A. In 1995, following the 1993 Faculty Gender Pay Equity Exercise,
a follow-up committee found upon analysis that the exercise had
eliminated any aggregate gender gap. Nonetheless, the follow-up committee
asked the university to "monitor, on a routine basis, the salaries of
faculty, and to hold each school/college accountable for achieving and
maintaining gender equity." The Committee on Women in the University was
designated as one of three committees to advise on such monitoring. In
spring 1997, the Committee on Women in the University requested that the
Office of Budget, Planning & Analysis (OBPA) conduct a follow-up
study. That office performed multiple regression analyses in 1998, using
the methodology of the 1992 pay equity study. The analyses use several
variables, including years since degree, years of service at UW, and
market value of area of specialty. The OBPA draft report of September
1998 shows an overall gender salary differential favoring women of - 0.5% when rank is included in the equation. Excluding rank from
the analysis, it estimates a 3% difference in salaries favoring men for
the university as a whole.
B. An outside consultant, Delores Conway (Associate Professor of
Statistics, University of Southern California), was hired to assist
UW-Madison in interpreting the OBPA analyses and to advise on next steps.
Professor Conway argues that, reliable regression analyses of faculty pay
need to include rank as a variable. Rank is an important variable because
the academic marketplace uses rank as a salary variable, UW does compete
in this marketplace and in this marketplace, rank heavily affects salary.
She demonstrates that the results may be confounded when one excludes rank
as a variable in salary regressions. Specifically, legitimate market
factors may be confounded with measured gender effects on salary. She
suggests that a separate study of hiring and promotion decisions is
necessary if it appears there may be gender discrimination in assigning
rank.
C. Professor Conway concludes that the OBPA regressions show no
statistically significant evidence of gender inequity in faculty salaries
across campus or across schools when rank is included as a variable.
Despite this finding, she cautions that regression analyses cannot show
whether individual cases of gender salary inequity exist in schools or
colleges. An individualized analysis is necessary in order to make a
judgment about the gender equity of an individual female faculty
members salary. Professor Conway considers graphs of salary
distribution, by rank, for individual departments, schools, and/or
divisions within a school to be important tools for flagging possible
problems in individual gender equity in faculty pay. Once individuals
with unusually low salaries are flagged, she cautions, they need to be
compared to individuals who are truly comparable - by rank, area of
specialization, and other factors indicating similar positions.
D. The Committee on Women, the Provost and the Administration accept
Professor Conways advice and conclusions.
E. The guidelines below are based on recommendations from Betsy
Draine, former Associate Vice Chancellor, the Committee on Women in the
University, and the report of consultant Professor Delores Conway.