Monkey Cage reader Ken Wedding forwarded me this post at The Economist, which piggybacks on Jeff Frankel’s post noting that states where majorities voted for McCain in 2008 received more federal spending per tax dollar in 2005 than did states where majorities voted for Obama. I re-did Frankel’s graph to rely on the actual spending per dollar rather than the state’s rankings, as Frankel did. (For more on why rankings are flawed, see here.)
The question is what this graph tells us, if anything. Here is The Economist:
This leads to situations where states that absorb huge amount of government aid (particularly for agriculture) are hotbeds of Tea Party activity, where voters decry the heavy boot of the federal government on their backs. It’s tempting to make this a huge gotcha point and slam Tea Partiers for cognitive dissonance, but the holding of conflicting beliefs is one of America’s deepest and most common traditions.
And so we add another chapter in this blog’s history of pointing out the ecological fallacy. Once again: from aggregate data, we have absolutely no evidence of whether any individual—whether a Tea Party supporter or not—holds “conflicting beliefs.” Perhaps the vast majority of residents of these states have entirely consonant beliefs: they supported McCain and oppose spending, or support both Obama and spending.
No one is revealed as a hypocrite by this sort of analysis.
{ 3 comments }
Once again: from aggregate data, we have absolutely no evidence of whether any individual — whether a Tea Party supporter or not — holds “conflicting beliefs.”
That’s true in this case, but in some cases you could use purely aggregate information to make reasonably definitive claims about individuals.
If the two beliefs in question are both popular enough that even if you assume the maximum possible disjointness there must still be a nonempty intersection, you can reasonably infer that at least some individuals hold both beliefs.
In this case, the pool of Tea Partiers is small enough that I imagine they could be fully disjoint from any other belief or group. But I would not be shocked if in at least some states, the pool of Obama voters is not large enough to swallow all of the (voting) spending-favorers, in which case there would have to be some McCain-voting spending-favoring individuals.
I’m much less interested in the political correlation, which I think is indirect (due to correlation with population density), than the variance. The difference between 65 cents on the dollar (CA, NY) and $2 on the dollar (NM, MS) is a factor of three (!), on federal taxes which themselves come to about 20% of GDP. This is like an average wage gap of 38%!
Sorry I just realized I did something stupid above — effective wage gap is more like 27%, which is still very high.
Comments on this entry are closed.