The Monkey Cage

The Texas legislature passed the Top 10% Law in 1996 guaranteeing automatic admission to any Texas public college or university for seniors who graduated in the top decile of their high school class. Using data on a representative sample of seniors (N = 12,029) enrolled in 96 Texas public high schools, we examine whether and how this law affects the educational aspirations and expectations of graduating seniors, as well as whether they apply to college. Hierarchical generalized linear models demonstrate that the knowledge of a percent plan has played an important role in raising the sights of students who might not otherwise consider college. This effect is particularly pronounced for minority students, although peer, family and high school context play pivotal roles.

That’s the abstract of a new study by Kim M. Lloyd, Kevin T. Leicht, and Teresa A. Sullivan (“Minority College Aspirations, Expectations and Applications under the Texas Top 10% Law,” Social Forces, March 2008).

So the law is having a beneficial effects, right? Well, maybe or maybe not. It depends on how you go about measuring success. For example, in a 2003 study (here) of before-and-after admissions patterns at the University of Texas at Austin and Texas A&M University, Brian Bucks concluded that:

Proponents of percent plans have pointed to the comparable percentages of minorities entering UT–Austin in say 1996 (under affirmative action) and in 1999 (under the Ten Percent Plan) as evidence that such measures are similarly effective in promoting campus diversity. This paper, however, suggests a more cautious interpretation of the success of the Ten Percent Plan. First, it appears that the Ten Percent Plan and institutional responses to the Hopwood decision have had varying success in restoring black and Hispanic enrollment at Texas’ flagship campuses. Diversity at UT–Austin is at or below pre-Hopwood levels, depending on the standard used, and minority representation at Texas A&M remains below its peak shortly before the court decision. Moreover, aggressive recruiting and financial aid efforts may explain a larger portion of any improvements in minority representation than changes in students’ admissions probabilities.

…[T]he fraction of black and Hispanic high school graduates in the state has grown over the period. Accounting for this fact in examining students’ college choices reveals that minorities—and particularly those with high college entrance exam scores—are now less likely than whites with comparable SAT scores to enroll at a selective Texas public university. Certainly one of the functions of a public university system, in particular, is to serve that state’s high school graduates, and this notion underlies the philosophy of the Ten Percent Plan itself. From this perspective, however, with an increasingly diverse population of high school graduates, the Ten Percent Plan has reduced minority graduates’ access to the state’s most selective universities by failing to even maintain the ethnic composition of pre-Hopwood classes at the flagships.

Moreover, in a 2006 study (“Capitalizing on Segregation, Pretending Neutrality: College Admissions and the Texas Top 10% Law,” American Law and Economics Review), Marta Tienda and Sunny Xinchun Niu found that the Top 10% Law was working to the disadvantage of minority students who graduated from racially integrated high schools:

We show that high levels of residential and school segregation facilitate minority enrollment at selective public institutions under the uniform admission law because black and Hispanic students who rank at the top of their class disproportionately hail from minority-dominant schools. However, qualifying minority students’ lower likelihood of college enrollment at the flagships reflects concentrated disadvantage rather than segregation per se.

So has the Top 10% Law been a Ten Percent Solution? The answer, it seems, is that the law’s effects, some of which would have been hard to anticipate, have been complex and cross-cutting. Here again, then, eliminating racial and ethnic inequalities in access to higher education has proven to be a problem that simply doesn’t lend itself to an easy, formulaic solution.

Posted by Lee Sigelman at 02:22 PM | Permalink | Comments (2)

President George W. Bush receives a briefing on hurricane damage from FEMA Director Mike Brown in Punta Gorda, Florida, Sunday, Aug. 15, 2004 (source ).

In the aftermath of the summer 2004 Florida hurricane season, the Federal Emergency Management Agency (FEMA) distributed $1.2 billion in disaster aid to Florida residents. This research presents two empirical findings that collectively suggest the Bush administration engaged in vote buying behavior. First, by tracking the geographic location of each aid recipient, the data reveal that FEMA treated applicants from Republican neighborhoods much more favorably than those from Democratic or moderate neighborhoods, even conditioning on hurricane severity, home value, and demographic factors. Second, I compare precinct-level vote counts from the post-hurricane (November 2004) and pre-hurricane (November 2002) elections to measure the effect of FEMA aid on Bush’s vote share. Using a two-stage least squares estimator, this analysis reveals that core Republican voters are easily swayed by FEMA aid - $16,800 buys one additional vote for Bush - while Democrats and moderates are not. Collectively, these results suggest the Bush administration maximized its 2004 vote share by concentrating FEMA disaster aid among core Republicans.

The paper is by Jowei Chen, a Ph.D. candidate in Stanford’s Department of Political Science. Chen has amassed some fascinating data, and the use of hurricane strength as an instrument in the two-stage least squares model is quite plausible. Ultimately, the Bush administration’s strategy is constitutes evidence for some political science theories (e.g., here) that emphasize the strategic value of mobilizing core supporters.

Two notes in response to anticipated questions:

1) This is not illegal. Chen writes:

FEMA enjoys wide statutory discretion in distributing disaster aid among applicants within each hurricane-affected county, once that county has received a disaster declaration. There are few formulaic rules, and the statutory language that authorizes the disaster aid program is sufficiently broad to afford FEMA significant latitude in distributing its money. Hence, there is nothing unlawful about the Bush administration’s apparent vote buying strategy documented in the data.

2) Vote-buying did not win Florida for Bush in 2004. Chen writes:

…there is no possibility that the pro-Republican bias in Florida FEMA aid altered the outcome of the election.

Of course, the story is provocative nonetheless. The paper is here. Highly recommended.

Posted by John Sides at 09:09 PM | Permalink | Comments (4)

A daughter with her senatorial father: Meghan and John McCain

In research published in the March 2008 issue of the American Economic Review, Ebonya Washington begins by noting that “Psychologists have demonstrated a link between offspring gender and parental beliefs on … issues of political significance.” Prior studies have established, for example, that parents of daughters are more likely to support pay equity, comparable worth, affirmative action in regards to gender in employment, and Title IX policies. (Click here for the abstract of a pertinent study.)

Might this relationship carry over the highest stratum of political decision makers?

Washington finds that it does. Based on her analysis of roll-call voting in the U.S. Senate, she concludes as follows:

While the notion that a legislator’s children influence his congressional voting behavior appears commonsensical, there has, to this point and to my knowledge, been no evidence to quantitatively substantiate this intuition. This paper begins to fill this hole in the literature. I find that conditional on number of children, parenting an additional female child increases a representative’s propensity to vote liberally on women’s issues, particularly reproductive rights. Such a voting pattern does not seem to be explained away by constituency preferences,
suggesting not only does parenting daughters affect preferences, but also that personal preferences affect legislative behavior.

Consequently these results [suggest that] to the realm of environmental effects, such as peers and neighborhoods, … we should add offspring effects.Not only should we consider the impact that parents have on children’s attitudes and behavior, but we should consider that there may be reverse causality in the parental/child attitude relationship.

A second contribution of this work is to the literature on congressional voting. This paper not only provides a robustness check on the finding that ideology impacts legislative voting, it also serves to identify an additional component of that ideology: child gender composition.

Posted by Lee Sigelman at 02:41 PM | Permalink | Comments (2)

M. Marit Rehavi, a PhD candidate in economics at Berkeley (yes, economics) explores this question in Sex and Politics: Do Female Legislators Affect State Spending? She examines over 20,000 bills and finds that the dramatic movement of women into US state legislators over the past 25 years was responsible for a modest shift in the rise of state health spending but not in other areas associated with women, such as education. She also finds that “[W]omen elected in close races have a larger effect on spending priorities than the average female office holder.” Read the complete working draft here.

Posted by David Park at 01:17 PM | Permalink | Comments (0)

Two recent analyses use current state-level polling data to forecast the Electoral College vote, assuming, of course, no systematic changes in support for the Democratic and Republican nominees.

The first is by Robert Erikson and Karl Sigman, and was posted over at Pollster.com. They take a 50-state SurveyUSA poll and simulate the outcome of a McCain-Obama or McCain-Clinton race; the simulations take into account sampling error in the polls. Their findings:

…our simulations yield a 88% chance of Obama beating McCain (with 306 Electoral College votes on average versus 233 for McCain), and a 74% chance of Hillary beating McCain (with 285 Electoral College votes on average versus 253 for McCain). About one percent of our simulated outcomes were Electoral College ties. (We ignored within-state variation in Maine and Nebraska, which divide their electoral votes by district.)

The second is by Josh Putman over at Frontloading HQ. (That is his map above.) His data are different: the latest polls posted on Real Clear Politics. His results mirror Erikson and Sigman’s in that Obama is more likely to beat McCain than Clinton. However, he estimates that Clinton would actually lose to McCain:

In the McCain-Clinton contest, the solid and leaning categories give McCain a 235-179 electoral college vote advantage with 124 electoral votes falling in “toss up” territory. If you allocate those states’ votes to the candidate with the leading average, McCain wins by a 90 electoral college vote margin, 314-224…

And he finds that Obama’s victory would be narrower:

Factoring in the toss ups, Obama has a 199-174 lead over McCain with 165 electoral votes to close to call. Again, if those electoral votes are allocated to the candidate leading in the average of post-Super Tuesday state polls, Obama claims victory by a 273-265 margin.

Obviously, much could change between now and November, but these are still interesting exercises in ascertaining the current state of play, if only for cocktail party conversation.

Posted by John Sides at 09:48 PM | Permalink | Comments (3)

What are American and European attitudes toward immigration? Do they differ? Clearly, the centrality of immigration in “settler societies” such as the United States — both in terms of the literal populating of the country and in terms of its founding myths — is greater than in most, if not all, European countries. But does this make the United States “exceptional” in how immigrants are viewed?

Jack Citrin and I have a recently published paper in which we examine attitudes toward immigrants and immigration in the United States and 20 European countries, drawing on the European Social Survey and the Citizenship, Involvement, Democracy Survey.

Our results suggest that Americans do not stand apart from Europeans in terms of the perceived consequences of immigration, the desired qualities of immigrants, and the preferred level of immigration. There is, however, one difference: attitudes toward cultural diversity more generally.

The two survey items that speak to diversity asked respondents whether they agreed or disagree with these statements:

It is better for a country if almost everyone shares the same customs and traditions.

It is better for a country if there are a variety of religions among its people.

Below are plots of the percent of respondents in each country who endorse homogeneity — i.e., agree that everyone should share customs and traditions and disagree that a variety of religions is better. (See the paper for fancier plots with means and confidence intervals.)

“American exceptionalism” emerges fairly clearly. Relative to almost all European nations, fewer Americans endorse cultural or religious homogeneity, and the differences between the U.S. and these other nations are almost always statistically significant (see Figure 1 of the paper).

The paper has more discussion of these results, as well as the individual- and country-level factors underlying attitudes toward immigration. Another noteworthy finding: few of the most obvious country-level attributes — unemployment, inflation, the size of the immigrant population, the growth of the immigrant population, etc. — explain the differences among countries in attitudes toward immigrants. I’d be happy to entertain other possibilities.

Addendum: In the comments, Chris Zorn asked for a scatterplot, and I am happy to oblige.

Posted by John Sides at 06:00 PM | Permalink | Comments (6)

Who knew?

Posted by Lee Sigelman at 08:44 AM | Permalink | Comments (1)

The short answer is no:

Political campaigns are just now learning how to put the Internet to best use. Low transaction costs and huge economies of scale tempt campaigns to move traditional activities online, but the effectiveness of virtual campaigns is unknown. This paper conducts 13 field experiments on 232,716 subjects to test whether email campaigns are effective for voter registration and mobilization. Both registration and turnout were unaffected, suggesting that email, while inexpensive, is not cost-effective.

The paper (here) is by David Nickerson. The key to successful mobilization, as Nickerson argues and other field experiments confirm, is personalized contact:

These results fit neatly into the pattern of voter mobilization results where the effectiveness of a technology is directly proportional to its personalized nature. High cost and relatively intimate face-to-face contact successfully moves people to the polls, whereas direct mail does little to change behavior. Given the ubiquity of unsolicited email and the low transaction costs associated with the medium, email should exhibit little success in mobilizing voters and this expectation is borne out.

Other posts on the internet and campaigns:
What’s in Candidate Websites?
Is the Internet “Democratizing” Campaign Donations?
Did “Macaca” Lose the 2006 Election for George Allen?

Posted by John Sides at 09:56 AM | Permalink | Comments (3)

15 years in the making! (Elections just kept happening, and every time we got caught up, here came another one.)

With a cast of thousands of campaign statements! (Laboriously but imaginatively assembled, coded, and analyzed.)

Dozens of tables for the quantoids! Hundreds of anecdotes for the rest!

Even if you hate the contents, you’re gonna love the cover!

Rush right down to your local supermarket and pick one up at the check-out stand! Well, probably not. But definitely do click here to purchase at amazon.com. While you’re at it, buy several copies and give ‘em to your friends.

Posted by Lee Sigelman at 12:01 PM | Permalink | Comments (3)

The ideal court system in a democratic society is supposed to operate “outside of politics,” reaching decisions solely by carefully weighing the law and the evidence while leaving political considerations in the hands of the legislative and executive branches. The extent to which that ideal is approximated in the day-to-day operation of the courts is something else altogether. At the highest level of the American court system, this issue shows up regularly in political science journals in the form of debates about whether Supreme Court decisions largely reflect the justices’ own policy preferences or are based on their earnest efforts to apply the law to the facts at hand.

Nowhere do political considerations intrude more overtly into the judicial process than in the practice of electing state court judges. A question naturally arises, then, about whether judicial election campaigns shatter the idealistic images of justice as blind and of the courts as isolated from the hurly-burly of politics.

In the current issue of the American Political Science Review, James Gibson uses a clever research design to probe the impact of the popular election of judges on public support for the courts (abstract here). Three different features of election campaigns, Gibson argues, can undermine the perceived impartiality of the courts: campaign contributions, negative camapigning, and policy pronouncements by the candidates. To gauge these effects, Gibson embedded a series of experimental vignettes within a statewide survey in Kentucky. In some vignettes, the focal candidate, an incumbent, was portrayed as receiving campaign contributions from organizations that regularly try cases before his court; in others, the judge was portrayed as declining such contributions, and similar manipulations were used to capture the effects of negative campaigning and policy pronouncements. The results of these survey-based experiments indicated that in these vignettes, campaign contributions and negative campaigning both produced significant declines in perceptions of the courts as fair and impartial; the issuance of policy pronouncements, however, had no such effect.

These results led Gibson to conclude that:

…when candidates for public office receive campaign contributions from those with direct business interests before the institution, many (if not most) citizens perceive policy making as biased and partial and the policy-making institution as illegitimate. Similarly, the use of attack ads causes many to question instititutional legitimacy. Just as certain are my conclusions that policy pronouncements by judicial candidates cause little harm to courts …

Those concerned about threats to the legitimacy of elected state courts would do well to turn their attention away from substantive policy pronouncements and focus instead on the corrosive effects of politicized campaigning, and especially campaign contributioins from those having business before the bench.

Posted by Lee Sigelman at 02:23 PM | Permalink | Comments (0)

In this Politics magazine piece, Nick Gillespie and Matt Welch of Reason magazine envision a future when a candidate like Ron Paul is President:

And when that future arrives — and it will, sooner or later — we’ll all look back at the failed campaign of a guy who said he didn’t want to run our lives, the economy, or the world, and wonder what took us so long.

How does one get to such a conclusion? Follow these easy steps:

#1. Extrapolate political trends from cultural trends. Use hackneyed cultural references when possible:

The key to such an optimism is recognizing that politics is a lagging indicator of American society, which has been moving with broadband-like speed [hackneyed #1] into an era of Do It Yourself culture and not-so-rugged individualism. Think of what Americans have come to expect and insist upon in their social and economic lives: increasingly individualized service, culture and consumer products at every level (“You want soy with that decaf mocha frappucino?” [hackneyed #2]); more and more control over education, healthcare and retirement; and a nearly full-throttled embrace of lifestyle tolerance and pluralism that was unimaginable in a pre-Netflix, pre-”Queer Eye for the Straight Guy,” pre-iPod America [hackneyed #3, 4, and 5].

Gillespie and Welch will have to explain to me the connection between individualism and “Queer Eye” — perhaps that’s what “not so rugged” implies? — but in any case I can give them one piece of disconcerting news. Every year since 1982, the National Election Studies has asked this question:

Some people think the government should provide fewer services, even in areas such as health and education, in order to reduce spending. Other people feel that it is important for the government to provide many more services even if it means an increase in spending. Where would you place yourself on this scale, or haven’t you thought much about this?

The percentage who wanted government to provide many more services has increased in the past 10 years to 43%, while those who want the government to provide fewer services has decreased to 20%. The data are here. Maybe “not so rugged” just means “not.”

#2. Fetishize exotic measures of political success and downplay the ones that count.

To be sure, by every conventional measure Paul’s presidential presidential campaign bid has been an abject failure — not a single primary win and only 14 delegates as of press time. Yet Paul managed to raise more than $20 million, virtually all of it online, and inspire an army of hyper-devoted and mostly youthful followers using a pitch — and a style — that will have much more to do with 21st century politics than whatever models of Buick and Oldsmobile the Democrats and Republicans eventually crank out this year.

Gillespie and Welch go on to mention that Paul had 67,000 people at MeetUp (20 times more than Obama! OMG!) and that Paul won raves from George Will, Johnny Rotten, and a group called “Strippers for Paul.” Yes, clearly the Democrats and Republicans are doomed if their candidates keep earning the support of delegates instead of aging punk rockers and strippers. Delegates are far less sexy, it’s true (although maybe not much less sexy than Johnny Rotten is now), but they actually provide political power. The question to ask is not, “How can the Democrats and Republicans be more like Ron Paul?” The question to ask is, in essence, “How can a candidate like Ron Paul learn to drive a Buick?”

#3. Fail to read political science.

The major political trend of the past 40 years is the inability of the two parties to grow, much less maintain, market share.

Gillespie and Welch cite Harris Polls showing that the percentage of people identifying with the Republican and Democratic parties has declined (from 80% of the public in 1970 to 63% in 2006). But lots of people who purport to be “independent” actually lean toward one of the two parties. Only about 10% of the population, as of 2004, is purely independent (see here). Moreover, the “independent leaners” behave like closet partisans. In 2004, 90% of Democrats voted for Kerry, as did 87% of independents who lean toward the Democratic party. Similarly, 95% of Republicans voted for Bush, as did 89% of independents who lean toward the Republican party. The “myth of the independent voter” has been well-known in political science since at least 1992; see here.

#4. Fail to see the lessons of your own analysis.

Gillespie and Welch cite Chris Anderson’s book The Long Tail.

In terms of goods and services, Anderson argues, we are turning “from a mass market into a niche nation.”

This is because half of Amazon’s sales comes from books that are ranked outside the top 130,000 titles. Etc. And here is the contradiction for Gillespie and Welch. Even if Anderson is correct, true political power is not to be found in the niches. Yes, passionate minorities can influence politics, but fundamentally, American politics is organized and structure by the two dominant parties, who of course collude to maintain their own power, as any self-interested political actor would. And parties are composed of diverse groups; they are coalitions, not niche organizations. Libertarians are not going to matter much unless they can transform themselves from a niche catalyzed only by the occasional candidate like Paul to a key player in either the Democratic or Republican coalition (or both!). (I’ve made a similar point in a previous post.)

How boring, I know. But sometimes an Oldsmobile is all you’ve got.

Posted by John Sides at 10:48 AM | Permalink | Comments (10)

Andy Gelman links to a new paper on money and UK politics. The abstract speaks for itself.

While the role of money in policymaking is a central question in political economy research, surprisingly little attention has been given to the rents politicians actually derive from politics. We use both matching and a regression discontinuity design to analyze an original dataset on the estates of recently deceased British politicians. We find that serving in Parliament roughly doubled the wealth at death of Conservative MPs but had no discernible effect on the wealth of Labour MPs. We argue that Conservative MPs profited from office in a lax regulatory environment by using their political positions to obtain outside work as directors, consultants, and lobbyists, both while in office and after retirement. Our results are consistent with anecdotal evidence on MPs’ outside financial dealings but suggest that the magnitude of Conservatives’ financial gains from office was larger than has been appreciated.

Andy isn’t sure about the substantive impact that this has for political science, given the disparities between the amounts of money that flows through politicians’ hands in functioning democracies and the amounts of money that they may personally derive from office. I’m not so sure about that, as the monies sticking to politicians’ hands do likely help shape their incentives (e.g. one can plausibly speculate that Tories who rock the boat too much aren’t going to have much luck cashing in on those directorships), but, in any event, the fact that Andy doesn’t spot any obvious methodological problems makes me at least think that the observed effect is likely real.

[Cross-posted at Crooked Timber]

Posted by Henry at 11:26 PM | Permalink | Comments (0)

Thinking about attending the NCAA Final Four this year, but don’t have tickets and don’t want to pay scalpers’ prices, which sometimes run up into the thousands of dollars? Help is on the way, in the form of (yet another) futures market. For a small fee, you can buy “reservations” to follow your favorite team to the Final Four in the event that they make it. After you purchase, you can, if you wish, sell your reservations if the team’s fortunes are rising as the field narrows down to four, and score a nice piece of change. If your team falls short and you’ve held onto the reservation, you’re out the nominal cost of the reservation, but that’s it. Or if you’re a steadfast fan and have held onto the reservation and if your team turns out to be one of the charmed four, then off you to to cheer them on to victory, with tickets you’ve purchased at face value plus the reservations charge. On the other hand, if you didn’t purchase a reservation in the first place, you can enter the reservations market and purchase one at whatever someone is willing to sell it for at that moment — and that will entitle you to buy the tickets themselves at face value.

Pretty sweet, eh? Well, yah. But, hmm, doesn’t this sound a little like, er, ticket scalping, and isn’t that illegal in most places? Or maybe it’s not scalping — just plain old gambling? Hmm, I can’t decide. So who’s running this ticket-scalping-or-is-it-gambling operation? Why, it turns out to be a money-making deal for that self-righteous overseer of the morality of college athletes, the NCAA itself. If the NCAA catches an athlete consorting with gamblers or if it finds a player’s complimentary tickets in the hands of scalpers, it’s bye-bye time for the player’s college hoops career, and it’s public humiliation for the transgressor. But let the NCAA itself run the show, and it’s just free enterprise, and who could object to something as all-American as that?

Read all about it here.

[Hat tip to Scott Adler]

Posted by Lee Sigelman at 09:30 PM | Permalink | Comments (1)

The mass media in China, though still largely state-owned and centrally controlled, are enjoying greater freedom, sometimes even to carry “very negative and critical stories” about government officials. “Although these negative reports [are] mostly restricted to local officials,…still it is a big step forward compared to the past.”

Here, in support of that conclusion, is a data snippet from a recent analysis (abstract here) of news reports, 1990-2003, that have won the China Radio and Television Awards, which were established as an incentive to excellence in Chinese journalism.

As the author of the study, Xi Chen (a Ph.D. candidate in the Planning, Governance and Globalization Program at Virginia Tech), concludes:

The facts that journalists are reporting ‘dark’ stories which are not supposed to increase support for the party and government … and are disclosing the frauds and deceptions practiced by local officials, and paying more attention to the harsh situation in the remote western and southwestern area of the country all demonstrate that they are becoming more professional and are playing more of a role of agent rather than the mouthpiece for the government as before.

Now, let’s not go overboard. No one is claiming that China now has anything approaching a free press that provides “fair and balanced” news coverage — see, e.g., this recent posting by James Fallows on Chinese coverage of the situation in Tibet. But a step forward is…well, a step forward.

Posted by Lee Sigelman at 08:30 AM | Permalink | Comments (1)

One way to think about the electoral landscape and make predictions for November is to focus on the “fundamentals,” such as the state of the economy. As I discussed earlier, those fundamentals tilt in Democrats’ favor.

A second way is to answer to two questions: what is the most important problem facing the country, and who can best handle that problem? Below are some recent data (from a March 7-10 NBC/Wall Street Journal poll and a Feb. 28-Mar. 2 ABC/Washington Post poll) that get at Americans’ priorities and their evaluations of the front-runners, respectively:

(Note: The ABC/WP poll did not ask comparable items about Hillary Clinton.)

Obama does just as well and sometimes better among independents; see here.

Attitudes about the parties show similar patterns:

(Source: Feb. 20-24 Pew poll.)

The issues that are most salient right now are issues where the public is more likely to trust Obama and the Democrats than McCain and the GOP. McCain tends to fare better than his party in that the “gap” between him and Obama is smaller than that between the parties. But nevertheless he trails with regard to the economy, health care, and immigration, and has a statistically insignificant lead with regard to Iraq.

Thus, the issue agendas of (some, perhaps crucial) voters may lead them to vote Democratic. The political science research undergirding this notion is by John Petrocik (article here, gated). Note also that this model of choice is less complicated that one based on issue positions — whose role Lee and I have complicated previously (here and here).

Posted by John Sides at 02:38 PM | Permalink | Comments (1)

Do the test. (And look out for cyclists.)

Posted by Lee Sigelman at 10:02 AM | Permalink | Comments (4)

Andrew Kydd has a short paper that uses game theory to analyze the benefits and disadvantages of ethnic profiling at airports etc. He starts with analysis of a simple Matching Pennies game (which I’ve always had in the back of my head as the most obvious simple way of modeling this process; I suspect the same is true of many people who have (a) taken a course in basic game theory, and (b) thought about profiling tactics), and moves on to a more complex specification. On the basis of a Colonel Blotto game, Kydd argues that some degree of ‘seemingly unfair profiling’ is likely to result from rational strategies employed by actors on both sides, except under relatively unlikely parameter values. He concludes that:

Profiing in the context of strategic terrorist groups is rational, when considered solely on the basis of what will stop more terrorist attacks. Other factors may constrain the United States from profiling, but the logic of interaction with terrorist groups suggests a powerful dynamic in its favor.

While the paper is still a work in progress, I worry that this is one of those cases where game theorists reach confident conclusions on the basis of ceteris paribus conditions, where the ceteris are anything but paribus if you look at them closely. Entirely apart from the ethical implications (which he seems interested in exploring in future iterations of the piece), the conclusion that profiling will stop terrorism rests on the assumption that increases in profiling of specific ethnic groups won’t have any consequences for future recruitment of terrorists from those groups. There’s an array of circumstantial evidence that the perceived humiliation that targeted searches and other such measures involve have significant implications for relations between the targeted group and other members of the population.

Jack Knight and I have some unpublished work where we try to tease this out in the context of trust relations - we argue that those who feel targeted by measures of this sort are rationally going to be more distrustful of those who belong to more privileged sections of the population for a variety of reasons. This may obviously have consequences for the ease of recruiting those aggrieved members to terrorist organizations. Obviously, there’s no good way of measuring the extent of this effect, assuming it exists. But equally obviously, this and related interaction effects mean that one should be quite careful before drawing broad conclusions about the merits of policies on the basis of simple game theoretic models - the unmeasured knock-on consequences may very likely swamp the specific effects that your model is designed to single out.

Posted by Henry at 06:27 PM | Permalink | Comments (2)

As a mere political scientist, I don’t pretend to have any expertise concerning the world of high finance. For folks like me, this article (gated) by Dean P. Foster and H. Peyton Young presents a crystal-clear explanation of the dismal economics of hedge funds. (An earlier version of this article appeared in the Washington Post on December 19, 2007. This more recent version is better, because it more explicitly addresses the perilous situation in which hedge fund investors put themselves.)

Posted by Lee Sigelman at 05:50 PM | Permalink | Comments (0)

As virtually any political campaign wears on, more and more sharp words get exchanged by the combatants and their handlers — a tendency exemplified last week by the sudden resignation of Obama foreign policy advisor Samantha Power, who had referred (off the record, she thought) to Hillary Clinton as a “monster,” and this week by the sudden resignation of Clinton supporter Geraldine Ferraro following her on-the-record remarks about Obama and race.

Some of this infighting is all in good political fun as campaigns go, for politics ain’t beanbag, as Mr. Dooley put it. Even so, sooner or later all the attacks and counterattacks become likely to leave a residue of hard feelings, creating a situation in which the vanquished may turn thumbs down on the victor. Thus the issue repeatedly arises of whether divisive primaries undermine a candidate’s prospects in the general election — an issue on which there is some good political science research, as John Sides recently posted here.

New grist for the mill has been delivered, in the form of some data that Joost van Beek, a dedicated psephologist, recently forwarded to us at “The Monkey Cage.” Drawing on exit poll results archived at msnbc.com, Joost calculated what he called the “Bitterness Quotient” between Obama and Clinton supporters in the Democratic primaries that have been contested so far. I don’t like that name, because what’s being measured isn’t necessarily bitterness; for example, a Clinton supporter could be dissatisfied by an Obama victory simply because the voter likes Clinton more on the issues and harbors no hard feelings toward Obama. Still, Joost’s data are valuable, for they tap directly into dissatisfaction with the prospect that one’s disfavored candidate might become the Democratic standard-bearer.

Question wording: “No matter how you voted today, would you be satisfied with [Clinton/Obama as the nominee]?”

The response frequencies are shown above for the 27 states for which the data are available. These data are up-to-the-minute, for they include results from the Mississippi exit poll that was conducted just three days ago. Excluded are three states in which Democratic primaries have been held: Hawaii, where no exit poll was conducted; Michigan, where Clinton was the only candidate on the ballot; and New Hampshire, where an exit poll was conducted but the dissatisfaction question wasn’t asked. Each entry in the table above is simply the percentage of Obama or Clinton voters who answered “dissatisfied” when asked whether, “no matter how you voted today,” they would be satisfied or dissatisfied if the other candidate turned out to be the Democratic nominee.

Some observations:

• Overall, Obama’s supporters have expressed less dissatisfaction at the prospect of having Clinton at the top of the ticket than vice-versa. The reason could be that Clinton has run the more negative of the two campaigns, creating doubts in her supporters’ minds about Obama; for example, according to a Wednesday news release by the Wisconsin Advertising Project, in the hotly contested Ohio primary more than one-fifth of Clinton’s advertisements but less than five percent of Obama’s were negative.

• The Opponent Dissatisfaction Index has varied considerably from state to state, especially among the Clintonistas, for whom it has ranged all the way from 30% to 72%. For the Obamaniacs, the range has been narrower (29% to 58%).

• If you stare at these data long enough, you may begin to see some patterns. But a better approach is to run some statistical analyses. Having done so, I’ll note that Obama’s supporters in the South have been a bit more receptive than his supporters in the rest of the country to the prospect of having Clinton as the nominee. On the other side of the coin, Clinton’s southern supporters have been less receptive than those elsewhere to a possible Obama victory.

• Since Super Tuesday (February 5), Obama’s supporters have expressed greater dissatisfaction about a potential Clinton nomination than they were doing before then — again, presumably as a result of Clinton’s attacks on their favored candidate. For Clinton’s supporters, there has been little movement over time in their dissatisfaction with Obama’s possible nomination.

• Putting the effects of region and timing together reveals that since Super Tuesday a major South versus non-South gap has opened up in Obama supporters’ dissatisfaction with a potential Clinton nomination, whereas earlier there had been no such gap. For Clinton supporters, though, neither region, timing, nor the two together, help account for the state-to-state variability in expressions of dissatisfaction with an Obama candidacy,

In general, then, there has been no systematic increase in opponent dissatisfaction with the passage of time in this campaign. In the particular case of Obama’s supporters in the South (a large prooportion of whom are African Americans), though, dissatisfaction with the prospect of a Clinton victory has indeed grown.

Several primaries are currently scheduled, and the prospect of do-overs in Michigan and Florida remains. Will these patterns remain in force, will they become even clearer, or will they be supplanted by new trends as the long campaign slog continues? Stay tuned.

[A doffing of the chapeau to Joost van Beek]

Posted by Lee Sigelman at 10:43 AM | Permalink | Comments (7)

Ezra Klein is surprised that a family income of $88,000 puts you in the top 20% of Americans. He thought it would take a larger income. This gets at an interesting question: what do people believe about the income distribution of the United States? And how do those beliefs accord with reality?

A colleague, Eric Lawrence, and I asked this question in a 2007 survey:

How much income do you think the average American household earns in a year? If you do not know, you can just give your best guess.

The median answer was $40,000. The actual median is $48,000. On average, people’s estimates are fairly accurate.

Actual income levels do affect median estimates:

This makes sense: people’s estimates in part reflect their personal circumstances and surrounding environment. But only in part. On the whole, I was impressed with how accurate estimates were.

This “analysis” is just a first cut. I’d love any questions and suggestions in the comments.

Posted by John Sides at 02:48 PM | Permalink | Comments (5)

George W. Bush has two, but Saddam Hussein had six. Bill Clinton has one, but Kim il-Sung (shown above with the infant Kim Jong-il) had seven. Ronald Reagan had three, but the Sultan of Brunei has 10. John F. Kennedy, Lyndon Johnson, and Richard Nixon had two apiece, but Idi Amin had somewhere between 30 and 45.

Do you see any pattern here?

I’m speaking, of course, of children.

On the assumption that what matters most to political leaders is survival, Dustin Beckett and Gregory D. Hess argue, in essence, that leaders are motivated to immortalize themselves or at least to secure the future of their regime (abstract here). This motivation, in turn, leads them to try to influence the selection of their successor. From the leader’s perspective, trust in one’s presumed successor is vital, or else a successor might pull a Khrushchev by disassociating himself from the legacy of the leader. (I’m aware that Khrushchev wasn’t Stalin’s immediate successor, but you get the point.)

How, then, to ensure that one can trust one’s likely successor to carry on as one would wish? Because blood is thicker than water, the safest route is to keep it all in the family. For the leader of a non-democratic system, then, the best way to assure a “trustworthy succession match” is to sire a whole lot of children, in hopes that by the luck of the genetic draw one of them will turn out to be a “person who can successfully maintain the regime and the expropriation of rents, and with whom the enforcement of ex-ante promises is less problematic.” (As you can tell from the way they talk, Beckett and Hess are economists.)

By contrast, leaders in democratic systems find themselves more hemmed in by those bothersome checks and balances, and so many other people have a say in the choice of their successor that democratic leaders can’t do all that much to determine who is next in the line of succession anyway.

From these considerations, Beckett and Hess extract the expectation that the leaders of non-democratic systems will sire more children than their democratic counterparts, in hopes of finding the trustworthy succession match that they so avidly seek.

Sounds pretty far-fetched, doesn’t it?

Well, Beckett and Hess’ empirics, based on the composition of the families of 221 leaders worldwide and a wide array of country- and leader-level data, indicate that the leaders of non-democracies do indeed produce more offspring than the leaders of democracies; Idi Amin may have been an outlier, but his autocratic peers also tend to be prolific begetters. On average, the leaders of fully non-democratic countries have sired between 1.5 and 2.5 more children than the leaders of fully democratic countries — and this relationship holds up when a wide array of statistical controls are instituted. (So it’s not just that, say, the leaders of Muslim countries, most of which are non-democratic, have lots of wives and sire lots of children. There’s more to it than that.)

To which I must caution that the empirical relationship that Beckett and Hess observe is a matter of differences of degree, on average, rather than a universal generalization. Exceptions come fairly quickly to mind. Stalin, for example, had only three children (which could, if one buys Beckett and Hess’ argument, help explain the destalinization that occurred soon after his death). And some democratic leaders do produce a large number of children. John Adams, for example, had five, Willliam Henry Harrison had nine, and George H.W. Bush had six. But perhaps these are just exceptions that prove Beckett and Hess’ rule: When it comes to succession (whether immediate, as in the Adams and Bush cases, or delayed, as with the Harrisons), the more children, the better.

Posted by Lee Sigelman at 09:33 AM | Permalink | Comments (4)

Click here to view.

[Hat tip to Matthew Yglesias]

Posted by Lee Sigelman at 08:56 PM | Permalink | Comments (0)

Over at Enik Rising, Seth Masket notes that in several Colorado counties, Hillary’s share of the county delegates exceeds the share of the vote that she got in the county’s caucus. The graph is above. He writes:

It’s hard to know if this is all due to chance, if she really does have an effective post-caucus strategy in the urban counties, or if the Obama folks are just flakier as the process goes on. It’s also hard to say just how much this will matter in the end. Each of the state’s seven congressional district will only send six or seven delegates to the DNC. So maybe she can flip two or three this way, and if she does that in the other caucus states, we’re talking about serious numbers, although obviously not enough to overtake Obama in pledged delegates. Still, every little bit matters right now.

Posted by John Sides at 01:29 PM | Permalink | Comments (0)

The concept of “red” and “blue” states is one of the worst tropes in conventional discourse about American political geography. Fortunately, James Gimpel, in collaboration with the Christian Science Monitor, has produced a much richer geographical portrait, one that captures interesting kinds of variation, such as “monied ‘burbs” (e.g., my hometown of Winston-Salem, NC) and “industrial metropolis” (my current home of Washington, DC), and “campus and careers” (Chapel Hill, NC, home of my alma mater, the top-ranked UNC Tarheels). The CSM will be profiling towns in each of 11 different types of communities during the campaign season.

Check out the website here.

Posted by John Sides at 12:25 PM | Permalink | Comments (2)

Posted by Sarah Binder at 03:04 PM | Permalink | Comments (3)

Back in the days when Henry Kissinger — nobody’s idea of a matinee idol — was frolicking with Hollywood starlets like Jill St. John (above) and Marlo Thomas, he explained his obvious but seemingly inexplicable appeal to the opposite sex with a pithy one-liner: “Power is the ultimate aphrodisiac.”

The other day I happened upon an article by sociologist John Levi Martin that was published in 2005 ago but seems to have attracted little attention at the time, titled, engagingly, “Is Power Sexy?” (abstract here). Although it’s tarted up in lots of fancy sociologese, the basic idea of the study is simple. Martin begins with the idea, based on both rational choice theory and evolutionary psychology, that women want to pair with high-status men, but he wonders whether the crucial factor is the male’s hierarchical position (which he calls “status”) or his specific relationship to to the women in question (which he calls “power”). To probe these possiblities, he analyzes data from a 1970s study of interpersonal relationships in a national sample of “naturally occurring communities” (aka communes). In that study, the respondents were asked to name fellow commune members they thought fit any of a number of descriptions, including “sexy.”

What Martin discovered, unexpectedly, was that although there was a link between status and perceived sexiness, it wasn’t because women found high-status men especially sexy. Rather, the status-sexiness link was in the minds of men, who seemed to be turned on by high-status women — but not by women who held power over them in particular. For their part, what women seemed to find sexy in men was their interpersonal power, not their more general status within the community.

I’m reluctant to assign any electoral meaning to this, as research on the traits Americans desire in their political leaders has never, insofar as I know, singled out “sexiness” as a desired trait (see here). Still, it bears thinking about in a year when not all the leading presidential contenders are men — especially when a disproportionate percentage of caucasian males have been turning out for lone woman candidate.

Posted by Lee Sigelman at 04:04 PM | Permalink | Comments (1)

Lane Kenworthy shows how it’s done.

Posted by Henry at 12:36 AM | Permalink | Comments (4)

Political scientists tend to be lousy political prognosticators. We speak in circumlocutions, carefully couching our every utterance in a web of qualifications. Our favorite answer to “what’s going to happen?” kinds of questions tends to be “It depends,” followed by a lengthy recitation of the reasons why things are just too complicated to be predictable. We prefer to do our “prediction” after the fact, when things are, after all, more predictable. To be sure, a few political scientists — Tom Mann, Larry Sabato, and the inevitable Norman Ornstein come immediately to mind — have mastered the art of speaking ostensible truth to the great unwashed, but the rest of us are far less skilled insofar as conveying our interpretations of current events and our expectations of future ones to the general public.

When most practicing political pundits are called upon to forecast what’s going to happen, they typically answer so forcefully and self-assuredly as to produce a sense of certainty in the listener, whose attention span, in turn, tends to be so short that he or she never remembers to check up a few days later on whether the pundit’s prediction was borne out.

Could it be that pundits are themselves dismal forecasters? Well, yes. That, at least, was the conclusion that Jarol Manheim, Susannah Pierce, and I reached several years ago in a study titled “Inside Dopes? Pundits as Political Forecasters” (abstract here). In that study, we assessed the accuracy of the predictions offered by members of TV’s noisiest shout-show, “The McLaughlin Group.” Those predictions turned out, in large measure, to be either untestable or, if testable, wrong.

Where does this leave us? For one thing, it leaves us wanting to know more about the bases of expert judgments, and in that connection I know of no better source than Philip Tetlock’s book, Expert Political Judgment.

For those who are interested enough to log a bit of time doing armchair fieldwork of their own on the punditocracy, there’s a new resource that could make things much easier. The folks over at CampaignCircus.com recently announced what they’re calling the “Pundit Accountability Project.” To keep tabs on what various pundits have been saying, all you have to do is click here and you’ll get a drop-down menu with the names of various pundits. As of now, 28 pundits are listed, but the intention is to keep adding pundits and clips as circumstances warrant. Click on a pundit’s name and you’ll get clips of his or her on-air statements that have some predictive component. The press release mentions that CampaignCircus.com plans periodic releases of clips by the pundit who made the most consistently wrong predictions — though it’s unclear what sort of “fact-checking” technique they’re going to use to determine who wears that crown. For example, the press release says “Howard Fineman is currently leading the pack” in predictive inaccuracy, but how that determination was made isn’t clear. In any event, this compilation of clips should at the very least be a fun entry to add to your “Favorites” menu, and it might even turn out to be a worthwhile source of data on the conventional “inside the Beltway” wisdom as the campaign plays out.

Posted by Lee Sigelman at 09:43 AM | Permalink | Comments (1)

It has become virtually certain that political candidates will have websites. And yet rigorous research about how candidates use the Web to present themselves is in short supply. Instead, we get breathless speculation about new technologies (“OMG, John Edwards is Twittering!”). Three recent studies are a welcome contribution.

This paper (.doc file), by Jamie Druckman, Martin Kifer, and Michael Parkin looks at 444 Congressional campaign websites from the 2002 and 2004 elections. Here are some notable findings:

• Only 44% of websites included audio or video content. Such content was more prevalent in Senate candidate websites than in House candidate websites.

• About 20% of websites were not updated.

• Personalized content for the user — such as taking a quiz or providing information for targeted marketing — was less common, but became more so in 2004, when 29% of websites had such content (vs. 18% in 2002).

• Only 10% of websites had interactive content, such as a chat with a candidate or a forum.

• Incumbents invest less in their websites, as measured by whether they are updated or have multimedia content. Incumbents have less incentive to make these investments because presumably they are less in need of votes.

• Democrats are more likely than Republicans to include personalization features and two-way communication.

• Candidates in competitive races are more likely to update their pages and to provide multimedia content, but less likely to provide a link to an external party website or to allow two-way communication. Candidates in close races have more incentive to control the flow of information, and external links and two-way communication reduce their control.

Druckman, Kifer, and Parkin conclude:

…we have shown that candidates have generally moved beyond an “electronic brochure” standard although they have had some trepidation in doing so. Moreover, our results show that their hesitancy in using these technologies is not only based on practical considerations of feasibility but also on critical political considerations that force candidates to weigh the strategic benefits and costs of each feature.

The second paper is by Tracy Sulkin, Cortney Moriarty, and Veronica Hefner. (Here is an earlier but ungated version.) They examine the issues candidates emphasize on their websites and in their televised advertising. Some of their findings:

• Because websites are far less costly than television advertising, candidates mention far more issues on-line than in their advertisements, where they are much more selective.

• Candidates are more likely to use their websites to appeal to core supporters, who likely make up most of the viewers of these websites. Thus, Republican and Democratic advertisements have more similarities than do their websites, where each party’s candidates devote more attention to issues that appeal to elements of their coalition. For example, Democratic candidates were more likely to discuss civil rights on their websites than in their ads.

• Candidates tend to take more specific positions on their websites than in their ads, perhaps because they simply have more space to expound on their ideas.

Finally, a Project for Excellence in Journalism study of the 2008 presidential candidates’ websites is here. These websites are more sophisticated, on average, than those examined in the Druckman et al. study, probably reflecting the greater penetration of technology in 2008 versus earlier elections and most certainly reflecting the greater resources of the presidential candidates versus congressional candidates.

Each of these studies is well worth reading. The chief lessons, to my mind, are three. First, although journalists will be quick to highlight new innovations in web-based campaigning when they first appear, most candidates will be late (or later) adopters. Second, some of the same incentives govern web-based communication as they do other forms of communication — e.g., it may behoove candidates in competitive races to keep tighter control over information. Third, websites are designed for a fundamentally different audience, which is likely to support the candidate whose website it visits. Thus, web content has less potential to change anyone’s mind, and, if anything, will tend simply to reinforce existing preferences.

Posted by John Sides at 08:17 PM | Permalink | Comments (1)