PBA Home > Institutional Research & Analysis > Comparing CU-Boulder with Peers > Staffing Levels Relative to Those of Other AAU Publics

CU-Boulder Staffing Levels Relative to Those of Other AAU Public Schools

July 2008

This is a July 2008 revision of the May 2008 posting, which had incorrectly calculated percentage differences.

Data from 31 AAU public schools show that CU-Boulder's staffing levels are below what would be expected, given total student FTE enrollment.  Our staffing levels are noticeably lower relative to our student FTE than those of other AAU publics for three of four employee groups.

• All employees.  Our actual staffing (n = 5,451)  is 24% below expected.
• Non-professional staff (technical, clerical, service/maintenance).  Actual staffing (n = 1,600) is 36% below expected.
• Tenured and tenure-track faculty.  Actual staffing (n = 1,044) is 14% below expected.
• Other professional staff.  The situation is quite different for this group (n = 2,807), which includes instructional and research staff not on the tenure track, administrators, librarians, others.  For this group our staff FTE is close to that predicted from student FTE and research expenditures.

Method.  Using IPEDS (Integrated Postsecondary Education Data System of the US Dept of Education) employees-by-assigned-position (EAP) data for fall 2006, we investigated relationships among actual staffing levels at 31 AAU US public schools and such variables as student total FTE enrollment, research expenditures, whether a school has a medical program, and whether it is a land grant institution.  Enrollment, research expenditures, and institutional characteristics are all from IPEDS collections as well, for fall 2006 or fiscal year 2005-06.  Three schools using FASB rather than GASB accounting principles were excluded:  Pennsylvania State University, University of Pittsburgh, and Rutgers.

The staffing data in the EAP contain full-time and part-time headcounts for 28 categories of employees.  The full- and part-time counts were converted to employee FTE, calculated as full-time plus part-time/3.  Based on work on the EAP reported in 2006, we collapsed the 28 categories into three employee subgroups: tenured and tenure-track (TTT) faculty; other professional staff (including administrators, instructional and research positions not on the tenure track, librarians, and other professionals);  and "non-professional" staff employed in technical, clerical, skilled crafts, service, or maintenance positions. We performed analyses by subgroup and for the total group of all employees. 

We developed linear models to predict staff FTE in subgroups and total, in each institution.  Staff FTE in TTT, non-professional staff, and total employees is a function of total student FTE only. The final model for other professional staff uses both student total FTE and research expenditures as predictors. Student total FTE was a statistically significant (p < .01) predictor in all final models. Total annual research expenditures was a statistically significant (p < .01) predictor for the other professional subgroup, only.

We used all 31 schools, including those with medical status, land-grant status, or both, in our analyses.  CU-Boulder has no medical program and is not land-grant. However, neither land-grant or medical status had any significant main or interaction effects on staff FTE for total or subgroups.     

Results.  The box plots below show distributions of the percentage difference between actual staffing levels and those predicted from student total FTE (and from research expenditures in the case of other professional staff). Features of each of the four plots:

• The horizontal line near the center of each box is the median of the distribution of differences.
• The ends of each box extend to the 25th and 75th percentiles.
• The vertical lines (whiskers) extend to the 5th and 95th percentiles.
• The schools with the most extreme values are labeled.
• CU-Boulder's percentage difference between actual and predicted staffing levels is denoted by a red circle.
• The orange dotted lines mark the area within 0.8 standard deviations of the mean. Values outside the dotted lines are said to be "noticeably different" from the mean.

The leftmost box plot, for example, shows that the total staff FTE for CU-Boulder employees is 24% below the predicted total. Moreover, this difference is "noticeable" in that the red circle is outside the area defined by the orange dotted lines -- that is, the CU-Boulder difference (-24%) is more than 0.8 standard deviations from the mean difference (which is about -5%). 

CU-Boulder's staffing levels are noticeably lower relative to our student FTE than those of other AAU publics for

• all employees (actual is 24% below predicted)
• non-professional (actual is 36% below predicted)
• TTT (actual is 14% below predicted)

In contrast, for "other professional staff" at CU-Boulder, the difference between actual and predicted staffing levels (actual is 9% below predicted) is near the median difference for all AAU publics.  Many of the "other professional staff" are paid from research grants. 

Actual employee totals and percentage differences from predicted totals are listed in the table below, by employee group, for each of the 31 public AAUs.  An Excel file of all data is also available, listing actual employee totals, predicted employee totals, residuals, percentage differences, research expenditures, land-grant and medical indicators, etc.

Three schools show more staff FTE than predicted for every subgroup, and are generally in the top 10 on actual vs. predicted: Michigan, Minnesota, and North Carolina.  Three are the opposite, with fewer staff FTE than predicted for every subgroup and generally in the bottom 10 on actual vs. predicted: Texas A&M, SUNY-Buffalo, and SUNY-Stony Brook.  Missouri shows the highest staffing levels relative to predicted for other professional and non-professional, but exactly average tenured and tenure-track numbers relative to student enrollment.  We suspect that reporting differences may be responsible for the apparent extremes seen at the SUNY campuses, Texas A&M, and Missouri.

Last revision 09/11/08