Skip to main content
. 2022 Apr 1;110(2):205–211. doi: 10.5195/jmla.2022.1289

Table 5.

Performance metrics for gender detection tools after removing all unisex names we were able to identify (n=9,077 given names)

Gender detection tool errorCoded errorCodedWithoutNA naCoded errorGenderBias
Gender API 0.6558 0.2454 0.5439 −0.0174
NamSor 0.4271 0.2863 0.1973 −0.1488
Wiki-Gendersort 0.9400 0.3724 0.9042 0.1885