Of the variables that were investigated as possible moderators of this particular practice (e.g., woman’s age, age difference between the husband and wife, race of the husband and wife, woman’s educational attainment, woman’s occupation, income share across the husband and wife), the overwhelmingly strongest predictor was educational level of the wives. Here are the odds ratios (as compared to women whose educational levels were less than a bachelor’s degree):
Women who hold a bachelor’s degree: 1.7
Women who hold a master’s: 2.8
Women who hold a professional degree: 5.0
Women who hold a doctoral degree: 9.8
The latter odds ratios (all of which are highly statistically significant; p < .001) capture the increased likelihood of not using a husband’s last name. For example, women who hold doctoral degrees are 9.8 times more likely to use a “nonconventional” surname (i.e., something other than simply their husbands’ family names) as compared to women whose educational level is less than a bachelor’s degree.