Using EAP to Look at Relative Staffing Levels -- Potential and Pitfalls

1
Using EAP to Look at Relative Staffing Levels
--Potential and Pitfalls

• Lou McClelland and Robert Stubbs
• University of Colorado at Boulder
• February 6, 2006, AAUDE

2
Who wants the comparisons?

• Staff Are we over or under-staffed relative to
  peers?
• Regents, administration Can we
• Plead poverty, need for more?
• Reduce staff and still be in line?
• Legislators, public

3
Issues in comparison

• Data source EAP
• Numerator
• Full-time, all, or FTE?
• Which subgroups?
• Denominator Per what?
• Student FTE, research dollars, ??
• Which peers AAU US public

4
Data source EAPEmployees by assigned position

• IPEDS winter submission
• Now driver of all HR surveys
• Employees as of 11/1, by
• Full-time vs. part-time
• Medical vs. not We excluded all medical
• 10 primary function/occupational activity
• Tenured, tenure-track, faculty status not on
  tenure track, w/o faculty status not fully
  crossed with functions

5
EAP matrix With Colorado row numbers and column
letters
28 valid cells 28 valid cells With faculty status With faculty status With faculty status D Without faculty status
28 valid cells 28 valid cells A Tenured B Ten track C Not TTT D Without faculty status
1 Instruction 1A 1B 1C 1D
2 IRPS 2A 2B 2C 2D
3 Research 3A 3B 3C 3D
4 Public service 4A 4B 4C 4D
5 Exec/admin/mgt 5A 5B 5C 5D
6 Other professionl 6A 6B 6C 6D
7 Tech/paraprof Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded 7D
8 Clerical/sec Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded 8D
9 Skilled crafts Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded 9D
10 Srv/maint Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded Not valid Grad assistants Col E, part-time only, excluded 10D
6
Numerator

• Full-time, all, or FTE?
• Used FTE Full-time plus 1/3 part-time
• Retains all data, easy, sensible to audience,
  used in Data Feedback Report
• Which subgroups?
• Comparisons using the 28 individual cells depend
  on comparable classification methods across
  institutions
• Check this

7
Checking cells in the matrix

• Used EAP 2005, with fall 2004 data
• Results very similar for EAP 2004
• Check raw distribution of counts over 34
  institutions for 28 cells
• Only 5 of 28 cells have 10 FTE for every
  institution

8
Cells where every school reported 10
Total FTE Total FTE With faculty status With faculty status With faculty status D Without faculty status
Total FTE Total FTE A Tenured B Ten track C Not TTT D Without faculty status
1 Instruction 1A 1B 1C 1D
2 IRPS 2A 2B 2C 2D
3 Research 3A 3B 3C 3D
4 Public service 4A 4B 4C 4D
5 Exec/admin/mgt 5A 5B 5C 5D almost
6 Other professionl 6A 6B 6C 6D
7 Tech/paraprof Not valid Not valid Not valid 7D
8 Clerical/sec Not valid Not valid Not valid 8D
9 Skilled crafts Not valid Not valid Not valid 9D
10 Srv/maint Not valid Not valid Not valid 10D
9
Check for paired columns or rows

• Every school has TTT tenured and tenure-track
  faculty, columns AB, minimum 600
• Look at distribution of counts over rows 1-6
• Institutions still reporting most TTT as
• Row 1 Instruction or
• Row 2 IRPS, Instruction, research, public
  service

10
TTT row 2 (IRPS) x TTT row 1 (instr)Clearly
must combine rows 1 and 2
11
Also not comparable for TTT in row 5Exec,
admin, management

• CO, NC, NE, IA, FL reported gt 10
• 13 schools reported none
• AZ, all UC, MI, Buffalo, OR, Pitt, Penn St, TX
  AM
• Suspect reporting practice or local terminology,
  not reality, is the difference
• Does it matter?
• It does in the IPEDS Data Feedback Report (DFR)

12
DFR Fig. 11 - of FTE professional staff by
assigned position
Exec/admin-gt
13
Categorizations matter in the DFR

• DFR lists pct of FTE in each of rows 1-6
• Not number per SFTE
• Easy to misread follows per-student-FTE figures
• Row 5 Exec-admin-mgt
• Peer median 6
• Colorado 14
• We said At other schools, tenured deans etc. are
  not in Row 5, so cannot compare this percentage

14
Do public AAUs have research staff?

• Row 3 is research Columns A B C D
• Sum of the columns, row 3
• Zero 10 schools
• Over 1,000 3 schools (Berkeley, CO, MD)
• And, those reported in row 3 may be
• TTT, Columns A/B
• Faculty status not TTT, Column C
• Without faculty status, Column D

15
Keep combining to fix Get 3 ultimate subgroups
With faculty status With faculty status With faculty status D Without faculty status
A Tenured B Ten track C Not TTT D Without faculty status
1 Instruction 1A 1B 1C 1D
2 IRPS 2A 2B 2C 2D
3 Research 3A 3B 3C 3D
4 Public service 4A 4B 4C 4D
5 Exec/admin/mgt 5A 5B 5C 5D
6 Other professionl 6A 6B 6C 6D
7 Tech/paraprof Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) 7D
8 Clerical/sec Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) 8D
9 Skilled crafts Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) 9D
10 Srv/maint Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) Pink TTT (col A, B) Blue All other professionals (1-6 C/D) Yellow Tech, clerical, skilled, service/maintenance non-professional (rows 7-10) 10D
16
Examine the 3 subgroups

• All schools have counts in all groups
• Average count about
• TTT 1500
• Other professional 4000
• Non-professional 3000
• Schools with more in one subgroup generally have
  more in all subgroups
• Correlations across 34 schools 70-80
• Plots show few obvious outliers

17
Other professional (vertical) vs. TTT
(horizontal)
Related but different. Far right Florida.
Top Ohio State
18
The numerator at last

• Staff FTE
• Excluding grad assistants
• For total plus three subgroups
• TTT Tenured and tenure track
• All professional staff not TTT
• Tech, clerical, skilled crafts, service,
  maintenance -- Non-professional

19
The denominator!

• Staff per what?
• Must normalize for size somehow
• What sensibly relates?
• Student FTE
• Research dollars
• Student or degree mix
• Student FTE alone seems insufficient
• So try multiple predictors

20
Predicting staff total and subgroup FTE

• AAU publics
• Without Pitt, Rutgers, Penn State (FASB so no )
• Without schools with medical
• N 13, model without Colorado
• Predictors
• Student FTE
• Research expenditures
• Pct of degrees that are doctorates
• Correlates .80 with research so proxies
• Land grant

21
Predictor combinations that work

• TTT SFTE land grant
• Other professional
  
• SFTE doc land grant
• Non-professional SFTE
• Total SFTE doc
• All R-squared .80-.91

22
Actual and predicted totals by student FTE
23
Punch line for Colorado

• CU staff FTE, pct different from predicted
• -11 for TTT
• 2 for other professional
• -29 for non-professional
• -7 to -12 overall 440 to 780 lt predicted
• These may make sense
• Cut the TTT last
• Many other professional paid with research

24
EAP and relative staffing levels

• Pitfalls
• Fine categorizations definitely not comparable
• Three subgroups may not be either
• Potential
• Available for all institutions
• Can readily see some of the incomparabilities
• Analyses like this show others
• But will there be any schools left if eliminate
  all?
• Probably related to reality
• Better than nothing