Arts and Sciences Tenure/Track Faculty Salary Analysis 1999-2000

Traditional methodology: model based on male population is applied to female population to predict female salaries.
Variables include: discipline groups, years here, years since terminal degree, squared terms, rank dummy variables.

Comparison of two sets of models: using 5 discipline groups versus 11 discipline groups
Mean female residuals, standard errors and z_scores

1999-00
using traditional discipline groups (11)
1999-00
using new discipline groups (5)
model sq
terms 		rank log
scale 		R2 standard
error 		difference:
mean
female
residual 		z-score R2 standard
error 		difference:
mean
female
residual 		z-score
models with rank variable
1 y y y 0.58 2.41% 0.11% -0.05 0.60 2.28% -0.17% -0.07
2 y y n 0.47 2,334 -608 -0.26 0.50 2,223 (253) -0.11
5 n y y 0.57 2.38% 0.09% 0.04 0.59 2.28% -0.50% -0.22
6 n y n 0.45 2251 -372 -0.17 0.48 2,171 (627) -0.29
models without rank variable
3 y n y 0.45 2.93% -0.89% -0.30 0.48 2.79% -0.91% -0.33
4 y n n 0.38 2,673 -925 -0.35 0.41 2,551 (565) -0.22
7 n n y 0.38 3.11% -3.96% -1.27 0.40 3.01% -4.29% -1.42
8 n n n 0.32 2,671 -2957 -1.11 0.36 2,581 (3,010) -1.17

Original discipline groups:
Math
Biology
Chemistry, Physics, Earth & Planetary Science
Psychology
Economics
Education
Anthropology, History, Philosophy, Political Science
English
ANELL, German, Romance Languages, Russian
Classics, Comp Lit, AFAM, Art History, Wom Studies
Performing Arts, Music
New discipline groups:
Math, Biology, Chemistry, Physics, EPS, Psychology
English, History
Economics
Anthropology, Political Science
ANELL, German, Romance Languages, Russian, Classics,
Comp Lit, Philosophy, Art History, Music, Performing Arts,
AFAM, Wom Studies, Education

Note: the original discipline groups are those used in the initial Washington University study of 1986-87 data supervised by Mary Gray, Professor of Mathematics and Statistics at American University. The original selection of department groups appears to reflect a perception of national faculty characteristics in the 1980's. The purpose of introducing the new aggregation groups is twofold: (a) to reduce the number of discipline variables and thereby increase adjusted R square and (b) to recognize the character of Washington University Arts & Sciences departments as of 1999-2000.

Error processing SSI file