Occam's razor slices through the myth that college women overachieve.

Abstract

College women have been believed to overachieve relative to college men. For a given ability profile, college women, on the average, achieve higher grades than men. A number of personality and work habit explanations have been invoked to account for this phenomenon. The present investigation sought to control sex differences in major field, through semipartial correlation, to determine whether female overachievement is an artifact of major field choice. The multiple regressions of grade point average on the Verbal and Mathematical sections of the Scholastic Aptitude Test were calculated for over 13,000 male and female students at four major universities. Differences in regression intercepts, which imply overachievement, were eliminated or drastically reduced when major field choice was controlled through semipartial correlations. Results indicate that the sex difference in overachievement is probably an artifact of the sex difference in major field choice.

Curricular Choice and the Differential Prediction of the Tertiary-Academic Performance of Men and Women

“…This meant that the semi-partial correlation between gender (men coded +1 and women -1) and GPA with the SAT variables partialled out -from gender -was -0.16. In a study by Hewitt and Goldman (1975) at four campuses of the University of California, the percentage of additional variance explained by gender varied between 0.45% and 2.05%. Pennock-Romn (1994) found that for two groups at the four institutions she studied, gender explained more than one per cent FGPA variance, namely 1.24% for non-Latino whites at a Texas university and 3.15% for an African-American group at a private California university.…”

“…Hewitt and Goldman (1975) divided the curricula at each of the four campuses of the University of California into a number of major fields and introduced the latter as a categorial (dummy) variable. Recall that these researchers found that gender explained an additional 0.45 to 2.05% ofFGPA variance at these campuses.…”

SEX DIFFERENCES IN SAT® PREDICTIONS OF COLLEGE GRADES

“…The mean over-and underprediction was -.03 for women and .02 for men (compared to -.05 for women and .04 for men in entire freshman classes). Sex differences in over-and underprediction of GPA by the SAT at four University of California campuses were reduced by about two-thirds, on the average, when college major was controlled (Hewitt and Goldman 1975).1 And sex. differences in over-and underprediction of twoyear GPA by the ACf occurred at the University of Iowa for students with business, liberal arts, and undecided majors but not for premedicine students (Gamache and.…”

“…It is obvious that underprediction for women could. arise if they took more leniently graded courses than men did (Hewitt and Goldman 1975). Indeed, studies that control for differences in grading standards have generally found less underprediction for women.…”

“…Selection, either explicit selection by the institution or self-selection by the applicants to the institution, can also create or increase over-and underprediction. Selection cannot cause the tendency for underprediction of women's college grades by admissions tests (Humphreys 1986), for women's admissions test scores are generally lower in the SAT and ACT test-taking population (American College Jesting Program 1988;Donlon 1984a), while women's grades are higher in most colleges (e.g., Clark and Grandy 1984;Hewitt and Goldman 1975). But selection can affect the extent of underprediction for women in individual institutions.…”