Which raises the question of whether it is even appropriate to “de-bias” one’s thinking when dealing with an individual who has grown into the stunted role of an unpronounceable name. (Anecdotally: those with dissonant names have a higher likelihood of being socially-awkward or disagreeable)

Laham and Alter appear to be heading in other (various) directions with their research, but is anyone looking into the personalities which grow around a “Sovanrithy” or a “Deolanda” ..?

Let’s look some of the results of that study, focusing on the data from their Table 8., which gives the call back rates for all the names used.

WHITE: BLACK

Female Female

Emily 7.9 Aisha 2.2
Anne 8.3 Keisha 3.8
Jill 8.4 Tamika 5.5
Allison 9.5 Lakisha 5.5
Laurie 9.7 Tanisha 5.8
Sarah 9.8 Latoya 8.4
Meredith 10.2 Kenya 8.7
Carrie 13.1 Latonya 9.1
Kristen 13.1 Ebony 9.6

Male Male

Todd 5.9 Rasheed 3.0
Neil 6.6 Tremayne 4.3
Geoffrey 6.8 Kareem 4.7
Brett 6.8 Darnell 4.8
Brendan 7.7 Tyrone 5.3
Greg 7.8 Hakim 5.5
Matthew 9.0 Jamal 6.6
Jay 13.4 Leroy 9.4
Brad 15.9 Jermaine 9.6

Does everyone see how the Black names that would be most difficult to pronounce from their spelling (e.g. Aisha, Keisha, Tremayne) got far lower callback rates than more pronounceable Black names (e.g. Leroy, Ebony, Kenya)? Prima facia, the easy-to-pronounce Black names look like they were called back at a rate typical of the White names. There were no difficult to pronounce White names.

In the section of the paper entitled “Interpreting Callback Rates,” the authors said “our results must imply that employers use race as a factor when reviewing resumes, which matches the legal definition of discrimination.” (21).

This claim must now be revisited.

http://www.economics.harvard.edu/faculty/mullainathan/files/emilygreg.pdf