Submitted by Eugen Tarnow, Jul 25, 2007 06:28

In the paper "What Drives Media Slant" one of the conclusions asserted is that "[newspapers] respond strongly to consumer preference".

While the paper is a great read and based on what seems to me to be correct and inventive research including plenty of data collection, I believe the conclusion above is NOT supported by the findings in the paper: in the paper there is NO investigation of time dependency and thus there is no way to say that newspapers "respond".

It seems to me that the authors and the news media quoting this article have committed the common statistical error of assuming that correlation implies cause and effect. Indeed, committing the same error one could draw the equally invalid conclusion that consumer preference is dictated by the newspapers.

In the paper "Media Bias and Reputation", the research is based on hypotheticals rather than gathered data. As such it really deserves no circulation at all in the common press. In any case, the model that emerges: "[newspapers] slant their reports toward the prior beliefs of their customers in order to build a reputation" might be incorrect and, for that matter, could go the other way as well - customers slant their beliefs towards the newspapers they read.

The question is still out there - do newspapers shape the readers or does it go the other way?

Note: Comments are screened, and in some cases edited, before posting. We reserve the right to reject anything we find objectionable.