Accuracy Gets No Notice

The December issue of The Atlantic Monthly contains a rather interesting article [“I was wrong, and so are you”] by Daniel Klein, a conservative/libertarian, who had published an op-ed piece in the Wall Street Journal in June of 2010 arguing that, based on a study that he and another economist had earlier conducted, liberals/progressives had a far poorer grasp of basic economics than did conservatives.  Right wing and conservative groups trumpeted the results, and comments on the study were the second-highest of anything published in the Journal for the month in which it was printed. Klein’s in-box was also filled with messages suggesting that he had rigged the study.

After considering the reaction and the criticisms of the analysis of that study [which had been designed for another purpose], Klein and his co-author designed a second study specifically for the purpose of evaluating the accuracy of people’s economic perceptions and comparing their political outlook to the accuracy of their economic views on various issues.  To Klein’s surprise, the second study indicated that [astonishing] that all across the political spectrum of the respondents, each group was equally wrong when evaluating the accuracy of economic statements at variance with their political beliefs. As Klein wrote, “the more a statement challenged a group’s position, the worse the group did” [in accurately evaluating the statement].

In short, in all cases, respondents were less accurate in economic judgments that conflicted with their underlying biases and views, and the greater the conflict, the lower the accuracy.  What was even more interesting was that the level of education seemed to matter very little or not at all.

To me, all this was scarcely surprising, but what was surprising was that, while scholarly reviewers found the new study accurate, there was essentially no public or media reaction to the release of the results of the follow-up study, even though Klein was very clear in declaring that the new study invalidated the results of the earlier work.  Given that the results of the second study were also at variance with Klein’s own political predilections, it would seem likely that there might be at least more than polite notice of the second study.

There wasn’t. The few academic/critical reviewers who did comment essentially said, “there’s a lot of confirmation bias out there.”  The conservative/right wing types have said nothing, in contrast to their trumpeting the earlier [and incorrect] work, and there seems to be little liberal reaction either.

In short, we all want to hang on to our biases, even in the face of information to the contrary, and the more that information challenges what we believe, the more strongly we dispute it.

Is it any wonder Congress can’t get anything constructive done?


6 thoughts on “Accuracy Gets No Notice”

  1. Joe says:

    Although I got most of their questions “right”, their answers seem more economist dogma than empirically proven facts.

    For instance, consider “minimum-wage laws raise unemployment”. In the abstract this is obviously true (if it costs more to employ someone, you’d expect fewer people to employ people). But, we’ve simplified the problem beyond reality.

    In the real world, each hour of life has a minimum cost (the cost of feeding oneself to live that hour). That minimum cost is a structural feature of the society one lives in. If the minimum wage law is below that number, it will have little effect since few people accept suicidal wages. If the minimum wage law is much greater than this number, it will have an impact. However there is a region in-between where increasing wages is unmeasurable, and may be beneficial (improving productivity since people worrying about their next meal don’t work well.)

    One has to be very careful simplifying reality into a mathematical model because it is very easy to prove something incorrect. For instance:

    Thus, I think the new study also appears to suffer from confirmation bias.

    1. Wayne Kernochan says:

      A quick add-on to your note, Joe: there was actually a detailed study in New Jersey a few years back that found that in that case, raised minimum wages had no or a positive impact on unemployment — and cited various apparent reasons. Michael Parkin’s textbook notes this study, and then cites several arguments from non-Keynesians as to why either the study might be inaccurate or might be too focused on one particular aberrant case. However, none of them actually, according to Parkin, did studies to verify their counter-arguments. Parkin concludes that the study is likely to be wrong and that therefore raised minimum wages will indeed cause unemployment to increase!

      That said, I think Klein is basically correct, and the liberals and conservatives I know have, in general, equally poor understandings of economics, often from a political bias — and I don’t exclude myself from that bias. However, I do find that at present, Neo-Keynesianism as represented by Krugman, Stiglitz, and deLong has a far better track record in both analysis and prediction, and a far better balance of real-world testing and attention to key factors poorly approximated by classical macroeconomics, than any alternatives, including the Lucas approach, the supply-side theories, and the classical economic theories represented by Fama, Mankiw, Taylor, Feldstein, and Parkin himself. To give one small example, based on their analysis I have been strongly maintaining for the past three years that inflation was nowhere near around the corner, to general economicallyl-informed disbelief — and I was right.

      That Neo-Keynesianism is so strongly denigrated as liberalism is, to put it bluntly, a problem of conservative bias, not liberal bias or “moderate” bias. I sure hope things change and I can go back to being humble 🙂

    2. Which tends to support the idea that everything suffers from one degree to another from confirmation bias.

  2. Wine Guy says:

    This is no surprise to anyone in the medical field. Take for instance the debunked idea that the MMR vaccine causes autism. THE STUDY’S OWN AUTHORS have printed at least 2 different redactions: people still cite the original study and seem surprised when the redactions are brought to their attention (despite being widely reported).

    Confirmation bias, indeed.

    1. Wayne Kernochan says:

      If you are referring to the same thing, the original English study was authored by an English doctor and there were questions even at the time about his conclusions. As a father of an autistic child who has had to put up with many obviously doubtful “cures” inflicted on my son, I have been following the whole business with interest but a lot of skepticism. A British reporter has been chasing down information about the study since then, and has apparently recently reported that he finally got access to the original data, showing that it did not at all support the conclusions in the paper.

      In other words, never mind the redactions; the whole thing should never have been published in the first place. Although obviously we can’t be sure, it does appear that most of the controversy stems originally from continuing dishonest behavior by the doctor, originating in and following from that original apparent deception.

      1. Steve says:

        Dr. Wakefield was a shill for lawyers representing a group of anti-vaccine groups. He also stood to make money from some alternative treatements he held patents on. He should be prosecuted for the number of preventable measles/mumps/rubella complications and deaths above the expected. The CDC should also seek damages from Jenny McCarthy, Oprah and the media companies that gave them airtime to spew falsehoods. It is the same as shouting fire in a public venue.

Leave a Reply

Your email address will not be published. Required fields are marked *