Hidden Agenda Interactive

We’ve added our interactive political web tool to this blog. You can click on it over on the left-hand margin under the book links. Basically, you enter various demographic info, and it tells you some average survey responses of Americans with those features.

I’ve also got a new post over at Psychology Today explaining how to think about the tool. In short, it’s not about trying to predict you; it’s an accurate description of a large database.

Big thanks to our programmer, Nate Weiss, who took my ridiculous spreadsheet of regression formulas and turned it into something genuinely cool.

Abortion and self-interest

Bryan Caplan, true to his word as usual, has a new post on my book with Rob Kurzban, this time focused on a topic near and dear to my heart — abortion attitudes, which were the subject of my doctoral dissertation a dozen years ago.

Oh, the stories we tell

Abortion attitudes are one of the hardest political areas to puzzle through. To begin with, there are powerful red herrings everywhere. Pro-choice people talk about it as an issue of women’s rights, even though women and men don’t differ much on abortion. Pro-life people talk about it as an issue of Biblical fidelity, even though the Bible is silent on abortion and ambiguous about when life begins.

On pages 60 to 64 of our book, Kurzban and I place these kinds of abortion stories in the wider psychological context. People quite frequently have nice-sounding stories (“I’m protecting women!” “I’m protecting children!”) about their own motives that, while genuinely believed by their conscious mental Spokespersons, nonetheless are inaccurate spin covering up the actual motives of their unconscious mental Boards of Directors, motives that often tend towards self-interest. The broad psychological point is one Caplan seems to agree with, though he thinks that politics is an exception to the general rule.

When trying to piece together an account of why people clash on something like abortion, then, knowing the content of someone’s verbal defenses isn’t the same thing as knowing their underlying motives. We have to look for other clues.

So what’s the behind-the-scenes mental Board of Directors up to when it comes to pro-choice and pro-life opinions? My view is that it has a lot to do (but not everything to do) with people’s conflicting sexual and reproductive lifestyles.

Abortion and one’s own sex life

To whom is the availability of abortion most useful? The initial answer is rather obvious, once I point it out. Some people are sexually very active at times in their lives when it’s really important that they not have kids. The availability of safe, legal, reasonably priced abortion services is a desirable backstop, even if they never use those services.

And, of course, there are other people for whom abortion services are much less useful. Though many in my own social circle doubt their existence, the data nonetheless show that there are plenty of Americans who wait to have sex until they’re in committed relationships, and then have lots of kids without doing much to explicitly plan the timing or number.

Lots of other people, of course, exist in places in between these prototypes.

Now, how do we find these contrasting groups in the data? Back when I studied the abortion attitudes of undergraduates, I was able to ask things like how old they were when they started having sex (or, if they hadn’t yet, what they expected for their sexual futures) as well as what age they expected to have their first child. And, indeed, I found that a major predictor of undergraduates’ abortion attitudes is, quite simply, how many years they expect to be sexually active before having kids. I had a sample of middle-aged college graduates and found the same thing retrospectively – their abortion attitudes had a lot to do with how many years they were sexually active before having children.

With larger surveys, though, one can’t pick the questions asked, and it’s tougher going. Few large surveys ask the right range of questions. We have to cobble together indirect clues.

The U.S. General Social Survey is the best large database I know for these purposes, though it’s far from perfect. We can look at number of past sex partners (seeing who has slept around a lot and who hasn’t), number of children (seeing who has had many children and who has had few), and related measures like going to bars (something strongly related to having more low-commitment sex partners and fewer current children). None of these fully captures the underlying ideas about the utility of abortion services, but add them together and they get you in the ballpark.

Another big clue is education level. The more educated people are, the more they delay having kids on average. And the more educated people are, the more is at stake from screwing up plans to delay having kids.

So far, though, we only have partial ideas about pro-choice attitudes – there are people for whom the availability of abortion services is more useful than for others. Those people should show markedly higher levels of pro-choice attitudes. But, then, why would anyone else care enough to oppose them?

Abortion and other people’s sex lives

One of the big points my co-author, Rob Kurzban, and his colleagues (especially Peter DeScioli), have emphasized is that moral attitudes have as much — probably more — to do with trying to regulate other people’s conduct as they do with regulating one’s own conduct. I’ve often thought of this as, for example: If your motivational systems don’t want you doing behavior X, then you’ll simply dislike doing behavior X; but if you’re motivational systems don’t want other people doing behavior X, then you’ll not only dislike doing it, but dislike when other people do it, and further try to get in their way and rally support by announcing that you think people who do behavior X deserve punishment.

So why would anyone care if other people got abortions? In the book, we trace this back to a wider set of moral and political issues that centrally are about regulating sexual promiscuity (including premarital sex, pornography, marijuana legalization, birth control, and, of course, abortion). For example, people trying to maintain committed relationships, particularly when they have more kids, have more to lose when other people treat sex lightly and present a constant stream of temptations. Pro-life attitudes are in part an arrow in a larger quiver meant to increase the social and legal costs of what we’ve called Freewheeler lifestyles.

Who are the people who would tend to benefit from increasing social and legal costs on promiscuity, partying, and related behaviors? People who are not themselves promiscuous, have more kids, and so on (people we’ve labeled Ring-Bearers). Who are the people who are harmed by increased social and legal costs on promiscuous partiers? Mostly it’s the Freewheelers.

So there’s clear overlap between the people who have a higher need for abortion services and the people who are harmed when others impose social and legal costs on Freewheeling. And there’s a clear overlap between the people who have a reduced need for abortion services and the people who benefit from reducing Freewheeling in others. Overall, on abortion, we’re looking for high-education Freewheelers on the liberal side, and less-educated Ring-Bearers on the conservative side.

When it comes to the GSS data, what I’ve shown (in the book and in yesterday’s post), is that we can combine a number of imperfect demographic items – sexual history, number of children, bar attendance, and education – to make pretty good predictions about abortion attitudes. Indeed, in yesterday’s post, I found that combining some of these items accounts for marginally more variance in abortion attitudes (10.6%) than does asking people whether they are liberal or conservative (9.5%).

What’s God got to do with it?

It’s also clear that abortion attitudes (and related lifestyle items) relate strongly to church attendance. The usual story is that people are taught in church that these things are bad, and so that’s why they oppose them.

But in the book (pages 81 to 88), Kurzban and I confront this story with another big database, the National Longitudinal Survey of Youth 1997. This study has tracked a big group of people born in the early 1980s with yearly surveys from their mid-teens to late 20s.

What we found was that lots of kids are raised religious, with parents going to church every week. But here’s the thing – by the time these kids reached their mid-20s, only 25% of these raised-religious kids were themselves going to church every week. Why the huge decline? We show that the likely answer involves, well, Freewheeling. Lots of raised-religious kids start partying and hooking up and shacking up in their late-teens and 20s. And when they do, their church attendance drops to about the same low level as kids not raised religious. The kids who remain in church are mostly those who largely abstain from Freewheeling and then get married.

Findings like these drive our view that churches are in large part support groups for Ring-Bearers. Churches often provide practical benefits especially helpful to young couples with children, help monitor and enforce their own member’s sexual traditionalism, and provide an organization that helps pressure the wider society to increase the costs of Freewheeling.

So, we argue, you can’t just say that going to church is the big cause of everything, because people really do stop going when it doesn’t fit their own lifestyles, and, we also show, often start going (despite having less religious parents) when it does fit their own lifestyles. The causality here is really complex.

As always: What are Caplan and I arguing about?

Caplan thinks that self-interest is usually a trivial factor in public opinion. I think that it’s often (but not always) a substantial factor (though not an exclusive one).

This is, as those of you following the discussion will know, a remarkably thorny dispute. One part of the dispute is simply over what demographic items in fact relate to issue opinions and whether those relationships are “substantial.” Another is whether those relationships indicate causal direction from the demographic items to the issue opinions. Another is whether those demographic items are in fact consistent with “self-interest,” which might be defined in different ways and might be accurately or inaccurately perceived. Another is whether, even if the demographics are causal and even if they are consistent with actual and perceived self-interest, it’s in fact evidence of self-interested motives — maybe, for example, as I seem to recall Andrew Gelman arguing (don’t ask me where), people are centrally motivated to do what’s best for everyone, but their own experiences affect their judgments in ways that create a sort of accidental selfishness.

So there’s no chance for motivated opponents to reach full agreement here — there are too many ways to defend essentially unresolvable differences.

But we can agree on some things. I say that certain sets of demographics in combination substantially correlate with certain political issues. I think Caplan agrees. (Though we disagree as to why — I think they are evidence of self-interested motives; Caplan does not.) On abortion, for example, we agree that sexual history is a substantial correlate.

Caplan says that lib-con labels substantially correlate with many issue opinions. I agree. (Though we disagree as to why — I think Caplan thinks these correlations are evidence only of causal flow from lib-con labels to issue opinions; I think they’re both causes and effects of issue opinions.) On abortion, for example, we agree that lib-con labels are a substantial correlate.

I say that there are ways to introduce non-casual correlates into a regression model that can suppress the coefficients for actual causes (see, e.g., pages 227 to 235 of my book). This makes me cautious about just popping everything into a regression model and hoping God sorts it out. I think Caplan would agree if he were to acknowledge this point. So far, it looks to me like he might think regressions reveal rather than assume causal directions, which is false.

So how do Caplan’s new regression models relate to our disagreements? It turns out, not at all. He shows something we agree about: Sexual history predicts abortion views. And something else we agree about: Lib-con labels are even bigger correlates of abortion views than sexual history. And something else we agree about: Predict abortion views with both sexual history and lib-con labels, and they both remain substantial. And something else we agree about: Religiosity is a big deal here as well, and suppresses (though doesn’t eliminate) sexual history as a predictor (we showed this in our book by first running our analyses of lifestyle issues without religious predictors, and then re-running them with religious predictors).

Caplan is trying to make hay out of the fact that lib-con labels are bigger correlates of abortion attitudes than sexual histories are. He needs to be really cautious here, however. I showed in my post yesterday, for example, (1) that combined demographics are a (marginally) bigger deal than lib-con labels on abortion, and (2) that for things like school prayer and immigration, the demographics are much bigger deals than lib-con labels. If he’s going to argue that the person with the biggest correlation wins, then he’s going to lose lots of the contests. In fact, though, these are the wrong contests. We agree that lib-con labels are often big correlates. The main disagreement is about whether self-interest-related demographics are also big correlates.

And we’re back to our starting point. I think the demographic correlates we’ve identified help show that self-interest is often a big deal in public opinion. Caplan doesn’t, not because he thinks these things don’t correlate, but because he thinks the correlations don’t signal self-interest. I think I’ve made some progress in helping Caplan to see what some of those correlates are (on abortion, for example, I don’t think it had ever occurred to him to check Freewheeler/Ring-Bearer features like sexual history). But we’re unlikely to ever agree on what those correlations mean.

Give the man what he wants

Caplan is back with more on my book with Kurzban. In this post, I’m going to try to give the man what he wants — regression models predicting issues with both key demographics and with people’s liberal-conservative labels. (FYI, some of this gets really technical; I’ll mark the worst paragraphs with the phrase “Technical points,” and those of you who don’t run data should feel free to skip those paragraphs.)

Taking stock of some things we’re agreeing and disagreeing about

Agreement: Political views are not 95% self-interest. Disagreement: Caplan thinks that self-interest occasionally plays a role but overall poorly predicts public opinion. I think it’s often (but not always) a substantial (but not exclusive) predictor.

Agreement: Self-placement on liberal-conservative labels correlates substantially with a range of issue opinions.

Technical points: We do appear to disagree on whether a simple multiple regression can solve disputes over the extent to which lib-con labels are causes or effects of issue opinions. Again, I’d point to pages 227 to 235 of the appendixes to my book for a straightforward demonstration that multiple regressions can do more harm than good in understanding things when an investigator has made mistaken causal assumptions about big correlates.

Disagreement: Whether issue opinions are one-dimensional, or, to use Caplan’s phrase, “boil down roughly to one big opinion, plus random noise.” I think that the high correlations among some issues and the near-zero correlations among others are precisely the evidence in question to show that there’s more than one statistical clump.

Disagreement: Whether entering liberal-conservative labels into a multiple regression would importantly undermine the demographic predictors of issue opinions. This point is the subject of this post. I’m going to see whether the major demographic patterns we found are importantly affected by entering lib-con labels in the models. Even if they aren’t importantly affected, however, this probably won’t help resolve my overall differences with Caplan, because he will surely stick to the next disagreement.

Disagreement: Whether or not the existing demographic predictors of public opinion count as “self-interest” (or some closely related concept).

Fine, here are some regressions with liberal-conservative labels as predictors

So, here, I’ll limit the investigation to something simple: Do the kinds of demographic correlates we talk about in the book largely go away when I control for overall liberal-conservative self-placement? I’ll leave to the side, for now, the further dispute over whether these demographic items do or don’t signal self-interest.

Technical points: For the book, we used categorical predictors and re-coded all issue variables as 100-point percentile scales. This helped make all the analyses comparable, given that, e.g., if a predictor had a coefficient of “5” in a regression model, in every case it meant “having this feature in the sample predicted a 5-percentile increase in liberalness on the issue in question, given everything else in the model.”

Abortion

One of our empirical points in the book is about how people who combine what we call Freewheeler lifestyle features with higher education tend to be pro-choice while people who combine what we call Ring-Bearer lifestyle features with less education tend to be pro-life (my prior post explains some of this).

And, sure enough, in a regression predicting abortion views in the General Social Survey, here are some results. Model 1 contains the coefficients without the lib-con “control” (i.e., the GSS’s POLVIEWS variable) and Model 2 contains the coefficients with lib-con in the model.

1 2
No sex partners since age 18 -5.7 -5.5
1 sex partner since age 18 -5.6 -4.3
5 or more sex partners since age 18 8.6 7.3
Not been to a bar in the past year -6.8 -6.1
3 or more children -5.1 -3.9
High-school diploma or more 9.2 9.9
Graduate degree 11.4 9.2

Technical points: My regressions often include overlapping predictors. E.g., if you want to know the typical responses for someone with a graduate degree, you’d include the coefficient for “High-school diploma or more” added to the coefficient for “Graduate degree.” And, e.g., in the income model later, if you want to know about someone in the top 10% of family income, you’d add the coefficients for both “Family income in the top 20%” and “Family income in the top 10%.”

So, in this case, people who have had five or more sex partners, go to bars, have fewer than three children (all Freewheeler features), and have lots of education are really likely to be liberal on abortion. People who have had one (or no) partner, don’t go to bars, have lots of kids (all Ring-Bearer features), and have less education are really likely to be conservative on abortion. And the results change very little by adding people’s overall lib-con self-placement into the model.

Technical points: The combined demographic features by themselves have an r-squared of 10.6. Lib-con labels by themselves have an r-squared of 9.5. Putting it all in the same model has an r-squared of 17.5. What this means is that the unique variance of the demographic features is 8.0 (i.e., 17.5 minus 9.5); the unique variance of lib-con labels is 6.9 (i.e., 17.5 minus 10.6); and then there’s an extra 2.6 that is overlapping variance that can be accounted for by either these demographics or by lib-con labels.

The bottom line is that both demographics and ideology are big predictors of abortion attitudes, and neither dominates the other in a regression model — indeed, there’s relatively little overlap between these predictors.

School prayer and immigration

Another empirical point in our book is that discrimination issues are often predicted by a combination of the group-based category under consideration with education and test performance.

And, sure enough, in a regression predicting views on school prayer, here are some results (again, Model 1 is without the lib-con predictor and Model 2 is with).

1 2
Not Christian 16.5 14.7
Bachelor’s degree or more 8.1 8.1
Test performance in top 20% 10.0 9.8

In short, for school prayer, we often find liberals among non-Christians and people who test well and have more education; the conservatives are especially likely to come from Christians with less brains. And it doesn’t change much by adding the lib-con label variable to the model.

Technical points: In line with the calculations from the prior technical note, the unique variance accounted for by demographics here is 12.0; the unique variance accounted for by lib-con is 1.5; the shared variance accounted for by either is 2.4.

The bottom line with school prayer is that demographics are a much bigger deal than lib-con labels.

Here are further results, this time from a regression predicting views on immigration:

1 2
Born foreign 9.2 9.4
Parents born foreign 6.7 6.2
White -7.7 -6.7
Latino 10.4 10.1
Bachelor’s degree or more 10.0 9.8
Test performance in top 20% 7.5 6.9

This time, the liberals tend to come from Latinos, immigrants, and those who test well and have more education; the conservatives are especially likely to come from native-born whites with less brains. And, again, the coefficients don’t move much when “controlling” for lib-con labels.

Technical points: The unique variance accounted for by demographics here is 11.7; the unique variance accounted for by lib-con is 1.5; the shared variance accounted for by either is 1.2.

The bottom line with immigration, as with school prayer, is that the demographics are a much bigger deal than lib-con labels.

Income redistribution

Another empirical point in our book is that income and race predict people’s views on whether government should redistribute income (here, the GSS’s EQWLTH variable). Here are regression models (again, without and with the lib-con variable):

1 2
Family income in bottom 20% 3.6 3.6
Family income in bottom 40% 3.6 2.8
Family income in top 20% -3.6 -3.3
Family income in top 10% -6.9 -7.0
Personal income in top 20% -3.9 -4.1
White -10.5 -8.6

Race is a big deal, but so is the cumulative effect of these income categories (i.e., people in the bottom 20% of family income and people in the top 10% of family income are 17.7 percentile points away from each other in Model 1 (adding together 3.6 + 3.6 + 3.6 + 6.9)). And, again, adding lib-con labels to the model doesn’t change things much.

Technical points: The unique variance accounted for by demographics here is 5.0; the unique variance accounted for by lib-con is 7.5; the shared variance accounted for by either is 1.3.

Unlike the prior models, here lib-con accounts for more variance as a stand-alone (8.8) than the demographics do as stand-alones (6.3). But if Caplan wants to argue that 6.3 is really small, he needs to simultaneously tell us why the stand-alone variance predicted by lib-con labels when it comes to school prayer (3.9) or immigration (2.7) are somehow not small.

My point all along has been that these demographic relationships are substantial by the usual standards, even if things like lib-con labels are also substantial correlates. Now I can report that it doesn’t matter much if we add lib-con labels to the demographic models — the demographics still get a substantial share of the variance (sometimes a bit less than lib-con, sometimes a bit more, sometimes a ton more).

And the point remains that, despite the lib-con correlations, general public opinion is really, really not “one-dimensional.” For example, while lib-con by itself predicts 9.5 percent of the variance in abortion and 8.8 percent of the variance in income redistribution, abortion by itself predicts only 0.2 percent of the variance in income redistribution — which is about as close to zero as these things get.

But, as I said earlier, none of this will settle my disputes with Caplan. I think he might have really thought that we weren’t putting lib-con in our models because we were hiding something. We weren’t hiding anything. I’ve explained at length — both in the book (e.g., pp. 15-21) and in replies to Caplan — why we didn’t “control” for the usual suspects like lib-con labels. The analyses in this post have added something else: even if we had put lib-con in all our models, it really wouldn’t have mattered much.

Having seen what will surely be disappointing results from these analyses, Caplan will probably retreat back to “but these demographics don’t say anything about self-interest.” He might also try a different range of “controls,” even though his comments so far have focused almost exclusively on liberal-conservative labels. Which is fine. I’m becoming resigned to the fact that I’m unlikely to convince Caplan of much.

Addendum

Caplan appears to still be worried that I’m hiding the ball, so here are a new set of regressions. In these, I’ve replaced the categorical predictors with more typical continuous measures for education and income. I’ve also combined lifestyle features into a single Freewheeler scale (if Caplan wants me to admit that the various components here (sexual history, drinking at bars, number of children, etc.) are small, I hereby admit it; the explanatory power is in combining these kinds of features into an overall lifestyle profile).

This time I’m reporting standardized coefficients for all predictors in the model. (All issue variables are coded such that higher values indicate more liberal positions.)

Abortion School Prayer Immigration Income redistribution
.199 Freewheeler

.170 Education

-.258 Lib-Con

-.242 Christian

.197 Education

-.119 Lib-Con

.189 Immigrant family

.097 Latino

-.092 White

.156 Education

-.125 Lib-Con

-.171 Family income

-.122 White

-.286 Lib-Con

 

Sexual politics and self-interest

Responding to Bryan Caplan has become part of my daily routine. In today’s episode, we’re talking about chapter 4 of my recent book with Rob Kurzban.

In that chapter, we discuss our work on how differences in sexual and family patterns relate to differences in religiosity and in sexual and family morals and politics. This is something we’ve been working on for a while now, including my dissertation on abortion attitudes, some material in Rob’s first book, and a series of journal articles on religion and on drug attitudes.

In the book, as Caplan says in his post today, Kurzban and I use the terms Freewheelers and Ring-Bearers to describe different broad lifestyle patterns. Freewheelers sleep around more, party more, get and stay married less, have fewer kids, and so on. Ring-Bearers go in a more traditional direction. We show that these kinds of lifestyle differences really do predict average differences in the public’s views on political and moral topics relating to premarital sex, pornography, abortion, birth control, and marijuana legalization.

Caplan doesn’t dispute our data. Instead, at issue is whether these political and moral differences count as self-interest. Kurzban and I argue that it’s self-interest when people who party and sleep around tend not to want others imposing higher social and legal costs on people who party and sleep around. We argue that it’s self-interest when people who want to sleep around while delaying having children tend to defend access to family planning. We argue that it’s self-interest when Ring-Bearers, whose own lives can be more substantially disrupted by Freewheeler behaviors, tend to seek to reduce those behaviors by making them more costly.

I’ve found that many people (from internet commenters to the op-ed board of a major newspaper) find this all rather “obvious” once it’s laid out. But not Caplan — he thinks it’s “bizarre.”

Of despots and ordinary folks

His main argument is about how it would really be better for any given promiscuous man to be the only promiscuous man. If it were just self-interest, each man would seek to minimize the promiscuity of other men while maximizing his own. Indeed, there’s a long line of evolutionary research showing something along these lines — when individual men achieve despotic power, in fact they often do impose rules that give themselves sexual access to lots and lots of women while imposing very strict limits on other men.

But the thing is that almost no one is a despot. Even in pretty small ponds, very few people are big enough fishes to set such one-sided rules. For ordinary folks, as we explain in our book, announcing to others that you think they should be punished for some behavior in fact makes it more likely others will increase their punishments if you engage in that behavior. Hypocrisy carries added costs. Ordinary folks can’t really have it both ways when it comes to setting general moral rules.

So, yes, I agree that individual men might want to be the only Freewheeler man. The problem is that that’s not usually a live alternative.

What’s self-interest got to do with it?

So let’s assume that people — for whatever reason — differ in their tendencies to lean in a Freewheeler or a Ring-Bearer direction. And let’s assume that, for most people, attempting to impose moral and legal punishment on others for Freewheeler behaviors makes it more likely others would increase punishment for the moralizer for those same behaviors.

The central issue then is whether or not we should expect self-interested Freewheelers to be more likely to support liberal lifestyle policies than self-interested Ring-Bearers.

Caplan says that, because each Freewheeler man wants to be the only Freewheeler man (surrounded by lots of Freewheeler women), our point is defeated. In fact, his argument misses the point. We’re comparing Freewheelers and Ring-Bearers (in a population where almost no one gets to set despotic rules). On average, if a man leans Freewheeler, it’s clear that, on balance, he has more incentives to support general rules lowering costs on Freewheeling than does a man who leans Ring-Bearer. Caplan may think it’s a close call for lots of Freewheeler men, but even if this were true (and I don’t think it usually is), this wouldn’t mean that Freewheeler men have the same average incentives as Ring-Bearer men. Caplan calls this a just-so story, but usually in social science we call it a theory. And, in this case, it’s a theory that is consistent with the data.

Hypocritical spouses

Caplan’s parting shot at undermining our argument is about how Ring-Bearers would be better off under liberal lifestyle rules because it helps distinguish between honest Ring-Bearers and Freewheelers-in-Ring-Bearer’s-clothing. Caplan and I presumably agree that Ring-Bearers usually benefit from others behaving like Ring-Bearers. But Caplan’s argument adopts a sort of fatalistic view of lifestyle behavior, as though there are no incentives that could make someone behave more like a Ring-Bearer or less like a Ring-Bearer.

My view is that incentives matter. If people adopt general societal rules that make Freewheeling more costly, fewer people will engage in Freewheeling behaviors. Further, as Kurzban and I discuss in the book, when religious conservatives loudly announce their support for punishing these behaviors, and embed themselves deeply within social networks that share this punitive attitude, their own incentives not to stray are tangibly increased.

There’s no great puzzle here. Ring-Bearers often want to minimize Freewheeling. So they tend to support conservative lifestyle policies that increase the costs of Freewheeling. The people on whom these costs are placed tend to oppose such costs. There are other details of course — there always are — but the core idea is pretty simple, and I don’t see anything that threatens it in Caplan’s arguments.

Caplan accuses Kurzban and me of bending over backwards to make our theory fit the data, but I think this is another one of those cases where Caplan doth protest too much. After all, he’s defending a view that says, despite self-interest mattering in lots of ways in human life, one of the big areas where self-interest doesn’t matter is (of all things) politics.

Caplan’s One Dimension

Bryan Caplan is back with more thoughts on my book. He tries a “noise” defense of his position that public opinion is “one-dimensional,” complains that we tested his actual statement rather than something else, and ends up ignoring important facts. Here we go.

We quoted Caplan, who said on p. 153 of The Myth of the Rational Voter: “There are countless issues that people care about, from gun control and abortion to government spending and the environment. …  If you know a person’s position on one, you can predict his view on the rest to a surprising degree.  In formal statistical terms, political opinions look one-dimensional. They boil down to roughly one big opinion, plus random noise.” (Caplan titles his recent post “Plus Noise!” and, sure enough, when he quotes our quote of his quote, the word random is somehow left out. I’m not sure this matters, though, given that calling it noise implies random.)

So, Caplan says it right there: “If you know a person’s position on one, you can predict his view on the rest to a surprising degree.” In our book, we set out to give an example that undermines this kind of statement. We show that if you know a person’s position on either income redistribution or same-sex marriage, you in fact can’t predict very well their view on the other.

Apparently our efforts have paid off, to a point. Caplan now agrees with us that if you know a person’s position on certain issues, you really can’t predict his views on certain other issues. For example, Caplan now admits that the correlation between gun control and abortion in the General Social Survey is near zero. This directly contradicts his statement in his book, unless what he meant by “to a surprising degree” is in fact “sometimes not at all.”

Noise and “one-dimensional”

But Caplan’s not backing down. He agrees with our take on the data, but maintains his overall point about “one-dimension.” His main defense is about noise. Yes, he says, everyone knows these are noisy relationships. He says that individual issue positions are especially noisy.

But here’s the problem with that. The correlation between views on abortion and views on school prayer is pretty big (around .27 in the 2002-2012 General Social Survey sample). The correlation between views on income disparity and views on spending levels on the poor is pretty big (.31 in the GSS). But the correlations between abortion or school prayer, on the one hand, and income disparity or spending on the poor, on the other hand, are really small. Indeed, for example, the correlation between school prayer and spending on the poor is actually zero in the 2002-2012 GSS sample.

So it’s not that individual issue positions never correlate with each other, but that a given issue often correlates with some issues but not others. (We talked about this on pages 12 to 13 of our book, in the section Caplan’s current post discusses.) Caplan says: “When you correlate two noisy things with each other, you get a really tiny correlation.” OK, so why does school prayer have a big correlation with abortion but a near-zero correlation with income disparity?

What this means, then, is exactly the point at issue: public opinion is not one-dimensional. There are some issues that hang together pretty well with each other (e.g., religious and lifestyle issues), and other issues that hang together pretty well (e.g., issues relating to redistributive economic programs), and yet these two sets of issues don’t relate strongly to each other. That’s what things look like when public opinion has more than one dimension.

And it’s not that there are just two dimensions. Views on immigration, for example, don’t relate very strongly to either views on abortion or views on income disparity.

Caplan wants his comments about noisy data to buttress his claim about one-dimension. I certainly agree that the data are noisy. But that’s not enough to defeat the obvious fact that diverse political issues form more than one statistical clump. And it’s not enough to defeat the obvious fact, which Caplan now admits, contrary to his statement that we quoted from his book, that some important political issues have near-zero correlations with each other.

Changing the subject

Caplan then proceeds to change the subject: “If Weeden and Kurzban really wanted to dispute the one-dimensionality of political opinion, they should have been correlating specific issue views with ideology, not specific issue views with each other.” In other words: To dispute Caplan’s point about one-dimensionality, we shouldn’t have tested his own claims about one-dimensionality, where his go-to point was that individual issues predict each other to a surprising degree. Instead, we should have tested a different point. Right…

The thing is that it’s not enough to just correlate individual issues with ideological labels. It’s perfectly possible that if pro-choice people tend to call themselves “liberal,” and if pro-redistribution people tend to call themselves “liberal,” you can still get ideology-issue correlations without the issues having much to do with each other. As I said in my earlier response to Caplan, these ideological labels can be in large part effects in addition to causes of issue opinions.

In short, the correlations between issues and ideological self-placement can’t address the point about dimensionality. Dimensionality is about the old political science notion of “constraint,” about “what goes with what,” about knowing that a person who leans in a given direction on issue X also leans in a predictable direction on issue Y. Which takes us right back, of course, to Caplan’s own original statement and our test of that statement.

Caplan’s Conspiracy Theory

Bryan Caplan, the GMU economist and leading libertarian thinker, has posted his review of my book with Rob Kurzban, The Hidden Agenda of the Political Mind. Given that Kurzban and I, to use Caplan’s phrases, “frontally attack” a position that Caplan “strongly endorse[s],” it’s not surprising that Caplan’s review is negative. He playfully suggests it’s in his own “self-interest for the book to be widely read” given that the book says nice things about him personally, but that’s probably not the whole story – while no sane person who has followed Caplan’s work can doubt his tremendous intellect and energy, I nonetheless think he gets some central things wrong about public opinion, a point that many would find convincing if they read the book.

What are we arguing about?

There’s a real danger here of talking past each other. In his review, Caplan points to his lecture outline on self-interest in politics. In it, he says he interprets “‘people are self-interested’ as ‘on average, people are at least 95% selfish.'” This is an incredibly high bar.

Kurzban and I are in fact arguing against claims like the one in the first paragraph of Caplan’s review: “While self-interest occasionally plays a role, it poorly predicts both issue positions and voting behavior.” Our response isn’t that self-interest always or almost always matters to the exclusion of everything else, but instead that self-interest does a good deal better than “poorly” predicting.

This difference leads to some confusion. Caplan, for example, claims that we do not admit that “non-interest-based ideology” affects people’s views. This is contradicted by our explicit remarks on this topic. For instance, on page 206 we wrote: “When it comes to party identifications and ideological labels, we think they can exert substantial causal influence on a range of judgments.” This admission doesn’t compromise our central point, though, since we never said that ideology doesn’t matter at all or is entirely about self-interest. Instead, our project was to look at whether self-interest has substantial effects on public opinion, regardless of what else might matter.

Based on our analyses, my view is that lots of people have self-interested issue positions lots of the time. It’s not nearly 95%, but there is vast space between that bar and “poorly.” Kurzban and I are responding to claims that self-interest rarely matters much with a rebuttal that it often (but not always) matters quite a bit (more in some cases than others). Self-interest is one of the big pieces in a complex puzzle.

So, Caplan and I agree that people aren’t 95% selfish. But Caplan and I disagree over whether self-interest is a trivial factor in public opinion.

Making one’s case

Caplan says that we “trumpet a strong, incredible thesis, then ‘interpret’ virtually every fact to fit it” and that we “never clearly state what would count as evidence against their view.” I could say the same of Caplan’s book, The Myth of the Rational Voter.

Our agenda was not exclusively to try to falsify our own view, just as Caplan’s agenda was not exclusively to try to falsify his own. I look forward to Caplan’s future posts indicating where we have gone wrong as an empirical matter. However, a key question we wanted to address is whether the evidence supports the strong claims that have specifically been made with respect to whether self-interest is a poor predictor of political opinions.

Here are some examples. In Caplan’s lecture outline, he says: “Unemployment policy – The unemployed not much more in favor of relief measures.” It’s an empirical claim. When we checked the data in our book, we found that the unemployed were about 29 percentage points more supportive of unemployment benefits than those working full time. Similarly, Caplan says: “National health insurance – The rich and people in good health are about as in favor.” In our book, we found that wealthier people with health insurance were 19 percentage points less supportive than poorer people without health coverage when it comes to opinions on whether it ought to be the government’s responsibility to provide health coverage.

Are these effect sizes “poor” or are they “substantial”? This question depends on how one attaches numbers to words. Still, we feel very safe using Caplan’s own view of things. Indeed, later in his outline he refers to the 22 to 35 point gap in party affiliation by race as “massive” and the 6-point gender gap in party affiliation as “marked.” So, on Caplan’s metrics, the unemployment gap ought to be considered “massive” and the health insurance gap ought to be considered somewhere between “marked” and “massive.”

If these sorts of analyses do not count as evidence against Caplan’s view, it’s unclear what would.

Definitions and tautologies

Caplan spends much of his time criticizing our view of “inclusive interests.” Kurzban and I have a new post over at This View of Life (written before we saw Caplan’s review) that addresses much of this, so I’ll point the reader there and just give some brief thoughts here.

A problem in thinking about self-interest in politics is that the standard definitions of “self-interest” and “group interests” don’t make sense. In his lecture outline, for example, Caplan says: “Drawing on evolutionary psychology, I interpret altruism towards blood relatives in proportion to shared genes as self-interest.” We agree, as do most political scientists. But this puts a “group” squarely in the middle of “self” interest. So we use the biological term “inclusive” in part to signal this.

We also take evolutionary work seriously on the breadth of typical human motivations. These include not only short-term economic outcomes, but also social status fights (which also have obvious economic implications) and fights over sexual and reproductive lifestyles (which are mostly not directly economic).

In short, what are we to call it when individuals express opposition to discrimination against their own race, or when individuals express opposition to discrimination against their own sexual orientation, etc.? How are these cases importantly different from individuals expressing support for lots of people getting unemployment benefits based on their own unemployment? And what are we to call it when people who sleep around and want to delay having children oppose condemnation of promiscuity and limits on family planning? Maybe these examples aren’t “self-interest” on some definitions, but they’re surely closely related to self-interest. More to the point, they’re obviously not things that negate evidence of self-interest.

Relatedly, Caplan says we believe that one’s allies and social networks consist of “millions of people.” This is false. Our point (pp. 38-40) was that individuals often have a tangible stake in what happens with the actual circle of non-related people with whom they share benefits and burdens (mostly including romantic partners and close friends, but also leaking over into co-workers, fellow church members, etc.). This may be indirect self-interest, but it’s certainly not the absence of self-interest.

In our data, the actual variables are all about effects of personal characteristics on political opinions; we don’t include measures of the characteristics of one’s social networks. Though, obviously, one can make statistical assumptions. For example, it’s clear that African Americans are poorer on average. Importantly, then, even when individual African Americans are wealthier, they typically have family members and friends and neighbors who are poorer than the social networks of white Americans who are similarly wealthier. I believe it would be nutty to think that this doesn’t relate to individuals’ self-interest (including family members and indirect effects of friends).

This doesn’t lead to a tautology. We trace through the evolutionary background that focusses our attention on social status and sexual matters in addition to short-term economic matters. And we freely admit in the book that there’s a definitional looseness at work. But it’s a looseness that has always existed in the self-interest literature – we’re just calling attention to it. Again, e.g., if individuals who are religious minorities tend to oppose discrimination against religious minorities, it’s incoherent to say that that’s evidence of an absence of self-interest. If individuals who engage in casual sex don’t want others imposing social costs on people who engage in casual sex, it’s incoherent to say that that’s evidence of an absence of self-interest. And so on. Further, Caplan quotes a section where we ourselves discuss issues we think don’t fit with a self-interest account – how can a tautology produce such results?

Our book is about connecting real-life circumstances to the details of public opinion. We find lots of these connections. Call it whatever you’d like.

Now you see it, now you don’t (nerd alert: skip this section if you’re not interested in methodological discussions)

How have so many political scientists missed these points? Mainly, I think, because they load up their regression models with lots of variables that are large correlates of public opinion but really might not be big causes of public opinion. We talk about this in the text of the book in chapters 1, 2, and 10, and provide an academic demonstration of one of the main points in part B of the appendix to chapter 2.

In short: Of course you can always find ways to make demographic coefficients shrink in models of public opinion by “controlling” for ideology, party affiliation, and various explicitly political scales (like right-wing authoritarianism, etc.).

Let’s say that income, race, and gender combine to predict to a degree views on redistributive issues. And then let’s say that these coefficients are greatly reduced by “controlling” for ideology, party affiliation, and so on. Does that mean that income, race, and gender never mattered in the first place? Does being a conservative Republican turn poor, minority women into rich, white men? At best, this doesn’t mean that income, race, and gender don’t matter, but suggests something about the manner in which they matter.

But the “at best” case has problems that are rarely acknowledged. Things like ideology, party, and political scales are often big correlates of issue opinions, but there are reasons to be skeptical that they are commensurately big causes of issue opinions. If lots of people choose ideological labels and party preferences in large part because of their issue opinions, then it’s really quite dicey to just pop them in a multiple regression as predictors, something we explain at length in part B to the appendix to chapter 2 in the book.

Caplan says that the book doesn’t produce multiple regressions including “a simple measure of left-right ideology as an explanatory variable” and that it “never ‘races’ its thesis against any competing view.” But there were reasons we chose not to run these races. First, I have big doubts about whether such correlates are in fact primarily causes rather than effects or non-causal siblings. Here, Caplan’s focus on the “magnitudes” of these correlations is beside the point – there are plenty of really big but non-causal correlates of political attitudes, something multiple regressions will often fail spectacularly to reveal (again, see our appendix). Second, even if these do predict issue preferences, and even if they suppress demographic coefficients in multiple regression, they can’t actually erase the demographic facts, given that most of the features we look at (e.g., race, gender, sexual orientation, education, income, etc.) are surely plausibly viewed as coming very early in the causal chain. Third, we’re not arguing that self-interest is the only or even the largest correlate of public opinion; again, we’re arguing against the view that self-interest hardly ever matters much with our position that it often matters quite a bit – the argument isn’t over whether self-interest beats all rivals in correlation size, but whether it manages to finish the “race” with a decent time.

The Hidden Agenda of Bryan Caplan

There are lots of other specific claims in Caplan’s review that I take issue with. Perhaps I’ll address some at a later date, but I’ll close now in the service of (relative) brevity.

If you’ve followed Caplan’s work, you’ll know that he is deeply committed to advancing libertarian policy objectives. His argument in service of this objective is, basically: (1) People mostly just want what’s best for society as a whole. (2) Libertarian economists know better than others what’s best for society as a whole. (3) Therefore, people should follow the policy advice of libertarian economists.

Step 2 might well be true. But, if it turns out step 1 is wrong, and that political disagreement has a lot to do with competing constituencies’ self-interest, then it’s harder to get to step 3. If low-education native-born folks tend to oppose libertarian immigration policies because they’re often worse off under these policies, then that’s a problem for the argument. Or if low-income people with low-income social networks tend broadly to favor greater income redistribution because their own circumstances are advanced, then that’s also a problem.

So I understand the impulse here to undermine the book. It interferes with his attempts to advance libertarian policies. Apparently Caplan also thinks of me as a self-deluded activist, indeed, a conspiracy theorist – someone who seeks to “replace decades of careful and curious social science with near-tautologies and just-so stories,” and that “the only ‘important’ thing about this book is that it might destroy a valuable body of knowledge.”

It’s certainly true that I think that a big wing of political scientists are wrong about self-interest (though another wing of political scientists, those who more routinely interact with pollsters and political professionals, are often much more knowledgeable about demographic interests). But this is what scholarly debate is all about. We showed our work in the book, including over 100 pages of statistical appendixes, including specific data-based refutations of key points in Caplan’s work.

Interestingly, Caplan has strongly defended the psychological position that people often follow self-interest but then cover it up behind nice-sounding stories. Yet he thinks politics is one of the key exceptions to this general rule. I see lots in the data to suggest that politics is far from a special exception, and little in Caplan’s review of the book’s “conceptual flaws” that convinces me that we didn’t find what we in fact found.