The Wall Street Journal falls foul of Freakonomics style analysis and errors
by Benjamin on September 2, 2009
The Wall Street Journal yesterday explained why football coaches vote Republican by revealing that college and NFL football coaches contributed a ‘massive’ $13,000 to the Republican presidential campaign, whereas they only contributed $4,600 to the Democrats.
For the WSJ this is grounds for some serious soul-searching as Obama appearantly “outraised Mr. McCain by more than a 5-to-1 margin”… Never mind that the figures (via Wikipedia even) were Obama $103m : McCain $42m or Democrat $321m : Republican $261m. Anywho, the WSJ proceeds to make the argument that football coaches have a winning mentality, discipline, down-to-earthness “and [are] just conservative, I think. It fits with the Republicans”. Right… Lets just check that bit about small samples, ‘un-contextual’ data and spurious results – which caused such trouble for the Freakonomics guys. 23 coaches donated money, 20 of which gave (an average $650) to the GOP, and 3 for the democrats. There are some 32 NFL teams and 64 ‘top flight’ College teams, so the sampling of individuals is far from representative – or more correctly, 76% of football coaches didn’t care enough for either party to donate – and when they did, democrats gave more on average than the republicans.
But lets forget all that – lets say the WSJ is on to something. Why might football coaches be more keen to vote republican than democrat? Maybe… Just maybe, they fit the republican voter base? Here I am not talking about the idea that according to the WSJ ‘many’ coaches are from the Southern States – is that true? – nor that “Republicans tend to revere strong, singular executive leaders—a pretty good description of a coach” – whatever that is supposed to say about the republicans and democrats… Maybe it’s just that the people who make more money vote for those who wish to keep the top bracket tax low? But are the coaches in the top brackets?
Yep, college football coaches make an average 11 times more than the average professor, clocking up $1m each on average well illustrated here or written about here, and the NFL coaches averaged some $2.5m in 2003, and by now it’s surely more, with Seattle’s Mike Holmgren leading the pack on an average $5.1m every year. So let me get this straight, of the 24% of football coaches who bothered to contribute to the presidential election, the 20 guys who gave to the republican party each parted with a hefty 0.026% of their annual income… That is equivalent to me buying a lottery ticket. Somehow I am not sure me buying the lottery ticket re-affirms my utter and complete belief in me winning the big prize… No, I’m sorry, the WSJ piece is a cute little piece of pop-economics, which one should probably not read very much into – oh, and for those curious about the relation between income and voter decisions, here’s a nice graph from one post 2008 election analyst:
Low income people on the left... Pun intended :)
Data running back to 1980 is available, and shows the same trend. The lowest income brackets 0-$15k and $15k-$30k, have consistently voted democrat, by almost 2:1 under Clinton and Bush. The highest income brackets $100k+ and $200k+ when available voted republican by almost the same proportion…
Tags: football, Freakonomics, income, Paul Krugman, PhDcomics.com, Politics, statistics, voter
I find the whole “Rich vote Republican, poor vote Democrat” argument rather spurious. Correlation =/= causation. What is the strength of any relationship between income and voter identity? Looking at electoral college results, those states with the highest median income regularly vote Democrat (CA, NY, MD, nearly all of New England, HA, etc) while those states with the lowest median income tend to vote Republican. Taking into consideration anecdotal evidence, our wealthiest legislators are nearly all Democrats. Hollywood, save a few B-listers that the Republicans like to trot out during campaign season, are nearly all Democrats. Warren Buffet, George Soros, etc, all Democrats.
I agree that there is (or historically was) a plausible relationship between income and political affiliation, but income does not necessarily cause that relationship. There are factors beneath the surface that manifest themselves in income as a quantifiable measure, and these less visible factors are the root causes of political identity.
“Football coaches aren’t the most diverse group, which may help explain their political similarities.” No kidding, who knew? Oh wait, we all did.
Joe, having checked the figures I have to agree that the republicans do indeed tend to win states with lower than average median income (18 in total, to the democrats 12) as on the below graph from this data:
That said, this doesn’t tell us anything about how the voters in the state voted, nor what the income levels of each voter is. Personally I don’t think anecdotal evidence is the best sort of evidence as the number of observations is too low. But if you believe in it, then you have to believe even more strongly in the anecdotal evidence given by the exit polls. As these represent thousands upon thousands of anecdotal evidence, they should presumably be more reliable. The Book “Red State, Blue State, Rich State, Poor State” is a pretty good read on election statistics, and they agree with your (and now also my) observation that lower than median income states tend to vote more republican. But they also emphasize that within these states, it is the richer ($50k+) segment of the population which tends to vote Republican… And that I think is a pretty interesting observation.
Hey Ben,
I followed the link to the review of Freakonomics, and I found it very weak (the review). So if I was picky, I wouldn’t say that the guys from freakonomics have been in trouble just because of this little piece of writing!
Best!
Clement
Granted Clement, granted.
I’ll be honest, there was a much longer debate about the issue, so I grabbed the first friendly review I could find. Something of more substance is below. I should probably add quickly that I enjoyed the book, as it was a great fun read, but I think there are some real issues which are glossed over…
Ariel Rubinstein (NYU) wrote a proper review of the book well worth a read.
and on the most controversial point that Legalising abortion reduced the crime-rate, there was a serious issue.
What do you make of these, especially Rubinstein