This post originally ran 1 July 2011. The authors never responded to my offer.
Democrats! I beseech thee! Do not let thy children fly the flag on the Glorious Fourth! The Red, White, and Blue are the devil’s colors. The bands that march down main are akin to crack to small ears. Exposing thy issue to fireworks might turn them into—gasp—Republicans! 
Or so says Harvard Kennedy School Assistant Professor David Yanagizawa-Drott and Bocconi University Assistant Professor Andreas Madestam. Neither, you will notice, has tenure. Not yet. I don’t know much about Bocconi U., but Harvard is infamously left wing to the point of falling off the cliff, so this new paper surely boosts Yanagizawa-Drott’s chances of gaining the prize.
But not if this work is judged on its statistical merits, which are well short of a what a sugar-addled 6th grader produces when trying to guess how many blue M&Ms are in a bag of candy for his science fair project. If I were the Dean of Kennedy and I saw this, I’d quietly slip brochures of Yale’s generous employment benefits under Yanagizawa-Drott’s door.
It’s being reported everywhere, that the pair of “researchers” said that “Fourth of July celebrations in the United States shape the nation’s political landscape by forming beliefs and increasing participation, primarily in favor of the Republican Party.”
They also claim:
The political right has been more successful in appropriating American patriotism and its symbols during the 20th century. Survey evidence also confirms that Republicans consider themselves more patriotic than Democrats. According to this interpretation, there is a political congruence between the patriotism promoted on Fourth of July and the values associated with the Republican party. Fourth of July celebrations in Republican dominated counties may thus be more politically biased events that socialize children into Republicans.
How did they discover these dirty little secrets? Why, by using statistics to “study the impact of Fourth of July during childhood on partisanship and participation later in life.” Sounds dull but plausible, in a typical academic way. But how did they measure childhood 4th of July participation? Are you ready to hear?
Many of you will say, “Yes! Tell me!” But I caution that if you are squeamish not to read farther (or is it further?). What is once read can never be un-read. I don’t want you to lose what little faith you had left in the academy.
Ready?
Our method uses daily precipitation data from 1920-1990 to proxy for exogenous variation in participation on Fourth of July as a child. The estimates imply that days without rain on Fourth of July in childhood increase the likelihood of identifying with the Republicans as an adult, voting for the Republican but not the Democratic candidate, and voter turnout.
Dammit, I should have added a legal disclaimer before that quotation. God knows how many people were eating when they read the quote, a conjunction of events which almost certainly lead to choking. To my reader’s lawyers: I warned them not to continue!
Understand: they did not actually measure anybody’s 4th of July participation, they merely asked if it rained when one was a kid. Forget the error in measuring rainfall on parade spots and whether small amounts of rain or large had the same effect; forget, too, of knowing where people were when kids, of whether they could or couldn’t go to parades regardless of weather, of whether the cities where it rained or shined had a history of Republican or Democrat activism, we’re using statistics here! This is science!
Our pair quotes newspapers to learn such things like, “Rain keeps crowds thin, ends some festivities…” Get it? From this they deduce, “Random rainfall therefore provides plausibly exogenous variation in participation in Fourth of July celebrations.” When is the last time you heard somebody casually drop “plausibly exogenous” in a conversation? Yet another benefit of attending Harvard.
Even though they admit that, “There are some issues with matching and measurement of rainfall for individuals during childhood” (using issues as an inappropriate synonym for impossible problems), our pair say their “findings are significant,” as in statistically significant. Do you now see why I claim this term should not be allowed? These main findings are:
[O]ne Fourth of July without rain before age 18 increases the likelihood of identifying as a Republican at age 40 by 2 percent the share of people voting for the Republican candidate at age 40 by 4 percent, and the share of people turning out to vote at age 40 by 0.9 percent.
If this is so, San Francisco, Los Angeles, and other sunny sumer cities should be hotbeds of rapid Republication roisterousness. Just like we see in real life, right? Right?
But enough. Even I, well used to grotesque pathological statistical specimens, find myself growing nauseous. And I still must make my 4th of July plans.
I call on my colleagues to admit that any paper can eventually be published in some journal and thus claim the label of “peer-reviewed” study. As idiotic as this work is, it will still find a home, which will allow it to be “cited” by other authors (many of whom will not be bothered to read the original), and so take on the appearance of “truth.” Very depressing thought, no?
I emailed both authors and offered them space for a rebuttal. I’ll let you know if they reply. In case you did not see it above, here is the link to their original paper.
Update I don’t know how many of you considered the full implications of this most frightening study. As of Saturday morning, the forecast for Monday at San Francisco, Los Angeles, New York City, probably in Boston and all along the east coast, and on the west, too, including Portland and Seattle, calls for bright sunny skies. No rain, friends. The weather is conspiring to unleash a wave of nascent Republicanism on these once Democrat strongholds.
Look out, DNC, it’s even worse! Because global warming—a.k.a. climate change, a.k.a. climate catastrophe—will bring even more sunny skies and parade-friendly weather. Waves of clement afternoons and of conservatism are set to sweep over the country. Quelle catastrophe!
Perhaps prior knowledge of this study is what has been driving Democrats to so vociferously advocate direct government action to combat global warming?
(Those who laugh at these projections have failed to understand they are exactly what are predicted by the paper.)
From the paper: To check whether the effects are truly driven by Fourth of July celebrations we run placebo regressions for all the outcome variables. I assume a placebo regression is one that generates a low p-value without containing any actual facts.
and
The notion that the family is an important agent of socialization ties in with a recent
theoretical economics literature emphasizing parents’ role in value transmission. D’you know, I would never have guessed.
My lawyers will be in touch Mr Briggs since you failed to post a warning ahead of the links you provided.
Already the fireworks start! Have the Republicans cornered motherhood and apple pie as well?
I call on you to publish a paper in a journal that backs up your claim that any paper can eventually published and earn the peer-reviewed label.
Doug Ransom,
Ha! Just look at my list of works to see that journals will take anything. Maybe I should submit to the Far East Indian Journal of Statistical Advances, Part IV (fictional, but only by a hair).
This all reminds me of the book “Full House†by Stephen Jay Gould. One of the points made in this book is that some fields of human endeavour have reached a point of diminishing returns. A similar claim is made in “The Rise and Fall of Modern Medicine†by James Le Fanu. Thus, with the pressure of producing ever new innovation in the fields of study affected with this problem and the large number of active researchers, we are afflicted with a growing number of bizarre, dubious, or just plain banal work. I believe that at any given point in time a field can support only so many researchers and usefully consume only so much funding. Anything beyond that is a corrupting influence. Of course, determining that, if you will excuse the expression, tipping point is difficult. I can see a follow up study now: The effect of July 1 celebrations in Canada on voting patterns. I will be watching the local fireworks tonight.
WAIT!! There are more republicans in the south and midwest than in the northeast and on the west coast? Unbelievable.
It’s quite interesting that those two learned academic doofusses [but I repeat myself] in their rush to publish have stumbled on a tiny segment of the political-party-selection determinative element they so casually seek. Fourth of July Parades are part of the process, true, but the science is settled that the true causative factor is not the parade itself, or the potential for rain thereon, etc., but the preponderance of the brands ormakes of fire engines in the parade itself that set in motion the once-in-for-all an adolescent’s political choice maturation.
If the engines were LaFrance or Van Pelt – foreign names, to be sure – youths standing agape on the sidelines were destined to become “Democrats”. If the names emblazoned in chrome on radiators were Pierce, Seagrave, Ward or several other common “all-American” variants, future “Republicans” would abound.
This is all quite simple. No rain records need be researched in making this report. But, upon reflection, what else does a Harvard man have to occupy his time?
Most people get more rightwing as they grow older. I blame taxes. Memory fades, I can rarely remember rain as a child.
This paper is s&@£.
Orwell commented that a certain amount of intellectual sophistication is needed to believe in nonsense. Those researchers are obviously very sophisticated.
San Francisco is a sunny summer city? Haven’t you read you Twain?
When we have fireworks shows, they launch the rockets into the fog, and it glows in pastel pinks and baby blues.
Doug M,
Twain was cold, not wet.
If I only knew then what I know now. To think, had I only saved all those regressions I ran in my first year of econometrics when we got access to stata, just to see what would happen when I put random data together. I could have published dozens of papers and become a famous economist, like “The Prophylactic Effects of Advanced Telecommunications” because there is an inverse correlation between the number of provisioned T1 lines and malaria rates in African countries.
William Sears: Tyler Cowen’s “The Great Stagnation” makes a similar argument. We’ve used up all the low hanging fruit, so economic growth is slowing down because now we’re down to refining the world changing technologies, not implementing new ones. It’s only a hundred pages and goes fast, makes for an interesting read.
“David Yanagizawa-Drott”: Wodehouse would have been quite proud of that.
Darn, I regret taking my children to the July 4th fireworks in several big cities to celebrate their birthday. They can be patriotic, but are not to become republicans. (^_^) Oh… on second thought, I shouldn’t be worried about the extra 2%, 4%, and 0.9%.
Mr. Briggs,
Maybe you are right. And mayhap the advocacy has something to do with the following.
I am not one of those Democrats so I am putting out another theory for fun.
(Can’t comment on the gay marriage post any more? I just got a chance to read it carefully.)
But not to worry, the environmentalists are moving to abolish July 4 fireworks celebrations, starting with California. In this way they will nip future Republcians in the bud. Too bad that the judge will allow fireworks in La Jolla, at least one more time http://www.10news.com/news/28120843/detail.html
I beg your pardon?!?! Perhaps a spoof? Rather like, perhaps, Jo Nova’s analysis of U.S. Postal rates and global warming, and her finding a better statistical correlation with the postal rates?
No? Really peer-reviewed and published? Goodness, what a finding! I’m sorry to say this, but I’m afraid I will retch all over my computer if I try to read the actual paper. Thus, I will ask if perhaps you can advise on how they addressed what I see as not insignificant confounding variables such as:
1) the evolutionary nature of the migration/movement of people across the nation during the past several generations, as they followed jobs, attended college, etc.;
2) the effects of immigration and naturalization;
3) the role of education;
4) the impact of discussions with family, trusted friend, neighbors and colleagues/co-workerson development of thoughts (re: “political leanings”);
5) the potential for religious teachings and beliefs (if any) to influence political leanings;
6) the effect of alien abductions on personal political thought; and
7) the role of relative family wealth on personal political thought.
Okay, alien abductions are probably too few to make a difference, but what about the other items???
Thanks for bringing this remarkable work to our attention. (I think…)
“stats” for AGW and Bomb Scares; please Fisk! Thanks!
http://www.jsedimensions.org/wordpress/content/bomb-threats-global-warming-and-decision-making-an-educational-experience_2011_05/
This is a serious question having skimmed the paper out of general interest. Now, my statistical education pretty well ended in the 1980s at (UK) “O” level – whatever that is in your money, with a bit of a top-up from a Computing degree as a mature student over the past few years.
All the same, from that humble level of understanding, the following statement appears to have no logical basis whatsoever:
“We first compare the contemporaneous change in preferences and behavior for adults, who experience a rain-free Fourth of July, with those that do not. This establishes whether Fourth of July affects individuals contemporaneously as adults.”
Establishes whether one affects the other??? Surely such a comparison does no such thing?
I clearly remember my high school stats teacher taking great pains to drum into us that finding a statistical correlation does NOT, of itself, establish any sort of cause and effect let alone establish in which direction such a relationship operates if it does exist. Yet here seems to be a statement to that effect on page 3 of a peer reviewed paper by two Professors!
So, my serious question is:
Has that fundamental principle, drummed (endlessly) into us by Mr Blackwell, changed in the past 25 years or did I misunderstand?
Alert the IPCC! They can use this sort of reference in the next report.
I’m residing my house on the fourth. Probably makes me Republican too. So does painting. Sad…..
Let me offer a small bit of counter evidence to the study in question. Bristol, Rhode Island, tomorrow will hold the 232nd consecutive celebration of the 4th of July where not only flags fly in abundance, but the line down main street is painted red, white, and blue for the occasion. Yet Rhode Island has been controlled by a Democrat legislature since 1937 (that’s 80 consecutive years for the math-challenged, which coincidentally many RIers are) and last year gave Hillary 54.4% of the vote. These Harvard guys either can’t be bothered to take a little field trip down Rt. 95 to test their claims … or they don’t want to.
References:
http://www.july4thbristolri.com/
https://www.nytimes.com/elections/results/rhode-island-president-clinton-trump
Harvard: Because not everyone can get into MIT
In my highly diverse neighborhood, the fireworks are mostly shot off by young blacks, who are probably not republican. I don’t know if the latinos shoot off fireworks, but if I hear their music, I’ll figure they’ve decided to celebrate the 4th. They are more likely to be La Raza than Republican.
Meanwhile, I look around, see America still hasn’t been made great again (whatever that actually means, I presume the change would be noticeable), and think, bah humbug. Harvard is behind the times.