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