Oh, where to start? How about right where we're supposed to, with the headline:
Harvard Says July 4th Parades Make Kids Republicans
Which, LOLOL, is this the most ridiculous example ever of the media sensationalistically misrepresenting the findings of an otherwise boring study which--oh, wait, here's the abstract for that study, by Andreas Madestam and David Yanagizawa-Drott:
Do childhood events shape adult political views and behavior? This paper investigates the impact of Fourth of July celebrations in the US during childhood on partisanship and participation later in life. Using daily precipitation data to proxy for exogenous variation in participation on Fourth of July as a child, we examine the role of the celebrations for people born in 1920-1990. We find that days without rain on Fourth of July in childhood have lifelong effects. In particular, they shift adult views and behavior in favor of the Republicans and increase later-life political participation. Our estimates are significant: one Fourth of July without rain before age 18 raises the likelihood of identifying as a Republican by 2 percent and voting for the Republican candidate by 4 percent. It also increases voter turnout by 0.9 percent and boosts political campaign contributions by 3 percent. Taken together, the evidence suggests that important childhood events can have persistent effects on political beliefs and participation and that Fourth of July celebrations in the US affect the nation's political landscape. [Emphasis added for obvious WTF? reasons]
So it's not here is the mother of all causation/correlation fallacies, cum GOP ergo propter GOP? Was this "study" done for some econometrics class where students drew an estimator and a regressor out of two hats, and had to connect them in 40 pages, double spaced?
Because it's been published, but not "published." It's been consulted upon, but not peer-reviewed. It's been--DUDE, it's not even Harvard; it's "Harvard's Kennedy School," which is, let's be frank, "Harvard," or maybe Harvard™. Granted, Kennedy's not Old Navy to Harvard's Gap, but maybe Banana Republic outlet to Banana Republic. There are some asterisks and fine print on that brand label.
Hoo wait, what? What did they say their assumptions were? Whether it rained on their parade? Yes. They ran a linear regression using ordinary least squares on July 4th rainfall data--their proxy for parade participation--and county-level presidential voting results through 2008. And they found statistically significant relationships of between 1-2%.
And what's a key interpretation? That "the republican [capitalization sic] bias is consistent with the idea that there is a political congruence between the patriotism promoted on Fourth of July and Republican beliefs." Why yes, it may be. And what better outcome than a statistical model that apparently confirms a preconceived idea or a particular political view point?
Because let's accept, for argument's sake, that, fundamental logical fallacies aside, the findings are accurate. Don't statistically significant differences of such non-magnitude really thwart any attempt to draw such simplistic conclusions? And for all the infallible, surely [!] climate-change-adjusted data that went into the model, is OLS really the best technique for analyzing what has to be considered, doesn't it, a heterscedastic time series? Does it not seem odd that there's no mention of changes over time or by age cohort, even though basically every graph in the paper looks like this?
OK, maybe not exactly like this, but isn't it possible that other factors, which go utterly unmentioned in the paper, may have a more direct relation to geographically based voting patterns over the last 90 years than whether it rained on America's birthday? Could one of those things be, oh, I don't know, tectonic shifts in the very interpretive mechanism applied to the results, namely that of a congruence between "the patriotism promoted on Fourth of July and Republican beliefs"? Because I don't think that a 2009 Rasmussen poll about Republicans' views of the Good Old Days is too helpful for understanding the meaning and history of parade&picnic-based patriotism.
Or maybe it's the perfect explanation. The "Harvard" study shows this parade impact steadily declining, to the point that, as seen above, it actually inverts. From Gen X onward, Fourth of July parades are turning everyone into Democrats. Maybe the data actually shows that Republicans are losing/have lost the 4th of July crowd. Where's that headline?
The real value in the study, I suspect, is in showing how pleased Harvard students can be with themselves, and how oblivious or unconcerned their onanistic model-massaging can be for the actual world beyond Cambridge.
While it must have amused the hell out of the class in a way Physics Brony could only wish for, once the study got into the wild, its pat headline and near-parodically smug conclusions only end up providing conveniently self-satisfying reassurance to lifelong GOP voters who don't want to believe that their party has gone batshit crazy the last, oh, 3-11 years or so, and is delivering major damage to the country, the economy, and the few, lucky national guard troops who happen to be in between tours in Iraq when the parade comes to town.
Ultimately, these Harvard geniuses demonstrate that being the smartest guy in the room only matters if you never venture out of the room.
And anyway, this study is flatly contradicted by another major recent release, by Malick, Pitt, et al, which proved pretty conclusively that sun-dappled, all-American, small-town childhoods actually turn your kid into Sean Penn.
Harvard Says July 4th Parades Make Kids Republicans [thefiscaltimes.com via dt sr boston correspondent sara]
Shaping the Nation: Estimating the Impact of Fourth of July Using a Natural Experiment* [hks.harvard.edu, pdf]