The Volokh Conspiracy

The Wall Street Journal recently published an article on widespread public ignorance about federal spending [HT: Andrew Varcoe]:

Many Americans have strong opinions about policy issues shaping the presidential campaign, from immigration to Social Security. But their grasp of numbers that underlie those issues can be tenuous.

Americans vastly overestimate the percentage of fellow residents who are foreign-born, by more than a factor of two, and the percentage who are in the country illegally, by a factor of six or seven. They overestimate spending on foreign aid by a factor of 25, according to a 2010 survey. And more than two-thirds of those who responded to a 2010 Zogby online poll underestimated the part of the federal budget that goes to Social Security or Medicare and Medicaid.

“It’s pretty apparent that Americans routinely don’t know objective facts about the government,” says Joshua Clinton, a political scientist at Vanderbilt University.

Americans’ numerical misapprehension can be traced to a range of factors, including where they live, the news they consume, the political rhetoric they hear and even the challenges of numbers themselves. And it isn’t even clear how much this matters: Telling people the right numbers often doesn’t change their views.

These are not new findings. I wrote about earlier survey data with similar results here and here. Despite the growing fiscal crisis that has emerged over the last few years, most of the public knows very little about federal spending.

The article suggests that this ignorance may not matter much because the majority of survey respondents don’t change their minds about policy priorities even when presented with correct information. It is certainly true that people are slow to change their minds about political issues, often even rejecting outright any data that conflicts with their preexisting views. In general, however, people with higher levels of political knowledge have much different views on many issues than those with low levels, even after controlling for partisanship, race, gender, income, and many other background variables. Knowing one key fact about the budget may not change your mind. But being generally knowledgeable about federal spending may well lead you to have different views from otherwise similar people who are mostly ignorant about it. Moreover, on some key issues where the balance of political power is close, there could be important effects on policy even if only five or ten percent of voters change their minds.

In this case, the public’s failure to understand that entitlements and defense constitute the lion’s share of federal spending probably makes them more reluctant to consider cuts in these areas. Conversely, the belief that foreign aid and payments to illegal immigrants are much greater than they actually are lead voters to focus their ire on these issues far more than is warranted.

Heather Mac Donald has an interesting column on the widely discussed data showing that some 87% of all Wikipedia contributors are male. She argues persuasively that it’s difficult to attribute the gap to discrimination, since most Wikipedia writers are anonymous, thereby making it virtually impossible for Wikipedia to discriminate against women even if they wanted to.

Wikipedia obviously covers many subjects. But to the extent that the underrepresentation of women there extends to the entries on politics and policy issues, it’s part of a longstanding pattern of lesser interest in such issues among women as compared to men. For example, decades of research show that there is a substantial gender gap in political knowledge, with men especially overrepresented among the 5% of the population who follow politics most closely, as measured by political knowledge levels (a group that is some 80% male). I present some of the relevant data in Part VI of this article, including the figure about the top 5%. As I describe there, the gender gap in political knowledge covers a wide range of issues. There are relatively few exceptions. Similarly, this recent Harris poll shows that 25% of men, but only 10% of women report reading at least one nonfiction book on politics over the last year. As in the case of Wikipedia, it’s hard to attribute these gaps to discrimination. Basic political information of the kind tested by pollsters is easily accessed in many different places, and booksellers would be more than happy to sell more political books to women.

Obviously, it would also be wrong to attribute the gap to “stupidity” on the part of women. Political ignorance is not stupidity. As I have often pointed out in the past, it is actually rational behavior for most citizens, assuming that their only reason to become informed is to be a “better” voter. Most of those who do acquire a lot of political knowledge do so primarily because they find it interesting, rather than because they are unusually intelligent or patriotic. Moreover, it is clearly not the case that women are generally less knowledgeable than men. Younger women today have higher average levels of educational attainment than men, and the Harris poll mentioned above shows that they also read more than men do overall.

Thus, the gender gap in political knowledge and interest in politics is likely due to lower interest among women in this particular field rather than “stupidity” or a general unwillingness or inability to acquire information. More men than women are interested in reading about politics and following policy issues closely.

Why the difference? The most obvious explanation is that politics was historically a male-dominated sphere from which women were largely excluded. Only in the last few decades has that begun to change in a major way. Elements of the old attitude surely persist, and they probably influence women’s propensity to become interested in politics. At the same time, that may not be the only factor. Despite massive changes in public attitudes on women’s role in politics over the last fifty years, the general gap in political knowledge has declined only modestly over time (I cite some of the relevant data in this article). The gender gap might therefore persist even after sexism in politics has been more fully overcome than it has been so far. One possibility is that women are simply more likely to have various nonpolitical interests than men are, which leaves less free time available to follow politics. For example, the Harris survey shows that many more women than men read various genres of fiction than men do (science fiction is an unusual, but predictable exception).

At this point, some readers will be tempted to dismiss the data by saying something like “I’m a woman and I love politics,” or “I know many women who follow politics all the time.” But the data are based on statistically representative samples of the population as a whole. For every such general pattern, there are many individual exceptions. The fact that I like Jane Austen novels doesn’t discredit the data showing that her readership is disproportionately female.

Our personal experience can be a poor guide to general patterns because we and the people we know are often unrepresentative. If you regularly read political blogs such as this one, you and your friends are likely to be much more interested in politics than the average person – male or female. The fact that women on average are less interested in politics than men is perfectly consistent with the existence of many individual women who follow politics closely and an even larger number of men who don’t. Indeed, the statistically average man is very far from being a political junkie. But among those people who do fall into that category, men are significantly more common than women.

UPDATE: In this 2008 post, I commented on the related fact that the vast majority of political bloggers are male, as are about 70-80% of political blog readers.

Although federal spending was a major political issue in the 2010 campaign and for many months before it, this recently released CBS poll [HT: Dan Mitchell] reveals widespread public ignorance about the distribution of spending between various programs. The detailed data reveal that only 23% know that Medicare and Medicaid take up between 20 and 30% of federal spending, and only 15% realize that Social Security takes up between 20 and 30%. Some 48% underestimate the extent of Social Security spending, with a much smaller percentage overstating it. Similarly, only 23% recognize that defense spending takes up between 20% and 30% of the budget. In this case, the most common error is to overestimate the extent of spending (a mistake made by 42%). Defense, Social Security, and Medicare/Medicaid, are by far the three largest items in the federal budget. And the vast majority of Americans don’t know how much of the federal budget is spent on them. Even if we count as “correct” answers that are close to the truth (on the grounds that all three programs are right around 20%, so both 10 to 20% and 20 to 30% might potentially be correct), the large majority still doesn’t know the answer in all three cases.

The majority overestimates the percentage of federal spending that goes to foreign aid, welfare, and earmarks. For example, only 9% realize that foreign aid is less than 5% of the federal budget, while 67% believe that it is higher than that, including 48% who believe that the true figure is a whopping 10% or more.

Knowing approximately how much federal spending goes to which program is not enough to have a reasonably informed discussion on spending policy. But it’s probably a necessary prerequisite to doing so. Therefore, it’s noteworthy that the majority of the public is ignorant on these points, despite the extensive public debate over the issue over the last two years.

This result is, of course, in line with previous data showing widespread public ignorance on a wide range of issues. As I have argued elsewhere, it is in fact rational for most voters to make little or no effort to acquire political information, because the chance of influencing electoral outcomes is so low. Few people find the details of federal spending policy interesting, and there is little incentive to learn about them just to be a better voter. Political ignorance is individually rational behavior that leads to bad collective outcomes. Voters may have a duty to become informed, but if so few take that obligation seriously.

It is tempting instead to blame politicians and the media for the extent of public ignorance on these issues. Certainly, politicians often talk about spending in very vague terms, and pretend that we can cut the deficit without touching entitlement programs or defense. No doubt, the media is also often uninformative. However, it is still easy for people to get accurate data on federal spending if they were so inclined. A quick google search turns up several easy to use sites such as this one and this one. It’s also not hard to find think tank studies, newspaper and magazine articles, and other information sources that discuss the budget in an accessible way.

Moreover, to the extent that politicians and the media don’t provide more information on spending, it is largely because they are responding to voter and consumer demand. If TV news viewers wanted more programs discussing the details of the federal budget, TV and radio news stations would be happy to provide it. It would be cheaper to produce than much of their current programming. Similarly, if voters punished at the polls politicians who fail to discuss budget specifics, more candidates would do so. In reality, of course, politicians are more likely to be rewarded at the polls for avoiding discussion of specific budget cuts than for embracing it. Doing the latter will earn them the ire of organized interest groups, while getting little reward from the general public. And TV news executives know that most viewers prefer entertainment to stories on the budget. In sum, politicians and the media are primarily responding to and exploiting public ignorance, rather than creating it. They may exacerbate the problem, but they didn’t cause it.

UPDATE: To my mind, the good news in the CBS poll is that it partly refutes the conventional wisdom that the public wants to cut spending in general, but opposes cuts in specific programs. Substantial majorities say they are willing to cut Social Security benefits for “retirees with higher incomes,” farm subsidies, defense spending, and money for “projects in your community.” There is a statistical dead heat (45% in favor, 48% against) on eliminating the mortgage interest deduction. The public isn’t nearly as willing to cut spending as I would prefer. But it is not correct to say that they oppose cuts in all important programs.

Of course the fact that a majority is willing to support cuts in these areas does not necessarily mean that any will actually be enacted. Social Security, farm subsidies, defense spending, and porkbarrel grants all have strong interest group constituencies behind them, which may be able to prevail against poorly organized majority public opinion.

Kelo v. City of New London, which ruled that government has the power to forcibly transfer property from one private owner to another in order to promote “economic development,” was one of the most unpopular decisions in the history of the Supreme Court. Polls conducted soon after the decision was issued in 2005 found that over 80% of the public opposed it.

Recent survey data compiled by Stephen Ansolabehere and Nathaniel Persily as part of a general study of public attitudes on constitutional issues sheds new light on public attitudes towards Kelo. The study, which is based on a poll taken in July 2009, finds that the public remains just as strongly opposed to economic development takings as in 2005. But it also indicates considerable public ignorance about the Court’s decision.

Question 215 in the 2009 survey asked respondents the following:

Governments sometimes use the power of eminent domain to acquire a person’s property at a fair market price for other uses. Recently, a local government transferred someone’s property to private developers whose commercial projects could benefit the local economy. Do you think the local government should be able to use eminent domain for this purpose or not?

This wording is quite favorable to the pro-Kelo side. It mentions the rationale for the taking (“benefit [to] the local economy”) and notes that the owners will get a “fair market price.” Respondents who are not experts in this field might believe that the latter actually means a “fair price” that takes account of the full extent of the owners’ losses, even though it only actually means “fair market value,” which is often not enough to fully compensate owners for the loss of “subjective value.” On the other hand, the question doesn’t mention any of the arguments against such takings, such as the strong likelihood that they will destroy more economic value than they create. Nonetheless, 81% of respondents said that government “should not be able” to engage in economic development takings, while only 16% concluded that it should have the power to do so. There was little disagreement between respondents with different partisan commitments or ideologies. This is almost exactly the same result as in the 2005 surveys. It suggests that public opposition to economic development takings is not a temporary artifact of the Kelo backlash, nor is it the product of question wording that favors opponents.

On the other hand, Question 301 in the same survey found that only 42% realized that the Court ruled that economic development takings were permissible, while 14% thought that the Court had ruled the other way and 43% were not sure. These figures likely overstate the true degree of public knowledge of Kelo because some people probably hit on the right answer by guessing without actually knowing it (a random guesser had a 50% chance of getting the right result). Previous research shows that a substantial minority of survey respondents prefer to guess rather than admit ignorance. Individuals who don’t even know which way the Court ruled in Kelo are probably also unlikely to keep track of post-Kelo reforms. As I argued in this article, this kind of public ignorance helps explain why so many of the latter have been ineffective. A 42% rate of correct answers is higher than we would get for most Supreme Court decisions. It also beats the mere 21% who knew (in a 2007 survey) whether their states had enacted post-Kelo reforms, and the mere 13% who both knew that and whether their state’s reforms were likely to be effective. But the 42% figure is still unimpressive for a ruling that drew such widespread press coverage and political opposition.

I do not believe that the strong public opposition to Kelo by itself proves that the case was wrongly decided or even that economic development takings are bad policy. After all, I have repeatedly argued elsewhere that public political attitudes are often the result of ignorance and irrationality. Nonetheless, the depth and persistence of public opposition to economic development takings is interesting, as is the extent to which it is coupled with widespread ignorance about the issue.

Just published at The New Ledger, an article describing the Democratic effort to label Ronald Reagan as an “extremist,” during his 1966 campaign for Governor of California. Thanks to all the VC folks who responded to my bleg a few days ago, and provided good research leads.

The Democratic Strategist (a website run by, you guessed it, Democratic strategists): “In recent days, as increasingly negative projections regarding the November election have appeared, a substantial number of Democrats have been seized with a genuine sense of panic.”

The article goes on to argue that such panic is unwarranted, as the recent polls merely reflect the movement of swing voters from the Democrats to the GOP, not a fundamental shift in political alliances by the great bulk of voters.

Earlier today, I received a robocall from a DNC-affiliated pollster asking whether I plan to vote for the Republican or Democratic congressional candidate in my district, and also asking my opinion of Barack Obama’s performance as president. The rub is that I live in Arlington, Virginia, represented in Congress by Rep. Jim Moran. Moran won reelection with almost 70% of the vote in 2008, and his district is one of the few areas that voted Democrat in last year’s gubernatorial election.

Maybe the DNC had some other reason to want survey voters in my district, but if, as the call suggests, the Democrats are worried about whether Moran is vulnerable, panic must be really setting in.

I tried using Survey Monkey for my two polls below, given the weird formatting issues I’ve been having with Pollhost, but I didn’t realize that the software puts very strict limits on how the free software is used (it only registers 100 responses per poll, and charges $200 per year if I want to use unlimited responses). That’s pretty useless for my purposes (or very expensive). So I am suspending the polls, and instead just posting the results of the first 100 responses. It’s not that revealing, given the low number of replies, but that’s the Internet for you.

First, conservatives had the following responses to why liberals have liberal views:

They lack the right values. 24%
They lack the right religious beliefs. 8.0%
They’re not as smart as conservatives. 8.0%
Liberals tend to engage in “groupthink.” 44%
They probably grew up around other liberals. 58%
Political issues are difficult, and liberal views are reasonable (even if I think they are wrong). 61%

Liberals had the following responses to why conservatives have conservative views:

They lack the right values. 19%
They have incorrect religious beliefs. 20%
They’re not as smart as liberals. 15%
Conservatives tend to engage in “groupthink.” 28%
They probably grew up around other conservatives. 63%
Political issues are difficult, and conservative views are reasonable (even if I think they are wrong).75%

For some reason, the only comments collected by the software were the comments by conservatives about liberals. These are of limited use without knowing the flip side of what liberals thought about conservatives, but here they are after the jump:

Continue reading ‘Oops, Bad Choice of Polling Company’ »

Here’s a quick poll for readers who self-identify as politically conservative. You can click on as many reasons as you think apply, and can enter any additional reasons not mentioned.

[Poll Suspended]

Here’s a quick poll for readers who self-identify as politically liberal. You can click on as many reasons as you think apply, and can enter any additional reasons not mentioned.

[Poll Suspended]

In an earlier reader poll, I asked readers who believe that the Constitution requires states to recognize same-sex marriage to indicate when the Constitution began to require it. The results were fascinating: Readers disagreed widely as to the answer, with responses spread pretty evenly among the different centuries and periods of U.S. constitutional history. What, if anything, does that suggest?

My guess is that we’re picking up the very different strands of what readers think it means to interpret the Constitution. Some readers believe that the Constitution never changes: Only interpretations change, as humankind becomes more enlightened. To most of those readers, the operative date necessarily would be when the relevant Constitutional provision was ratified. Other readers think of constitutional interpretation as interpreting precedents, and look to when cases were handed down that establish the doctrinal case for same-sex marriage. To those readers, the key date would be when the most important precedents appeared. Still other readers see the Constitution as mirroring enlightened social attitudes (or what is thought to be enlightened attitudes, anyway). I gather they focus largely on when opposition to same-sex marriage began to appear irrational to enlightened people, which is only quite recently.

Those are my guesses, anyway. Actual mileage may vary. Void where prohibited.

I had a sense of deja vu reading over the results of the VC reader poll on same-sex marriage. Here are the key results. First, as a matter of policy:

Favor same-sex marrriage: 61%
Oppose same-sex marriage: 33%
Don’t know: 6%

Second, as a matter of constitutional law:

Laws banning same-sex marriage are constitutional: 54%
Laws banning same-sex marriage are unconstitutional: 39%
Don’t know: 9%

So VC readers who responded and who have an opinion on the topic think that a law banning same-sex marriage is a bad idea by about 2-1, but a slight majority thinks such laws are constitutional.

That rough breakdown reminded me of another VC reader poll on another big constitutional question, albeit on a totally different issue. Back in January 2008, I asked readers about their policy and constitutional views about a hypothetical law permitting the death penalty for child rape (in light of a then-pending Supreme Court case, Kennedy v. Louisiana). To be clear, obviously I’m not suggesting that there is any connection between gay marriage and either the death penalty or child rape. What’s interesting to me is the relative split in how VC readers approached their views of policy versus constitutional law. Here was how VC readers voted as a matter of policy:

Percent that opposed such a law: 62%
Percent that favored such a law: 38%

Notably, there was no “I don’t know” for this option. Here was the breakdown for whether such a law is constitutional:

Constitutional 54%
Unconstitutional: 20%
Not sure: 25%

While the results are not identical, some similarities exist. 2-1 thought such laws were a bad idea, while a slight majority thought they were constitutional. I wonder if that same policy/constitutional breakdown exists for other controversial issues involving legal challenges to laws that touch on the culture wars. That might be the case: I wouldn’t be surprised if a certain percentage of readers find that their policy and constitutional views generally align and a certain percentage of readers find that their policy and constitutional views often diverge. Of course, it could just be a coincidence. It’s hard to know.

Among the interesting results of the VC readership polls on same-sex marriage, I was particularly intrigued by the answers to when the Constitution began to require gay marriage. This question was only supposed to be answered by those who believe that the Constitution presently requires state recognition of same-sex marriage. 526 readers responded so far.

On to the results. First, 24% of readers who responded believe that the Constitution began to require states to recognize same-sex marriage before or during the 1700s. I assume this crowd is mostly thinking that state recognition of same-sex marriage is a natural-law right, or perhaps just more generally that it goes back to the Founding era.

Next, 29% of readers who responded believe that the Constitution began to require states to recognize same-sex marriage in the 1800s. This was the most popular answer. The 14th Amendment went into effect in 1868; I gather that is the date most readers in this group have in mind. In contrast, only 3% of readers picked from 1900 to 1954 as the key date.

After that, 13% of readers who responded believe that the Constitution began to require states to recognize same-sex marriage in the Warren Court era from 1954-1969. I gather that many of the readers in this group are focused on Loving v. Virginia in 1965 as a key precedent. In contrast, only 6% of readers who responded thought that the Constitution began recognizing the right in the Burger Court years, from 1969 to 1986.

Moving on, 16% of readers who responded believe that the Constitution began to require states to recognize same-sex marriage in the Rehnquist Court era from 1986 to 2005. This period included Romer v. Evans and Lawrence v. Texas, and I would guess that most readers that picked this window would focus on those cases as the key precedents that helped create the right.

Finally, 10% of readers who responded believe that the Constitution began to require states to recognize same-sex marriage only in the last five years. I’m not sure, but I would guess that at least some readers in this group are focused on the changing social attitudes toward same-sex marriage in the last five years.

Anyway, I realize some readers aren’t interested in public perceptions of the Constitution. And I realize that the polling question will be ambiguous to some, which is inevitable when trying to tease out the wide range of different theories different people may have. But still, I thought the results were pretty interesting.

I wanted to do a few reader opinion polls on matters related to gay marriage. Of course, different readers will have different views of how useful these polls are. But I’m curious, so here goes. (Oh yeah, and sorry for the strange formatting: There’s some sort of incompatibility between our software and Pollhost’s, and right now I can’t figure out what it is.)

Let’s start with the basic question of whether readers favor or are against same-sex marriage as a matter of policy. For purposes of this first poll, assume that favoring same-sex civil unions but not same-sex marriage counts as not favoring same-sex marriage.

Now let’s turn to a legal question, the constitutionality of not recognizing same-sex marriage:

Now let’s get a little more specific. If the standard of review is rational basis, is there a rational basis for opposing same-sex marriage?

Continue reading ‘Reader Polls on Same-Sex Marriage’ »