Persistence to Graduation for
Transfer and "Native" Students:
A Case Study
John W. Miller, Ph.D.
President, Central Connecticut State University
Braden J. Hosch, Ph.D.
Assessment,
Director of Institutional Research & Assessment CCSU
Association for Institutional Research Annual Forum
Seattle, WA
May 26, 2008
Overview
Institutional Profile
p
Overview of the problem → to what extent do
differences in earned credit hours contribute
to differences in graduation rates?
Methodology and Results
Conclusions and Implications
Institutional Profile: Central
Connecticut State University
p y
Public – part of Connecticut State Univ. System
Carnegie 2005 Master’s-larger programs
Fall 2007 Enrollment:
12,106 headcount (9,704 undergraduate, 23% residential);
9,288 full-time equivalent enrollment
52% female; 16% minority y
Full-time, first-time students: 1,469 (56% residential)
Full-time, new transfer students: 678
Six-year
Six year graduation rates:
44% full-time, first-time students
56% transfer students (full-time upon entry)
Methodology
Divide subjects by origin
Full-time first-time student
Full time
Full-time transfer students
Group subjects by credits earned rather than
by fall of entry
y y
Measure time to graduation (Fall-to-Fall),
regardless of time at institution
e.g. after earning 76-90 credits, how many
graduated in 1, 2, 3, 4, 5, 6, and 10 years?
Produce descriptive statistics and run logistic
regressions
Study Population
Unduplicated, Full-Time Students in Study Population, Fall 1991-2006
Native Students Transfer Students Grand
Race/Ethnicity Women Men Total Women Men Total Total
Non-Resident Alien 125 141 266 64 49 113 379
African American/Black 664 757 1,421 370 378 748 2,169
American Indian 43 40 83 25 18 43 126
Asian 168 210 378 146 184 330 708
Hispanic 542 504 1,046 283 195 478 1,524
White 7,612 7,455 15,067 3,675 3,679 7,354 22,421
Unknown 397 434 831 234 243 477 1,308
Grand Total 9,551 9,541 19,092 4,797 4,746 9,543 28,635
Total observations (duplicated individuals): 86,342 over 16 years
Graduation Rates by Credits Earned
Credits Earned at Beginning of Fall Semester
Yr. Of 16-30 31-45 45-60 61-75 76-90 91-105 106+
Entry* FTFT Transf FTFT Transf FTFT Transf FTFT Transf FTFT Transf FTFT Transf FTFT Transf
In 1 Yr Graduated (%) 0% 0% 0% 0% 0% 0% 0% 0% 14% 9% 46% 32% 72% 56%
91-'06 Basis Obs. (N) 10,640 2,403 4,017 2,706 7,012 3,905 4,101 3,802 4,881 3,679 3,693 3,308 4,738 5,598
In 2 Yrs Graduated (%) 0% 1% 1% 3% 13% 11% 37% 27% 64% 46% 77% 64% 84% 75%
91-'05 Basis Obs. (N) 9,858 2,217 3,648 2,503 6,481 3,595 3,671 3,481 4,408 3,390 3,286 3,025 4,160 5,160
In 3 Yrs Graduated (%) 11% 10% 25% 26% 54% 41% 66% 53% 78% 64% 83% 76% 87% 81%
91-'04 Basis Obs. (N) 9,108 2,041 3,257 2,324 5,899 3,273 3,257 3,177 3,883 3,090 2,894 2,759 3,606 4,739
In 4 Yrs Graduated (%) 41% 31% 50% 44% 67% 54% 72% 61% 81% 70% 85% 79% 89% 83%
91-'03 Basis Obs. (N) 8,389 1,854 2,899 2,092 5,254 2,994 2,855 2,888 3,410 2,805 2,506 2,517 3,105 4,312
In 5 Yrs Graduated (%) 52% 41% 56% 49% 71% 59% 74% 64% 83% 73% 86% 80% 90% 84%
91-'02 Basis Obs. (N) 7,520 1,683 2,507 1,916 4,674 2,719 2,459 2,613 2,974 2,535 2,167 2,293 2,685 3,923
In 6 Yrs Graduated (%) 55% 44% 58% 53% 73% 62% 76% 66% 85% 75% 88% 82% 91% 85%
91-'01 Basis Obs. (N) 6,765 1,513 2,128 1,721 4,096 2,464 2,106 2,346 2,552 2,314 1,839 2,059 2,260 3,522
Differences in Graduation Rates
CCSU Native Students vs. All Transfer Students : Graduation within Six
Years from Time of Earning Designated Number of Credits
100%
90%
80%
70%
Percent graduated
60%
%
50%
40%
30%
20%
10%
0%
16-30 31-45 45-60 61-75 76-90 91-105 106+
p g g
Credits Completed as of the Beginning of the Fall Semester
First-Time Full-Time Students Transfer Students
Logistic Regression (1)
Two variable model (Forward conditional)
B S.E. Wald Exp(B)
Total C ed ts Earned
ota Credits a ed 00
0.021 0 000
0.000 5,9
5,921 0
1.022
Transfer -0.437 0.024 343 0.646
Constant -0.449 0.015 886 0.638
Probability of graduating from within 6 years of
completing specified credits
All variables significant at
p<0.001
Probabilit of graduating within 6
100%
90%
80%
w
70% Model accuracy:
60%
50% 68.9% correct
years
40%
30%
Native
Cox & Snell R2: 0.138
Transfer
20% Nagelkerke R2: 0.188
0 188
ty
10%
0%
103
111
119
15
23
31
39
47
55
63
71
79
87
95
Credits completed at beginning of fall term
Logistic Regression (2)
Multivariate model (Forward conditional)
B S.E. Wald Exp(B)
Total Credits Earned 0.020 0.000 3,518 1.020
Transfer
T f 0 413
-0.413 0 026
0.026 248 0 661
0.661
GPA Earned 1st Fall 0.240 0.018 170 1.271
Hrs Attempted 1st Fall 0.075 0.005 224 1.078
GPA Earned in Fall of Analysis 0.893 0.016 3,299 2.443
Constant -4.026 0.070 3,274 0.018
sig. p<0.001
All variables sig at p<0 001 Additional Significant Variables:
Female
Model accuracy: African American/Black
75.1% correct Hispanic
Cox & Snell R2: 0.273
0 273
Nagelkerke R 2: 0.371 But model accuracy improved by only
0.1% or less with inclusion
Major Findings
Total credits earned is most significant
predictor of graduation
g g
Native students graduated at higher rates
than transfer students after controlling for
credits earned
Other significant predictors:
First semester GPA
Credits attempted first semester
GPA in term of analysis
Conclusions and Implications (1)
1 At CCSU, and I suspect at other institutions,
1.
we need to recognize the lower success rate
of our transfer students by comparing their
progress to their equivalent credit
counterparts.
2. While recognizing and respecting the
i f f d
experiences of transfer students, we must
appreciate that they still need similar support
(e.g.
to FT/FTs modified for them (e g FYTE
course).
Conclusions and Implications (2)
3 More careful exploration of our transfer
3.
compacts to determine how to better pave
the way for transition is required.
py y g
4. We must more deeply study our “counting”
of credits as meeting requirements, used as
electives, or not counted at all to ensure
i i
maximum appropriate progress toward d
completion at time of transfer.
Persistence to Graduation for
Transfer and "Native" Students:
A Case Study
Contact Information:
John W. Miller, Ph.D.
President, Central Connecticut State University
millerjw@ccsu.edu
Braden J. Hosch, Ph.D.
Director of Institutional Research & Assessment, CCSU
hoschbrj@ccsu.edu