Monday, November 27, 2017

Lee Jussim on Liberal Bias in Social Science

From Lee Jussim, a blogger at Psychology Today:
Citation counts are one very common measure of how “important” a scholarly publication is. When others cite one’s work they are usually acknowledging its importance and drawing on its ideas. More citations, more influence and importance.

Now consider the storybook image of the scientist as someone who strives for objectivity. If it were true, studies of comparable scientific quality will be similarly influential, even if they produce different outcomes, because they both have comparable claims to reveal something true. But this is not the case. Papers in my home discipline of social psychology that can be used to craft narratives advancing social justice are generally cited far more than papers of equal or even higher scientific quality that contest those narratives. Here are two concrete examples. 
When a paper finds stereotype bias, it gets nearly 1,000 citations but when a failed replication of that same study gets published, it gets 30.

When a paper reporting a single study finds evidence of bias against women in STEM it gets 600 citations; when another paper reporting five studies finds gender bias favoring women, it gets 70 citations.

See this paper for several other examples.