ArticlePDF Available

Institutional Retention Strategies at Historically Black Colleges and Universities and Their Effects on Cohort Default Rates: 1987-1995. Monograph Series

Authors:

Abstract and Figures

This study analyzed institutional factors that significantly affect cohort default rates at historically black colleges and universities (HBCUs) and assessed the potential for reducing these rates. It has been hypothesized that the HBCUs most successful in lowering student cohort default rates are those that have also succeeded in increasing student retention rates. Swail's (1995) conceptual framework for student retention, which focuses on barriers and issues facing minority students, was used to identify and assess factors that might contribute to reducing cohort default rates. A mixed time-series/cross-sectional model examined the effectiveness of institutional strategies designed to reduce cohort default rates; data was analyzed for financial aid, recruitment and admissions, curriculum and instruction, academic services, and student services at 80 HBCUs for the period 1987-95. Analysis revealed that increases in some non-academic components (grants and student services) of an institution's budget are statistically associated with increases in cohort default rates. However, increases in instructional services are associated with lower cohort default rates, with such increases having their biggest effect at relatively low levels of spending. The study concludes that, although the data explain about 40 percent of the variation in cohort default rates, student-based characteristics and other factors account for the majority of the variation. (Contains 12 references.) (CH)
Content may be subject to copyright.
N O V E M B E R 1 9 9 9
M O N O G R A P H S E R I E S
Institutional Retention
Strategies at Historically
Black Colleges and
Universities
and Their Effects on
Cohort Default Rates:
1987 - 1995
By Fred J. Galloway
Associate Dean
School of Education
University of San Diego
Watson Scott Swail
Associate Director for Policy Analysis
The College Board
This publication provided courtesy of
epiepi
Educational Policy Institute
Washington, DC
Watson Scott Swail, Ed.D.
President
www.educationalpolicy.org
Improving educational policy & practice through research
Washington Office • 25 Ludwell Lane • Stafford, VA 22554 • 1 (877) e-POLICY
Los Angeles Office • Occidental College • 1600 Campus Road • Los Angeles, CA 90041 • 1 (877) e-POLICY
Institutional Retention Strategies at
Historically Black Colleges and Universities and
Their Effects on Cohort Default Rates: 1987 - 1995
By Fred J. Galloway
Associate Dean
School of Education
University of San Diego
Watson Scott Swail
Associate Director for Policy Analysis
The College Board
Published by the Sallie Mae Education Institute
901 E Street NW
Washington, DC 20004
November 1999
Copyright © 1999 Sallie Mae Education Institute, all rights reserved.
Portions of this Monograph may be reproduced with written permission
from the publisher.
The Sallie Mae Education Institute (SMEI) is a non-profit corpo-
ration that focuses its attention on issues in higher education
finance, especially student and family ability to pay for postsec-
ondary education. SMEI conducts research and publishes find-
ings, works with higher education leaders to offer creative solutions
to problems, and furnishes objective analyses to policy-makers and
other interested parties.
A B O U T T H E A U T H O RS
Fred J. Galloway is currently associate dean and tenure-track associate
professor in the School of Education at the University of San Diego. Prior to
joining the University faculty, he was project director for the national Direct
Student Loan Evaluation project at Macro International, as well as director of
federal policy analysis at the American Council on Education in Washington,
D.C., where he represented the interests of the higher education community
before the Executive and Legislative branches of the federal government. Dr.
Galloway received his bachelor’s and master’s degrees in economics from the
University of California, San Diego, and his doctoral degree in the econom-
ics of education from Harvard University.
Watson Scott Swail is associate director for policy analysis of The College
Board, where he provides data and analysis on issues relating to three prima-
ry areas: academic preparation, access to college, and postsecondary success.
In addition to his research projects, he directs and produces The College
Board’s annual Trends in Student Aid, Trends in College Pricing, and Trends
in Academic Preparation. Dr. Swail is also adjunct professor in the
Educational Leadership Department at George Washington University. He
received his bachelor’s degree from the University of Manitoba, his master’s
degree from Old Dominion University, and his doctorate in educational poli-
cy from George Washington University.
As college prices have risen, so has student borrowing to pay for them. These growth
patterns have led to concerns about borrower defaults, as not all students successful-
ly complete their education programs and, consequently, some default on their loans.
Defaults cost the taxpayers, the government, and the lenders. They also cost the stu-
dent borrowers a great deal of money in lost credit and related negative effects on
their lives. Thus, there is great interest in studying the causes of defaults and using
this research to reduce default costs and probabilities. The literature of student finan-
cial aid abounds with research reports on defaulter characteristics and how they dif-
fer from student borrowers who do not default, accounts of successful actions taken
by loan program administrators to avert defaults, and distinguishing characteristics of
institutions whose students have above average default rates.
The literature contains very little discussion of the relationships between what insti-
tutions do, particularly how they allocate their resources among manifold activities,
and the effects their acts might have on student borrowers' repayment experiences.
This study employs sophisticated statistical methods to assess the effects of institu-
tional spending (resource allocation) in such areas as instruction, academic support
services, student services, and student financial aid, on students' subsequent cohort
default rates. The subjects of this research are colleges whose students have experi-
enced above-average default rates in federal loan programs, the Historically Black
Colleges and Universities. These colleges were chosen as subjects for two reasons.
First, there are great individual differences in their students' default rates, the
resources they have to apply to general and education expenditures, and in the ways
they divide funds among various program activities. Second, some of these colleges
are in danger of losing institutional eligibility to participate in the federal student
loan programs because of their students' higher default rates. So the researchers
hoped that their analyses would help the colleges to address defaults and the root
cause of defaults, student attrition, more effectively.
The Sallie Mae Education Institute is pleased to publish this important study. We
believe it offers institutions guidance in how they might better allocate resources to
reduce their students' default rates. And we believe it offers student default
researchers ideas and hypotheses to explore when they conduct their next studies.
There is still much to learn.
Jerry S. Davis
President
Sallie Mae Education Institute
November 1999
F O R E W O R D
During the latter half of the 1980s, annual costs to the federal government for defaults
in the nation’s largest student loan program, the Federal Family Education Loan
Program (FFELP), began to soar. Between 1984 and 1986, annual default costs near-
ly doubled, rising from about $712 million to over $1.3 billion, before reaching a high
of $3.2 billion in federal fiscal year (FFY) 1991 (Conner, Saab, and Cicmanec, 1997).
Congress, the federal government, postsecondary education institutions, lenders, and
all FFELP participants became concerned about the threat default costs represented to
continued support of student loans. During those years, and in the years hence, these
parties cooperated in many ways to help reduce default costs and suppress default
probabilities. One of those strategies was to use borrower cohort default rates to
restrict institutional participation in federal student aid programs.
Unfortunately, America’s Historically Black Colleges and Universities (HBCUs) have
much higher cohort default rates than the national average. At the time of the 1998
Higher Education Act amendments, about one-third of the HBCUs had cohort default
rates above the legislated threshold rate. Therefore, many of those colleges were in
danger of losing eligibility to participate in the federal loan programs. A few remain in
such danger today.
This study focuses on an analysis of the factors that significantly affect the cohort
default rates at HBCUs, and on the potential for reducing these rates. It is hoped that
ndings from the analysis will help pave the way for further analysis, and ultimately
develop a better understanding of which institutional practices have greater success in
reducing the higher default rates for these colleges and universities.
INTRODUCTION
M O N O G R A P H S E R I E S 1
Background
Three discoveries contributed to the implemen-
tation of institutional cohort default rates as eligibil-
ity criteria. First, the data on default payments
showed that a disproportionate number of defaulting
borrowers had attended proprietary (private, for-prof-
it) business, trade, and technical schools (Merisotis,
1988; General Accounting Office, 1989). Second, it
was discovered that many defaulting borrowers spent
only a year or two in postsecondary education and
dropped out before completing their academic pro-
grams (Beanblossom and Rodriguez, 1989). It was
concluded that colleges and schools were at least par-
tially responsible for borrower defaults. The logic
was as follows: If schools enrolled only students who
had the ability to benefit from their curricula, if
schools offered programs and services that helped
retain admitted students until completion of their
education goals, and if schools helped their students
nd satisfactory employment after graduation, then
student borrowers would be much less likely to
default on their loans. Congress reasoned that if the
evidence indicated that the schools and colleges
could not do these things, then their students should
no longer be eligible for federal student loans.
A third discovery, that the vast majority of
defaults occur in the first year or two after borrowers
leave school or college (Davis and Knapp, 1990), led
Congress to consider their students’ cohort default
rates as primary evidence that certain postsecondary
institutions could not successfully serve borrowers.
The term “cohort default rate” was first defined in
the Omnibus Budget Reconciliation Act of 1990
(P.L. 101-508) as the percentage of students at a par-
ticular institution who enter repayment during a fis-
cal year who default before the end of the following
scal year.
For example, here is the formula for the most
recent cohort default rate (CDR) available, the one
for FFY 1997:
The Number of Students Who Entered
Repayment in FFY 1997 and Defaulted Before
the End of FFY 1998 (Numerator)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _x100= CDR
The Number of Students Who Entered
Repayment in FFY 1997 (Denominator)
Using such a simple measure as evidence of suc-
cess with students as a criterion for continued partici-
pation in the federal student loan programs is not
without complications. First, many factors that are
beyond the control of a school or college contribute
to defaults on student loans. For example, a borrow-
er’s unwillingness to seek any job or a job that earns
enough to meet loan payments, or a borrower’s deci-
sion to spend earnings on consumer goods rather
than loan payments, can lead to defaults. Changes
in the economy or local job market can inhibit bor-
rowers’ employment opportunities and subsequent
ability to repay their student loans.
Second, many institutions are committed to the
national goal of providing access to postsecondary
education and training to all who might benefit from
it. However, it is not always easy for such institutions
to determine which applicants will benefit.
Sometimes institutions that are committed to the
access goal admit and enroll students who are unable
to meet their curricular requirements. Such students
may drop out and, because their education and train-
ing were not successful, may have difficulty finding
jobs. If these students were borrowers, they may
default on their loans. Other “open door” institu-
tions, such as community colleges, that are commit-
ted to offering all students the opportunity to try to
succeed in postsecondary education will probably
enroll some who will fail, drop out, and subsequently
have difficulty repaying a student loan.
Third, the mission of some institutions is to serve
populations of students whose socioeconomic status-
es, previous education and training, and lack of pre-
paredness make it difficult for all of them to succeed,
regardless of what the institutions do for them. Thus
their student borrowers very likely will have higher
default rates than will the student borrowers at elite,
selective colleges. Only one of the HBCUs in the
study sample is selective in its admissions practices.
Congress was faced with choosing a specific
cohort default rate which would not discriminate
against colleges and schools that were trying to
achieve laudable goals and sometimes failed to do so,
and yet a rate which would restrict institutions who
were doing disservice to their student borrowers.
Congress decided that the threshold for loss of insti-
tutional eligibility would be a cohort default rate of
35 percent for FFY 1991 and FFY 1992, 30 percent
for FFY 1993, and 25 percent for any succeeding year.
Because of their special historical mission, and
the characteristics of the students they serve, the
HBCUs were exempted from these limitations until
M O N O G R A P H S E R I E S2
July 1, 1993. The Higher Education Amendments
Act of 1992 extended that exemption to July 1,
1998. (The loss of eligibility sanctions also does not
apply to tribally controlled community colleges and
Navajo community colleges.) The current law man-
dates that a school with an official cohort default rate
of 25 percent or greater for the three most recent fed-
eral fiscal years (in this case, FFY 1995, FFY 1996,
and FFY 1997) lose eligibility to participate in the
FFELP and the Direct Loan Program [Higher
Education Act, Section 435(a)(2)]. Such a school
becomes ineligible to participate in the FFELP and/or
Direct Loan program for the current and two suc-
ceeding fiscal years, unless the institution successful-
ly appeals its loss of eligibility.
Were it not for the exemption, when this
research was begun 18 months ago, fourteen HBCUs
would have lost their eligibility to participate in the
federal student loan programs because their cohort
default rates for FFY 1993, FFY 1994, and FFY 1995
each exceeded 25 percent (General Accounting
Office, 1998). Another nineteen institutions were in
some potential danger of losing eligibility because
their cohort default rates exceeded 25 percent in one
or two of the three most recent fiscal years. Still
another 13 institutions had FFY 1995 cohort default
rates in excess of the Department of Education’s
“cautionary” level of 20 percent.
Many HBCUs have achieved success in lowering
their cohort default rates. For example, nineteen
HBCUs whose FFY 1988, FFY 1989, and FFY 1990
rates exceeded the statutory threshold of 25 percent
have fallen below the threshold for the most current
cohort years (General Accounting Office, 1998).
However, many have not done as well as Congress
initially hoped they would. Their students continue
to default at a much higher rate than students at most
other types of postsecondary institutions. The aggre-
gate 1995 student loan cohort default rate for HBCUs
was 18.5 percent; the rate for non-HBCUs was only
7.8 percent (General Accounting Office, 1998).
Congress and the federal government remain
concerned about cutting federal budget costs. Due to
this concern, the recently-enacted Higher Education
Amendments of 1998 (P.L. 105-244) contains a pro-
vision which extended the exemption for HBCUs for
just one year (to July 1, 1999). After this date, all
HBCUs and tribally-controlled colleges that have
had cohort default rates of 25 percent or higher for
the past three federal fiscal years must submit default
management plans to the U.S. Secretary of
Education in order to remain eligible for participa-
tion. Thirteen HBCUs were required to submit these
plans. There is reason to worry that these HBCUs
will no longer be able to provide federal loans to their
students if their default rates remain high.
That most HBCUs have successfully lowered
their students’ cohort default rates indicates that they
changed their policies, programs, and practices to
achieve that goal effectively. Research has shown
that only 5 percent of all student borrowers who
receive their degrees default on their FFELP loans
(Davis, 1996). For this reason, it was hypothesized
that the HBCUs that have been most successful in
lowering their students’ cohort default rates were
those that were also successful in increasing their stu-
dent retention rates. Put another way, if their cohort
default rates went down, then the colleges’ retention
(and graduation) rates must have gone up.
Swail’s Conceptual Framework for Student
Retention (Swail, 1995) was used to help identify
and assess which factors might have contributed to
success in reducing cohort default rates. In doing so,
it was assumed that factors positively related to
improvements in retention are positively related to
improvements in default rates. The research goal was
to identify and quantify primary factors that are sta-
tistically associated with cohort default rate reduc-
tions. The intention is that identifying these factors
at colleges that have successfully cut their cohort
default rates will lead to potential replication by
other relatively unsuccessful institutions.
For example, if it is determined that a certain
base level of per student expenditures on academic
support services is necessary to achieve reductions in
cohort default rates, then colleges must be able to
achieve that level to expect to meet the federal cri-
teria. Or, for another example, if it is determined that
an increase in expenditures on student support ser-
vices is more closely related to a reduction in cohort
Figure 1. Five components of Student Retention
Framework (Swail, 1995)
M O N O G R A P H S E R I E S 3
default rates than an increase in expenditures on aca-
demic support services, then HBCUs will know better
how to target limited resources to reach their default
reduction goals. It is also intended that the research
ndings will help federal policymakers understand
whether it is realistic to expect every single HBCU to
reach the target cohort default rates for all schools
participating in the FFELP and Direct Loan program.
The Student Retention Model
Swail’s Conceptual Framework for Student
Retention (1995) focuses on the barriers and issues
facing minority students at HBCUs and other minor-
ity-serving institutions. Originated from a synthesis
of previous retention research, specific retention
strategies were then formed into a five-part frame-
work. A validation process by way of delphi tech-
nique resulted in the final framework, whose
components are presented in Appendix A. The five
basic components include student services, academic
services, curriculum and instruction, financial aid,
and admissions and recruitment.
Swail’s framework pulls together many of the
widely-held tenets of minority student retention, pre-
viously illustrated through landmark studies by Tinto
(1975), Astin (1977), Beal and Noel (1980), Bean
(1986), and others. While some may argue the
specifics of the models developed by these and other
researchers over the years, all subscribe to a similar
philosophy about the problems facing disadvantaged
students in higher education and the basic strategies
to increase student retention.
Figure 1 illustrates how the five components of
the conceptual framework act as a basic foundation
for student retention at a college or university. While
the geometric frame simplifies the nature of the stu-
dent experience in higher education, it does offer a
visual interpretation of how students and institutions
interact.
Figure 2 shows the relationship between the
institutional factors or practices of the college and
the academic and social skills students bring with
them to college. If one conceptualizes the triangle as
the student college experience, one can think
abstractly about the various factors that may impact
that experience. Academic preparation, readiness for
college, maturity, social awareness, ability to get
along, and the myriad of other developmental issues
all have some impact on a student’s ability to persist
in school. While the factors presented in Figure 2 are
not exhaustive, they do provide some idea of what
students bring with them from high school and/or
the workforce.
Figure 2.
Factors impacting
upon the geometric
model of student per-
sistence
M O N O G R A P H S E R I E S
4
The ultimate responsibility of the insti-
tution is to provide a quality education for
its students. Through each of the five com-
ponents of the framework, institutions have
an opportunity to build a support system to
help students stay in school and receive a
degree. For example, an institution may
provide specialized orientations (student
services) to help freshman students accli-
mate to their new surroundings, or provide
tutoring and study assistance opportunities
(academic services) to assist the academic
development of their students. (See
Appendix A for details of the original
framework).
One challenge facing HBCUs is the set
of issues students bring with them to cam-
pus. It was mentioned earlier in this paper
that the mission of many HBCUs is to serve
populations of students whose backgrounds
may make it difficult to succeed. Generally
speaking, the academic preparation of stu-
dents attending HBCUs is not commensu-
rate with their peers at other institutions.
These students, on average, have not taken
the same rigorous course work in high
school, have not been exposed to high-level
education, have not always had the support
of family and community, and lack some of
the general skills required to excel in post-
secondary education (e.g., study, time man-
agement, etc.). Additionally, research
suggests that these students also have a dif-
cult time making the social adjustment to
postsecondary education.
These factors alone provide a formida-
ble barrier to success for many students.
However, academics represent only part of
the challenge. Many of the students attend-
ing HBCUs require financial aid to attend.
How students and their families pay for col-
lege varies widely depending on their acad-
emic ability, current income status, and
available savings or other financial
resources. Low-income students may be eli-
gible for Pell Grants, but these provide only
modest support at most colleges. The
remainder is often made up by loans, many
of which are subsidized and offered through
the federal government. In 1996, about 53
percent of undergraduate students at
HBCUs received Stafford Loans of $3,629, on aver-
age. While average tuition costs at HBCUs are lower
than national averages, many students will have
accumulated loan debt in excess of $10,000.1
For graduates of HBCUs, repaying their loans is
not particularly difficult. Most will find jobs that will
allow them to begin repayment soon after gradua-
tion. And while no one likes to repay loans, a
$10,000 debt is not a formidable barrier to most grad-
uates. The bigger problem is with students who leave
college before completing their degree programs.
These students do not have the advantages afforded
bachelor’s degree recipients, including increased
skill, knowledge, and a degree to flaunt around the
job market. In a society that rewards credentialism,
the degree itself—regardless of skill in many cases—
is valued by employers. Unfortunately, students with-
out degrees will have a much more difficult time
nding a job (let alone a decent paying job) than
those with degrees. Still, the loan debt persists, and
many of these students face a difficult situation:
choosing between paying rent or paying their student
loan. Understandably, repaying student loans quickly
becomes of secondary importance.
It was hypothesized that the answer to decreasing
institutional cohort default rates lies in increasing
student persistence to a degree. While persistence
rates are affected by admissions policies, what insti-
tutions do with their students once on campus may
be the only direct strategy of retaining students from
year-to-year.
Research Design and Methodology
A mixed time-series/cross sectional model was
used to examine the effectiveness of various institu-
tional strategies designed to reduce cohort default
rates. The pooled model contained data for 80
HBCUs2over the entire period for which cohort
default rates have been calculated by the U.S.
Department of Education (1987 - 1995).3The model
itself can be thought of as a series of cross-sectional
models over time describing changes in both cohort
default rates and the institutional strategies designed
to reduce them. The financial and enrollment data
for the HBCUs came from the U.S. Department of
Education’s Integrated Postsecondary Education
Data System (IPEDS) annual finance and fall enroll-
ment surveys, the undergraduate admissions selectiv-
ity data came from the 1998 Peterson’s Guide to
Four-Year Colleges, and the cohort default rates were
provided by the U.S. Department of Education.
M O N O G R A P H S E R I E S 5
For each of the five strategic areas identified in
Swail’s retention framework, there were at least at
least two conceptual ways in which each strategy
could be measured. For those strategies best repre-
sented through institutional expenditures, a “bud-
getary shares” and a “spending per FTE” framework
was developed. In this manner, four expenditure-
based strategies (academic services, student services,
instruction, and financial aid in the form of grants)
could all be measured as either their respective share
of total educational and general expenditures in a
given year, or in terms of annual expenditures per full-
time equivalent student. To test the inferential
robustness of the theory, estimates were developed
under both frameworks. For the fifth and final type of
strategy, institutional selectivity, an ordinal measure
describing the selectivity of the institutions’ admis-
sions policies based on the SAT scores of admitted
freshmen was developed. The precise definitions of
the variables, along with their sources, are presented
in Table 1.
These measures were then used to first estimate
the effects of each one of the institutional strategies
independent of the other strategies, beginning by
using ordinary least squares to linearly regress each
one of the measures against the institutional cohort
default rates. However, since it takes time for these
institutional changes to effect an institution’s cohort
default rate, lag structures of between one and four
years were tested, to represent the length of time it
might take for a change to work its way through the
system.
After measuring the magnitude of the individual
linear effects, tests were conducted for non-linearities
in the estimated effects. Since in particular, many of
Table 1
Variable Definitions and their Sources
Variable Name Definition Source
CDF Cohort Default Rate U.S. Department
of Education
ISPCT Percentage of Total Educational and General IPEDS
Expenditures Devoted to Instructional Services
ISFTE Expenditures on Instructional Services Per IPEDS
Full-Time Equivalent Student
ASPCT Percentage of Total Educational and General IPEDS
Expenditures Devoted to Academic Services
ASFTE Expenditures on Academic Services Per IPEDS
Full-Time Equivalent Student
SSPCT Percentage of Total Educational and General IPEDS
Expenditures Devoted to Student Services
SSFTE Expenditures on Student Services Per Full-Time Equivalent Student IPEDS
AIDPCT Percentage of Total Educational and General IPEDS
Expenditures Devoted to Financial Aid
AIDFTE Expenditures on Financial Aid Per Full-Time Equivalent Student IPEDS
SELECT Categorical Variable Describing the Selectivity of the Institutional
Admissions Policy (Noncompetitive, Minimally
Difficult, Moderately Difficult, Very Difficult, Most Difficult ) Peterson’s Guide
M O N O G R A P H S E R I E S6
the expenditure-based strategies are hypothesized to
have their greatest effects at low levels of spending,
several functional forms designed to capture this rec-
iprocal effect were tested. Together, the linear and
non-linear effects were then used to estimate both
the relative size of the effects produced by each insti-
tutional strategy, as well as to identify the most effi-
cient lag structure for each strategy.
In the final part of the analysis, the individual
effects were combined into a more general model
describing all five of the institutional strategies. To
determine the optimal lag structure and specification
for the final “budgetary shares” and “expenditure per
FTE” models, the results from the first two parts of
the analysis were used to sequentially narrow the
range of acceptable lag structures and specifications.
This sequencing was done until it created a set of
models that produced the same inferences and
explained the same variation in cohort default rates.
From among these few models, ones with con-
formable lag structures were selected to allow com-
parisons of the results of the two conceptual models.
Results
The first part of the analysis involved estimating
the effects of each one of the institutional strategies
independent of the other strategies. Each one of the
“budgetary shares” and “spending per FTE” variables
was linearly regressed against the cohort default rates.
The elasticities, significance levels, lag structures,
and explained variation for both models are present-
ed in Tables 2 and 3.
Examination of these tables shows that three of
Table 3
Elasticities, Lag Structures, and Explained Variation for the
Independent Variables in the Linear Expenditure per FTE Model
Elasticity Significance Optimal Lag Range of Explained
Level Structure Variation (R2)
Academic Support -.119 1% 1 Year 3% - 5%
Student Services .040 Insignificant 4 Years 0% - 0%
Instructional -.294 1% 4 Years 14% - 15%
Grants .253 1% 4 Years 8% - 11%
Admissions -.147 1% 4 Years 12% - 14%
Table 2
Elasticities, Lag Structures, and Explained Variation for the
Independent Variables in the Linear Budgetary Shares Model
Elasticity4Significance Optimal Lag Range of Explained
Level Structure Variation (R2)
Academic Support -.393 1% 1 Year 1% - 2%
Student Services .543 1% 4 Years 1% - 3%
Instructional -.412 1% 4 Years 12% - 14%
Grants .528 1% 4 Years 19% - 21%
Admissions -.147 1% 4 Years 12% - 14%
per FTE results in a smaller decrease in the institu-
tion’s cohort default rate (.29). To a large extent, the
same holds true for academic expenditures: a 1 per-
cent increase in the share of the budget devoted to
academic services results in a decrease of slightly less
than half a percentage point (.39) in the institution’s
cohort default rate in one year, while a similar
increase in academic expenditures per FTE results in
a smaller decrease in the institution’s cohort default
rate over the same time period (.12). For both mod-
els, tightening up admissions policies by admitting
more “select” students results in a decrease in the
institution’s cohort default rate by .15 percent over
four years.
Two strategies result in higher cohort default rates.
the institutional strategies led to decreases in the
institution’s cohort default rates, and the other two
led to increases or had no effect at all. The three
strategies that led to decreases involved tightening
up admissions policies, increasing spending on acad-
emic services, and increasing spending on instruc-
tional services. For all three of these institutional
strategies, the effect sizes are modest, and take from
between one and four years to work their way
through the system.
For example, a 1 percent increase in the share of
the budget devoted to instructional services results in
a decrease of slightly less than half a percentage point
(.41) in the institution’s cohort default rate in four
years. A 1 percent increase in instructional services
M O N O G R A P H S E R I E S 7
Table 4
Elasticities, Lag Structures, and Explained Variation for the
Independent Variables in the Non-Linear Budgetary Shares Model
Elasticity Significance Optimal Lag Range of Explained
Level Structure Variation (R2)
Academic Support -.090 1% 1 Year 2% - 3%
Student Services .075 1% 4 Years 0% - 0%
Instructional -.385 1% 1 Year 13% - 19%
Grants .313 1% 4 Years 15% - 16%
Admissions -.147 1% 4 Years 12% - 14%
Table 5
Nonlinear Effect Sizes, Lag Structures, and Explained Variation
For the Independent Variables in the Expenditure per FTE Model
Elasticity Significance Optimal Lag Range of Explained
Level Structure Variation (R2)
Academic Support -.084 1% 3 Years 3% - 5%
Student Services .039 1% 3 Years 1% - 2%
Instructional -.236 1% 1 Year 9% - 17%
Grants .234 1% 4 Years 6% - 9%
Admissions -.147 1% 4 Years 12% - 14%
M O N O G R A P H S E R I E S8
These are increasing expenditures on student services
and financial aid, both of which take four years to
work through the system and, in the “budgetary
shares” model, result in slightly greater than half a
point increases in an institution’s cohort default rate
(.54 for student services and .53 for financial aid).
However, in the “expenditure per FTE” model, there
is no effect associated with increases in student ser-
vice expenditures, and a more modest increase in
cohort default rates is associated with a 1 percent
increase in expenditures on grants per FTE (.25).
Although these results are only preliminary, one
possible explanation for the positive, and possibly
counter-intuitive, effect of financial aid may be that
increasing expenditures on grants encourages more
marginal students, who are heavily dependent on
loans as well as grants, to enroll and then subse-
quently drop out. As for the positive results associat-
ed with student services, perhaps increases in this
share of the budget are spent on non-academic ser-
vices that actually decrease the student’s chances of
completing their degree (e.g., expenditures on frater-
nities and sororities, or for purely social events).
In the next stage of the analysis, non-linearities
were tested for in ways that institutional strategies
worked to change the institution’s cohort default
rates. This was done because many of the expendi-
ture-based strategies were hypothesized to have their
greatest effect at low levels of spending, and this sort
of reciprocal effect can only be captured through the
use of non-linear functional forms. This information
is presented in Tables 4 and 5, and suggests that, for
several strategies, the non-linear approach explains
more variation in cohort default rates than does the
linear specification. For example, in the “budgetary
shares” model, the reciprocal specification associated
Table 6
Specifications, Elasticities, Significance Levels, and Lag Structures for
The Independent Variables in the Full Budgetary Shares Model (R2= .41)
Specification Elasticity Significance Level Lag Structure
Academic Support Reciprocal .032 Insignificant 1 Year
Student Services Linear .426 1% 4 Years
Instructional Reciprocal -.269 1% 1 Year
Grants Linear .341 1% 4 Years
Admissions Linear -.001 1% 4 Years
Table 7
Specifications, Elasticities, Significance Levels, and Lag Structures for
The Independent Variables in the Full Expenditures per FTE Model (R2= .36)
Specification Elasticity Significance Level Lag Structure
Academic Support Reciprocal .038 5% 1 Year
Student Services Linear .070 10% 4 Years
Instructional Reciprocal -.258 1% 1 Year
Grants Linear .248 1% 4 Years
Admissions Linear -.001 1% 4 Years
M O N O G R A P H S E R I E S
with instructional expenditures explains between 14
and 19 percent of the variation in cohort default
rates, while the linear specification explains only
between 12 and 14 percent. Similarly, the reciprocal
specification for academic support explains between
2 and 3 percent of the variation in cohort default
rates, compared with between 1 and 2 percent for the
linear specification.
In the “expenditures per FTE” model, the non-
linear approach represented an improvement over
linear specifications for several of the institutional
strategies. In particular, the reciprocal specification
associated with instructional expenditures, academic
support expenditures, and student service expendi-
tures all produced models that explained more of the
variation in cohort default rates than did their linear
counterparts. Although the student services strategy
seems to have little overall effect on cohort default
rates, increases in both academic support expendi-
tures and instructional expenditures seem to have
their greatest effect at low levels. This suggests that
institutions that spend relatively little in academic
support and instructional areas should be able to see
significant reductions in their cohort default rates by
increasing funding in them.
Taken together, the results of the first two parts
of the analysis suggest that the strongest effects, both
in terms of explained variation and size of effects, are
for spending on grants and instructional services.
However, they operate in different directions. A 1
percent increase in the share of the budget devoted
to grants causes about half a percentage point
increase in an institution’s cohort default rate (.53).
A similar increase in the share of the budget devoted
to instructional expenditures produces slightly less
than half a percentage point decrease in an institu-
tion’s cohort default rate (.39).
Although expenditures in both academic ser-
vices and student services are associated with fairly
large effect sizes, they explain a relatively small
amount of the variation in cohort default rates. The
institutional selectivity measure explains a relative-
ly large amount of variation in cohort default rates,
but its effect size is quite small. As such, it appears
that increased spending on instructional services,
especially at low levels, seems to offer the best
chance for HBCUs to reduce their cohort default
rates effectively.
However, before making any conclusive state-
ments about the size of the effects associated with the
ve institutional strategies, it is appropriate to com-
bine these effects into a more general model describ-
ing all five of the institutional strategies. As was
noted above, the results from the first two parts of the
analysis were used to narrow the range of lag struc-
tures and specifications to arrive at final “budgetary
shares” and “expenditure per FTE” models. These
models have the same lag structures and specifica-
tions and are presented in Tables 6 and 7.
Examination of these tables yields some inter-
esting and surprising findings. Although spending on
both grants and instructional services still yields the
strongest and most robust effects, their magnitude is
diminished when all of the other institutional strate-
gies are statistically controlled for. Specifically, a 1
percent increase in the share of the budget devoted
to grants increases an institution’s cohort default
rate by about a third of a percent (.34). At the same
time, a 1 percent increase in the share of the budget
devoted to instructional services decreases an insti-
tution’s cohort default rate by slightly less than a
third of a percent (.27). As expected, similar
increases in expenditures per FTE for both of these
strategies produce smaller changes than those associ-
ated with the 1 percent increase in the share of the
budget -- specifically, .25 percent for grants and .26
percent for instructional services. However, from a
policy perspective, increasing spending on instruc-
tional services produces a declining cohort default
rate within a year, while increasing spending on
grants takes four years to produce an increase in
their cohort default rate.
The results for the three other institutional
strategies are mixed. For example, although increas-
es in admissions selectivity are responsible for signif-
icant decreases in cohort default rates, the size of the
effect is so small as to be of no practical use to insti-
tutions struggling with high cohort default rates.
Put another way, tightening up admissions criteria is
unlikely to significantly cut default rates. When
expenditures on academic support are included in
the full model, the effect size washes out. This sug-
gests that other institutional strategies are explain-
ing the same thing. Finally, the results on student
service expenditures are mixed, but they do suggest
that increases in this type of spending may indeed
lead to increases in cohort default rates four years
down the line.
9
M O N O G R A P H S E R I E S
Conclusions and Policy Implications
For several year some HBCUs have been strug-
gling with persistently high cohort default rates.
Despite the fact that many have been successful in
reducing their default rates below the 25 percent
threshold that applies to most other types of institu-
tions, some HBCUs remain in danger of losing their
eligibility to participate in the federal student loan
programs. Since students who drop out of college
have been shown to default at much higher rates
than those who persist, this paper has examined sev-
eral retention strategies available to HBCUs to see
what effects they might have on lowering the insti-
tutions’ cohort default rates.
Through the use of both a “budgetary shares” and
“expenditure per FTE” model, the effectiveness of
several institutional strategies designed to increase
retention were examined to see what effect, if any,
they may have on an institution’s cohort default
rates. Although the results were mixed, several
important facts surfaced.
First, increases in such non-academic compo-
nents of an institution’s budget as grants and student
services are statistically associated with increases in
the cohort default rates. These findings suggest that
increases in grant spending may actually encourage
marginal students heavily dependent on grants and
loans to enroll and then subsequently drop out, and
that spending increases on student services may help
produce a campus climate that actually reduces stu-
dents’ chances of completing their degrees.
Second, increases in at least one academic com-
ponent of an institution’s budget, instructional ser-
vices, are associated with lower cohort default rates,
which suggests that spending more on instructional
services may help motivate students to persist in
degree completion. Furthermore, since increases in
spending on instructional services have their biggest
effect on cohort default rates at relatively low levels
of spending, the HBCU’s that are struggling with
high cohort default rates while devoting a relatively
small share of their budgets to instructional services
are strongly urged to increase spending in this area.
Finally, the institutional strategies examined in
this study together explained about 40 percent of the
variation in cohort default rates. This suggests that
institutional decision-making about allocation of
resources among different alternative functions
affects cohort default rates. But student-based char-
acteristics and other factors, not to mention chance,
are left to explain the majority of the variation in
cohort default rates.
Taken together, the results of this study should
provide new insights for many HBCUs struggling
with stubbornly high cohort default rates. Although
only one institutional strategy, increasing expendi-
tures on instructional services, was especially effec-
tive for both increasing retention and reducing
cohort default rates, the fact that the biggest “bang”
for the instructional services “buck” comes at low
levels of instructional services spending should help
many HBCUs reduce their cohort default rates.
For policymaking purposes, it is important to
note that all the strategies available to HBCUs to
increase retention explain less than half of the varia-
tion in HBCU cohort default rates. This suggests
that many HBCUs may be doing all they can under
their present circumstances to keep their students in
school and to handle their default rates. Therefore,
it seems that Congress would be well advised to allow
these institutions to continue to participate in the
federal student loan programs, allowing them to serve
their students.
10
M O N O G R A P H S E R I E S
APPENDIX A-1
Institutional Components for Student Persistence
12
M O N O G R A P H S E R I E S
APPENDIX A-2
Institutional Components for Student Persistence
13
M O N O G R A P H S E R I E S
APPENDIX A-3
Institutional Components for Student Persistence
14
M O N O G R A P H S E R I E S
APPENDIX A-4
Institutional Components for Student Persistence
15
M O N O G R A P H S E R I E S
APPENDIX A-5
Institutional Components for Student Persistence
16
M O N O G R A P H S E R I E S
APPENDIX B
Instructional Expenditures: This category includes general academic instruction, occupational and vocational
instruction, special session instruction, community education, preparatory and adult basic education, and remedi-
al and tutorial instruction conducted by the teaching faculty for the institution’s students.
Academic Support Expenditures: This category includes expenditures for the support services that are an inte-
gral part of the institution’s primary mission of instruction, research, or public service, including expenditures for
libraries, museums, galleries, audio/visual services, academic computing support, ancillary support, academic admin-
istration, personnel development, and course and curriculum development.
Student Services Expenditures: This category includes funds expended for admissions, registrar activities, and
activities whose primary purpose is to contribute to students’ emotional and physical well-being and to their intel-
lectual, cultural, and social development outside the context of formal instruction, including expenditures for
career guidance, counseling, financial aid administration, and student health services.
Scholarships and Fellowships: This category includes all expenditures given in the form of outright grants and
trainee stipends to individuals enrolled in formal course work, as well as aid to students in the form of tuition
remission.
17
REFERENCES
Astin, A.W. 1977. Four Critical Years. San Francisco, CA: Jossey-Bass, Inc.
Beal, P.E. and L. Noel. 1980. What Works in Student Retention.
Iowa City, IA: American College Testing Program.
Bean, J. P. 1986. “Assessing and Reducing Attrition.”
New Directions for Higher Education 14(1): 47-61.
Beanblossom, G. and B. R. Rodriguez. 1989.
Characteristics of Stafford Loan Program Defaulters: A National Study.
Washington: U.S. Department of Education, Office of Student Financial Assistance.
Conner, D, R. Saab, and K. Cicmanec. 1997. Federal Student Loans Programs
Data Book, FY 94 - FY 96. Washington: U. S. Department of Education.
Davis, J. S. 1990. Correspondence with Gisela Vallandigham, August 30.
Davis, J. S. And L. G. Knapp. 1990. “How Federal Subsidies to the Stafford Loan
Program Are Distributed Among Pennsylvania Borrowers.”
Journal of Student Financial Aid, 20(3): 24-43.
General Accounting Office. 1989. Guaranteed Student Loans: Analysis of Student Default
Rates at 7,800 Postsecondary Schools. GAO/HRD-89-63BR.
Washington: U.S. General Accounting Office, 1989.
______________________. 1998. Student Loans: Characteristics of Students and Default
Rates at Historically Black Colleges and Universities.
GAO/HEHS-98-90. Washington: U.S. General Accounting Office.
Merisotis, J.P. 1988. "Default Trends in Major Postsecondary Education Sectors."
Journal of Student Financial Aid. 18 (3): 18-28, .
Swail, W. S. 1995. “The Development of a Conceptual Framework to Increase Student
Retention in Science, Engineering, and Mathematics Programs at Minority Institutions
of Higher Education.” Ed.D. dissertation, The George Washington University.
Tinto, V. 1975. “Dropout from Higher Education: A Theoretical Synthesis of
Recent Research” Review of Educational Research, 45: 89-125.
ENDNOTES
1Data extracted from the National Postsecondary Student Aid Study (NPSAS) 1995-96.
2Although there were 98 HBCUs in existence during this study period, this analysis was lim-
ited to the 80 HBCUs that had four-year undergraduate programs.
3Given the lag structure in the general model, non-cohort default rate data were also collected
for the 1984 to 1987 period.
4The elasticities that appear in Tables 2-7 are from the models with the greatest explanatory
power. However, since the “admissions” variable is categorical, the concept of an elasticity is unde-
ned. As such, this number represents the reduction in the default rate associated with moving
from one level of institutional undergraduate admissions selectivity to the next higher level.
... To the extent that college graduates are more likely to find employment and earn higher wages, repaying student loans will be less difficult for them than for students who dropped out or stopped out prior to graduation. These students do not have the advantages afforded college graduates, including increased skill, knowledge, and a degree to present to prospective employers (Galloway & Swail, 1999). However, student loan debt remains whether one is employed, underemployed or unemployed. ...
... NCES data indicates that repaying student loans is more an issue for students at Georgia's HBCUs than PWCUs with average federal loan 3-year student loan defaults rates of 20 percent and 10 percent respectively. Considering the correlation between college graduation and repayment of student loans, one answer to decreasing institutional cohort default rates, especially at HBCUs is to increase students' persistence to degree attainment (Galloway & Swail, 1999). ...
Article
Full-text available
Since their inception 150 years ago, Georgia’s historically black colleges and universities (HBCUs) have provided African American students with the best mechanism, and for some, the only opportunity to receive a collegiate education. Justifications for and against the continued relevance of HBCUs in Georgia have been raised and argued since the Brown decision in 1954 and the passage of the Civil Rights Act of 1964. More than 50 years after these monumental decisions Georgia’s HBCUs remain relatively segregated, with lower endowments, budgets, graduation rates and higher cohort default rates than the state’s Predominantly White Colleges and Universities (PWCU). In defiance of political tactics to close and merge them, and despite the confluence of lower average admission requirements, less funding for institutional scholarships, limited technological resources and smaller operating budgets, Georgia’s HBCUs have persevered in providing higher educational opportunities not only for African Americans, but for students of all races. This entry focuses on HBCUs in the State of Georgia and adds to the literature on their history, purpose, effectiveness and continued relevancy.
... Significant research has been compiled on the subject of minority student retention over the past few decades, among which the leading ones include Stewart et al (1997), Galloway and Watson (1999), Good et al (2000), Greenwald andDavis (2000), andSt. John (2004). ...
Article
Since their inception 150 years ago, Georgia's historically black colleges and universities (HBCUs) have provided African American students with the best mechanism, and for some, the only opportunity to receive a collegiate education. Justifications for and against the continued relevance of HBCUs in Georgia have been raised and argued since the Brown decision in 1954 and the passage of the Civil Rights Act of 1964. More than 50 years after these monumental decisions Georgia's HBCUs remain relatively segregated, with lower endowments, budgets, graduation rates and higher cohort default rates than the state's Predominantly White Colleges and Universities (PWCU). In defiance of political tactics to close and merge them, and despite the confluence of lower average admission requirements, less funding for institutional scholarships, limited technological resources and smaller operating budgets, Georgia's HBCUs have persevered in providing higher educational opportunities not only for African Americans, but for students of all races. This entry focuses on HBCUs in the State of Georgia and adds to the literature on their history, purpose, effectiveness and continued relevancy.
Chapter
Retention is a major issue at Texas Southern University (TSU), as it is at most universities. As a historically Black public institution that has an open admissions policy, the effort to increase retention rates occupies considerable attention. This chapter is based on an institutional study of a planning process designed to improve the retention of freshmen at TSU.
Article
Conclusions Asynchronous Distance Learning courseware offers several advantages to multiple disadvantaged learning populations. While asynchronous Distance Learning Interactive Multimedia Instruction offers financially strapped HBCUs lower operating costs, it also offers HBCUs and non-traditional student populations improved (1) accessibility, (2) user acceptance. Many Distance Learning students chose online sections as an alternative to not taking the course at all rather than to the traditional classroom section. Among technologically savvy populations, user acceptance for Distance Learning is high. Also interesting was how traditional learning populations benefited from some Distance Learning strengths, (1) hypermedia external resources for self-directed learning, and (2) frequently graded assignments improving study habits. In converse, the non-traditional student populations taking the Distance Learning sections reported less studying time, and suffered from slightly lower learning retention on the final examination.
Article
This chapter examines the unique contributions and strategies of minority-serving institutions that successfully serve and graduate students of color.
Article
Full-text available
Although a number of studies have focused on the identification of factors impacting upon minority student persistence at four-year colleges and universities, few have studied the specific factors that affect minority persistence in science, engineering, and mathematics departments at predominantly minority-serving institutions. The main objectives of this study were to: identify causal factors regarding minority-student attrition, identify successful retention practices and programs, and develop a retention framework for administrators and practitioners to utilize during the planning and implementation stages of program development. It utilized a modified Delphi process to validate a conceptual framework for student retention that was developed from an extensive review of related literature. A panel of 16 experts from across the country rated and commented on specific objectives within the framework during the 2-round Delphi technique. The result of this was a student retention framework incorporating five components: financial aid, recruitment and admissions, academic services, curriculum and instruction, and student services. A collection of approximately 80 specific objectives were identified during this process and are categorized within the 5 components. Contains 156 references. (JRH/Author)
Article
Distribution of federal expenditures for interest benefits, special allowance payments, default claims reinsurance, and administrative cost allowances among over 61,000 borrowers is described, including in-school, grace and repayment periods by institution type, years enrolled, dependency status, and family financial circumstances. (Author/MSE)
Article
Information on GSL defaults in five states is reviewed: California, Illinois, Massachusetts, New Jersey, and Pennsylvania. Default rates are defined and levels of default are examined using a variety of analytical methods. (Author/MLW)
Article
This study examined characteristics differentiating defaulters on Stafford student loans (formerly Guaranteed Student Loans) from non-defaulters. Data were drawn from a sample of 100,000 records randomly selected from the fiscal year 1987 guarantee agency cumulative borrower file. A cohort of 7,382 borrowers taking out their last loan in FY 1983 was divided into defaulters and non-defaulters. The relationships of five variables (institution type, adjusted gross income, dependent/independent status, enrollment status--graduated or withdrew, and years of schooling completed) with each other and with default/non-default status were examined. The study also provides cohort default rates for various borrower subpopulations based on a large, national random sample, furnishes a comprehensive summary and critique of the literature, and suggests potentially fruitful avenues for further research. (MSE)
Article
Improving retention calls for methodical investigation and program design.
Federal Student Loans Programs Data Book, FY 94 -FY 96
  • D Conner
  • R Saab
  • K Cicmanec
Conner, D, R. Saab, and K. Cicmanec. 1997. Federal Student Loans Programs Data Book, FY 94 -FY 96. Washington: U. S. Department of Education.
Correspondence with Gisela Vallandigham
  • J S Davis
Davis, J. S. 1990. Correspondence with Gisela Vallandigham, August 30.