order to
[03:19] assess whether given piece of evidence
[03:21] is making a credible leap of faith or
[03:23] not because as much as some of these
[03:25] methods are extremely advanced,
[03:28] they're still are trying to make a leap
[03:30] of faith in terms of what they're
[03:32] claiming, right? They're all going to be
[03:34] wrong in some way at describing the very
[03:36] complex social reality. They have to
[03:39] decomplexify, sorry, that's a very weird
[03:41] word that I chose to say in a podcast,
[03:43] decomplexify reality. And so I think
[03:47] it's like really focusing on those
[03:48] intuitions that we wanted to build the
[03:50] module around or that I wanted to build
[03:52] the module around. And yeah, so that's
[03:54] what I did.
[03:55] >> Well, that's fantastic. And go a little
[03:57] bit further into how this might be
[04:00] useful for people. So, what are you
[04:01] imagining of the kinds of situations in
[04:03] which the material that people are
[04:05] covering in this module might turn out
[04:07] to be helpful?
[04:07] >> Yes, so especially in the UK government,
[04:10] there's been a big push for doing
[04:11] evidence-based policy making. I think
[04:13] the UK has been generally quite good at
[04:15] that. So, that means that there's a lot
[04:17] of impact evaluations that are produced
[04:19] by different parts of ministries and so
[04:20] on. And some of these people are
[04:22] themselves data scientists and know how
[04:24] to do the analysis, but ultimately they
[04:27] then bring that evidence forward to
[04:29] other parts of the government who
[04:31] themselves, like other people in the
[04:33] civil service or higher-ups or
[04:35] politicians themselves who then need to
[04:36] make decisions based on those things.
[04:38] And a lot of our students who do a
[04:40] master in public policy or a master in
[04:42] international public policy, they might
[04:44] end up in these kinds of positions where
[04:46] they're presented with these reports and
[04:48] have to make decisions based on them.
[04:50] And the danger often, especially when we
[04:52] talk about quantitative evidence, which
[04:54] this module focus on slightly more or
[04:57] quite a bit more because that's more my
[04:58] personal expertise than qualitative
[05:00] methods, is that there's a danger that
[05:02] like, oh, there's numbers, it's maths,
[05:04] it's objective, and I can just believe
[05:05] what comes out of it. But there's
[05:07] actually a lot of things that we need to
[05:09] question around that. And it's basically
[05:11] giving students the confidence in the
[05:13] tools to know where to look without
[05:15] having to look at the mathematical
[05:17] formula in those. And so, the idea that
[05:20] this would then empower them to be able
[05:22] to make decisions even though they feel
[05:25] that they haven't understood the depth
[05:27] of the methods.
[05:28] >> And you're suggesting there that this
[05:30] may be useful for students if they go on
[05:31] to careers in the civil service. But
[05:34] presumably also it can be useful if they
[05:36] themselves become politicians or
[05:37] journalists, commentators, and you know,
[05:40] a lot of maybe people like us are often
[05:43] quite kind of disparaging about
[05:45] politicians and the degree to which
[05:47] politicians actually understand the
[05:48] evidence that is before them. But I
[05:50] guess part of the point of this module
[05:52] is that yes, those politicians, it would
[05:55] be great if they did have the confidence
[05:57] to engage with this sort of evidence
[05:59] even if they don't have the full grasp
[06:02] of the methods.
[06:02] >> Exactly. And ultimately, the idea is
[06:05] that this would lead to better policy
[06:07] making in the future, that people are
[06:09] able to better assess also the state of
[06:11] the evidence because it's one thing to
[06:13] only understand one study very well, but
[06:16] it's another thing being presented with
[06:18] a whole like loads of different studies
[06:20] and try to make informed decisions on
[06:22] that basis. And I think
[06:24] the idea of the module is to give people
[06:26] at least a starting tools on like how to
[06:28] develop those skills beyond just the
[06:30] analysis and try to um yeah, feel more
[06:32] confident because also a big problem
[06:34] with these methods and the pressure on
[06:35] the stats and everything. Like people
[06:37] feel oh, that's just not my thing and
[06:40] therefore I'm not going to engage with
[06:41] it and I'm going to trust other people
[06:42] to do it, but ultimately, we should all
[06:45] be thinking about what we can learn from
[06:46] that. And yes, policy makers, civil
[06:49] servants, but like politicians
[06:50] themselves,
[06:52] also think tanks, people who work in
[06:54] like regulation as well, so public
[06:57] service adjacent positions as well. I
[07:00] think that there's a broad
[07:02] like a very broad population of people
[07:04] who this might be useful for.
[07:05] >> And do you think there's acknowledgement
[07:07] of these ideas among people in those
[07:10] positions? And do you get a sense that
[07:12] people in the civil service or in public
[07:15] service more broadly, that they see that
[07:17] there is a deficit and that something
[07:19] like this could be helpful in filling
[07:20] that?
[07:21] >> Yeah, I think so. I mean, I don't have
[07:23] direct personal experience from that,
[07:24] but one of our colleagues, Dr. Jack
[07:26] Blumenau, who's like has been doing a
[07:27] lot of work with the cabinet office,
[07:29] he's told me that even people in there
[07:31] are conscious that they are lacking They
[07:34] have people who are data scientists who
[07:36] can do the evidence-based like the
[07:39] impact evaluations and so on, but they
[07:40] really lack the capacity actually
[07:42] understand what can be done with that
[07:44] with those methods. So, while there's
[07:46] been a huge development of doing impact
[07:48] evaluations all around, we are missing
[07:52] the link between those impact
[07:53] evaluations being made and then people
[07:56] actually being able to assess the like
[07:58] we're missing I mean that link needs to
[08:00] be strengthened so that we then go to
[08:02] like actual decisions.
[08:04] >> Yeah, so you've said that you're seeking
[08:06] to develop various skills and intuitions
[08:08] on the module. Do you want to go a bit
[08:10] further into the content? What are these
[08:11] skills and intuitions that are at the
[08:13] core of the module?
[08:14] >> Yes, so I structured the content around
[08:16] three different blocks. So, the three
[08:18] different types of inferences that we
[08:20] generally do with evidence. So, one
[08:22] would be measurement inference is trying
[08:25] to characterize what the things we
[08:27] observe might mean about the things we
[08:29] observe. So, the example that I often
[08:32] use is in the context of survey methods,
[08:34] we think about we ask people a series of
[08:37] questions about what their attitudes to
[08:39] certain things in society are and then
[08:41] based on their patterns of responses we
[08:43] infer that they are more or less
[08:44] right-wing. So, that would be trying to
[08:46] under like characterize what these
[08:48] people are and we do the same thing with
[08:50] you know, democracies. We observe
[08:52] democracies or like we observe countries
[08:54] doing a certain type of things,
[08:56] elections that they're free and fair,
[08:57] that there's a regular change of
[08:59] government and we infer that they are
[09:01] democracies and that measurement
[09:03] inference is really at the core of
[09:05] quantifying a lot of the things that we
[09:07] then use in subsequent analysis. So,
[09:09] that's the first type of inference.
[09:12] Then there is the causal inference. So,
[09:13] once we have measured certain things is
[09:15] like assessing the causal relationships
[09:17] between those two things and essentially
[09:19] it's all about the age-old adage at this
[09:22] point correlation is not causation and
[09:24] it's trying to give people an intuition
[09:26] when can we credibly claim that two
[09:29] things that go together go together
[09:31] because one causes the other and it's
[09:33] really structured around the idea that
[09:35] we need to think in terms of alternative
[09:38] universes to help us identify whether we
[09:41] are actually comparing things that are
[09:43] nearly the same except for the one thing
[09:45] that's different whose effect we want to
[09:47] assess.
[09:48] And then the third block is about
[09:50] population inference. So, most often we
[09:52] only have access to data from a slice of
[09:55] reality. We only have access to a sample
[09:58] of respondents. We only have access to a
[10:00] sample of
[10:02] participating local authorities in some
[10:03] kind of program and so on. And so, we
[10:06] try to generalize from what we observe
[10:08] on those smaller populations or the
[10:10] samples onto a broader population. So,
[10:12] what would happen if we
[10:14] applied this policy that we applied in
[10:17] the subset of local authorities to all
[10:18] of them and so on. Or what can we
[10:21] conclude from a sample of 2,000
[10:23] respondents saying that they would vote
[10:25] for a particular party in one election
[10:26] to the next one. So, those are the three
[10:28] inferences that I focus on and how I
[10:31] structure the course. And the really
[10:32] important point that I try to make from
[10:35] beginning on and throughout that the
[10:37] fact that you're doing inferences means
[10:39] that you're trying to make claims about
[10:40] something that we never directly
[10:42] observe. And regardless of how
[10:44] complicated and how complex and correct
[10:48] in the quotation marks or analysis seem
[10:51] is that we're always still making a leap
[10:53] of faith. We're always still trying to
[10:55] just
[10:56] say something about something we can
[10:58] never directly observe. And
[10:59] understanding that any analysis makes a
[11:02] little bit of that leap of faith, I
[11:04] think also helps.
[11:06] I think that that's what students really
[11:07] took away in some ways that they felt
[11:09] like, "Okay, so just because those
[11:11] people are like better able to do these
[11:13] statistical analysis doesn't necessarily
[11:15] mean that they're more correct. There's
[11:17] still is uncertainty for them as well."
[11:19] And basically structuring the course not
[11:21] on getting things correct, but getting
[11:23] people to a point where you say like,
[11:25] "My leap of faith is as credible as it
[11:28] gets. I try to do my best to to an
[11:30] educated guess about what actually
[11:33] something that might not say.
[11:34] >> Mhm.
[11:34] >> great. Thank you. Really good overview.
[11:36] Let's go into the first of those areas
[11:38] of inference. So, the measurement
[11:40] inference about what is the thing. So,
[11:42] as you were describing that there, I
[11:44] guess I was particularly thinking about
[11:46] survey research where we have questions
[11:48] that we put to people. And certainly
[11:50] something that I'm always thinking about
[11:51] whenever I look at a survey, I always
[11:53] want to see exactly what was the
[11:54] question. What was the question wording?
[11:56] And sometimes you're not given that and
[11:57] you're just kind of told kind of vaguely
[11:59] what was the question about. But
[12:01] actually the question wording can make
[12:03] such a big difference. So, is that the
[12:04] kind of thing that you're talking about
[12:06] when you're exploring that aspect of the
[12:08] overall module?
[12:09] >> Yes, absolutely. So, I because I use
[12:11] surveys quite a lot in my own research,
[12:13] I actually often use survey questions as
[12:15] examples of those. And I mean, I
[12:18] completely agree. I also have some
[12:20] research ongoing with one of our
[12:21] colleagues, Tom O'Grady, where we
[12:23] actually assess the effect of question
[12:24] wording. Maybe it might be useful to um
[12:27] say how we structure the seminars in
[12:29] this course. So, like me with the two
[12:31] PGTAs who joined me on this module, they
[12:33] helped me a lot. So, Joel Tasca and
[12:34] Marco Capelluti. One of them was a PhD
[12:37] student at Essex and Marco a PhD student
[12:38] in our department. They were really
[12:40] helpful at helping me design the
[12:42] seminar. So, basically our idea was that
[12:44] in the seminars, the students would take
[12:46] a piece of evidence. It generally
[12:48] wouldn't be a pure research paper, but
[12:50] actually something produced by a think
[12:52] tank or a research institute that was
[12:55] trying to produce research or a report
[12:58] whose audience is generally like the
[12:59] wider public. And the students in the
[13:02] module in the seminars, first like a
[13:04] group of students were asked to present
[13:05] it and explain what the measurement
[13:07] exactly was. And then the students would
[13:09] critically engage with all of the
[13:11] aspects of that evidence report. And in
[13:14] the measurement weeks, we had a lot of
[13:15] survey-based
[13:17] research. And it's exactly that where
[13:19] the students at the beginning
[13:21] you often write in the report is like,
[13:23] "People's attitudes to democracy is
[13:25] this." Or "People's attitudes is that.
[13:27] So, immediately it going rather than
[13:30] saying our measure of this thing, they
[13:32] say like this thing. They are not
[13:35] explicit about the fact that they have
[13:36] only a measure of the thing that they
[13:38] want to measure. And students really
[13:41] started looking into the how the
[13:43] questions were phrased and started
[13:44] thinking about like, "Oh, but does that
[13:46] really mean what the people who are
[13:48] reporting on it actually um say it
[13:50] means?"
[13:51] >> And actually this
[13:52] democracy, I mean you mentioned
[13:54] democracy as an example of that. This
[13:55] always seems to me to be one of the
[13:57] clearest examples where this is
[13:59] problematic, that people are often
[14:00] making claims about what people think
[14:02] about democracy on the basis of a
[14:04] question that is worded something like,
[14:07] you know, "Do you agree that we need a
[14:08] strong leader who is able to get things
[14:10] done?" And, you know, we in our world
[14:13] think of strong leader, "Aha, that means
[14:15] something alternative to democracy." But
[14:17] there's absolutely no reason to think
[14:18] that someone who's never come across
[14:19] that question, who's just got a survey
[14:21] question in front of them, is thinking
[14:22] of it in those terms.
[14:23] >> Yeah, exactly. And I think I often try
[14:26] to tell the students, especially when
[14:28] we're talking about surveys, which I
[14:30] often find really interesting, is like,
[14:32] "How easy do you find it to agree with
[14:34] that statement?" Because a lot of these
[14:36] statements are in this direction also,
[14:37] like trying to measure populism for
[14:39] instance, like questions such as like,
[14:41] "The will of the people is important,"
[14:42] which we associate in political science
[14:45] with adhering to populist ideology, but
[14:47] it's extremely easy for anyone to agree
[14:49] with that statement. So, it's trying to
[14:50] think in those terms. But what I
[14:53] also try to make the point is like
[14:55] there is no perfect measurement. So, I
[14:57] think students then start engaging
[14:58] critically with those things. And of
[15:00] course, we need to really think about
[15:02] like, "Where's the wording of a question
[15:04] or any kind of measurement tool we
[15:05] devise
[15:07] too leading and therefore not useful?"
[15:10] But ultimately, none of these questions
[15:12] will be able to capture the complex ways
[15:14] in which people think about democracy or
[15:16] relate to democracy on a day-to-day
[15:18] basis, which might not be really that
[15:21] present in their minds. And when we use
[15:23] survey questions, we try to get a survey
[15:24] question in order to distinguish between
[15:27] people. And I try to focus on the fact
[15:29] that
[15:30] ultimately, those survey questions are
[15:32] not to get at the right answer, because
[15:34] we don't know exactly what the right way
[15:36] of measuring people's attitudes to
[15:38] democracy is. What we want to do is
[15:39] devise measurements that are useful for
[15:42] our purposes. And the purpose often in
[15:44] survey research is to try to find out
[15:46] where do people disagree. Where's there
[15:49] variation in how people respond to those
[15:51] things? And yes, we might say like the
[15:53] problem with the questions are like, do
[15:54] you think a country needs a strong
[15:56] leader? That you might not get enough
[15:58] variation to be useful, because it's
[16:00] extremely easy to agree with. But it
[16:02] might help you to identify those who are
[16:05] very much in support of this kind of
[16:07] thing, or it might help you identify
[16:08] those who are very much against that.
[16:10] And then, ideally, you would ask also
[16:12] other types of questions that cover
[16:14] different aspects of these ideas. And
[16:16] none of them by themselves actually gets
[16:19] close to what we want to measure, but
[16:21] the combination of them might be useful
[16:23] enough for us to identify patterns and
[16:26] try to get a sense of like, where is
[16:29] there differences in the population in
[16:30] how they feel about it. And then,
[16:32] obviously, we then go on and think, does
[16:34] that also correlate with how they vote,
[16:37] or how they would react in the face of
[16:39] certain events, and so on. So, I try to
[16:41] get them at the same time to critically
[16:43] engage with those things, but also say
[16:45] like, at one point, all of these will be
[16:47] wrong. We need to think about like,
[16:50] where do we stop to try to make them
[16:51] useful enough?
[16:53] >> Yeah.
[16:53] >> For all the purposes.
[16:54] >> We could keep on talking about
[16:56] measurement and inference for the whole
[16:57] of our half hour, and you take 3 weeks
[17:00] and just focusing on that in the module.
[17:01] But I'm also really keen that we talk
[17:03] about causal inference. So, you said
[17:05] earlier that there's a kind of
[17:06] fundamental intuition to get there
[17:07] around correlation not being causation.
[17:11] Where further can we go in thinking
[17:13] about causal inference without going
[17:16] into the depth of the statistical
[17:18] analysis.
[17:19] >> So,
[17:21] yes, I think it's clarifying for a lot
[17:23] of people and like for a lot of the
[17:25] students like obviously we have methods
[17:26] in which
[17:27] develop methods such as where we use
[17:29] regression to assess those correlations
[17:31] and then we think about like what are
[17:33] the things we could control for to do it
[17:35] or do we design experiments in a
[17:37] particular way and so on. And I think
[17:39] what I tried to really focus the
[17:41] intuition on is that
[17:43] regardless of how you do it, ultimately
[17:45] to any causal statements regardless of
[17:47] how complex the method is, the
[17:49] underlying idea is really are you
[17:51] comparing
[17:52] like because obviously we cannot observe
[17:54] two different states of the world for
[17:56] any given person unless, you know,
[17:58] you're an author or a movie maker and so
[18:00] on and then you world is your oyster.
[18:03] And the idea is then to say that okay,
[18:05] everyone though implicitly makes those
[18:08] kinds of statements. So, when people
[18:10] said like during COVID for instance like
[18:12] oh, if our country only implement to
[18:15] these policies that this other country
[18:16] did, they're implicitly saying like
[18:18] those countries those two countries are
[18:20] the same and the only difference is the
[18:22] policy that they're implementing. But
[18:24] that obviously doesn't seem very
[18:26] intuitive because those countries are
[18:28] not the same. There many other things
[18:29] that are not the same that might have
[18:32] consequences on how effective those
[18:33] policies will actually be.
[18:35] So, I really try to focus on this like
[18:37] what is called the potential outcomes
[18:39] framework but without all of the maths
[18:41] and everything. Just really thinking
[18:43] like okay, ultimately for a causal
[18:45] statement to be truly credible, you need
[18:47] to say
[18:48] the thing that I'm comparing like the
[18:50] control group thinking in terms of like
[18:52] one group gets treated and one is like
[18:54] something doesn't happen need to be as
[18:57] comparable to each other as possible
[18:59] except for the difference in the status
[19:02] of being treated or getting some benefit
[19:05] or getting some policy or not and so on.
[19:08] And so, the idea then would start with
[19:10] randomized control trials because that
[19:12] always clarifies their intuition quite
[19:14] well is that if we randomly assign a
[19:16] treatment, obviously, the people who
[19:18] were assigned a treatment cannot choose
[19:20] whether they're treated or not. And
[19:22] therefore,
[19:23] in general, a randomized experiment will
[19:25] generally ensure that the treatment and
[19:28] control group or like different levels
[19:29] of treatment are comparable to each
[19:32] other. So, are essentially similar in
[19:34] every other respect other than the
[19:36] treatment.
[19:37] The problem is though, obviously, we
[19:39] cannot randomize everything and it would
[19:41] be unethical. Sometimes it's impractical
[19:44] and well, the really important bit
[19:46] really is the ethics of it. And
[19:48] obviously, as much as we social
[19:49] scientists would want that, I don't
[19:50] think it would be a great for the world
[19:53] if we randomized everything. And so,
[19:55] then we go on like what can we do when
[19:57] we don't have a
[19:58] if we don't have access to randomization
[20:00] as researchers, how can we take the data
[20:03] we observe in the world and try to find
[20:06] out where can we make good comparisons
[20:08] basically? And that generally involves
[20:10] trying to rather than
[20:13] doing complicated statistical models is
[20:15] actually trying to think about ways in
[20:17] which by chance certain things are more
[20:20] comparable than something else. So, we
[20:22] go through various different designs
[20:23] that are often used in impact
[20:24] evaluation. There's what's called
[20:26] difference in differences where the idea
[20:28] is that we compare two groups over time
[20:31] that yes, they are different to begin
[20:33] with, but because we just observed the
[20:34] trajectories, we can argue that like the
[20:37] trajectories would look the same had
[20:39] they not been treated. And obviously
[20:41] then someone who makes claims based on
[20:43] that will have to
[20:44] convince the reader that that's the
[20:46] case.
[20:46] >> Do you want to give an example of that
[20:47] just to help us get our heads around how
[20:50] that difference in differences design
[20:52] works?
[20:53] >> Yes, so I'll take the kind of classic
[20:55] example that Jack Blumenau and I also
[20:57] use in a
[20:58] module that's dedicated only to that.
[21:00] So,
[21:01] there's a paper from Ladin Lens from
[21:03] 2009, I think. And they looked at the
[21:07] effect of The Sun endorsing Tony Blair
[21:11] in 1997. And The Sun quite famously
[21:14] claimed that they had won it for Tony
[21:17] Blair because they had endorsed him,
[21:19] whereas they hadn't endorsed the Labour
[21:20] candidate in the previous election. And
[21:23] basically, what those people in that
[21:25] paper did is that they used survey data
[21:27] from the British Election Study, which
[21:29] has a panel of people. So, basically,
[21:31] the same people get interviewed at
[21:33] regular intervals, and their vote choice
[21:35] gets observed at these different
[21:37] elections. And they compared those
[21:39] respondents who are Sun voters who at
[21:41] one point they were asked, "What kind of
[21:43] newspapers do you read?" And a set of
[21:45] people said, "I read The Sun regularly."
[21:48] And then they compared the trajectory of
[21:50] vote choice for Labour among those
[21:52] people who said they were reading The
[21:54] Sun and those who didn't. And the idea
[21:57] is that they can compare what's the
[21:58] baseline difference already before The
[22:01] Sun changed their endorsement. So, when
[22:03] actually they weren't pro-Labour or like
[22:05] didn't say they were endorsing the
[22:06] Labour candidate in 1992. And then
[22:09] looked at how the rate of the Labour
[22:12] vote increased for The Sun voters
[22:14] relative to those who are non-Sun
[22:15] voters. And what they argue is that
[22:19] obviously, both of these vote shares
[22:20] increased because 1997 was a big
[22:23] landslide election for Labour, but it
[22:24] increased more for those Sun voters. And
[22:27] then what the authors tried to explain
[22:29] like tried to argue is that like nothing
[22:31] else except The Sun endorsement changed
[22:34] for that group of people who read The
[22:36] Sun. So, that's an example of a
[22:37] diff-in-diff. A more policy-related
[22:40] application is generally observing, for
[22:42] instance, unemployment outcomes in
[22:45] different local authorities, and then
[22:47] some kind of policy gets put in place,
[22:49] but only in some local authorities. And
[22:52] what we compare then the trajectories of
[22:54] unemployment, for instance, for those
[22:56] where there wasn't a policy put in
[22:58] place, and for those where it did. And
[23:00] then just whether the rate of change is
[23:02] different helps us get a sense of what
[23:05] the policies effect might have been. But
[23:07] again, while this is like technically
[23:09] quite clever, a big part of those papers
[23:12] or like those research reports like what
[23:14] they need to really argue and it's not
[23:15] that much a stats problem than an actual
[23:18] common sense problem is try to convince
[23:20] that
[23:21] really nothing else has changed other
[23:23] than the treatment for those local
[23:25] authorities where the policy was put in
[23:26] place. So for instance, if any other
[23:28] policy was put in place in those same
[23:30] local authorities, you cannot really
[23:32] argue that it's due to that policy
[23:34] happening.
[23:34] >> Shall we say a very quick word as well
[23:36] about population inference? So that's
[23:39] the third kind of leg of the module and
[23:41] we don't have terribly much time on
[23:43] this, but you want to just very quickly
[23:44] give us a sense of what are the key
[23:45] things that students learn?
[23:47] >> Yes, so I try to get them to send really
[23:50] this idea that often especially in
[23:52] survey data or something, we only make
[23:53] claims based on a small subset of
[23:55] observations and people say like, well,
[23:57] we find this in this
[23:59] you know, in this sample of respondents
[24:01] and that means that in the sample of
[24:03] respondents the vote share like in the
[24:05] next general election like what they
[24:06] would vote in general election for labor
[24:08] is so and so much percent. We then infer
[24:11] that this would be the same in the
[24:12] population because the sample itself and
[24:15] they argue then is representative of the
[24:17] population. So the module the population
[24:19] inference leg is really around the idea
[24:21] of representativeness because
[24:23] basically we take samples because we
[24:25] cannot observe everything. So we only
[24:28] observe a sample and we say that sample
[24:30] is representative, so therefore can
[24:32] stand in for the whole of the
[24:34] population.
[24:36] And what we focus on in here is first of
[24:38] all thinking about like are the people
[24:40] or are the units that are used in the
[24:42] sample representative of the broader
[24:43] population that the producers of that
[24:46] research say it is, right? So the
[24:49] students are encouraged to think about
[24:50] the differences between what the
[24:52] population might be and those people who
[24:53] end up in a sample, but we also think in
[24:56] terms of representativeness of for
[24:58] instance of how we measure those things
[25:02] on them, right? Like for instance, we
[25:03] can say like a survey question about
[25:05] democracy is not really representative
[25:07] also of like all of the ideas that
[25:10] someone might have in the head about
[25:11] democracy. So, I try to talk about
[25:14] representativeness not just in the pure
[25:15] statistical terms of like random
[25:17] sampling and therefore we have a sample
[25:19] that's representative, but also in a
[25:21] slightly more qualitative way as well.
[25:23] So, we also focus on how focus groups
[25:25] can be extremely representative of a
[25:27] wider range of views, whereas
[25:30] a sample of 2,000 respondents with a
[25:32] survey question that are like more
[25:33] contrived and therefore like they're
[25:34] going to be less representative in a
[25:36] slightly different way. So, I also try
[25:38] to make the case of also mixing the
[25:40] methods to get a different like looking
[25:42] at the problem from different way
[25:43] because we only observe one slice of
[25:46] reality.
[25:47] We need to think about how can we like
[25:49] pierce through that like veil in
[25:51] different ways and try to get a fuller
[25:53] picture of what might be actually
[25:55] happening in the public's mind.
[25:57] >> This is sounding terrific. I'm feeling
[25:58] very envious of all the students who get
[26:00] to take this module. How does it
[26:01] actually go in practice? How do you find
[26:04] students respond to it? Do they find
[26:05] things particularly challenging or
[26:08] eye-opening or empowering?
[26:11] >> Yes. So, I did as we mostly do in all of
[26:14] our modules the continuous module
[26:16] dialogue where throughout the term once
[26:17] at specific moments we ask them well, I
[26:20] do a survey
[26:22] where they can put in like what they
[26:23] feel like and I ask them what they feel
[26:25] like they've learned the most at that
[26:27] point in the module. And one thing that
[26:29] I always find and that I found great
[26:31] this year is like after a couple of
[26:33] weeks when I first talked about
[26:34] regression and tried to give them a
[26:36] sense of what really is regression,
[26:37] literally it's just drawing a line
[26:39] through a cloud of dots. I mean, that's
[26:41] what it is. And then someone wrote and
[26:44] when I asked like what's the most
[26:45] interesting thing you've learned and
[26:46] someone responded, "Regression and how
[26:48] it's actually not that scary." And I
[26:51] think that that was generally a theme
[26:52] where they felt like
[26:54] it felt great to be able to critique all
[26:56] those pieces of evidence that they read
[26:58] and I observed also the teaching of my
[27:00] PGTAs and in that I saw how the students
[27:03] said like, "Oh, here are the methods. I
[27:05] don't understand all of it, but here's
[27:07] what I take from it." And they still
[27:09] felt that they could critique and
[27:11] critique what those people had done
[27:13] because ultimately I asked them to focus
[27:15] what they should focus on is like, "What
[27:16] is the data that goes into those really
[27:19] complex calculations?" Because that's
[27:21] where all of these inferences really
[27:23] hinge on. And I think it's that
[27:25] empowering moment where first they read
[27:27] that and go like, "Well, these are
[27:28] methods I don't understand. Let me just
[27:30] like put that paper away and forget
[27:32] about it." And it's the moment where
[27:33] they say like, "No, actually I have
[27:35] things to say about this." And I think
[27:37] at one point in the module they got all
[27:39] really into the critiquing part and me
[27:41] and the TAs were like, "Okay, now we
[27:43] need to go from the critiquing part to
[27:45] like
[27:46] the useful enough kind of part of our
[27:48] philosophy of teaching there where it's
[27:50] really like, "Okay, all of these things
[27:52] are wrong. We still need to see what do
[27:54] we learn from that?" And I think yeah,
[27:56] and I think students at the end felt
[27:57] like
[27:58] you they wrote they wrote their
[28:00] assessment was choosing a piece of
[28:02] evidence like anything of their choice
[28:04] could be report from
[28:06] from a think tank or it could be report
[28:08] from a government and so on and then
[28:09] they critique that evidence and honestly
[28:11] it was great to see a large majority of
[28:13] the students really engage really well
[28:15] with it and a lot of them came to talk
[28:17] in office hours either one of us at one
[28:19] point about their ideas. And they had
[28:22] wonderful ideas and had really
[28:23] interesting things to say and I would
[28:25] say like it's their first step on
[28:27] actually like "Okay, I'm giving this
[28:29] evidence and writing down like what they
[28:31] think they can learn from that or what
[28:33] they think should be done better and so
[28:34] on." And I feel like a lot of them felt
[28:38] really scared like that's why they chose
[28:40] to take this module and not the
[28:41] introduction to quantitative methods
[28:42] where I have to do all the coding was
[28:44] like, "I'm not going to like any of
[28:45] this." And then in the end, they'd like
[28:47] covered so much ground, and yeah,
[28:49] overall students, at least those who
[28:52] talked to me and answered the surveys,
[28:54] were generally like the survey, were
[28:55] generally very positive about that
[28:57] feeling.
[28:58] >> Fan- fantastic. Of course, one might
[29:00] think there's a population inference
[29:01] there from the people who choose to
[29:02] answer the survey.
[29:03] >> That's always a problem.
[29:05] >> Yes. I guess we should also think a
[29:06] little bit about the limitations of this
[29:08] approach. And I suppose
[29:10] I'm maybe someone who is a bit like the
[29:13] students at the end of this module in
[29:15] that my stats training was many decades
[29:17] ago, and so it's all very rusty, and
[29:21] frankly quite a lot of the stats methods
[29:22] that are used now hadn't even been
[29:24] invented when I was being taught stats
[29:26] in the 1990s. And so, you know, I think
[29:29] I feel that I can read lots of
[29:32] stats-heavy papers and get a sense of
[29:34] what I should be getting out of them,
[29:35] but I think also, you know, if there are
[29:38] specific problems in how they are
[29:41] applying the stats in those papers, then
[29:44] I'm not going to notice notice that,
[29:45] frankly. Someone else has noticed that
[29:47] for me. Is that a limitation that we
[29:49] have to be aware of, and that students
[29:50] have to be aware of? Are there other
[29:52] aspects of limitations that you want the
[29:54] students to be aware of at the end of
[29:56] the module as well?
[29:57] >> Yes, of course, the students obviously
[30:00] will or if they ever become in come into
[30:02] these positions where they'll read loads
[30:04] of evidence, they also have to trust
[30:05] that the people who are doing the
[30:06] analyses are competent in doing it well.
[30:09] So, that also means that like the other
[30:11] type of quant methods training that we
[30:12] do, where we do more like actual coding,
[30:15] and I teach like a causal inference
[30:17] methods module that's more advanced. We
[30:19] have other colleagues teaching quant of
[30:20] text analysis. I teach an undergraduate
[30:22] module only about measurement inference,
[30:24] where like those people who do more data
[30:26] science, like they need to also be
[30:28] trained very well, but also trained with
[30:29] this kind of conceptual thinking as
[30:31] well, and not just the methods.
[30:33] I think another limitation of this is
[30:36] again something that I've talked a lot
[30:37] with my colleague Jack Blumenau, so I
[30:39] want to give a shout out to him, about
[30:41] the fact that there is a danger
[30:43] especially with evidence-based policy
[30:45] making that we will focus on things that
[30:47] we can measure and evaluate. And in some
[30:49] ways policy making requires creative
[30:51] thinking and making bold policies and
[30:54] changing people's lives will mean that
[30:57] you will again make a leap of faith
[30:59] terms of like we might not have
[31:02] evidence about whether this is going to
[31:03] work or not, but we want to take that
[31:06] step. And when you think about like the
[31:07] big advances of the last 20 years
[31:10] century when it comes to like the NHS or
[31:13] I mean the the New Deal in the US and so
[31:16] on. Like all of these like bold social
[31:18] policy who really transformed people's
[31:20] lives for the better,
[31:21] those were things where there was no
[31:23] evidence to know whether it was going to
[31:24] work or not. And there's a little bit
[31:26] the danger that focusing too much on
[31:27] these like small like not small, they're
[31:30] big and when in the accumulation, but
[31:32] like focusing on small steps and small
[31:34] policies and testing all of that. There
[31:36] are good things to it, but it also is
[31:38] the danger that therefore also policy
[31:40] makers become themselves a little bit
[31:42] too cautious in what they try. And I try
[31:45] to talk about that in the last part of
[31:46] the module to say there's limitations to
[31:48] it. First of all, obviously politicians
[31:50] or policy makers might not listen to
[31:52] that because they have their own
[31:53] ideological agenda, but that's not
[31:55] necessarily always a bad thing.
[31:57] And I think I trying to make them aware
[31:59] that
[32:00] understanding better evidence still
[32:02] means that there is a leap of faith and
[32:04] that doesn't mean that we shouldn't try
[32:06] to make the leap of faith in those
[32:07] circumstances where we can't close the
[32:09] gap that well between what we think it
[32:12] might do and like the evidence that we
[32:14] have and the thing we might want to do
[32:17] if that makes sense. So I think that
[32:18] that really is a limitation I tried to
[32:21] bring home. And and I guess that that
[32:22] would be dedicated a whole other module
[32:24] or a whole If I had more weeks, I would
[32:26] talk about like doing decisions under
[32:28] uncertainty and also how that relates to
[32:31] people's risk preferences and values and
[32:33] norms and so on. But unfortunately, we
[32:35] only have 10 weeks.
[32:36] >> So all of those are the constraint.
[32:37] Final question, have you learned
[32:39] anything yourself from teaching the
[32:41] module?
[32:43] >> Yes, what have I learned? First, I mean,
[32:45] one thing that I really enjoyed is the
[32:47] thinking about also the seminars. Like,
[32:49] a lot of the modules I teach involve
[32:51] coding, so the seminars are kind of
[32:53] standard in
[32:54] not standard. I mean, I find it
[32:55] exciting, but it's they get a data set,
[32:58] they get some questions, they have to
[33:00] code it in front of the computer. And
[33:02] this one was really one where there was
[33:04] much more discussion. And sure, I've
[33:05] taught like a substantive module on
[33:07] welfare states before where there is
[33:08] discussion, but not on methods as much.
[33:10] And I really liked also involving the
[33:12] PGTAs who like came both from different
[33:15] they they have their own different
[33:16] experience, and they were the ones like
[33:18] we together found the materials to talk
[33:20] about in the seminars. And what I
[33:22] realized is how great it was that it's
[33:24] not only me choosing those seminar tasks
[33:26] to discuss because it meant that they
[33:28] had their own ideas, and they found like
[33:29] very creative things. And the students,
[33:32] one thing they really praised is like
[33:33] the materials that we had chosen for the
[33:35] seminars because they covered a lot of
[33:36] ground, and that is something that me on
[33:39] my own like would not have been able to
[33:41] give that set of diverse ideas. So, like
[33:44] the collective effort of us three really
[33:46] helped there.
[33:48] Then, I think I learned I mean,
[33:49] obviously this was the first time of
[33:51] like really writing lectures from
[33:52] scratch for that is like learning how to
[33:55] do the pacing. And also, I was always
[33:58] quite ambitious in all of the material
[34:00] that I could cover that I wanted to
[34:01] cover, and so I had to sometimes
[34:03] moderate a little bit, but I realized
[34:05] that if you really focus on those
[34:08] intuitions, like students get more
[34:10] excited about it than if you go into the
[34:12] details of the maths, which I personally
[34:14] obviously find exciting, but understand
[34:16] that others do not. Yeah, and I feel
[34:18] like obviously it's very empowering for
[34:20] myself also like
[34:22] having done really a complete module
[34:24] from scratch and feeling that you have
[34:26] yourself empowered people, which I do
[34:28] feel with the other modules, but like
[34:30] the students were more vocal about how
[34:31] they said that. I mean, one student in
[34:34] one of those CMD things said like it's a
[34:36] it's crazy how
[34:37] >> module dialogues. Yeah, the process by
[34:39] which we assess how the module's going.
[34:41] Yeah.
[34:41] >> Yes, so in that where I asked the
[34:43] students how things were going, one
[34:45] comment was that I can't believe that
[34:46] this module hasn't existed before, which
[34:49] obviously is very flattering, but also
[34:51] makes me very proud that like it feels
[34:53] like it was actually filling a need that
[34:55] was there. And I had been thinking about
[34:57] this for a while, but it took me a while
[34:59] to go like, "Okay, I'm going to do it."
[35:00] It's like that gut a little bit that I
[35:02] have. I think that that that is
[35:04] something I learned as well.
[35:05] >> Well, it sounds like you're doing really
[35:06] fantastic work on this module. As I
[35:08] said, I'm feeling very envious of all
[35:10] the students taking it. You've given us
[35:11] a wonderful little taster over the last
[35:13] few minutes, but it would be a great to
[35:15] hear more. So, thank you so much, Julia.
[35:17] Normally at this point, I give you the
[35:19] details of the publication that we've
[35:20] been discussing. Clearly, I can't do
[35:22] that this time cuz we've been discussing
[35:23] a module instead, but I can say that if
[35:25] you enroll on one of our master's
[35:27] courses here in the UCL Department of
[35:29] Political Science and School of Public
[35:30] Policy, then you too can opt to take
[35:33] Julia's module in full, and then you
[35:35] will thereby gain from all of her
[35:37] wonderful insights and energy. Well
[35:40] worth doing. We will be back next week
[35:42] when we'll be looking at the tradition
[35:44] of political parades in Northern
[35:47] Ireland.
[35:48] To make sure you don't miss out on that
[35:49] or other future episodes of UCL
[35:51] Uncovering Politics, all you need to do
[35:53] is subscribe. You can do so on Apple,
[35:55] Google Podcasts, or whatever podcast
[35:57] provider you use. And while you're
[35:59] there, we'd love it if you could take a
[36:00] moment of time to rate or review us,
[36:03] too.
[36:04] I'm Alan Renwick. This episode was
[36:06] produced by Matthew Dean. Our theme
[36:08] music is written and performed by John
[36:09] Man. This has been UCL Uncovering
[36:12] Politics. Thank you for listening.

Introduction to the Podcast

Hello and welcome to UCL Uncovering Politics, a podcast from the School of Public Policy and the Department of Political Science at University College London. In this episode, host Alan Renick introduces the topic of electoral politics concerning housing, posing the question of whether homeownership or renting influences voting behavior and how this relationship may be evolving over time.

Historical Context of Class and Voting

Alan Renick begins by discussing traditional views on voting behavior, which historically linked voting patterns to occupational class. The prevailing notion was that individuals in manual working-class jobs tended to vote for left-leaning parties, while those in white-collar, middle-class occupations leaned towards right-leaning parties. However, over recent decades, this class-based model has faced challenges as other factors, such as education and age, have emerged as stronger predictors of voting behavior. This shift indicates a new political divide centered around cultural and identity issues rather than purely economic concerns.

Despite these changes, a recent study suggests that economic status continues to play a significant role in shaping voting patterns, albeit in a different manner than in the past. The study, conducted by Josh Godard, a PhD student at UCL, posits that housing has become a more critical determinant of economic status and, consequently, voting behavior.

Josh Godard's Research Background

Josh Godard elaborates on the traditional understanding of class within political science, which has primarily focused on labor market positions, particularly occupation, income, and employment status. He notes that while earlier literature emphasized a binary distinction between working-class and middle-class occupations, recent studies have identified more nuanced divides within these groups. For instance, within the middle class, there is a distinction between managerial roles and sociocultural professionals, with the latter increasingly leaning towards left-leaning parties.

Godard's research aims to extend this analysis by suggesting that to understand contemporary social stratification, one must look beyond the labor market to include housing as a significant factor. He argues that the relationship between class and voting behavior is evolving, necessitating a re-evaluation of how we measure class in the context of modern societies.

Defining Social Class

When asked to define social class, Godard explains that it encompasses an individual's economic position within society, which includes their level of affluence and security. He acknowledges that historically, occupation was a primary determinant of economic status, particularly in the UK, where class has been a significant factor in political behavior. However, he notes that similar patterns can be observed across various democracies, with occupational class still influencing voting behavior, albeit to a lesser extent than in the past.

The Importance of Housing in Class Structure

Godard highlights several transformations in housing markets over the past few decades that have impacted different social groups unevenly. He identifies three key trends: first, significant house price inflation relative to income; second, the financialization of housing, leading to increased mortgage debt; and third, changes in rental markets characterized by a reduction in social housing and increased insecurity in private rentals. These trends have contributed to growing inequalities and have made housing a crucial determinant of class position and economic security.

For example, an individual with a low income in a historically working-class occupation may own a home purchased at a lower price in the past, thereby enjoying economic security. In contrast, a younger individual with a higher income but renting may face greater economic insecurity due to the rising costs of homeownership.

Disconnect Between Labor Market and Housing Market Positions

Godard emphasizes the disconnect between traditional measures of class based on labor market positions and the growing importance of housing market positions. He argues that understanding this disconnect is essential for reconciling the evolving nature of class and voting behavior. His research tests the hypothesis that housing market position significantly influences voting behavior and that this relationship has intensified over time.

Hypotheses and Expectations

In his study, Godard outlines several hypotheses regarding the relationship between housing market positions and voting behavior. He draws on existing literature that suggests housing circumstances influence political attitudes along two primary dimensions: economic policy attitudes (left-right dimension) and pro- versus anti-system sentiments. Homeowners, particularly those with high-value properties, tend to hold right-leaning views on economic policies, while renters often exhibit left-leaning views and higher levels of political disaffection.

Godard anticipates that center-right parties will attract more support from homeowners, while radical left parties will mobilize renters due to their shared left-leaning economic positions and populist rhetoric. He also expects that the relationship between housing market positions and voting behavior will have strengthened over time, reflecting the increasing importance of housing in determining class status.

Methodology of the Study

Godard discusses the methodological challenges he faced in conducting his research, particularly the lack of consistent data on housing tenure in cross-national surveys. To address this, he created two datasets: one to measure the relationship between homeowners and renters and another to explore finer distinctions within these groups. He utilized national election studies from 14 countries over the past 40 years and household panel data from six countries to analyze these relationships.

The analysis primarily employs logistic regression to assess the relationship between housing tenure and voting behavior, considering both bivariate and multivariate settings to control for other demographic factors that may influence voting patterns.

Key Findings of the Research

Godard's findings indicate that housing tenure significantly influences voting behavior. Homeowners are more likely to support center-right parties, while renters tend to favor radical left parties. Furthermore, he finds that this relationship has grown stronger over time, with housing tenure becoming a critical factor in explaining support for these party families. When comparing housing tenure to other demographic factors such as income, education, and age, housing tenure emerges as the most important predictor of voting behavior for both center-right and radical left parties.

In examining subcategories within housing tenure, Godard confirms his expectations that high-value homeowners and outright owners show strong support for center-right parties. Conversely, low-value homeowners exhibit a tendency to support radical left parties. Among renters, social renters are more likely to support radical left parties compared to private renters, who show a slightly higher inclination towards green parties.

Implications for UK Politics

When discussing the implications of his findings for UK politics, Godard notes that the relationship between housing and voting behavior has historically been significant in Britain. Despite expectations that this relationship would have intensified due to rising housing unaffordability, the data suggests that housing has always been a crucial predictor of voting behavior in the UK. This enduring significance may reflect the historical context of housing as a class marker in British society.

Future Research Directions

Looking ahead, Godard aims to expand his research beyond class voting to explore broader aspects of class politics. He intends to investigate how attitudes towards economic policies and welfare have evolved over time and whether political parties have adapted their platforms to reflect changing class dynamics. This future analysis may provide insights into the causal mechanisms underlying the observed relationships between housing tenure and voting behavior.

Conclusion

In summary, the podcast highlights the evolving relationship between housing and electoral behavior, emphasizing the importance of housing as a determinant of class position and voting behavior. Godard's research contributes to a deeper understanding of how economic factors, particularly housing, shape political affiliations and voting patterns in contemporary societies. The findings underscore the need for political scientists to integrate housing into discussions of class and voting behavior to reflect the complexities of modern electoral dynamics.

According to the transcript, the speaker, Alan Renick, discusses the evolving factors that influence voter behavior, particularly emphasizing the role of economic status and how it is increasingly shaped by housing rather than traditional measures such as occupation. The speaker outlines that in the past, one could predict voting behavior primarily based on class, specifically whether individuals belonged to manual working-class or non-manual middle-class jobs. However, recent trends show that education and age have become more significant predictors of voting behavior.

The discussion highlights a shift in the political landscape, where the relationship between class and voting is becoming less pronounced, giving way to the influence of cultural and identity issues, such as gender, immigration, and morality. Despite this shift, the speaker asserts that economic status continues to be a key predictor of voting behavior, albeit determined through housing status rather than occupational status. The speaker mentions the work of Josh Godard, who argues that housing has become a more crucial determinant of class position and voting patterns in contemporary society. This implies that individuals who may have low incomes or insecure occupations can still have significant economic security if they own their homes, contrasting with younger individuals who may earn more but lack housing security.

Overall, the transcript suggests that voters often do not simply make rational choices based on their material interests but are influenced by a complex interplay of factors including their economic situation, their social identity, and the cultural environment surrounding them.

• [01:22] "Does all of this mean that economic status no longer matters in shaping voting patterns? While a new study argues that it does not, it finds that such status remains a key predictor of voting behavior."
• [01:14] "Indeed, today in many countries, education and age are better predictors of voting behavior than is occupation."
• [10:29] "Housing is a very strong factor in people's economic prospects, economic fortunes, economic status."

The transcript does not directly address how voter information or election debate formats affect the quality of voters' choices. There is an emphasis on the changing dynamics of voter behavior, particularly the influence of housing on voting patterns, but specific details regarding how information or debate formats shape voters' decision-making processes are not mentioned.

However, the conversation implies that the complexities of class and economic status may lead to varied interpretations of information among voters. As voters navigate their housing situations and their implications for economic security, their decision-making could be influenced by the quality of information they receive, though this is not explicitly stated in the transcript. The speaker focuses more on the relationship between housing, economic status, and voting behavior rather than the mechanisms of information dissemination or debate formats.

The transcript does not specifically address the role of moral convictions or economic circumstances in driving ideological disagreement or far-right leanings. The discussion primarily centers around the shifts in voter behavior based on changes in class, housing status, and economic security. Although the speaker notes the emergence of new cultural and identity issues that influence voting behavior, the transcript does not provide explicit insights into how moral convictions or economic factors contribute to ideological divides or support for far-right parties.

Instead, the focus is on the changing nature of class as it relates to voting, particularly highlighting how housing has become a critical determinant of class status and voting behavior. While it is clear that economic circumstances play a role in shaping voters' perceptions and preferences, the transcript lacks a detailed exploration of the interplay between moral beliefs and economic factors in relation to far-right ideologies.

The transcript does not provide explicit recommendations for how citizens can better navigate misinformation or evaluate political promises. However, it implies the necessity of understanding the evolving relationship between housing, economic status, and voting behavior. The speaker suggests that voters need to recognize the importance of their housing situation in relation to their economic security, which in turn shapes their political preferences. By being aware of these dynamics, citizens may be better equipped to assess political promises and the motivations behind them.

Furthermore, the discussion reflects on how traditional measures of class and voting behavior are shifting, indicating that voters should critically evaluate the information they receive regarding housing policies and economic measures proposed by political candidates. Although the transcript does not detail specific strategies for navigating misinformation, the emphasis on the importance of housing suggests that a deeper understanding of economic and housing issues may help voters make more informed decisions.

The transcript does not explicitly mention how voters judge governments retrospectively or based on economic outcomes, nor does it discuss the implications for fair accountability. The focus of the conversation is primarily on the relationship between housing status and voting behavior. However, it can be inferred that since housing has become a significant indicator of economic security and class position, voters may indeed assess governments based on their effectiveness in addressing housing issues and economic inequalities.

This retrospective judgment could imply an expectation of accountability on the part of elected officials to provide solutions to the housing crisis and to ensure economic stability for their constituents. The lack of direct commentary on retrospective judgment in the transcript suggests that while it is an important topic, it was not a primary focus of the discussion.

[00:05] Hello, this is UCL Uncovering Politics
[00:07] and this week we're looking at the
[00:08] electoral politics of housing. Does
[00:11] whether you own or rent your home affect
[00:14] how you vote? And is this relationship
[00:17] changing over time?
[00:24] Hello, my name is Alan Renick and
[00:26] welcome to UCL Uncovering Politics, the
[00:28] podcast of the School of Public Policy
[00:30] and Department of Political Science at
[00:32] University College London. It used to be
[00:35] said that if you wanted to predict how a
[00:37] person would vote, the most important
[00:39] aspect of their lives to look at was
[00:41] their job. People in manual workingclass
[00:44] occupations tended to vote left. People
[00:46] in white-collar middle-class occupations
[00:49] tended to vote right. Recent decades
[00:52] have seen such class-based models of
[00:55] voting increasingly challenged. Indeed,
[00:57] today in many countries, education and
[01:00] age are better predictors of voting
[01:02] behavior than is occupation. And the
[01:05] rise of factors such as education and
[01:07] age is often associated with the
[01:09] emergence of a new dividing line in
[01:11] politics based not around questions of
[01:14] economic policy, but rather around
[01:16] cultural and identity issues such as
[01:19] gender, immigration, and morality. But
[01:22] does all of this mean that economic
[01:24] status no longer matters in shaping
[01:27] voting patterns? While a new study
[01:29] argues that it does not, it finds that
[01:32] such status remains a key predictor of
[01:35] voting behavior. But what determines
[01:37] economic status has shifted. Whereas in
[01:40] the past it was decided mainly by
[01:43] occupation, now that role is performed
[01:45] to a much greater degree than before by
[01:48] housing. The author of that study is
[01:51] Josh Godard, a PhD student here in the
[01:53] UCL department of political science. And
[01:56] I'm delighted that Josh joins me now.
[01:58] Josh, welcome to UCL Uncovering
[02:01] Politics. And let's start with the
[02:03] background to your research. What have
[02:06] we traditionally known? I gave a bit of
[02:07] a hint of it in in the introduction
[02:09] there, but really be useful to have it
[02:11] in your own words. What have we
[02:12] traditionally known or thought that we
[02:14] have known about the role of class in
[02:17] shaping people's voting behavior?
[02:19] >> Great. Thanks, Alan. Yeah. So as you
[02:21] alluded to there's a very big literature
[02:24] on the relationship between class and
[02:26] voting in political science and the kind
[02:29] of traditional
[02:31] the traditionally this literature has
[02:34] conceptualized your class position in
[02:37] terms of your position within the labor
[02:38] market. So whether that be primarily
[02:40] your occupation or also potentially your
[02:42] income or your employment status. But
[02:45] yeah, the the kind of the most
[02:47] significant section of this literature
[02:48] has looked at occupation. And the
[02:50] traditional literature measured this in
[02:51] terms of a binary distinction between
[02:54] people with manual jobs, working-class
[02:56] jobs, and people with non-manual white
[02:59] collar middle-ass jobs. And as you
[03:02] identify showed that there's been a
[03:03] decline in the relationship between
[03:05] membership of those groups and electoral
[03:08] behavior. Now more recently a kind of
[03:10] strand of literature emerged which
[03:13] challenged this measurement of class or
[03:17] occupation structures and suggested that
[03:20] within these two groups so within these
[03:21] two broadly defined middle class and
[03:23] workingclass groups there is there are
[03:26] other significant divides. So and this
[03:28] literature is focused mainly on the
[03:30] middle class saab and shown that within
[03:33] kind of white collar occupations we can
[03:35] see a divide between people in kind of
[03:37] managerial positions and people in more
[03:40] kind of interpersonal professional
[03:41] positions and this is commonly these are
[03:43] these roles are commonly called
[03:44] socioultural professionals in the
[03:46] literature and managers tend to remain
[03:48] more likely to support parties of the
[03:50] centerides in particularly and
[03:52] socioultural professionals have become
[03:55] more or are now more likely to support
[03:57] left-leaning parties and I guess the
[04:00] literature has also interacted with
[04:02] changing party structure party systems.
[04:04] So the increasing success of radical
[04:07] right parties and has identified that
[04:09] they have become they have more support
[04:11] within the working-class groups and also
[04:14] green and left libertarian parties and
[04:15] showing that they've become more popular
[04:17] amongst the kind of these socioultural
[04:18] professionals. But even that kind of new
[04:22] literature which critiqued the
[04:23] traditional literature retains a focus
[04:25] on the labor market and occupation. So
[04:28] it says this traditional binary
[04:30] operationalization of labor markets is
[04:33] appropriate maybe for kind of industrial
[04:36] post-war societies but now labor markets
[04:39] are far more complicated in a kind of
[04:41] the kind of post-industrial context. So
[04:43] we should look at them and measure them
[04:45] in different ways. My study looks to
[04:48] kind of take this I guess a step further
[04:50] and say that to understand contemporary
[04:52] patterns of inequality or I guess what
[04:55] sociologists would call social
[04:57] stratification we we might need to look
[04:59] beyond the labor market.
[05:00] >> That's really helpful and really
[05:01] interesting. Just on that last point
[05:03] that you were getting to there, what
[05:04] actually is social class? What do we
[05:07] mean by social class? And why is it that
[05:09] traditionally that has been linked so
[05:11] strongly to labor market position
[05:14] specifically? Well, I think there's
[05:16] different ways to think about it. So,
[05:19] what I take it to mean in the paper, I
[05:22] think, is your kind of broader or your
[05:24] kind of position within the or your kind
[05:28] of economic position within society. So,
[05:31] how affluent or how secure you are or
[05:33] how poor or insecure you are in society.
[05:36] And I think there has also been
[05:39] traditionally this has interacted with a
[05:42] sense of identity. So I think maybe that
[05:44] is where the traditional focus was on
[05:46] occupation. But also I think that I
[05:49] think I do think the traditional focus
[05:50] on occupation is was justified in so far
[05:52] as your job really was the main
[05:54] determinant of your level of affluence
[05:57] of your economic position if you like.
[05:59] Um so and does that was that the case
[06:02] everywhere? I guess I'm I'm coming from
[06:04] a perspective of mainly thinking about
[06:05] UK politics and it's often been said
[06:08] that class is particularly important in
[06:10] UK politics or historically has been
[06:12] particularly important in UK politics.
[06:14] But do we find similar sorts of patterns
[06:16] across democracies? Yes. So we certainly
[06:19] see in terms of the relationship with
[06:21] voting, we we certainly see a stronger
[06:24] relationship between your occupational
[06:26] class and voting across my paper looks
[06:30] mainly at kind of or looks at
[06:33] post-industrial western societies and
[06:35] yeah that is the pattern we
[06:36] traditionally see.
[06:37] >> Okay. Okay. So why is it that you think
[06:40] that we should also look at housing? You
[06:42] were beginning to hint at that when you
[06:43] were suggesting that traditionally your
[06:46] job has been the main factor that shapes
[06:47] your economic fortunes, but I guess you
[06:49] were suggesting that that has changed in
[06:51] recent decades.
[06:52] >> Yeah. So, I think there's been a few
[06:54] kind of important transformations within
[06:57] housing markets over the past probably
[06:59] three or four decades again in
[07:00] post-industrial societies
[07:03] which have had uneven impacts on
[07:06] different social groups. So some groups
[07:08] they've really helped, they've really
[07:10] advantaged and some groups they're
[07:11] disadvantaged and that therefore kind of
[07:15] h housing has become a significantly
[07:17] more important kind of driver of
[07:19] inequality or determinant of one's kind
[07:22] of class position in their societies.
[07:24] Now I identify three of specific
[07:27] transformations in housing markets and
[07:29] they are house price inflation. Pretty
[07:31] much I show in the paper that house
[07:32] prices have grown massively relative to
[07:34] income across advanced democracies over
[07:36] the past three or four decades. Also
[07:39] housing financialization. So the massive
[07:42] expansion of housing debt and then also
[07:45] mortgages. So people have much bigger
[07:47] mortgages now than most.
[07:48] >> Yeah. So people are like or or a
[07:50] significant group of people are heavily
[07:52] exposed to or are exposed to
[07:54] significantly higher levels of debt in
[07:57] the in mortgage debt now. And then the
[07:59] third thing was the transformation of
[08:01] rental markets. So we've seen kind of a
[08:03] shrinking of social housing sectors and
[08:05] also the private rental sectors or
[08:08] private rental sectors have generally
[08:09] generally become more deregulated and
[08:12] insecure over time. And I argue in the
[08:15] paper that these three trends have
[08:18] interacted to drive significant
[08:21] inequalities between or increase
[08:23] significant inequalities between housing
[08:25] market groups
[08:27] and such that now yeah your kind of your
[08:30] level of economic security or your
[08:32] affluence or if you like I guess a
[08:34] sociologist would say your position in
[08:35] the class structure is increasingly
[08:38] determined by your housing market
[08:40] position and that has become more
[08:42] important over time relative to your
[08:44] labor. a market position.
[08:45] >> So that could mean, for example, you're
[08:46] on a person, you're a person who's on a
[08:48] fairly low income, perhaps in a an
[08:51] occupation that would have been regarded
[08:54] as working-class occupation
[08:55] historically, but you own your own house
[08:58] because maybe you're a little bit older.
[08:59] You got into the housing market at an
[09:01] earlier stage. And therefore, you're
[09:03] suggesting actually you're economically
[09:06] quite secure. And you might contrast
[09:08] that with someone who's younger,
[09:10] potentially on a higher income,
[09:11] potentially in a more kind of high
[09:13] status occupation, again speaking in
[09:16] kind of traditional terms, but who
[09:18] doesn't own their own home, maybe
[09:20] they're renting and therefore it faces
[09:22] much greater insecurity economy.
[09:24] >> Yeah, I think that's right. And I think
[09:26] yeah, you're right to identify the kind
[09:27] of time-sp specific nature of this
[09:29] change in order to explain these.
[09:31] Um I think yeah what you've identified
[09:33] there is a growth in groups whose kind
[09:35] of labor market positioning is
[09:36] inconsistent if you like with their
[09:38] housing market position and yeah as you
[09:40] identify there's a significant group in
[09:42] societies now who have maybe lower
[09:44] incomes or have never had particularly
[09:45] high incomes but because they bought
[09:46] house at an earlier point when they were
[09:48] affordable in a in particularly in an
[09:51] area of an area of their country which
[09:53] is showing significant house price
[09:55] growth over recent times. They're now in
[09:57] relatively affluent housing market
[09:59] positions. But if we just looked at
[10:00] their labor market position, they they
[10:02] show up as being relatively insecure.
[10:04] And equally, yes, kind of younger
[10:06] people, maybe professionals often in in
[10:09] cities might earn quite good income, but
[10:11] increasingly find it difficult to obtain
[10:13] housing security or enter home
[10:15] ownership. So housing is a very strong
[10:18] factor in people's economic prospects,
[10:20] economic fortunes, economic status. And
[10:23] then you're you're arguing that we
[10:25] should expect that that would have an
[10:27] impact upon voting behavior. Yeah,
[10:29] that's right. So I I I I kind of
[10:31] identify there's a kind of disconnect
[10:33] between these two things we've been
[10:35] talking about. So there's a disconnect
[10:36] between on the one hand the literature
[10:39] on class voting measures a class in
[10:42] terms of or overwhelmingly in terms of
[10:44] your labor market position whereas we've
[10:46] actually seen the growth of housing as a
[10:48] determin of class. And I kind of want to
[10:50] in a way begin to reconcile this kind of
[10:53] contradiction and integrate housing into
[10:56] these kind of broader debates about the
[10:58] shifting kind of social basis of
[11:01] politics, this shifting relationship
[11:02] between demographics and class and
[11:04] voting behavior. So yeah, I test the
[11:08] theory that because it's become more
[11:10] important as a determinant of your class
[11:12] position or your level of economic
[11:13] security or affluence, we might expect
[11:15] your there to be a significant
[11:17] relationship between housing market
[11:19] position and voting behavior and for
[11:21] that this will have grown over time.
[11:24] >> Yeah. So one of the hypotheses that you
[11:26] set out at the start of the paper is is
[11:28] simply that we should expect to see some
[11:30] kind of relationship and that we should
[11:32] and then another of the hypothesis is
[11:33] that we should expect that relationship
[11:35] to have increased over time but you also
[11:39] then go on to hypothesize specific
[11:41] relationships that you're expecting to
[11:43] find.
[11:44] >> Yeah, absolutely. So and to kind of come
[11:46] up with these expectations I rely on the
[11:48] on the literature which we always
[11:50] already have a significant literature on
[11:52] housing and politics. I just don't think
[11:54] it I don't think it's yet being
[11:56] integrated into these broader debates.
[11:58] But from this literature, we we know
[12:01] that your housing market position or
[12:04] your housing circumstances influences
[12:06] people's political attitudes and
[12:08] primarily it does this along to what I
[12:11] identify as attitudinal dimensions in
[12:13] the paper. So the first is on attitudes
[12:17] towards economic or welfare policy and
[12:19] we might kind of conventionally refer to
[12:22] this as the kind of traditional
[12:23] leftright dimension and typically people
[12:26] in more affluent housing market
[12:27] positions. So homeowners but especially
[12:30] high value homeowners, outright owners
[12:32] and landlords as well. They
[12:34] traditionally have more rightle leaning
[12:37] views on this dimension with those in
[12:40] more insecure housing market positions.
[12:41] So renters having more lefting views.
[12:44] There's also a second attitudal
[12:45] dimension which is of importance and I
[12:47] refer to this as the pro versus
[12:49] anti-system dimension and it captures
[12:52] levels of political disaection.
[12:55] So I show that or the literature had
[12:57] traditionally had shown that those in
[12:59] more affluent housing market positions
[13:00] again homeowners especially high value
[13:02] homeowners etc they tend to have more
[13:04] pro-system positions whereas renters
[13:07] tend to have higher levels of
[13:08] disaection. So I think we can think of
[13:11] these different housing market groups as
[13:13] mapped out on these two dimensions. And
[13:16] we can also think of the in terms in
[13:19] order to generate expectations about how
[13:21] they vote. We can also think of
[13:24] political parties or party families as
[13:26] occupying spaces on these dimensions as
[13:29] well. And those that are closer to the
[13:30] housing market positions to the housing
[13:32] market groups, you might expect to to
[13:34] attract greater support from those
[13:36] housing market groups. So for instance,
[13:39] center right parties I hypothesize
[13:42] appear particularly well suited to
[13:44] attracting disproportionate support from
[13:46] those in more affluent housing market
[13:48] groups. So from home owners because they
[13:51] tend to be right right leaning on
[13:53] economic policy and also as kind of
[13:55] longstanding
[13:57] parties of government they seem well
[13:59] suited to attract pro-system voters. On
[14:02] the other side, in in terms of renters,
[14:05] I I theorize that it will be radical
[14:07] left parties that are best placed to
[14:08] disproportionately mobilize them. And
[14:10] that's because again, they have the most
[14:12] left-leaning positions of the of the
[14:14] party families, but also often display
[14:17] populist rhetoric and are well placed to
[14:19] appeal to anti-system voters. As I as I
[14:23] said earlier, renters also tend to
[14:24] exhibit anti-system attitudes as well.
[14:27] >> Yeah. So you have kind of strong
[14:29] expectations for the relationship
[14:31] between the housing market position and
[14:33] support for those two families of
[14:34] parties. So centerright parties and
[14:37] radical left parties. By contrast, the
[14:39] other various types of parties that that
[14:42] they just don't quite match so closely
[14:44] any of these housing market groups. So
[14:47] you're expecting that they they support
[14:49] for them would be less affected by
[14:51] housing market position.
[14:52] >> Yes. So exactly. So I I don't expect the
[14:54] other party families to kind of
[14:56] consistently attract disproportionate or
[14:59] or significantly disproportionate
[15:01] support from any housing market.
[15:03] >> Yeah. So just to give an example of that
[15:05] social democratic parties you see as
[15:07] being on the left in terms of economics
[15:09] but actually being quite kind of
[15:11] pro-system being traditional parties
[15:13] that have been operating within the
[15:14] system for for many decades. And because
[15:17] that doesn't map on to any of the
[15:19] housing groups very closely, you just
[15:21] think that support for social democratic
[15:23] groups would be mainly predicted by
[15:25] other factions.
[15:26] >> Yeah, exactly. And also in the kind of
[15:28] mirror of that, we could think about
[15:30] radical right parties as being as
[15:32] exhibiting kind of anti-system positions
[15:34] but also not being not being leftwing on
[15:36] economic policy. So not being we I
[15:39] wouldn't necessarily attract them to
[15:40] expect them to attract disproportionate
[15:42] support from renters. Yeah. And just one
[15:43] other aspect of this kind of set of
[15:45] expectations that we should just clarify
[15:47] before we move on to how you investigate
[15:50] these things is so you've set out two
[15:52] dimensions there of attitudes on
[15:54] different aspects of policy and I think
[15:56] probably quite a few of our listeners
[15:58] will be familiar with two-dimensional
[16:00] models of the political spectrum but
[16:03] with slightly different dimensions. So
[16:06] people will be used to the left right
[16:07] dimension which was the first one that
[16:08] you set out there. But then we often
[16:10] speak of a second dimension of politics
[16:13] being liberal versus conservative on
[16:15] more kind of social issues. There are
[16:17] attitudes on gender questions, attitudes
[16:19] on immigration as well. That's different
[16:22] from your second dimension, isn't it? So
[16:23] you're not suggesting that housing is
[16:25] going to be a strong predictor of um
[16:28] attitudes and preferences on that
[16:30] dimension.
[16:31] >> Yeah, that's right. So the evidence on
[16:33] that we have suggests that housing is
[16:35] not strongly predictive of attitudes on
[16:38] this as you as you write the kind of
[16:40] cultural dimension we call it if you
[16:42] like
[16:43] >> and that's also another reason to expect
[16:45] that parties that are kind of primarily
[16:47] if you like compete on that dimension.
[16:49] So whether that be radical right parties
[16:52] or social liberal parties won't attract
[16:54] particularly disproportionate support
[16:55] from housing market groups because
[16:56] housing market groups again housing
[16:59] market position is not particularly
[17:00] determinative of support on that
[17:03] dimension.
[17:04] >> Yeah. Okay. It's good just to clarify
[17:05] that. Great. Thank you. So there are
[17:07] various other nuances in your
[17:08] hypotheses. you have quite a kind of
[17:10] rich structure of expectations. But
[17:12] maybe we'll deal with those hypotheses
[17:14] that we've discussed so far. And maybe
[17:16] we've got time, maybe we'll get into
[17:17] some of the others when we're discussing
[17:19] the results. But before we get to the
[17:21] results, of course, we need to know
[17:23] about the methodology that you use, how
[17:25] you go about investigating these
[17:27] questions. So what's the approach that
[17:28] you take in this paper? Well, I think a
[17:31] key challenge of this paper is simply
[17:33] the fact that the kind of typically used
[17:38] cross-national
[17:40] survey sources don't consistently or
[17:43] often at all collect data on housing
[17:46] tenure. And I think that's part of the
[17:47] reason why we have a lot of these kind
[17:49] of studies for occupational class but
[17:50] not for not for housing. So I had to
[17:54] basically create two data sets and the
[17:57] first one allow allowed me to identify
[18:02] the kind of or measure the main
[18:04] divisioning housing markets that I
[18:05] identify that between home owners and
[18:07] renters. And for this I use I collect
[18:10] data from national election studies. So
[18:13] about a hundred of these surveys from 14
[18:15] countries over the past 40 years. And
[18:18] that allows me to measure the how the
[18:21] relationship between housing tenure
[18:23] measured in terms of owners versus
[18:24] renters um has and voting has changed at
[18:28] the time and across countries. I also
[18:31] identified in the paper other divides
[18:33] within housing markets within these
[18:35] groups. So within the homeowner and
[18:36] renter groups relating on the owner side
[18:39] to house value, whether you're an
[18:41] outright owner or have a mortgage and
[18:43] whether you're a landlord or not. And on
[18:46] the renter side in terms of whether
[18:47] you're a social or private renter and to
[18:51] capture these more because the national
[18:53] election studies don't tend to capture
[18:54] these kind of more fine grain detailed
[18:56] if you like dimensions of your housing
[18:57] market position. I
[19:00] um use household panel data from six
[19:04] countries to identify those. So there's
[19:06] one data set which I use to measure the
[19:09] relationship between owners and renters
[19:10] and a second one which I use to go at
[19:11] these more fine grain housing market
[19:13] divides. And in terms of the methods
[19:16] used, it's a primarily descriptive
[19:17] paper. So it's I use mainly logistic
[19:20] regression analysis. But importantly I
[19:22] kind of look at the relationships in
[19:26] both a barriiate setting. So just
[19:28] looking at the raw kind of simple
[19:30] relationship between the two these two
[19:32] variables.
[19:32] >> So housing and voting
[19:33] >> housing and voting. But I also consider
[19:36] it in a multivariate setting by which I
[19:38] mean in including control variables. And
[19:41] this is important because it could be
[19:43] that for instance if I identified that
[19:45] the relationship between housing and
[19:47] voting has become more important over
[19:48] time. People might say well is this not
[19:50] just the relationship is this not just a
[19:52] reflection of the relationship between
[19:54] age and voting becoming more important
[19:56] over time because we know aging age is
[19:58] becoming increasingly associated with
[20:00] both housing and voting housing position
[20:01] and voting. I also have to look at I
[20:03] also have to control for these other
[20:05] factors to just check that this that the
[20:08] the findings that I get are not
[20:09] reflective of electoral realignments
[20:11] around other demographic
[20:13] characteristics.
[20:14] >> Great. So on then to the key question
[20:17] what do you actually find and shall we
[20:19] shall we start off with that basic
[20:20] relationship between uh housing
[20:23] dichotomized between owners and renters
[20:26] on the one hand and voting behavior on
[20:28] the other hand.
[20:29] >> Yeah, Grace. So I find so your housing
[20:33] tenure is important for voting and the
[20:36] specific relationships between tenure
[20:38] groups and parties that are
[20:40] hypothesized. I did find support for
[20:42] these. So homeowners are significantly
[20:45] more likely than renters to support
[20:47] center-right parties and renters are
[20:50] significantly more likely than
[20:51] homeowners to support radical left
[20:54] parties.
[20:55] And for both of these I mean today in
[20:56] the in the last in recent times for the
[20:59] other party families. So for social
[21:02] democratic parties, radical right
[21:03] parties and green parties, renters are
[21:07] more likely are slightly more likely to
[21:08] support them than homeowners even once
[21:10] you control for other characteristics
[21:12] but the relationship is far weaker. So
[21:15] it is the case that we see a significant
[21:17] housing based surviving electoral
[21:18] behavior and that is kind of given
[21:20] expression by two party families. the
[21:22] center right on one on the one hand and
[21:23] the radical left on the other.
[21:25] >> And has that changed over time?
[21:27] >> Yeah. So this is I think this is kind of
[21:29] potentially the most striking finding I
[21:31] found in the paper is that the the these
[21:33] two relationships have become
[21:34] significantly stronger over time. So for
[21:37] the center right in the kind of 1980s
[21:40] there's a relatively weak relationship
[21:42] between housing tenure and voting and
[21:44] over time this has grown significantly.
[21:46] So that housing tenure is now very
[21:48] important in explaining support for the
[21:50] center right. And you see the mirror
[21:51] image for the radical left for a kind of
[21:55] measure of how important it's become if
[21:57] you like. I I find that center I I
[22:00] compare the effect of housing tenure to
[22:03] the effects of other demographic divides
[22:05] on voting for these party families. So I
[22:07] compare housing tenure to income,
[22:10] education,
[22:12] age and occupational class. And I find
[22:14] that for both these party families, it
[22:16] is the most important thing. So it is
[22:19] absolutely essential I think to
[22:21] understanding the kind of social bases
[22:23] of these two party families.
[22:26] >> And then you look into
[22:29] the the subcategories of housing tenure.
[22:32] So the different groups within
[22:34] homeowners and the within renters, the
[22:37] socially and privately renting groups as
[22:39] well. What did you what I I guess maybe
[22:42] we should just first clarify what you
[22:43] were expecting to find once you broke
[22:46] that down and then what you
[22:48] >> Yeah. So I was expecting there not just
[22:51] not just to find this divide between
[22:52] home owners and renters but also to to
[22:54] find divides within this group within
[22:56] these groups. So I was expecting
[22:59] again my expectations revolved around
[23:02] these two the two these kind of two key
[23:04] party families. I expected highv value
[23:07] homeowners to be especially supportive
[23:09] of center right parties. I expected
[23:11] outright homeowners, homeowners who have
[23:13] paid off their mortgages to be
[23:14] especially supportive supportive of
[23:16] centeright parties. And also I expected
[23:19] private landlords to be especially
[23:20] supportive of centeright parties. I also
[23:23] hypothesized that social renters should
[23:26] be or I expected them to be especially
[23:28] likely to support radical left parties
[23:31] as well
[23:31] >> compared with private renters. And
[23:33] what's the logic there? So the logic
[23:35] there is that again we haven't got as
[23:38] much work on on renters to draw on here
[23:41] but one thing we do know about private
[23:43] renters is they are more likely to
[23:45] anticipate becoming homeowners over in
[23:48] the future. So I therefore expected them
[23:50] to behave slightly more like like
[23:52] homeowners and they do but they still
[23:55] they still behave differently from
[23:56] homeowners. So yeah and in terms of what
[23:58] we find we actually find that so looking
[24:01] at these three divides amongst home
[24:04] owners house value is the most
[24:05] important. So all of my expectations are
[24:08] actually I find support for so I I'd say
[24:10] center right parties do attract
[24:12] particularly strong support amongst high
[24:14] value homeowners outright owners and
[24:16] landlords but the effect is especially
[24:18] strong for house value. Um, I also find
[24:21] that low value homeowners are are more
[24:24] likely are significantly more likely
[24:25] than high value home owners to support
[24:28] radical left. So radical left is their
[24:31] support isn't limited to renters. It's
[24:34] they also attract a good level of
[24:35] support amongst low lowv value
[24:36] homeowners. They still attract more
[24:38] support than amongst private and social
[24:40] renters than amongst low value
[24:42] homeowners. on the renter side again the
[24:45] the relation so there there are
[24:48] differences between social and private
[24:49] renters but they're slightly smaller so
[24:51] private renters are more likely than
[24:54] social renters to support green parties
[24:56] I find with social rent and also center
[24:58] right parties but they're they show
[25:00] quite low levels of supports for both
[25:02] group for the center right and social
[25:04] renters are more likely than private
[25:06] renters to support social democratic and
[25:09] radical left parties
[25:11] >> and we should remember that these There
[25:13] are findings that are across either 14
[25:16] countries for for the kind of dichotomy
[25:18] of owner versus renter or across is it
[25:21] six countries you said for the the the
[25:24] more fine grained analysis there.
[25:26] Speaking personally I suppose I'm
[25:28] particularly interested the results for
[25:29] the UK. We're sitting here in the UK. I
[25:31] guess many of our listeners might be
[25:33] particularly interested in the UK as
[25:35] well. And I'm kind of thinking, gosh, I
[25:37] really want to put this article on the
[25:39] reading list for my British politics
[25:41] module because it's very directly
[25:43] related to some of the questions that we
[25:44] ask in relation to the UK in that
[25:47] module. But so far, you've been talking
[25:49] about findings that are kind of pulled
[25:51] across all of these various countries.
[25:55] If you did look simply at the UK, would
[25:58] you find the same picture that housing
[26:00] matters and that it has become more
[26:01] important in shaping voting behavior
[26:03] every time? Well, this is a very
[26:05] interesting question. So, first of all,
[26:07] I might just UK as a case kind of within
[26:10] the within my study. So I I do predict
[26:15] that the relationship between housing
[26:17] and voting will be stronger in certain
[26:19] types of housing markets and especially
[26:22] those within within which the housing
[26:24] market transformations that I discussed
[26:26] at the start that we discussed at the
[26:27] start were kind of most acute. in which
[26:30] the house price inflation has been
[26:31] greater, rental market insecurity is
[26:33] greater just because I expect
[26:34] inequalities in h between housing market
[26:36] groups to be greater in those societies.
[26:38] And typically these societies within the
[26:40] literature I are identified as kind of
[26:42] the kind of English-speaking countries
[26:45] if you like. I label them the Anglo
[26:46] liberal group in the in the in the paper
[26:49] and these are Britain, Ireland, Canada,
[26:52] United States, Australia and New
[26:54] Zealand. So I I my my prior expectation
[26:57] was for the relationship between housing
[27:00] and voting to be particularly strong in
[27:01] these countries and also probably to
[27:03] have grown most significantly in these
[27:05] countries. So
[27:08] what I actually find is that there isn't
[27:10] a strong relationship or that the
[27:11] relationships the strength of the
[27:13] relationships between housing and voting
[27:15] doesn't vary significantly across
[27:17] different housing market types
[27:19] within this the
[27:20] >> so there is a relationship between
[27:22] housing housing type and voting. It's
[27:25] just that that relationship is broadly
[27:26] the same yet across these different
[27:28] countries.
[27:29] >> Yes, absolutely. So housing your housing
[27:31] tenure for instance seems to be roughly
[27:32] as important in different so in Anglo
[27:36] Liberal housing markets as it is in
[27:37] other as as it is in other housing
[27:39] markets.
[27:40] >> Yeah. The UK itself is an outlier. So I
[27:45] find that so within all different types
[27:46] of housing markets there has been a
[27:48] significant growth in the relationship
[27:49] between housing tenure and voting
[27:51] behavior over time. We don't see this in
[27:54] Britain. Britain's an outlier in that it
[27:57] housing has always been very important
[27:59] or as or at least in so far as my data
[28:02] goes back so it goes back 40 years. It
[28:04] has always been a a hugely important
[28:06] predictor of voting in Britain. I shared
[28:10] your expectation. I guess that I thought
[28:11] that this it would be I thought it would
[28:13] be I I didn't really expect this. I
[28:15] expected because the housing kind of
[28:18] unaffordability is so acute in Britain.
[28:20] I expected this and it's become more
[28:22] acute over time. I expected this to be
[28:24] important. But I think maybe it reflects
[28:26] the fact that housing has always been an
[28:28] important kind of class marker in
[28:30] British society. And maybe this is
[28:32] getting more into kind of social
[28:34] history. But I think it's quite
[28:35] intuitive that
[28:38] housing has always been an important
[28:39] kind of I guess market of status in
[28:41] Britain. We've always had a kind of
[28:43] large social housing sector in Britain
[28:45] at least since the war for instance. and
[28:47] that's kind of been tied up, I guess,
[28:48] with working-class identity. I guess
[28:50] middle-class aspirations have always
[28:52] been tied up within home ownership. So,
[28:54] I think maybe it reflects at that that
[28:56] the kind of unusual degree to which
[28:59] housing is important for class
[29:00] historically in Britain. And I guess
[29:02] it's also the case and your data go back
[29:04] to the mid1 1980s and you can't remember
[29:07] the 1980s, but I can. And I mean I can
[29:10] remember the 1980s as a period when I
[29:14] mean it may be that housing inequality
[29:16] hadn't quite reached the levels that it
[29:17] has reached now but it was growing very
[29:19] rapidly or at least it was perceived in
[29:21] the public discourse as growing very
[29:23] very rapidly dur during that per people
[29:25] period and people were amazed by what
[29:26] was happening to house house prices and
[29:29] the nature of the housing market during
[29:31] that period and then that was
[29:32] immediately followed by the crash of
[29:35] 1989 and I mean you were talking earlier
[29:38] about the increasing importance of
[29:41] mortgage debt as a factor. And I mean I
[29:44] remember the early 1990s as a time when
[29:46] >> when talk of negative equity so people
[29:50] whose mortgage mortgages their
[29:52] outstanding mortgages were bigger than
[29:54] the size than the value of their house
[29:57] because the house prices values had gone
[29:59] down. That was a big big thing in that
[30:01] period to a degree that has never quite
[30:03] recurred actually since then. So I guess
[30:06] it may simply be that the the UK in some
[30:09] ways moves earlier than some other
[30:11] places and therefore it's already moved
[30:13] by the time of your early data in this
[30:16] analysis. But if you were if you were
[30:18] able to go back into the 1970s and 1960s
[30:21] perhaps you would see a different
[30:22] picture in the UK as well. Yes, I think
[30:24] that's yeah I think that's probably
[30:25] right. I also think you identify how
[30:28] kind of housing was important even in
[30:31] the in the 1980s and or growing as an
[30:33] important as a kind of predictor of your
[30:36] social position. That also meant that it
[30:37] was also kind of very politicized in the
[30:40] 1980s as well. Right. So we had a the
[30:41] right to buy scheme uh of the Thatcher
[30:44] government being a highly kind of
[30:46] contested political issue. So I guess
[30:49] housing was always a kind of central
[30:50] political question in in a way that
[30:52] maybe it wasn't in other countries. So
[30:53] that might be another way to explain
[30:55] that.
[30:56] >> Yeah.
[30:57] >> Fascinating. Fascinating. We've learned
[30:58] so much from this study already. What
[31:01] more is there to explore? You're this is
[31:03] part of your wider PhD project, but what
[31:05] what other questions are you seeking to
[31:08] explore in relation to this? So more
[31:10] broadly, I want to
[31:14] extend this analysis not just to look at
[31:16] class voting, but class politics more
[31:19] generally. So in this paper we I've kind
[31:22] of identified that housing has grown
[31:25] significantly relative to occupation as
[31:27] a predictor of voting and maybe
[31:29] therefore the nature of uh class voting
[31:31] has changed from revolving around
[31:33] occupational labor markets to revolving
[31:35] around housing markets. I'm also going
[31:37] to consider that in relation to other
[31:39] other aspects of class politics. So
[31:41] political attitudes and looking at
[31:43] whether we don't have much analysis on
[31:45] how the relationship between
[31:48] your attitudes towards for instance kind
[31:50] of left right issues like redistribution
[31:52] welfare have changed over time. So I'm
[31:55] going to consider whether this has also
[31:57] grown over time and whether it's growing
[31:59] relative to occupation and this might
[32:01] perhaps help to explain some of the
[32:02] relationships we see in this paper. And
[32:05] I'm also going to look at the role of
[32:08] political parties.
[32:10] So if the kind of underlying kind of if
[32:13] you like social base of class politics
[32:15] is changing is that reflected in how
[32:18] parties politicize economic policy or
[32:22] welfare politics. So, and I'm going to I
[32:25] explore whether um whether parties
[32:29] increasingly politicize housing as an
[32:31] issue, increasingly politicized divides
[32:33] between housing market groups and and
[32:36] whether they've become more polarized on
[32:39] housing policy over time and will that
[32:42] help us understand the why in all of
[32:44] these relationships? because you're
[32:45] you're very modest in how you describe
[32:47] your findings so far in the paper and
[32:49] you've indicated it during this episode
[32:52] as well. You talk about the findings as
[32:54] descriptive findings in which you see a
[32:56] relationship, a correlation between
[32:59] housing tenure and voting patterns, but
[33:02] you avoid making any causal claims about
[33:04] why it might be exactly that this
[33:07] relationship exists. Will that further
[33:09] analysis allow you to go into that why
[33:11] question?
[33:12] >> Yes, I think so. So I think that if it's
[33:14] true that housing has become more
[33:17] important as a kind of determinant of
[33:19] affluence, it we'd also expect it to be
[33:21] true that it's become more important in
[33:23] terms of whether you support
[33:27] redistribution. For instance, if we kind
[33:28] of hypothesize that homeowners have
[33:30] become relatively better off over the
[33:31] years, we might expect them to become
[33:33] relatively more right-wing on economic
[33:35] policy and this would help to explain
[33:37] why their voting behavior has also
[33:39] become more right-wing. Um and equally
[33:42] yes in the role of political parties. Um
[33:45] so there's some theories in political
[33:47] science which suggest that kind of
[33:48] social cleavages or social divides need
[33:51] to be in order for them to become to be
[33:54] expressed politically they need to be
[33:56] mobilized by political parties. So is it
[33:58] the case that parties have kind of
[34:01] noticed this change in the class
[34:02] structure that's happened and said oh
[34:04] there's these growing groups at the kind
[34:06] of polarized ends of housing markets if
[34:08] you like. So there's a growth of group
[34:10] kind of affluent housing market groups
[34:11] and a growth of those in insecure
[34:13] housing market groups. these are
[34:14] potential electoral groups for me to
[34:16] mobilize for us to mobilize and have
[34:18] they moved or polarizing their housing
[34:20] market positions in order to mobilize
[34:22] these groups and is that playing a part
[34:25] in the growth of housing as an electoral
[34:27] divide which I've identified here and
[34:29] indeed we do seem to see some parties
[34:32] particularly central right parties and
[34:34] perhaps liberal parties as well kind of
[34:37] associating themselves if you like with
[34:39] nimi approaches to housing and green
[34:43] parties radical ical left parties to
[34:44] some extent social democratic parties
[34:46] such as labor in the UK associating
[34:48] themselves with the kind of yimi idea of
[34:51] wanting more housing to be built.
[34:54] >> Yes. So I have already done some
[34:56] analysis for this and interestingly I
[34:58] kind of shared your I shared your prior
[35:01] if we look at party manifestos they have
[35:03] not actually become more polarized in
[35:04] their in their positions on house
[35:06] building. Huh.
[35:07] >> They have become more polarized in their
[35:10] positions on other um facets of housing
[35:13] policy. So for instance on t the
[35:16] taxation of wealth and housing assets.
[35:19] So on this center right parties have
[35:21] moved more to the right in their policy
[35:22] positions and in their rhetoric over
[35:23] time whereas parties of the left have
[35:26] moved toward to the left on these. We
[35:28] see the same thing in social house
[35:30] building specifically but also private
[35:33] rental market regulation. So we're
[35:35] seeing that we're seeing the parties of
[35:36] the left, especially radical left and
[35:38] green parties becoming more in favor of
[35:40] rent control was for instance and
[35:43] stronger tenency protections and center
[35:46] rights, but also interestingly radical
[35:47] right parties becoming more opposed to
[35:50] these.
[35:51] >> Well, thank you so much Josh. You're
[35:53] giving us a tantalizing glimpse there
[35:54] for some further work beyond the study
[35:56] that we've been examining today. So
[35:58] we'll look forward very much to having
[36:00] you back on the podcast when you've
[36:02] developed that further. But this has
[36:04] been terrific. I've learned so much and
[36:05] this is definitely going on my reading
[36:07] list for the British politics module.
[36:09] So, thank you so much. We have been
[36:11] discussing the article Housing and
[36:13] Electoral Behavior: The Changing Face of
[36:15] Class Voting in Advanced Democracies by
[36:17] Josh Godard, published in the February
[36:20] 2026 issue of the journal Electoral
[36:23] Studies, but it's already available free
[36:25] of charge online. And as ever, we'll put
[36:28] full details in the show notes for this
[36:31] episode. We will be back with more
[36:33] cuttingedge political science research
[36:35] next week. I think we're going to be
[36:36] talking about ways of tackling
[36:38] disinformation on social media, but I'm
[36:40] not completely sure. So, watch this
[36:41] space. And to make sure you don't miss
[36:43] out on that or other future episodes of
[36:46] UCL Uncovering Politics, all you need to
[36:48] do is subscribe. You can do so on Apple,
[36:50] Google Podcasts, or whatever podcast
[36:52] provider you use. And while you're
[36:55] there, we'd love it if you could take a
[36:56] moment of time to rate or review us,
[36:59] too. I'm Alan Renick. This episode was
[37:02] produced by Adaner Kingwell Banham. Our
[37:04] theme music is written and performed by
[37:06] John Man. This has been UCL Uncovering
[37:08] Politics. Thank you for listening.
[37:17] >> [music]

Introduction to Moral Politics

In the latest episode of UCL Uncovering Politics, the discussion centers on how political leaders can utilize moral issues to their advantage, particularly in times of political trouble. Alan Renick introduces the topic by referencing the historical phenomenon where governments facing unpopularity can rally public support through perceived threats to national security. This concept has been exemplified in events such as the Falklands War in 1982 and the aftermath of the September 11 attacks in 2001, where leaders like Margaret Thatcher and George W. Bush experienced significant boosts in approval ratings. However, Renick emphasizes that while governments can exaggerate existing threats, deliberately creating security crises would be reckless.

Exploring Alternatives to Security Threats

Renick introduces a new study by Dr. Daniel Schulty, which investigates whether moral issues can serve as alternatives to security threats for leaders seeking to maintain power. Schulty, an associate lecturer at UCL, explains that his research is rooted in his PhD thesis, which examines how morality policy issues—such as abortion, LGBTQ rights, and religious education—can be leveraged by would-be autocrats to consolidate their power. The study provides experimental evidence that moral issues can indeed be utilized similarly to security threats in the context of regime maintenance.

Understanding Moral Issues

Schulty defines moral issues as non-economic, value-based topics that involve conflicts over societal norms regarding the right way of living. He focuses on two specific types of moral issues in his study: conflicts surrounding religious buildings and traditional family values. The research is particularly relevant in the context of electoral authoritarian regimes, where leaders may exploit these moral issues to bolster their popularity among a religiously affiliated populace.

Mechanisms of Moral Politics

Renick and Schulty discuss the mechanisms by which moral issues can be used to distract citizens from pressing economic or political challenges. Similar to the rally-around-the-flag effect seen in security contexts, moral agenda-setting can invoke loyalty to religious communities and instill fear regarding the erosion of religious values in society. Schulty posits that this strategy can effectively unite citizens behind their leaders, allowing them to overlook other grievances such as corruption or economic instability.

Case Studies in Turkey

Schulty's study is grounded in the political landscape of Turkey, particularly during a period marked by economic crisis and the COVID-19 pandemic. He highlights how the Turkish government has politicized moral issues, specifically the conversion of the Hagia Sophia from a museum back into a mosque and the withdrawal from the Istanbul Convention, which aimed to combat domestic violence. These cases illustrate how the government has framed moral issues to distract from economic hardships and maintain political support.

Experimental Design of the Study

The research employs a survey experiment conducted online with a sample of over 1,000 Turkish citizens. Respondents were divided into four groups, each receiving different prompts regarding the economic situation, security threats, or moral issues. The study aimed to measure how these different treatments influenced perceptions of the economy and support for government policies. Schulty hypothesized that reminding citizens of security or moral threats would decrease the perceived importance of economic issues and increase support for the government.

Findings and Interpretations

The results revealed that the security treatment significantly increased support for government policies, while the Hagia Sophia morality treatment effectively diminished the perceived importance of economic issues among citizens. Interestingly, the Istanbul Convention treatment did not yield statistically significant results, suggesting that different moral issues can have varying impacts on public perception. Schulty theorizes that the Hagia Sophia issue resonated more with the general population, while the Istanbul Convention was less popular due to its association with gender ideology and Western values.

Implications for Future Research

Schulty emphasizes the need for further research to explore the generalizability of these findings beyond Turkey, particularly in countries with similar religious nationalist contexts, such as India, Israel, Poland, and Russia. He notes that while the mechanisms of moral politics may apply in various settings, the effectiveness of these strategies could differ based on the political environment and the level of media control in authoritarian regimes.

Conclusion and Future Directions

In conclusion, Schulty's work sheds light on the strategic use of moral politics by authoritarian leaders to maintain power and distract citizens from pressing issues. His upcoming book aims to expand on these findings, exploring how morality politics can neutralize institutional opposition and serve as a diversion during crises. As the political landscape continues to evolve, understanding the dynamics of moral politics will be crucial for analyzing the behavior of leaders in both authoritarian and democratic contexts.

Listeners are encouraged to subscribe to UCL Uncovering Politics for future discussions on political strategies and their implications in contemporary society.

In the transcript, the speaker discusses the intricacies of voter behavior, particularly emphasizing that voters do not always act based on purely rational choices. While traditional political theory often suggests that voters make decisions based on material interests, policy preferences, and identity, the speaker highlights that emotional and moral factors can significantly influence these decisions. This perspective is rooted in the understanding that voters' choices are often shaped by their moral convictions and the broader socio-political context, which may lead to deviations from rational calculations.

For instance, the speaker points out how political leaders can capitalize on moral issues to influence public sentiment and bolster their power. This suggests that voters might rally around leaders or parties that resonate with their moral beliefs, even if these choices may not align with their immediate material interests. This phenomenon can lead to ideological adherence that transcends rational economic considerations. Additionally, the speaker alludes to mechanisms such as fear of losing moral values or a sense of patriotism that can influence voters to prioritize moral issues over economic ones, reinforcing the idea that emotional and identity-based factors play a crucial role in shaping electoral outcomes.

• [01:44] "...a threat to the country's security...the public will then rally around the flag and the government's fortunes will revive."
• [05:49] "Morality issues are non-economic value-based topics that involve conflicts over norms about the right way of living."
• [06:17] "There needs to be high religiosity citizens or a significant portion of the population kind of affiliated with a dominant religion that provides the source for these religious values."

The transcript does not provide specific details regarding how voter information or election debate formats affect the quality of voters' choices. Thus, this aspect remains unaddressed within the conversation.

• [00:49] "...we have seen big increases in support for example for Margaret Thatcher's government amidst the Faulland's war of 1982 or for the Bush administration in the United States in the wake of 911."
• [01:02] "But for a government to engineer such a threat against its own country would be to say the least foolhardy."

In the discussion, the speaker articulates the influence of both moral convictions and economic circumstances on ideological disagreements and the rise of far-right leanings. They suggest that moral issues, such as those related to religious values and family traditions, can serve as powerful motivators for voters, particularly in contexts where there is a strong religious nationalist sentiment. The speaker notes that leaders can exploit these moral concerns to distract citizens from pressing economic issues.

This manipulation of moral framing is particularly relevant in electoral authoritarian regimes, where leaders may leverage moral narratives to consolidate their power. By shifting public discourse toward moral issues, leaders can create a sense of urgency or threat to social values, prompting citizens to rally behind them against perceived threats. This can lead to an increase in support for far-right ideologies, as leaders frame these moral issues as integral to national identity and security.

• [06:01] "Morality issues are non-economic value-based topics that involve conflicts over norms about the right way of living."
• [06:46] "The literature has not put, in my opinion, enough emphasis on the role that morality policy issues can play for these electoral authoritarian regimes..."
• [08:44] "...fear of the diminishing role of religious values in society."

The transcript does not explicitly mention recommendations for how citizens can better navigate misinformation or evaluate political promises. Therefore, this aspect is not covered in the discussion.

• [03:39] "...Bush's support and popularity spiked to approximately 90% after the events of 9/11 and other governments around the world often turn to security."
• [04:12] "...the Indian government highlighted an alleged terrorist threat within Pakistan before the 2019 election."

The speaker touches on how voters often judge governments based on retrospective evaluations that hinge on economic outcomes, particularly during crises. This retrospective judgment implies that voters tend to hold governments accountable for their economic performance. As seen in the examples provided, such as the Turkish government's handling of economic issues during a crisis, the public's perception of government efficacy can shift dramatically based on economic indicators. This suggests that voters' expectations for accountability are closely tied to their immediate experiences with economic conditions.

This reliance on economic outcomes for judgment raises questions about the fairness of accountability, especially in contexts where external factors may influence economic performance. The speaker implies that while citizens may wish to hold governments accountable, the complexities of economic realities can muddy this process, leading to potential disparities between public expectations and actual governmental capabilities. Thus, the accountability of political leaders becomes a nuanced issue, dependent not only on their policies but also on the broader economic context.

• [19:30] "...respondents receiving any of these different treatment conditions...should have a higher support of government policy and a lower importance of economy than those in the control group."
• [20:30] "...citizens want it. They’re hungry for a message trying to explain what is going on."
• [20:50] "...while we might think that agenda setting might not be very effective because citizens can go to the grocery store and see that prices of vegetables and fruits are high."

[00:05] Hello, this is UCL Uncovering Politics
[00:07] and this week we're looking at how
[00:08] political leaders can play moral
[00:10] politics to their own advantage. How and
[00:13] when can leaders mobilize concern about
[00:16] moral issues to bolster their power.
[00:24] Hello, my name is Alan Renick and
[00:26] welcome to UCL Uncovering Politics, the
[00:28] podcast of the School of Public Policy
[00:30] and Department of Political Science at
[00:32] University College London. It has often
[00:35] been said that if a government is in
[00:37] trouble with voters, one thing that can
[00:39] help it out is a threat to the country's
[00:42] security. The public will then rally
[00:44] around the flag and the government's
[00:46] fortunes will revive. And certainly we
[00:49] have seen big increases in support for
[00:51] example for Margaret Thatcher's
[00:52] government amidst the Faulland's war of
[00:55] 1982 or for the Bush administration in
[00:58] the United States in the wake of 911.
[01:02] But for a government to engineer such a
[01:04] threat against its own country would be
[01:06] to say the least foolhardy. And let's be
[01:08] clear there's no suggestion that at all
[01:10] that was the case for the Faullands or
[01:12] 911. Government can talk up or
[01:15] exaggerate existing security threats. We
[01:18] might think of President Trump's
[01:19] deployment of military forces in various
[01:22] US cities, for example. But are there
[01:25] alternatives to this securityization
[01:27] route? Can those in power distract
[01:30] publics from political troubles in other
[01:33] ways? Well, a new study explores one
[01:37] such possibility. It looks at the degree
[01:39] to which governments can use moral
[01:41] issues as alternatives to security
[01:44] threats. And drawing on evidence from
[01:46] experimental research conducted in
[01:48] Turkey, it concludes that they can. The
[01:52] author of that study is Dr. Jan.
[01:56] You say, "I was thinking about your
[01:57] surname and thinking, oh, I got to get
[01:58] this right." Look at the
[02:01] I should have said at the start. By the
[02:02] way, don't worry about if you want to
[02:04] repeat anything that I've said in the
[02:06] introduction. That's absolutely fine. So
[02:08] don't worry about Alan's already said
[02:09] stuff so I can't just feel free to
[02:11] answer as you wish. We're almost there.
[02:14] The author of that study is Dr. Daniel
[02:17] Schulty, associate lecture.
[02:20] The author of that study is Dr. Daniel
[02:22] Schulty, associate lecturer in protest
[02:25] revolution and qualitative methods.
[02:27] There's a title here in the UCL
[02:29] department of political science. And I'm
[02:32] delighted that Daniel joins me now.
[02:34] Welcome Daniel to UCL Uncovering
[02:37] Politics. It's great to have you on. And
[02:40] before we go into your own analysis, can
[02:43] we maybe explore the background a little
[02:45] bit further? I suggested there that
[02:47] leaders can sometimes use security
[02:50] threats, perhaps also exaggerate, even
[02:52] create such threats in order to advance
[02:55] their own interests. What do we already
[02:58] know about how that's done, how often
[03:01] it's done?
[03:03] Hi Alan, thank you so much for having me
[03:06] on the podcast. Yes, there is a large
[03:09] literature called rally around the flag
[03:11] or diversionary war literature which
[03:14] argues that leaders can turn to security
[03:17] threats and you're right to point out
[03:19] that that often these threats are real
[03:22] ones that occur, but other times they
[03:25] can manipulate information about these
[03:28] threats and bolster them. and you gave
[03:31] some great examples from the literature
[03:33] already with the Fulan wars and 9/11. So
[03:36] just to turn to the 9/11 example, Bush's
[03:39] support and popularity spiked to
[03:41] approximately 90% after the events of
[03:44] 9/11 and other governments around the
[03:48] world often turn to security. So an
[03:50] example I use in the paper is the
[03:52] Turkish government often politicizing
[03:54] Kurdish security issues. So a minority
[03:58] ethnic group in the southeast of Turkey
[04:00] and northern Syria and they often
[04:02] politicize these during election periods
[04:05] specifically between 2015 and 2019. And
[04:08] another example in the paper comes from
[04:10] India where the Indian government
[04:12] highlighted an alleged terrorist threat
[04:14] within Pakistan before the 2019
[04:16] election. So those are two two examples
[04:18] that the paper discusses in terms of the
[04:20] security literature. So this idea of
[04:23] security being used and security threats
[04:27] being used by leaders in power in order
[04:29] to strengthen their position is quite a
[04:32] kind of familiar one. In the paper that
[04:35] we're talking about here, you want to
[04:37] look at something slightly different
[04:38] from that and you want to study whether
[04:40] moral issues can be used in a similar
[04:43] way. Why was it that you wanted to get
[04:45] into that question?
[04:46] So um my PhD thesis and book project
[04:51] focuses on the role that morality policy
[04:54] issues like abortion, LGBTQ rights and
[04:57] religious education can play for wouldbe
[05:00] autocrats coming to power and trying to
[05:03] stay in power. And so this article was
[05:07] one of the chapters of my PhD thesis and
[05:10] it provides experimental evidence for
[05:13] one of the uses that would be autocrats
[05:16] can use morality policy which is to
[05:18] maintain power or what we call regime
[05:20] maintenance. They try to stay in power
[05:23] and morality policy issues is one of
[05:26] multiple different tools in a toolbox
[05:28] that they can turn to. So we'll get on
[05:30] to exploring the rest of that the PhD
[05:33] thesis and the book project towards the
[05:34] end hopefully. But for the moment let's
[05:36] focus in on this particular segment. So
[05:39] you're talking about the way in which
[05:40] they can use moral issues in these sorts
[05:42] of ways. What do we actually mean by
[05:44] moral issues here?
[05:45] >> Yes. So morality issues are non-economic
[05:49] valuebased topics that involve conflicts
[05:52] over norms about the right way of
[05:54] living. And two types of morality issues
[05:58] that I focus on in the paper are ones
[06:01] over religious buildings or religious
[06:03] sites and the other one is focused on
[06:05] traditional family values.
[06:08] >> And why would we expect so so well first
[06:10] of all before I ask that you're you said
[06:12] there that you're focusing on
[06:14] authoritarian leaders in this work. Is
[06:16] that right?
[06:17] >> Correct. So in in Turkey the term that's
[06:20] often used is electoral authoritarian.
[06:22] So somewhere between a full
[06:24] authoritarian and a democracy, sometimes
[06:27] called like a hybrid regime.
[06:29] >> And okay. And your argument is that or
[06:32] your hypothesis is your expectation is
[06:34] that leaders of that kind are likely to
[06:37] use moral issues in order to try to
[06:39] shore up their own popularity. Yeah,
[06:42] >> that that's the basic argument.
[06:44] Obviously, there's contingency and
[06:46] there's different issues that the
[06:47] leaders can turn to, but the literature
[06:50] has not put, in my opinion, put enough
[06:52] emphasis on the role that morality
[06:54] policy issues can play for these
[06:57] electoral authoritarian regimes, but
[06:58] especially ones that have uh a religious
[07:02] nationalist context where a large
[07:04] percentage of the population would be
[07:06] self-identifying and affiliated with a
[07:08] dominant religion. And can you just
[07:10] spell out for a little bit for us why
[07:14] you would expect
[07:16] authoritarian leaders to use morality
[07:19] politics in this way? Is it is the
[07:21] underlying mechanism the same sort of
[07:23] thing as we see with security issues? Is
[07:25] it a different kind of mechanism?
[07:27] >> So the literature on the security agenda
[07:31] setting points to kind of patriotism. So
[07:34] loyalty to the nation among citizens as
[07:37] being like a major mechanism by which
[07:39] this rally around the flag or rallying
[07:42] around the leader dynamic plays out. So
[07:45] there's like fear of national death or
[07:46] humiliation that can motivate citizens
[07:49] in the face of a security threat like
[07:52] 9/11 to kind of come together to put
[07:54] aside partisan divides and unite amongst
[07:57] the leader for this specific fight
[08:00] against an alleged enemy. Morality
[08:03] agenda setting works in terms of
[08:05] religious devotion. So loyalty to the
[08:07] religious community and an important
[08:09] scope condition as I mentioned is there
[08:12] needs to be high religiosity citizens or
[08:14] a significant portion of the population
[08:17] kind of affiliated with a dominant
[08:19] religion that that provides the source
[08:21] for these religious values. And another
[08:24] mechanism is the fear of the diminishing
[08:27] role of religious values in society.
[08:30] >> Yeah. So it's actually a similar kind of
[08:32] mechanism if I understand correctly to
[08:34] the case of a security threat. It's a
[08:36] kind of threat to the moral values of
[08:39] the society that that is that people are
[08:42] thinking of here. Would that be fair?
[08:44] >> In general terms that that's correct. So
[08:47] morality is about the fear of the
[08:48] diminishing role of religious values in
[08:50] society and also pride in the shared
[08:53] religious heritage.
[08:54] >> Okay. In essence, what you're thinking
[08:58] is that just as with the case of
[09:01] security threat, if we say see a leader
[09:03] is facing unpopularity because of
[09:06] corruption scandals or economic
[09:08] troubles, whatever it might be. In the
[09:11] case of a security threat, the idea of
[09:12] rallying around the flag is the idea
[09:14] that well, if you get people to think
[09:16] that there's a big external threat
[09:17] threatening the nation, then the nation
[09:19] has to come together and come together
[09:20] behind the leader in order to to focus
[09:24] on repelling that threat. and people
[09:25] kind of stop thinking so much about the
[09:27] economic challenges or the corruption
[09:29] allegations or whatever it might be were
[09:31] the source of unhappiness and similarly
[09:33] in the case of morality politics and the
[09:37] use of morality politics in the way
[09:38] you're describing the idea if I
[09:41] understand correctly but I can correct
[09:42] correct this if I'm wrong is that
[09:45] leaders are facing challenges again it
[09:47] may be they're facing economic woes
[09:49] whatever it might be and they think how
[09:52] can we kind of ring people around behind
[09:55] me as leader. And the idea again is that
[09:58] you get people to fear that there's some
[10:01] kind of threat to the moral order and
[10:04] then they will to some extent forget the
[10:06] other concerns and just focus on
[10:08] confronting that threat as a community
[10:11] as a whole. That's correct Alan. So an
[10:14] important literature that I'm building
[10:16] upon is the authoritarian propaganda
[10:19] literature and within that literature
[10:21] there's emphasis on what is called
[10:23] agenda setting. So agenda setting is
[10:25] about shifting the agenda or what the
[10:29] government is talking about in the news
[10:30] or in speeches to an issue that is
[10:33] favorable to the government. And in my
[10:36] paper I'm focusing on a time in Turkey
[10:39] where there's an economic crisis as well
[10:41] as the COVID pandemic. And these are
[10:44] crises facing the Turkish government.
[10:46] And the idea is that governments want to
[10:50] shift from unfavorable topics like poor
[10:52] economic performance, rising costs of
[10:55] living, inflation, plummeting value of
[10:58] the Turkish lera, and they want to shift
[11:00] to issues that they can own and that
[11:02] that are that they can show as being
[11:05] victories for them.
[11:06] >> Can you say a little bit about what
[11:08] kinds of moral issues might work here
[11:11] and that you I mean, are there different
[11:12] kinds of moral issues? Are they all sort
[11:14] of the same? Right. So in my paper I
[11:18] test two types of morality issues that
[11:22] were already being politicized by the
[11:25] Turkish government at the same time. And
[11:28] so the paper goes in the history of
[11:30] those two issues to say that this is not
[11:33] just an experiment that is kind of
[11:35] making up kind of different vignettes or
[11:37] scenarios, but they're [clears throat]
[11:38] actually based upon what the Turkish
[11:40] government was actually already kind of
[11:42] politicizing in the news. So the first
[11:45] issue is the Hagi Sophia mosque issue.
[11:48] So for the listeners the Hagia Sophia
[11:52] you know was originally a Byzantine
[11:54] church and this is hugely famous
[11:57] extraordinary building in Istanbul.
[11:59] >> Correct. Yeah.
[12:00] >> So during the Republican period in
[12:02] Turkey. So this building was a museum
[12:04] from the 1930s up until 2020. And in
[12:09] 2020, the government changed the status
[12:12] of the Hagia Sophia from a museum into a
[12:15] mosque. So that's one. And I situate
[12:18] this very specific case within the
[12:21] literature on religious buildings and
[12:24] religious worship centers and how there
[12:26] are conflicts around the world between
[12:28] different groups of people who are
[12:30] claiming different buildings as part of
[12:32] their religious heritage and culture.
[12:34] The second one which I call and I
[12:37] situate into the broader family of
[12:40] traditional family values is the
[12:42] withdrawal from the Istanbul convention
[12:44] issue. The Istanbul convention was a
[12:47] treaty an international treaty which
[12:49] Turkey signed in 2010 which was to fight
[12:53] domestic violence. So in a way to help
[12:56] protect and strengthen women's rights
[12:58] issues in in Turkey. So the government
[13:00] signed on to this but in 2021 they
[13:04] withdrew from the treaty citing language
[13:07] saying that this treaty was imposing
[13:09] gender ideology and forcing kind of
[13:12] western values related to homosexuality
[13:15] on onto Turkey a place where such
[13:17] western values were foreign. And do you
[13:20] have these two moral issues in the study
[13:25] just because it's useful to have several
[13:28] issues because it's always possible that
[13:30] one issue that you choose might have
[13:32] some particular issue, some particular
[13:33] features that you haven't really thought
[13:35] of that would skew the results. So it's
[13:37] just useful to have several issues or is
[13:39] it that actually you're theorizing a
[13:41] difference between different kinds of
[13:43] moral issue? Based on my research, there
[13:46] are reasons to think that different
[13:48] morality policy issues are not the same
[13:51] and different issues have different
[13:53] strengths and weaknesses or costs and
[13:55] benefits. And by including two different
[13:59] morality issues, I could find test
[14:03] whether different issues would have
[14:04] different effects. At the same time,
[14:06] these were two important and issues that
[14:08] were being discussed by by the Turkish
[14:11] government. and I wanted to include them
[14:15] in my survey.
[14:16] >> Yeah. Okay. So, you were kind of
[14:19] theorizing that there might be
[14:20] differences between issues, but you
[14:21] didn't kind of specifically hypothesize
[14:24] that one would have a stronger effect
[14:25] than the other one, for example.
[14:27] >> That's correct. There's there was no
[14:29] formal hypothesis from the outset when I
[14:31] pre-registered this to to have
[14:34] differences between the morality issues,
[14:36] but that was an interesting finding that
[14:38] came up which we will get to in just a
[14:40] moment. So then a good tease. Then
[14:42] before we get to the findings though,
[14:43] let's just set out the design of this
[14:46] study a little bit more. We've given
[14:48] some hints. We've talked about how this
[14:50] is done in Turkey. We've talked about
[14:52] the fact that you're looking at these
[14:55] two morality issues and you're comparing
[14:59] them with a security threat as well. We
[15:03] I think we've hinted at the fact that
[15:05] this is done on the basis of a survey of
[15:07] public opinion. Do you want to just set
[15:09] out in a little bit more detail what is
[15:11] the design of the study? So the main
[15:14] method here is a survey experiment that
[15:17] was fielded online to an online panel in
[15:20] Turkey. And this survey was taken by
[15:23] Turkish citizens and it was fielded by a
[15:27] reliable data company called DATA which
[15:30] was formerly survey sampling
[15:32] international and the total sample is a
[15:35] little bit more than a thousand
[15:37] different respondents. So 1,125
[15:41] and this was fielded in 2022.
[15:44] >> Yeah. And we should just make sure
[15:46] everyone's up to speed with this. So by
[15:47] a survey experiment, you mean a survey
[15:50] in which essentially you vary the
[15:52] wording of some of the questions between
[15:54] different respondents and then you look
[15:56] at how does the variation make a
[15:58] difference to their responses to further
[16:01] questions thereafter.
[16:02] >> That's correct. So the sample was
[16:05] divided into four different groups and
[16:08] each group had about 270 or 280
[16:10] respondents in them. One was one one of
[16:14] the groups was the control condition
[16:15] which the respondents getting that that
[16:19] condition were reminded about the poor
[16:22] economic situation. So recently the
[16:24] economic situation in Turkey has
[16:26] garnered headlines with increasing
[16:27] prices and inflation and the decreasing
[16:30] value of the Turkish lera. And then in
[16:32] the security issue treatment, the
[16:34] respondents also read that same prompt
[16:37] about the poor economy. But then they
[16:40] were told to imagine that the government
[16:43] made a statement on the matter. Turkey
[16:45] and the region has been facing serious
[16:47] security challenges. Turkeykey's strug
[16:49] struggle for survival is more important
[16:51] than the increased cost of living. So
[16:53] that's the security issue treatment. And
[16:55] then the morality issue treatments give
[16:58] the respondents information about the
[17:00] economic situation. But instead of
[17:03] talking about the security situation,
[17:05] they say, "Imagine that the government
[17:06] has been saying the government has been
[17:08] restoring dignity to Turkeykey's Ottoman
[17:10] heritage by converting the Hagya Sufia
[17:12] back into a mosque. Turkeykey's struggle
[17:14] to restore dignity to Turkeykey's
[17:16] religious heritage is more important
[17:18] than the increased cost of living."
[17:20] >> Yeah. And then for the the last morality
[17:23] issue is quite similar but instead
[17:24] focused on the Istanbul convention
[17:26] issue.
[17:27] >> Fantastic. Yes. You ask people to read
[17:30] these little bits of text and then so so
[17:33] the first text just about the economy,
[17:36] second text little bit on the economy
[17:37] but also the security threat and then
[17:39] the third and fourth little bit about
[17:41] the economy and then the two different
[17:43] moral issues. And essentially you're
[17:46] just kind of reminding people or
[17:47] prompting people to think, oh, security
[17:49] is an issue or this moral concern might
[17:52] be an issue. And then you're asking them
[17:55] a few questions about their perceptions
[17:58] of economic policy. Is that right?
[18:00] >> Correct. So the two outcomes that were
[18:02] measured in the study were the first the
[18:05] importance of the economy and second
[18:08] their respondents overall support for
[18:11] the government's economic policy.
[18:13] >> Yeah. And what you're wanting to see is
[18:16] if you remind people of the security
[18:18] threat or the moral threat, does that
[18:21] lead to a difference in how they they
[18:23] respond to those further questions about
[18:26] the economy relative to if they didn't
[18:28] get any of that and they just got the
[18:30] information about the economic
[18:32] situation. That's correct, Alan. So I
[18:36] thought that these or I hypothesized
[18:38] that these different treatments would
[18:41] decrease the importance of the economy
[18:43] and increase the support for
[18:46] government's economic policy.
[18:47] >> Yeah. So the first of those ideas is
[18:50] that you're as you you're kind of
[18:53] changing what's on people's agendas if
[18:56] you like. So you're reducing the
[18:57] salience, reducing the importance of the
[18:59] people attached to the economy relative
[19:02] to other issues and therefore you're
[19:04] also causing people perhaps to forget
[19:07] some of the economic problems and
[19:09] therefore be more supportive of the
[19:11] government position on the economy.
[19:13] >> That's correct.
[19:14] >> Great.
[19:16] Should we say anything more about the
[19:18] design of the study before we go into
[19:20] the findings? Do you want just to tell
[19:22] us what the specific hypothesis were?
[19:25] >> Yes. So the first hypothesis is that
[19:28] respondents receiving any of these
[19:30] different treatment conditions, whether
[19:31] they're security or morality, should
[19:33] have a higher support of government
[19:35] policy and a lower importance of economy
[19:38] than those in the control group. And
[19:40] this is important important because at
[19:43] the time that this was fielded, Turkey
[19:44] was going through a huge e economic
[19:47] crisis with approximately the official
[19:50] stat stats say that there was like 80%
[19:54] inflation in in the year 2022 and the
[19:58] news was constantly talking about the
[20:00] rising cost of products at the grocery
[20:03] stores. And so while we might think that
[20:06] agenda sending setting might not be very
[20:10] effective because citizens can go to the
[20:13] grocery store and see that prices of
[20:17] vegetables and fruits are high. On the
[20:20] other hand, you know, citizens want it.
[20:22] They're hungry for a a message trying to
[20:25] explain what is going on. And so the
[20:28] literature I was building upon was
[20:30] predicting that even in this economic
[20:32] crisis, agenda setting should be
[20:33] effective.
[20:34] >> Okay. So that's the core hypothesis.
[20:36] You've got a few other more specific
[20:37] hypotheses, but let's perhaps get into
[20:39] the results on the sort of core question
[20:43] of whether mentioning these security
[20:46] issues and moral issues makes a
[20:48] difference or not. On that key point,
[20:51] what did you find? So I find that
[20:54] security the security treatment had a
[20:57] strong effect on the support for
[20:58] government policy outcome while the
[21:00] hagio sophia morality treatment had a
[21:02] stronger effect on decreasing citizens
[21:05] evaluation of the importance of the
[21:07] economy.
[21:09] >> So it's a kind of curious kind of
[21:11] mixture of effects there. Do we have do
[21:13] you have any sense of what's going on
[21:15] there? Can how should we interpret those
[21:16] findings? So the Haga Sophia morality
[21:19] treatment
[21:21] had a strong effect on decreasing
[21:22] citizens evaluations of the importance
[21:24] of the economy whereas the Istanbul
[21:27] convention and treatment did not have
[21:30] statistically significant results. So
[21:32] when comparing these two issues I felt
[21:35] like that these two issues kind of
[21:37] represent different types of morality
[21:40] issues. So the Hagio Sophia issue was
[21:43] more kind of popular amongst the broader
[21:46] population. But there was a significant
[21:49] portion of the population in Turkey that
[21:51] supported the Istanbul convention and
[21:54] was against the government overturning
[21:56] it. And so that issue was less popular
[21:59] than the Hagi Sophia Morati treatment.
[22:01] when it comes to why I didn't find
[22:04] effects on for both outcomes for for
[22:07] security I talk about how so a previous
[22:10] paper that that I'm extending by Salem
[22:13] Erdom itch foundings for security when
[22:16] that survey when his survey was fielded
[22:18] in 2019 during an economic crisis but
[22:22] 2022 was even more of an economic crisis
[22:24] and so while I was expecting to have
[22:27] results for security in for both
[22:30] outcomes
[22:31] In the paper I suggest that that first
[22:35] 2022 was a larger economic crisis than
[22:38] in 2019 and also the discussion on
[22:41] security issues. You have to remember
[22:43] that Russia had just invaded Ukraine at
[22:45] that time. And so this was a changed
[22:48] kind of environment in terms of kind of
[22:51] security issues. And so this is an
[22:53] important kind of context that I think
[22:55] might explain but future work can really
[22:58] probe in to try to better understand
[23:00] these. But those are two of the
[23:02] explanations that I try to offer in the
[23:04] paper.
[23:04] >> It's really interesting though. It kind
[23:06] of it gives us at least very preliminary
[23:09] suggestions that there are some
[23:11] circumstances in which politicians
[23:13] leaders can use morality politics and
[23:16] indeed security threats in order to
[23:20] distract attention from other matters.
[23:24] But there's only so far they can go
[23:26] here. If the other matters are really
[23:28] strong and really kind of forceful in
[23:30] people's minds, then people are not
[23:33] going to be distracted away. So it
[23:35] sounds like you're suggesting that
[23:37] combining your study with previous work
[23:39] done by others, we can start maybe to
[23:41] see a pattern of that kind. The other
[23:44] point to emphasize is that I also had a
[23:47] hypothesis too that was looking at kind
[23:50] of these treatment effects on specific
[23:53] subsets of the population. And I did
[23:56] find evidence for among high religiosity
[23:59] citizens that all three treatments of
[24:02] the security and both the morality
[24:04] issues amongst high religiosity citizens
[24:08] had an effect a statistically
[24:09] significant effect. Yeah. on on shift on
[24:12] shifting their support for government
[24:14] economic policy.
[24:15] >> Yeah. Which is kind of what you would
[24:16] expect given the mechanism that you
[24:18] outlined earlier on that if it's that
[24:20] that people's sense of religious
[24:22] identity in a way is being threatened by
[24:24] by what the government is saying it is
[24:27] working against then more religious
[24:29] people would have stronger sense of
[24:30] religious identity and therefore they
[24:32] would be more affected by that. You
[24:33] would expect that.
[24:34] >> That's correct.
[24:35] >> Yeah. Okay. Really interesting. So that
[24:38] is the specific study. We could go a
[24:40] little bit further. There are some more
[24:41] hypotheses there, but time is slightly
[24:44] against us. So, let's just think about
[24:46] do these findings generalize. So, you're
[24:49] focusing in on Turkey at, as you've
[24:52] said, a very particular point in time in
[24:55] 2022.
[24:56] Would you expect these kinds of findings
[24:58] to extend to other places to other times
[25:02] as well? What are you able to say about
[25:04] that? So an important scope condition
[25:07] for the argument of the ability of
[25:10] morality issues to to have an effect in
[25:13] terms of agenda setting is I think that
[25:15] we need a religious nationalist context
[25:18] right where morality issues have
[25:20] salience and so in in the paper I say
[25:23] that future work should explore the
[25:25] generalizability of these findings
[25:26] beyond Turkey perhaps in places such as
[25:29] India, Israel, Poland, Russia and other
[25:32] countries with a dominant religious
[25:33] tradition and highly religious
[25:35] populations.
[25:36] >> Yeah. And would you expect these
[25:38] mechanisms to extend beyond kind of
[25:41] authoritarian leaning settings as well?
[25:43] I mean, you know, the examples that that
[25:45] I talked about right at the start of the
[25:47] the episode were from the UK in the
[25:51] 1980s, the US in 2001, not cases that we
[25:54] would particularly think of as examples
[25:56] of democratic backsliding or anything
[25:58] like that. So the idea that rallying
[26:01] around the flag effects function in well
[26:05] functioning democracies is I think
[26:07] clearly established. Would this would
[26:11] morality politics be usable in those
[26:13] settings too? I think an important kind
[26:16] of difference between full democracies
[26:19] and electoral authoritarian or fully
[26:21] authoritarian systems or regimes are
[26:23] that in authoritarian regimes the
[26:25] government has greater control over the
[26:27] media and because of that I think that
[26:30] the and some work that I've read by
[26:32] others have suggested that these agenda
[26:35] setting effects can have longer
[26:37] longerlasting duration than those in
[26:40] democracies.
[26:42] >> Okay. Interesting. Good. something
[26:44] further to study, no doubt, in in future
[26:46] work. Final question then. You mentioned
[26:49] at the start that this work comes from
[26:52] your PhD thesis and that you're now
[26:54] turning that PhD thesis into a book
[26:57] looking at morality politics in
[26:59] autocratizing societies.
[27:01] What's the overall thesis of the book?
[27:05] So the overall idea is that morality
[27:08] politics can play two vital strategic
[27:10] purposes for authoritarian governments.
[27:13] Allowing them to first neutralize
[27:15] institutional veto players when they're
[27:17] coming to power. So in an initial phase
[27:19] when they're trying to establish
[27:20] themselves and centralize and
[27:22] consolidate power and second is to serve
[27:24] as a tool of diversion to distract
[27:27] citizens from crises such as the
[27:29] economic recessions which I discussed in
[27:32] this paper today. And this is during
[27:34] like a second phase which I call the
[27:36] regime maintenance phase.
[27:39] >> Will listeners have to wait long until
[27:41] the book comes out?
[27:42] >> Well, I'm working on on a book currently
[27:45] and looking to secure a contract and so
[27:48] I'm hoping that the book would come out
[27:50] in the next year or two.
[27:51] >> Well, let us hope so. Daniel, this has
[27:53] been really interesting. I've found this
[27:55] such an engaging conversation. you've
[27:57] you've taken a concept that I guess is
[27:59] quite familiar to quite a few people
[28:01] around the idea of the security threat
[28:02] being used in these ways and you've just
[28:04] kind of extended that and explored that
[28:06] in new ways and revealed important
[28:08] evidence on these matters. So, thank you
[28:10] so much.
[28:11] >> Thank you so much for having me. We have
[28:13] been discussing the article rallying
[28:15] around the mosque or flag the effects of
[28:17] morality and security agenda setting on
[28:20] policy performance in Turkey by Daniel
[28:23] Schulty which was published over the
[28:25] summer in the journal Mediterranean
[28:28] politics and we will as ever put details
[28:31] in the show notes for this episode.
[28:34] Next week, amidst the latest COP climate
[28:37] summit, we will be looking at a new book
[28:40] that examines how the international
[28:42] community is trying and failing to
[28:45] address the worst effects of climate
[28:47] change. To make sure you don't miss out
[28:50] on that or other future episodes of UCL
[28:52] Uncovering Politics, all you need to do
[28:54] is subscribe. You can do so on Apple,
[28:56] Google Podcast, or whatever podcast
[28:58] provider you use. And while you're
[29:00] there, we'd love it if you could take a
[29:02] moment of time to rate or review us,
[29:04] too.
[29:06] I'm Alan Renick. This episode was
[29:08] produced by Elanar Kingwell Banham and
[29:10] Kaiser Kang. Our theme music is written
[29:12] and performed by John Man. This has been
[29:15] UCL Uncovering Politics. Thank you for
[29:17] listening.
[29:25] >> [music]