Arnold Kling  

Papers that look interesting

PRINT
India's Achilles Heel... Biases in Evaluation Scales...

Glaeser and Sacerdote, suggesting that when rich people are around other rich people, they vote Democratic, but when they are around less-rich people, they vote Republican. Not having read the paper, I'm guessing on the basis of the abstract that rich people want to vote for Republicans for low taxes, and they are inclined to vote for Democrats on social issues; but people are somewhat chameleon-like on social issues, so the more social liberals you're around, the more socially liberal you become.

Krueger and Schkade, testing 229 women on subjective well-being ("net affect") two weeks apart, and finding a correlation of .64, with a correlation of life satisfaction of .59. They think this is OK. I'd say that most personal economic variables are more highly correlated at two weeks apart. I'd like to think that if you measured wealth at two-week intervals, the correlation would be something like .999.

In fact, I think it's worth thinking about how the difference in relative measurement error affects studies of the relationship between wealth and life satisfaction. Because life satisfaction is measured on a finite scale, measurement errors and "true" life satisfaction are likely to be negatively correlated. (On a scale of 1 to 10, you truncate a measurement error that otherwise would cause you to measure somebody as 11.) This might be sufficient to explain the observation that life satisfaction shows diminishing returns with respect to wealth--it may be purely an artifact of these measurement biases.


Comments and Sharing


CATEGORIES: Economic Methods



COMMENTS (3 to date)
conchis writes:

I think you're right about the correlation in the biases, but I doubt it's enough to explain diminishing returns as a statistical artefact.

Oddly enough, I'd always thought diminishing marginal utility of income/wealth/etc. was something most people had fairly little problem with, before anyone started trying to measure it. Now we're getting evidence consistent with it, and some people's priors appear to be shifting the other way, which doesn't make a lot of sense to me.

Matthew c writes:

Now we're getting evidence consistent with it, and some people's priors appear to be shifting the other way, which doesn't make a lot of sense to me.

Perhaps people who buy into a highly materialistic paradigm of human behavior tend to see their beliefs as applicable almost everywhere, in the same way that people obsessed with a fundamentalist religious paradigm tend to view every phenomena within the lens of sin, faith, and religious fervor?

Dog of Justice writes:

My interpretation of the Democrat vs. Republican result is that most people estimate the effects of policies by extrapolating from themselves and the people they see every day. If you're around lots of other rich people, you're likely to assume people you haven't met think similarly to you and them; and vice versa.

Comments for this entry have been closed
Return to top