Report comment

Report comment

Report comment

More to the point, what we think of as “stereotypically black-sounding” names are actually stereotypically lower-class-black-sounding names, so that introduces an obvious confounding factor. I wonder how you’d control for this.

Well, Bertrand & Mullainathan (2004) use level of education of mothers of children with a given first name, which they had direct data on, as a proxy for class status associated with a given first name. Unsurprisingly, the black first names in their study (Aisha, Latonya, Rasheed, Leroy, etc.) had uniformly lower levels of mother’s-education than the white first names (Emily, Meredith, Brendan, Jay, etc.). However, within each racial name group, higher levels of mother’s education did not correlate with higher rates of job-interview callbacks. (In fact, in each case there was a weak negative correlation—e.g., Aisha and Rasheed had the lowest callback rates but the highest mother’s-education scores among the black names.)

There are various plausible interpretations of this—e.g., that employers do judge by class stereotype associated with a given name, but level of mother’s education isn’t a good proxy for that; or that employers are misinformed about relative class status associated with some names; or that employers’ are sensitive to differences in class status between “higher” and “lower” but not within the categories—but one of them is that class status isn’t what employers are responding to and race is.

A study that actually paired black and white names of similar class status would be the best way to resolve this question, I agree.

Report comment

Report comment

Report comment

Deckbuilding and strategy patterns of thought also mixed with real world elements in these moments, but I don’t think I can explain how that felt, sorry.

Report comment

I’m now confused. Quoting directly from the abstract:

>… inclusion of
Arab characters in a nonviolent game was sufficient to increase anti-Arab attitudes…

and

>…Third, playing a terrorism
themed game even without Arab characters led to higher anti-Arab attitudes…

I don’t have time to read beyond the abstract, but that’s a direct contradiction to the portions of the passage you quoted. What’s going on? How did this pass review?

EDIT: Oh, I get it. There are 3 experiments. You’re quoting from passages which describe individual experiments, whereas I’m quoting from the abstract, and if any one of three experiments found a positive result they put it in the abstract.

I guess how you feel about this study depends on how you think multiple comparisons should be reported.

Just eyeballing the figures though, I think (assuming the non-displayed error bars are tiny, which might be a bad assumption) it does suggest that that people ARE primed towards “Arabs are terrorists” via contexts which are only about Arabs or only about Terrorists (rather than about both Arabs and Terrorists), which does indicate a long term problem rather than just a short term priming effect.

Report comment

Report comment

I also observe that someone, having found data that listed the zip codes of where the mothers were born in some birth data, went and did the heavy crunching to determine that actually the names didn’t make any difference in reality. (Cited in one of the Freakonomics books.)

Report comment