How Not to Report on Social Science Research

The Seattle experiment with a very high minimum wage has presented a rare opportunity for economists to study the impact of large minimum wage increases. This, in turn, could potentially contribute to a better informed public debate over the costs and benefits of minimum wage increases. Unfortunately, a report today in the Hutchins Roundup (published by Brookings Institute) provides a glimpse into how the interaction between social scientific research and media reporting is likely to play out. Based on that glimpse, I would have to say the likelihood of that interaction resulting in a better informed citizenry is negligible.

It is important to keep in mind that Brookings is a think tank, and therefore is more likely to report on the social science in a responsible manner than is a typical news outlet. And yet the Hutchins Report article begins with this headline: “MINIMUM-WAGE INCREASES LEAD TO EMPLOYMENT AND EARNINGS LOSSES FOR LOW-WAGE WORKERS.” If this were a report on a new consensus understanding among economists, then the headline would be entirely appropriate. Unfortunately, the headline simply summarizes the tentative conclusions  of one contested working paper published online recently by the National Bureau of Economic Research. One has to read to the bottom of the Hutchins Report summary to learn that the paper “drew immediate criticism from Berkeley’s Michael Reich.” And one has to follow the link to Reich’s criticism to learn that the authors of the highlighted NBER paper privately consider their study to be “a work in progress and that they are open to suggestions to improve it.” In other words, in private the authors of the study are adhering to the true spirit of social scientific inquiry — realizing that current findings are only tentative, and being open to evidence that they have erred. That is how social science is supposed to work. Social science requires studies like the NBER working paper to be made public so that social scientists can engage in the process of criticism and reassessment that drives us closer to the truth. But when the news media — and even think tanks — are quick to report preliminary tentative results from a “work in progress” with a headline that reads as though the work is widely accepted scientific claim, the impact on public perceptions is entirely predictable. Those who want to believe minimum wage increases are a bad idea will accept the tentative conclusions of this working paper as indubitable economic fact. Those who want to believe such minimum wage increases are a good idea will dismiss the study entirely and/or convince themselves that such evidence — even if widely accepted by economists — is irrelevant. A more responsible headline would read, “Contested Working Paper Finds Evidence that Large Minimum Wage Increases Reduce Hours Worked and Earnings for Workers.” If a think tank feels a need for something more attention-grabbing, we can be certain all mainstream news outlets will as well.

UPDATE: One article reporting on the same two minimum wage studies mentions the scholarly disagreement in its headline (“A Tale of Two Studies: Poor Research Leads to Poor Findings on Minimum Wage”), but then it goes on to argue one study is clearly superior to the other. This is considerably better than the Hutchins Report headline, but still a less than ideal approach to reporting on social science. This article in the Atlantic is even better.

UPDATE 2: On July 5th, 30dB reports that the UW study is dominating social media coverage, with “minimum wage+increase” getting only a 34% positive score over the last seven days.

Leave a Reply

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