Biased impartiality in the media

Half the internet seems to be talking about the US Presidential primaries at the moment, so Jeremy Burke’s work on media bias seems particularly pertinent. I might talk about his other papers another time but the one that’s captured my attention today addresses the myth that balanced media coverage is equivalent to fair coverage.

It is usually accepted that the media should provide unbiased coverage of events and issues. Unfortunately it is difficult for the reader to tell how biased the coverage is if they have no prior information about the issues being reported. In the absence of further information it can appear that coverage which reports both sides of a contentious issue is the least biased. In the competitive world of large media conglomerates a reputation for accurate, unbiased coverage is essential. The competition drives reporters to appear impartial by simply reporting the arguments provided on each side of a debate. Of course, providing both sides of an issue and reporting it in an unbiased fashion are not always the same thing.

Consider the issue of tax cuts in the current US Presidential race. Predictably, the Republican candidates favour them in order to stimulate investment, and the Democratic candidates generally oppose them because they necessitate cuts in public spending. There are undeniably good arguments to be made on both sides of that political divide. However, many Republican candidates have tried to defend their tax cut policies by claiming that they increase revenues and won’t require any cuts in federal spending. Now this is plainly false (linked to two very good economists cos I know that even factual statements are not uncontroversial once they’ve been politicised). So did the media coverage point out that the candidates had made an error by claiming we were on the right hand side of the Laffer curve? Not at all: they provided ‘both sides’ of the issue by reporting what the candidates said as a legitimate point of view.

This graphically illustrates the distinction made above: to report in an unbiased fashion would require the media to tell readers that the Republicans’ statements were not supported by reputable economists. No doubt this would have been interpreted by many who are unfamiliar with the economic research as exhibiting ‘liberal bias’. In order to be perceived as unbiased, the media were forced to omit important information from their coverage. Burke’s paper shows that the desire to appear impartial leads to information losses in the media, and that competition between media outlets only exacerbates the problem. Sometimes the only way to get the extra information is to go to partisan media outlets who have a vested interest in reporting it and use that information to fill in the blanks for yourself.

  • The benchmark study on media bias remains Groseclose and Milyo: brilliant paper. Start with ADA scores for Congressmen. Impute an ADA score for think tanks based on which congressmen cite them. Then, impute an ADA score to media outlets based on which think tanks they cite. Instead of having to benchmark against some external measure of unbiasedness, they measure it against median voter preference as indicated by the ideology of the median member of Congress. One of my favourite articles of the last decade.