2002-03

advertisement
University of Wisconsin-Green Bay
Faculty Salary Review, 2002-03
In Fall 1997, UW-Green Bay prepared an analysis of the structure of compensation for full-time
legal faculty. That study focused on compression of salaries across the ranks, competitiveness
of salaries here with national and system-wide averages, and individual factors impacting
salaries. This report will replicate several of those analyses using faculty salary data from 200203 in order to assess the current status of the compensation structure.
Describing the Faculty
The data set used for this analysis contains salary and demographic data for 150 full-time legal
faculty who taught during 2002-03. A wave of retirements in the late 1990’s has transformed
the faculty, in some respects, in the seven years since1996-97. Assistant Professors comprised
the smallest faculty category in 1996-97, and are the largest category in 2002-03 (see figure 1).
In comparison to the 1996-97 report, the faculty has become more female (37%, up from 28%
seven years earlier), more racially and ethnically diverse (14% faculty of color, up from 11%),
and younger (averaging 47.6 years, down from 49.2 years) (see table 1). Table 1 also divides
Legal Faculty by Rank
40%
1996/97
2002/03
30%
20%
10%
0%
Assistant
Associate
Full
Figure 1
Table 1. Demographics of Faculty by Rank and By Division, 2002-03
Rank
Assistant
Associate
Full
Division
Lib. Arts & Sciences
Professional Studies
All
Institutional Research
UW-Green Bay
Number
Women
Minorities
Age (mean)
58
52
40
27 (47%)
20 (38%)
9 (23%)
8 (14%)
9 (17%)
4 (10%)
40.3
49.3
55.9
Years Here
(mean)
3.3
13.6
24.5
123
27
150
43 (35%)
13 (48%)
56 (37%)
19 (15%)
2 (7%)
21 (14%)
46.7
51.5
47.6
13.1
9.6
12.5
Page 1 of 7
Printed: 7/1/2016
the faculty according to the two voting units, Liberal Arts and Sciences (LAS) and Professional
Studies (PS). The two units differ on all dempgraphic elements. Although faculty in PS average
almost five years older than their peers in LAS, faculty in LAS have taught here for longer. This
reflects the fact that the PS faculty were 43 years old, on average, at the time of hire compared
to 35 years old, on average, in LAS. However, age at hire may be increasing for LAS. Full
professors in LAS were hired at an average age of 32, Associate Professors were hired at an
average age of 35 and the Assistant Professors were hired at an average age of 37. In PS, the
average age at hire is the same for all three ranks.
Table 2 describes the faculty salaries by rank and by division. As was true seven years ago, all
ranks have positively skewed salaries, with a few high salaries pulling the means up over the
median values. Although the average salary is higher in the Professional Studies than in LAS,
both divisions have at least some individuals earning near the campus minimum and near the
campus maximum salary values.
Table 2. Faculty Salary Information by Rank and by Division, 2002-03
Mean
Minimum
Median
Maximum
Rank
Assistant
Associate
Full
$45,643
$54,056
$63,934
$39,395
$44,190
$47,800
$42,858
$51,284
$63,259
$73,929
$92,009
$88,371
Division
Lib. Arts & Sciences
Professional Studies
All
$51,665
$61,510
$53,437
$39,395
$41,159
$39,395
$48,151
$62,686
$50,322
$88,371
$92,009
$92,009
Compression in the Salary Structure
Compression occurs when salaries paid to junior level faculty within a particular academic field
approach or exceed those paid to more experienced faculty within that same field. Table 3
shows the ratio between the average salaries at the various ranks for each field present in the
data. Several of these ratios are based on single data points, or individual salaries. Of 35
compression ratios calculated, only one shows total compression, that between associate and
full professors in the field of physics.
Campus-wide, associate professors earn 18% more than assistant professors. This reflects
significant 'decompression' since 1996-97, when associate professors earned approximately
10% more than their junior colleagues. Compression between full professors and associates
has slipped some, from a ratio of 1.21 in 1996-97 to 1.18 today. As was true seven years ago,
salaries at UW-Green Bay are more compressed between the ranks than they would be if our
faculty all earned the national average paid to faculty in the same rank and field. If our faculty
earned the national averages, the ratio between associates and assistants would be 1.22
(rather than 1.18) and the ratio between full professors and associates would be 1.26 (rather
than 1.18). If the average assistant faculty salary is held constant, it would take an additional
expenditure of $333,800 in associate and full professor salaries for UW-Green Bay to have
compression ratios that match the national averages.
Institutional Research
UW-Green Bay
Page 2 of 7
Printed: 7/1/2016
Table 3. Rank Compression by Field
Field
Ratio of Associate to
Ratio of Full to Associate
Assistant Salary Average
Salary Average
Accounting
1.13
Not Applicable
American Indian Studies
1.14
Not Applicable
Art
1.19
1.17
Biology
1.05
1.48
Chemistry
1.24
1.19
Communication Proc.
1.20
1.48
Economics
Not Applicable
1.29
Education
1.18
1.23
English
1.12
Not Applicable
Engineering
1.10
Not Applicable
Finance
1.13
Not Applicable
Geography
1.14
1.27
History
1.27
1.25
Languages
1.13
1.34
Management
1.08
Not Applicable
Marketing
1.08
Not Applicable
Math
1.08
1.20
Music
1.12
1.32
Physics
1.11
0.91
Political Science
Not Applicable
1.54
Psychology
1.25
1.11
Social Work
1.25
1.11
All faculty
1.18
1.18
Note: The following fields are not listed because they do not have faculty spread across
the ranks: Anthropology, Computer Science, Earth Science, Environmental Design,
Nursing, Philosophy, Public Administration, Sociology and Theatre.
National Comparisons by Rank and Field
Table 4 places UW-Green Bay’s salary averages into state and national contexts. In 2001-02,
the assistant professor average at UW-Green Bay was the lowest among Wisconsin’s public
comprehensive institutions. The associate professor average was the median of those peer
institutions, and the full professor average was tenth out of the eleven. In 1996-97, Green Bay
Assistants were ranked 3rd among the comprehensives, associates were ranked 7th, and full
professors were ranked 10th.
In comparison to what our faculty would receive if they were paid the national average for their
rank and field, salaries at UW-Green Bay are 4% low for assistants, 8% low for associates and
15% low for full professors. These findings are similar to national comparisons made in 199697 (see also Appendix A). Paying all UW-Green Bay faculty the average salary for their rank
and field would require an additional salary expenditure of $727,900 (plus additional related
expenditures for fringe benefits), or reducing the number of faculty by approximately 8% (from
150 to about 138).
Institutional Research
UW-Green Bay
Page 3 of 7
Printed: 7/1/2016
Table 4. Local, State and National Salary Comparisons by Rank
Rank
Assistant
Associate
Full
All
Green Bay
Mean,
2002-03
$45,643
$54,056
$63,934
$53,437
Rank of UW-Green Bay
average within the UWComprehensives, 2001-02 (1)
11th (of 11)
6th (median of 11)
10th (of 11)
Not available
National
Mean (2),
2002-03
$47,686
$58,342
$73,598
$58,290
Gap between
Green Bay and
National
-$2,043 (-4%)
-$4,286 (-8%)
-$9,664 (-13%)
-$4,853 (-8%)
(1) 2002 UW System Fact Book, “Average AAUP Salaries of Full-time Faculty as of 200102”, page 56.
(2) These “national” averages are based on what the average salaries at each rank would
be if each individual at UW-Green Bay was paid the average salary for his or her rank
and field, as reported in the 2002-03 annual salary survey of the College and
University Personnel Association (CUPA).
Appendix A lays out average salaries at the level of the particular disciplinary field. However, at
UW-Green Bay, faculty are neither hired nor paid through traditional disciplinary structures.
Table 5 documents salary competitiveness broken down according to the Budgetary Unit
responsible for each individual’s salary. The table lists Units in the order of their salary gaps.
Only one Unit has an average salary higher than the national comparison
Table 5. Competitiveness by Budgetary Unit
Budgetary Unit
Faculty
(N)
Information & Comp. Sci.
8
Nursing
2
Social Work
3
Humanistic Studies
22
Communication & Arts
25
Public & Environ. Affairs
7
Human Biology
9
Business Administration
13
Education
9
Human Development
10
Social Change & Devel.
8
Natural & Appl. Sciences
26
Urban & Reg. Studies
8
Source: See notes on Table 4.
Green Bay
Mean Salary
National Mean
Salary
$65,316
$55,490
$54,812
$49,049
$47,197
$59,746
$45,313
$72,035
$49,877
$51,495
$52,883
$52,006
$57,130
$63,945
$57,152
$56,753
$51,642
$50,036
$63,633
$49,450
$77,177
$56,130
$57,854
$59,860
$59,252
$70,421
Gap between
Green Bay and
National
$1,372 (+2%)
-$1,662 (-3%)
-$1,941 (-3%)
-$2,593 (-5%)
-$2,839 (-6%)
-$3,887 (-6%)
-$4,137 (-8%)
-$5,143 (-7%)
-$6,253 (-11%)
-$6,359 (-11%)
-$6,977 (-12%)
-$7,246 (-12%)
-$13,291 (-19%)
Individual Factors
Researchers typically use a regression model to explore salaries within an institution,
considering such factors as gender an ethnicity once factors like rank, field, years in rank, time
at the institution and merit have been taken into account (see Richard D. Howard, Julie K.
Snyder and Gerald W. McLaughlin, “Faculty Salaries”, in The Primer for Institutional Research,
ed. Meredith A. Whiteley, John D. Porter and Robert H. Fenske [Tallahassee, Florida: The
Institutional Research
UW-Green Bay
Page 4 of 7
Printed: 7/1/2016
Association for Institutional Research, 1992], 51.).
following eight variables.



The model for Green Bay included the
Four variables measured at the interval level:
o the market value of the instructor’s field, estimated as the average salary paid to
instructors in that same field and rank;
o years in his/her current rank;
o years employed at the University;
o age (eventually removed from the model);
One ordinal level variable:
o the instructor’s merit, categorized as high, medium and low; [Instructors were
sorted by Budget Unit and by their most recent merit amount divided by their
base salary. New assistant professors were ranked based on the average merit
rating of 1.9%. In each Budget Unit, the top third of instructors received merit
scores of 3, the middle third received scores of 2 and the bottom third received
scores of 1. Campus wide, the bottom 10% received scores of 0. Those scores
were used in the regression model.]
Three categorical, or “dummy” variables:
o degree status, specifically the additional value, if any, of having completed the
highest degree for the field;
o ethnicity, specifically the additional value, if any, derived from being “white”;
o gender, specifically the additional value, if any, derived from being male.
Collinearity diagnostics revealed modest multicollinearity between age and years employed at
the University. Age was not statistically significant, and its removal had little impact on the
remaining variables. The statistical results from the regression model appear in Table 6.
Table 6. Regression Results
Variable
Intercept
Market Value
Highest Degree
Merit Level
Years in Rank
Years at UWGB
White
Male
Estimate T for HO:b=0
177.78
0.06
0.77
17.95
1123.60
0.52
1199.13
2.61
804.27
6.88
-371.08
-4.34
4919.89
4.05
30.42
0.03
Prob> or <T
0.9544
<.0001
0.6035
0.0102
<.0001
<.0001
<.0001
0.9744
Standardized B
0
.9002
.0190
.0953
.4470
-.3399
.1471
.0013
The adjusted R-square indicates that this model accounts for 81% of the variance in Green
Bay’s 2002-03 salaries (the same degree of fit as achieved in 1996-97). The remaining 19% of
variance in salaries could be due to the following kinds of factors:



meritorious performance not adequately predicted by the four-level merit variable;
regional and campus-specific market differences between fields that are not reflected in
the national market variable.
market variability within particular disciplines. For example, political science includes
both political theory, typically a relatively low-paid area, and public administration, often
a relatively high-paid area; geography includes both cultural geography and GIS;
Institutional Research
UW-Green Bay
Page 5 of 7
Printed: 7/1/2016


other issues relating to how faculty are coded as opposed to how they function currently
(e.g., faculty who function in a truly interdisciplinary manner may be poorly represented
by the disciplinary fields used by CUPA).
and random differences.
As anticipated, the most important variables (as indicated by the relative magnitude of
the standardized estimates) are the current market value of the instructor’s discipline for
his or her rank and the length of time that he or she has been in that rank. One's merit
level also contributes significantly. Possessing the highest degree for one's field does
have a statistically significant impact on salary. This may stem from the fact that
virtually all -- 96% -- of the cases in this study have the highest degree for their field.
Years at UW-Green Bay has a negative and statistically important impact on salary.
Holding all other included variables constant, each additional year of service to the
University relates to a decline of ones salary by $371. The time one spends in rank
accrues salary value; the time it takes to reach that rank decreases salary value. The
best example one can give to illuminate this might be the case of two individuals who
started at the University in the same year, in the same field and who are both currenty
equally meritorious full professors. The one who progressed to the rank of full professor
soonest will have the higher salary.
In this type of analysis, researchers hope to find statistically insignificant coefficients for
ascribed characteristics such a gender and ethnicity. In 2002-03, male faculty at UW-Green
Bay did not receive statistically higher salaries than females. The 1996-97 study also failed to
find any gender descrimination in the faculty salary structure. The slope on the factor
measuring ones racial or ethnic background is not insignificant. In 1996-97, the slope on this
factor was $4,055 with a standardized beta of .134, findings very similar to this year’s results.
This does not mean, necessarily, that UW-Green Bay’s salary practices and structures are
racially discriminatory. In this sample at this time, if all other measured factors are held
constant, one would predict that a white instructor would have a salary $4,920 higher than an
instructor who is a person of color.
Explorations of these variables along racial/ethnic lines suggest that the relationship between
the four-level merit variable and salaries differs between faculty who are white and those who
are persons of color. In the high-performance levels of merit (values of three and four), there is
no statistically significant impact for being white. It is only at the lower levels of merit that one
finds the negative impact associated with being a person of color. Table 7 shows that for white
faculty how close one is to the national average for one’s rank and field does not seem related
to the length of time at the University nor to last year’s merit level. As long as minority faculty
have performed to a medium or high standard, they have maintained the same kind of
relationship to the average salary for their rank and field as their white peers. However, when
minority faculty have failed to perform to a medium or high level, their salaries appear to be
suffering to a much greater extent than their white peers. The lowest performing minority faculty
have all been here an extremely long time and have extremely low salaries relative to their
markets. For those individuals (n=3), it may be the historic accumulation of poor performance
that explains their very low salaries, and it may be that the four-level merit score is inadequate
to account for their experience. The experience of the eight minority faculty at the “low”
category is more disconcerting because their low salaries are less obviously explicable.
Institutional Research
UW-Green Bay
Page 6 of 7
Printed: 7/1/2016
Table 7. Merit, Color and Pay
Merit Level
Faculty
Very Low
Low
Medium
High
All
3
8
5
5
21
Faculty of Color
Years
Salary as a Percent
Here
of National Average
(Mean)
(Mean)
34.7
65%
9.8
86%
8.6
93%
11.2
94%
13.4
87%
Faculty
9
35
49
36
129
White Faculty
Years
Salary as a Percent
Here
of National Average
(Mean)
(Mean)
14.3
93%
11.1
94%
11.4
93%
14.4
94%
12.4
94%
Conclusion
This report has examined compression, competitiveness and individual factors related to faculty
salaries in 2002-03. Salary compression has generally improved since 1996-97, with the gap
between assistants and associates increasing and the gap between associates and full
professors remaining about the same is seven years earlier. The overall competitiveness of
faculty salaries has remained virtually unchanged in this period. Aggregate salary averages
here have not moved any closer to comparative national data.
The relationships between various individual factors and faculty salaries have changed very little
since 1996-97. It is good news that the increase in the number of women in the faculty, from
28% to 37%, has not decreased their salaries vis-a-vis male instructors. In fact, the gap
between men and women has narrowed. In 1996-97, the average salary paid to female faculty
was 17% lower than the average male salary. By 2002-03 that gap had fallen to 14%. The
analysis showed that the 14% difference is due to choice of field and tenure at the University,
not gender per say. There is good new, too, for minority faculty. In 1996-97 then average salary
paid to minority faculty was 14% lower than the average minority salary. By 2002-03, that gap
had fallen to 9%. Unfortunately, that gap is not, apparently, entirely due to choice of field and
tenure at the University. The continuing presence of statistically significant differences between
minority and white faculty salaries when other factors are held constant should not be ignored.
The “Campus Diversity Plan 2008” calls for annual data gathering on turnover rates and tenure
in position minority and non-minority employees. It would be useful to consider adding salary
averages to that annual report, at least for employees in the faculty lines.
Institutional Research
UW-Green Bay
Page 7 of 7
Printed: 7/1/2016
Download