A few pages into the book, I was quite surprised to see a highlighted passage of the book, along with a small note with the number of people who had highlighted that passage! Sure enough, the book I downloaded now contains the most popular highlights from other readers of the book. A little poking around discovered two other features: 1) I could turn off this feature (thanks Amazon!); and 2) You could link to a collection of the most popular highlights. Somewhat disturbingly, I did not find either an opt in or an opt for sharing my own highlights.

I’m trying to figure out what to make of this new feature. On the one hand, as a social scientist, I am always curious as to what other people are thinking and doing, so voilà: Now I get to see what my fellow readers think is important in the book I’m reading. In a nutshell, intellectual voyeurism (at least at the aggregate level). Also thinking as a social scientist, my next thought was: DATA. Are we far away from an analysis of whether Glenn Beck readers highlight more passages than Paul Krugman readers? And will Amazon give me this data disaggregated by age, region, or — dare I dream? — party affiliation?

On the other hand, the obvious concern is the one I just raised, which is the issue of privacy. Now not only does Amazon know what books I downloaded, they also know what passages of those books I find most interesting. I can see it now: “We see you highlighted ‘economic voting in Eastern Europe,’ perhaps you would be interested in “Regional Economic Voting” (available for purchase here, also available for the Kindle!).”

But I have to say, it was the collection of social highlights at the end that most troubled me as a professor. I mean, why bother reading the book if you can just skip to the end and figure out what everyone else thought were the most important parts? I remember taking books out of the library when I was in college (for today’s undergraduates, the library is that big building where they keep books) and finding out that someone else had scribbled notes in the margin previously; you could get the same effect by buying used books. There was always the temptation to just skim the book by reading what the previous person had highlighted, especially in the more tedious sections. I always thought one of the nice side benefits of moving to electronic versions of books and articles was the removal of this temptation; when the work is in electronic format, everyone gets a clean copy. But alas, no more; now it is worse. You not only get what a couple people have scribbled in the margins, you can get what everyone (collectively, at least) has decided is important! Again, another excuse not to actually read.

As an author, though, the opportunity for feedback is kind of enticing; I do write for a blog, after all. It would be interesting to be able to watch on your electronic copy of your own book what it is that readers are highlighting. I imagine as this technology improves, we might also get the opportunity to see what notes people are putting in the margins, and ultimately, perhaps, even be able to answer them with our own comments. In this world, books would start to look a little more like blogs. Not sure if that’s a good thing or a bad thing, but it would be pretty wild to scribble some question down in the margins in your Kindle or the like, only to come back an hour later and find that the author had answered it. On the downside, would anyone ever want to write a statistics textbook ever again in that sort of a world?

Consider a simple game with three players (no pun intended!) who all have to choose where to play basketball next year. They can each choose to stay in their own city or move to the city of another player. (So for the moment we leave out the option of them all going to New York or Chicago for simplicity’s sake, but I think the logic still holds fine if we expand the number of destinations.) The players get utility from the following four factors:

1. The money they are paid in salary.

2. The chance to win a championship.

3. The quality of life in the city in which you choose to live.

4. Not being perceived as a villain.

Let’s now assume the following:

All three were going to get roughly the same salary no matter where they signed.

The chances of winning a championship increase monotonically the more of the big three sign in the same place for those who sign there (i.e., probability of winning a championship is highest if you sign with both of the other three, which is greater than if you sign with one of them, which in turn is greater than if you don’t sign with either of them). As a corollary, the chances of winning a championship decrease for the one person who signs elsewhere if two of the big three sign in one place.

One gets perceived as a villain for abandoning one’s current team, but this “villain penalty” increases 1) if you originally grew up in the region of the team you are playing for, and 2) the more you say, “well, I might stay; I just don’t know.” To put this another way, LeBron will pay the biggest “villain” penalty, and Bosh, who made no secret of his desire to get out of Toronto, will pay the lowest. (Note: For those who care about these sorts of things, we could also call this the A-Rod effect, who still gets booed in Seattle to this day.)

I’m also going to go out on a limb and assume the quality of life in Miami is higher than in Cleveland or Toronto for our young stars (and especially so for Wade and Bosh — LeBron maybe a little less so because of friends and family in Cleveland).

OK, so where does this get us from the perspective of Wade? His top preference is to stay in Miami (no villain penalty, best quality of life) with Bosh and LeBron joining him (highest chance of winning championship). His decision gets more difficult as Bosh and LeBron go elsewhere, and especially if they go elsewhere together.

What about Bosh? He clearly revealed from statements leading up to the free agency period that he was willing to pay the “villain” penalty to gain the quality of life benefit of being somewhere else. So in our simple game, he too prefers Miami (higher quality of life than Cleveland, given he wants out of Toronto), and he prefers to be there with Wade and LeBron. Again, things get more difficult for him only when the “championship” option looks better elsewhere, especially if Wade and LeBron both go somewhere else together.

Which brings us to LeBron. He is going to be pay a heavy “villain penalty” for leaving Cleveland. If that penalty is sufficient enough to outweigh the benefit of living in Miami vs. Cleveland, then he would likely prefer to be in Cleveland with Bosh and Wade joining him there. (Plus, remember LeBron’s quality of life benefit from Miami is probably lower than that of the other two.) Indeed, he might prefer to stay in Cleveland even without Bosh and Wade, provided those two remain in separate cities. Where things become more difficult for him is if Bosh and Wade both go to the same place outside of Cleveland; in the probability of winning a championship this then makes the difference  between joining them (highest chance, all three in one place) and staying put (lowest chance: He’s in a different city and the other two are in the same city) the largest it could be, making the cost of staying in Cleveland the highest it could be.

So here’s the question: What do you do if you are Wade and Bosh and you want to achieve your most preferred outcome (i.e., all three in Miami)? Once LeBron makes his decision, you simply solve for the highest utility, balancing playing with LeBron vs. quality of life vs. the villain penalty. However, if you can make your decision before LeBron, then perhaps you can influence LeBron’s choice, especially if you have an idea how much LeBron values winning a championship.

Thus the optimal strategy for Bosh and Wade, I think, is for both to credibly commit to going to Miami before LeBron makes his decision. This puts the maximum possible pressure on LeBron to come to Miami. Now, all of a sudden, LeBron is making his decision with complete information: If he sufficiently values the chance to win a championship, he has to come to Miami also. Whereas if LeBron makes his decision first, he might choose to go to Cleveland on the expectation that there is at least some non-trivial chance that Wade and Bosh will join him there.

So this leads to the interesting question: By allowing Bosh and Wade to make their decisions first, did LeBron possibly get himself into a situation where he ended up with a suboptimal outcome? If so, it would certainly put that ESPN special in a whole new light: not just obnoxious, but possibly even counterproductive. By committing himself to a specific timetable — and remember, the demands of the ESPN show called for absolute secrecy regarding his decision — he gave Wade and Bosh a chance to both 1) move first and 2) have a little time to think through the strategic value of moving first. So, in the end, the need for the King to play to the public may have led to the King himself getting played — surely not the first time in history this has happened!

This is a bold and provocative claim, and therefore ought to be subjected to serious statistical scrutiny. The first step would be to validate their coding of the 1,643 reported sexual assaults. They cite this Department of Defense report as their source; if anyone is interested in attempting to replicate their coding I would be happy to post the results. However, for the moment, let’s just posit that the claim that that 134 of these assaults were indeed same-sex in nature is correct, and instead turn to the inference made by the authors of this report: that homosexuals serving in the army are three times more likely than heterosexuals to commit sexual assaults. What are some factors that could potentially confound this inference?

1. First and most significantly, the study provides no evidence of the proportion of same-sex assaults that are committed by homosexuals. This is crucial to the study, because the authors want to leverage the information in the study to argue that homosexuals should not be allowed to serve in the military. But their data measure assaults by men against men or women against women, not the number of assaults by homosexuals. Thus without any understanding of the proportion of same sex assaults that are committed by homosexuals, the inference that homosexuals are more likely to commit sexual assault is invalid.
2. Second, we don’t know if the proportion of homosexuals in the military matches the proportion in the general population. The authors of the study assume that the proportions are similar, but if homosexuals are overrepresented in the military relative to the general population, then the inference is invalid. Moreover, it is not even clear that the general population is the right reference group; the military is overwhelmingly made up of young men. So even if we think the demographic composition of the military reflects the general population — which it may very well not — we’d still want to know something like the prevalence of homosexuality among 18-30 year old males, not among the population as a whole.
3. Moreover, even if we assume that the proportion of homosexuals in the military mirrors the proportion in the general population, the conclusions of the study are dependent on a low estimate of homosexuals in the general population (less than 8.15/3, or less than 2.7 percent). Other studies have found higher estimates.

There are other more detail oriented data-related questions to worry about, such as how the results would be affected if a single individual committed multiple assaults or how the authors of the study coded cases where the accused was not found guilty of misconduct (see for example USMC Case #61). But overall, any one of the three big concerns listed above should raise serious questions about the study’s conclusion that homosexuals serving in the army are three times more likely than heterosexuals to commit sexual assaults, taken together they ought to lead us to approach this claim with a very healthy degree of skepticism.

One other point I can’t help making — the report also provides a bunch of (fairly explicit) descriptions of same sex sexual assaults (not sure why these needed to be in the report?), and in one section goes on to point out that a Persian-Farsi linguist was accused of committing a same-sex sexual assault (see p.14). The heading of that section: Do we really need more homosexual linguists? So while we’re on the subject of inference, I assume from this argument that the Tailhook Scandal would lead the Family Research Council to conclude that we no longer need more heterosexual Marine Corp aviation officers?

[Hat tip to Neal Beck and Dan Kselman.]

In recent years, Britain has been characterized as a 2  1/2 party system: It has two viable parties that can win elections (Labour and Conservative), and then a third party that wins a non-trivial amount of seats but has not really been considered a legitimate contender to win an election (Lib-Dems). While strategic voting ultimately occurs at the district level, one would have to assume that in the past, the Lib-Dems have been disproportionately hurt by strategic voters: With Labour and the Conservatives assumed to have viable shots at winning elections, we would have to guess that more often it is potential Lib-Dem voters that abandon their party to choose between Labour and Conservative than the other way around.*

Which brings us to the 2010 elections. As I noted in a response to Henry’s previous post, one effect of the recent British debates and the explosion of Cleggmania could be that it frees up Lib-Dem supporters who have in the past voted strategically to actually vote sincerely in 2010. This would suggest some sort of underlying tipping model: As long as Lib-Dem support stays below a certain level, added popularity in the polls might still not transfer into that many additional votes if strategic voters defect on election day. However, once the party reaches the level of the big three (i. e., Lib-Dems are presumed to be just as viable an option as the other two parties), then there might be a rather dramatic increase in actual votes for Lib-Dems as the strategic voters come home.

My question for readers is the following: How could we know if this was actually occurring? Two options seem fruitful to me, but I am interested in other suggestions (especially from anyone who has surveys in the field!).

First, anyone with repeated surveys in the field could compare thermometer scores (0-100 rankings of how much you like a particular party) for the three parties with vote intention. The hypotheses would be that up until some “tipping point” there should be a significant gap between the proportion of respondents who rank the Lib-Dems highest on the thermometer scores and those intending to vote Lib-Dem. On the individual level, if we regress vote intention on ranking the Lib-Dems highest on thermometer rankings, we would expect to see the size of the coefficient on ranking the Lib-Dems highest increase after the tipping point. What’s great about the 2010 British elections is that we have a good guess at where this tipping point should be: in the days immediately after the first P.M. debate. Moreover, I think this type of analysis could work either with panel data or repeated new surveys, although with panel data we could of course track actual switchers.

A second strategy would be to compare the proportion ranking Lib-Dems highest on thermometer scores who go on to vote Lib-Dem in 2010 in post-election surveys (or immediate pre-election surveys) with previous elections. If the proportions are roughly constant, then we could conclude 2010 was no different, and — provided there was a drop-off between the thermometer scores and the vote intention/choice — the strategic voters stayed away. If the proportions increased, however, it would be evidence that strategic voters came home.

What are some drawbacks of this approach? The most serious seems to be that voters who are planning on voting strategically might need to rationalize that decision in their own heads by changing their thermometer rankings to reflect vote choice. Does anyone have any previous research suggesting whether or not this is the case?

Other thoughts? Ideas for how to proxy for preferred party other than vote intention or thermometer rankings? Anyone have surveys in the field attempting to test for strategic voting in this election?

*************************

*It is possible that in individual districts, pre-election polling could convince a possible conservative supporter that their particular race was going to come down to the Lib-Dem candidate and the Labour candidate, and therefore strategic voting in that instance could actually help the Lib-Dem. That being said, I would still maintain that if the national election was presumed to come down to Labour and Conservative, the Lib-Dems should be disproportionally disadvantaged by strategic voting. Although the more I think about this, the more this strikes me as an interesting question: Has anyone tested whether strategic voting at the district level is moderated by national implications of elections in parliamentary systems with single member districts?

By the term “Colored Revolution,” I am referring to particular pattern of events (which Professors Valerie Bunce and Sharon Wolchik have perhaps more accurately labeled the electoral model, whereby government change took the follow format: the opposition coalesced in preparation for an election; the election was held and the government implicated in (usually massive) electoral fraud effecting the outcome of the election*; protests were held following the election; and eventually a change of government of some form occurs. I have previously written as to why I think fraudulent elections are a particularly useful vehicle for allowing individual citizens who oppose the government to overcome collective action problems associated with protesting against the government, and there was also clearly a learning effect as these Colored Revolutions spread from country to country.

However, the last Colored Revolution took place in Kyrgyzstan five years ago. Since then, attempts at Colored Revolutions in the post-Soviet space — especially in Belarus but also in Azerbaijan and I think elsewhere — have not succeeded. Moreover, the most well known attempt at a Colored Revolution outside of the post-Soviet space — last summer’s attempted Green Revolution in Iran — has not succeeded (yet!).

From this framework, what is notable about yesterday’s events in Kyrgyzstan is that they do not embody the electoral model. Indeed, there was no election, and consequently no fraudulent election to galvanize protesters. Moreover – and I think this will become apparent as an important distinction in the coming days — this means there is no election to present an aura of democratic legitimacy to the new government (if it holds). What I think is similar, however, is the focus on corruption, which I have argued helped galvanize the Colored Revolutions (and indeed was epitomized by the massive electoral fraud, which can be conceived of as corruption writ large ) and which has been cited as a cause of yesterday’s protests in Kyrgyzstan.

So the big picture question I want to throw out is the following: do yesterday’s events represent the end of the Electoral Model of regime/government change in the post-Soviet world? Has the failure of the electoral model to spread since 2005 led to its demise? Will this reverberate outside of the post-Soviet space? As I noted in my previous post, Iranian Tweeters have clearly taken notice (as one Iranian Tweeter wrote “We must learn #Kyrgyzstan. It took em 1day 2drive away the gov. They occupied gov buildings,Weapons & Fought back #IranElection”.). As the sun comes up on the United States, I’ll see if I can get some guest posts from some of the other scholars who have written on the Colored Revolutions, but for now I welcome any and all comments on the topic.

———————————

*In Bunce and Wolchik’s version of the electoral model they include cases where opposition involvement in the elections (and especially in monitoring the election) was sufficient to ensure the current government could not actually steal the election. So the steps of fraudulent results and massive protest are missing, but the actual results are the same: a new regime/government brought to power via an election. They thus include Croatia and Slovakia as examples of such cases. It is also worth noting that the Kyrgyz Tulip Revolution did not actually lead to the results of the election that inspired it — a parliamentary election — being overturned. Instead, a deal was cut whereby the president was replaced, but the parliament stayed put. For more, see here and here.