August 26, 2004

Top 1% Income Poll Was Urban Legend

In comments at Asymmetrical Information, Daniel Davies debunks the story about how 19% of Americans think that their incomes are in the top 1%:

What happened is that there was a tax proposal which would only (or chiefly) benefit a particular set of high earners, about 1% of the population. Someone asked "Do you think this proposal will benefit you?" to a sample of people. Nine [recte: 19] per cent said "yes". This was then reported as "10% [recte: 19%] think they're in the top 1%", rather than "8% [recte: 18%] don't understand this tax proposal properly".

And this seems to be exactly right, according to this op-ed by Brookings Institution folks Carol Graham and H. Peyton Young:

An October 2000 Time-CNN news poll showed that 19 percent of Americans thought that they were in the high income group that would benefit from proposed tax cuts - defined as roughly the top 1 percent of the distribution.

If the poll had asked "Are you in the top 1%?" there would be no need to define this group at all; so it seems like the poll must have been as D-squared describes.

This disturbs me from the point of view of the epistemology of testimony (to put it in a highfalutin way)--I believed this non-fact, and saw it passed around by commentators I usually trust (DeLong, Kleiman), but it just didn't hold up. It's David Brooks' fault, but we should all have known better than to believe him.

One problem is that (as Kleiman notes) it's durn hard to find the results of specific old polls on the net; at least, I couldn't do it with this one, and I tried fairly hard. Anyway, I wanted to record this, in the hope of fighting the urban legend a bit.

Two notes:
--There probably are theories of propositional attitude ascriptions on which, given that only the top 1% would benefit from the tax cuts, "19% of Americans think they're in the top 1%" comes out true in some contexts. But it should come out mostly false.
--I'm not sure it's quite fair of D-squared to say that the claim entailed "that 22.5 million people held wildly delusional views about matters which were right in front of their noses." My income is right in front of my nose, but the percentile of my income is not; all I can say with confidence right now is that it's below the median household income, and I follow these things more than the average.

Posted by Matt Weiner at August 26, 2004 02:35 PM

I suspect there's something to the intuition that many taxpayers overestimate their relative income. But, ask a sloppy question, get an unilluminating answer. I wonder what percentage of Americans literally believe they are among top 1% of income earners. Someone should ask them explicitly.

Other interesting questions: How many people do you know in the top 1%? How would you rate your (your children's/your grandchildren's) chances of joining the top 1%.

I'd also like to know what percentage of Americans believe they belong to the bottom 1% of income earners.

Posted by: Lindsay Beyerstein at August 27, 2004 07:35 PM

I'm not sure I would say that I had an intuition that many taxpayers overestimate their relative income--I would call that argument more of "a desperate attempt to save face after Daniel Davies, who is usually right, pointed out that I was being an idiot."

OTOH, D-squared himself often points to studies showing that people overestimate their own capabilities unless they're clinically depressed.

One of the interesting things about this poll is that it was usually reported as saying that 20% of people expected to be in the top 1% sometime during their lifetimes. I wonder how that was framed in the poll. Anyway, I think "Do you think you'll benefit from this tax plan?" was a good question to ask--if David Brooks hadn't got hold of it for his own nefarious purposes, everything would be fine.

Posted by: Matt Weiner at August 30, 2004 10:30 AM