| I.
Introduction
Since
the early 1980s, welfare policymakers and program operators
have debated the role of adult education in program strategies
to help welfare recipients make the transition from welfare
to work. The so-called Human Capital Development, or HCD,
strategy focuses on providing education and increasing welfare
recipients’ basic academic skills and education credentials,
following research evidence that these skills and credentials
are prerequisites to obtaining stable employment.
[1] This HCD strategy became popular especially
during the mid to late 1980s. For example, the California
Greater Avenues for Independence (GAIN) program, initiated
in 1986 and evaluated by Manpower Demonstration Research Corporation
(MDRC), provided education and training to large numbers of
welfare recipients. During the same time, the federal government
also placed a greater emphasis on adult education, as evidenced
in the Family Support Act (FSA) of 1988, Public Law 100-485.
However,
during the early 1990s, alternative program strategies gained
popularity, seeking rapid job entry for welfare recipients
instead of providing them with education first. Such strategies
- known as Labor Force Attachment (LFA), or “work first,”
strategies - are supported by research findings
suggesting that quickly entering the labor force is a promising
trajectory to long-term self-sufficiency. [2] For example, in 1995, California ended testing
of literacy and math skills as welfare recipients enter the
state’s welfare-to-work program, thereby eliminating these
tests as a way to determine who initially needs education
services. This turn toward more work-focused welfare-to-work
programs was reinforced by the Personal Responsibility and
Work Opportunity Reconciliation Act (PRWORA) of 1996, P. L.
104-193, which placed time limits on welfare receipt, making
longer stays in education and training programs less attractive
for program operators and welfare recipients.
However,
the debate surrounding HCD programs for welfare recipients
is not settled. Although LFA programs appear to be more effective
than education-focused programs in the short term, there is
no proof that offering job search programs to all welfare
recipients (regardless of their education needs) leads to
long-term self-sufficiency for a majority of the welfare caseload.
A significant number of programs currently operated for welfare
recipients offer education classes among their array of possible
services, and such services are likely to remain important
under Temporary Assistance for Needy Families (TANF), the
funding structure for PRWORA.
Comparisons
of HCD and LFA strategies are not a major focus of this report.
Instead, this report mostly addresses key questions about
how HCD programs in general, and adult education activities
in particular, affect the educational and economic outcomes
of welfare recipients. Specifically, these analyses move beyond
overall HCD program effects and focus on participation in
adult education and the effects of such participation. This
is important, because many welfare recipients who are assigned
to HCD programs do not enroll in adult education classes or
they drop out of these classes after only a brief spell of
participation. As a result, the impacts of the larger programs
do not necessarily reflect the full potential of the adult
education services provided to those who receive them. Learning
more about how those specific services affect participants
is the primary focus of this report.
|
Background Information: Research Design
for the National Evaluation of Welfare-to-Work Strategies
This
report is part of the National Evaluation of Welfare-to-Work
Strategies (NEWWS)¾an evaluation of programs
begun under the Family Support Act, conducted by the
Manpower Demonstration Research Corporation, and funded
by the U.S. Department of Health and Human Services,
with support from the U.S. Department of Education.
The evaluation includes programs in seven sites across
the country: Atlanta, Georgia; Columbus, Ohio; Detroit,
Michigan; Grand Rapids, Michigan; Oklahoma City, Oklahoma;
Portland, Oregon; and Riverside, California. It uses
a random assignment research design to estimate the
overall effects of the welfare-to-work programs being
studied.
In each site, individuals who
were required to participate in the program were assigned,
by chance, either to a program group that had access
to education, training, and other employment services
and whose members were required to participate in the
program or risk a reduction in their monthly welfare
grant (a “sanction”), or to a control group, which received
no services through the program but whose members could
seek out services from the community. This random assignment
design ensures that there were no systematic differences
between the background characteristics of people in
the program group and those in the control group when
they entered the study. Thus, any subsequent differences
in outcomes between the groups can be attributed to
the program.
It is important to note that
the programs in the evaluation were not subject to the conditions
and requirements of what are currently referred to as “welfare-to-work”
programs as defined under TANF. During the follow-up period analyzed
in the report, individuals in the 11 studied programs did not face
a time limit on eligibility for welfare assistance as they would
now under TANF. All the programs, however, shared TANF’s primary
goal of moving welfare recipients into paid work and off assistance.
Furthermore, among the 11 programs are some which are strongly employment-focused—the
welfare-to-work strategy favored under TANF—as well as some which
are strongly basic education-focused. The programs varied in many
other ways as well, including how broadly the participation mandate
was applied to the welfare caseload and how strictly it was enforced,
the amount of child care support provided for program participation
or employment, and methods of case management. The programs also
served different welfare populations and operated in a variety of
labor markets. Lastly, it is important to point out that the programs
being evaluated here are state-operated welfare-to-work programs
originally developed under the Family Support Act of 1988. These
programs are unrelated to the “welfare-to-work” programs currently
being operated by local Workforce Investment Boards, supported by
the Workforce Investment Act, and administered by the U.S. Department
of Labor.
The
research design for the study presented in this report often diverges
from that of the larger NEWWS study. Instead of comparing randomly
assigned program group members and control group members, this study
often compared the experiences of welfare recipients who participated
in certain adult education activities with those of recipients who
did not and sought to estimate how varying degrees of participation
affected education and employment outcomes. A detailed discussion
of these “nonexperimental” comparisons is featured in a separate
text box on page 7.
|
A. Purpose of This Study
The
analyses presented here help answer many important policy questions surrounding
adult education for welfare recipients. These questions concern (1) the
quality of the education services provided, (2) the extent to which welfare
recipients participate in education, (3) the extent to which welfare recipients
earn education credentials, (4) the value of the education services provided,
and (5) the value of basic skills and education credentials in the labor
market during the mid-1990s. These issues are addressed primarily by comparing
the experiences of recipients who participate in adult education with
those of recipients who do not and by assessing the relative effectiveness
of different levels of participation in adult education.
B. Overview
of This Report
These
analyses of how adult education works in the context of welfare-to-work
programs were conducted for a large sample of welfare recipients
who entered one of the 11 programs studied in the National
Evaluation of Welfare-to-Work Strategies (NEWWS) without
a high school diploma or General Educational Development
(GED) certificate.
[3] Thus, the findings do not generalize to welfare
recipients who do have a high school diploma but who
still may be served by HCD programs that provide more advanced
levels of education and training. The types of adult education
examined in the report encompass adult basic education (ABE)
classes, programs preparing students for the GED exam, regular
high school classes, and classes in English as a Second Language
(ESL). Among these, ABE and GED preparation accounted for
most of the adult education in the 11 mandatory welfare-to-work
programs studied. These 11 programs operated in seven sites,
and each program was operated under the federal FSA and its
Job Opportunities and Basic Skills Training (JOBS) program.
(Program intake for this study began in June 1991 and ended
in December 1994; data presented cover June 1991 through December
1997. See the accompanying box for further information about
this study and report.)
-
This
chapter summarizes most, but not all,
[4] of the analyses presented as a collection
of papers in this report, specifically addressing the
following questions:
-
What
are the characteristics of adult education providers in
welfare-to-work programs? What are typical attendance
patterns in these classes?
-
To
what extent, and for whom, do welfare-to-work programs
increase participation in adult education services and
increase educational attainment and achievement?
-
Do
education-focused welfare-to-work programs improve education
outcomes?
-
What
is the payoff to additional participation in adult education?
-
How
do education outcomes and milestones affect the employment
outcomes and self-sufficiency of welfare recipients?
-
Among
those who participate in adult education, who moves on
from adult education to receive postsecondary education
and training, and how does this contribute to their earnings
and self-sufficiency?
The
analyses in the report take a unique perspective on adult
education, one that will be of interest to the adult education
community as well as to those involved in welfare policy.
First, our analyses reflect on the effectiveness of adult
education services provided to a highly disadvantaged group
of students: low-income, mostly jobless, single-parent women
who lack a high school diploma or GED and are receiving welfare.
This group represents a significant share of all adult education
students. One study found that, in 1992, 22 percent of all
new students in U.S. adult basic education, high school completion,
and GED programs had received public assistance in the year
before enrollment; about 11 percent of all new ESL students
met this criterion.
[5] This same disadvantaged group is likely to
be of increasing concern to welfare policymakers. Drastic
reductions in welfare caseloads since their peak in 1994 are
also changing the face of caseloads, which now increasingly
consist of “hard-to-serve” recipients. It is likely that many
of those left on the welfare rolls will lack an education
credential and will have poor reading or math skills.
[6]
Second,
the adult education programs studied and their effects on
students reflect the fact that these programs operate within
the context of welfare-to-work programs. Such programs provide
other services, such as counseling, child care, job search
assistance, and postsecondary education and training. Although
many issues facing adult educators are essentially unrelated
to the welfare status of the students they serve, the context
of available supports, expectations, and requirements is different
for those enrolled in adult education as part of a welfare-to-work
program. Thus, the measured effects of these programs reflect
not only the payoff to adult education but the effects of
a larger package of services and requirements that included
adult education. As part of such a package, the adult education
provided could be strengthened to produce greater effects
(for example, if students receive help with child care or
transportation). However, the effects of adult education also
could be weakened by other program components (for example,
if program rules limit the time that students may be enrolled
in adult education or if the program emphasizes a quick transition
from welfare to work).
Third,
for welfare recipients in our study, participation in adult
education was mandatory. While “traditional” adult education
students enroll on a voluntary basis and can therefore be
presumed to be motivated to learn, such motivation may sometimes
be lacking when students are compelled to participate by mandatory
welfare-to-work programs. Like most other adult education
students, those mandated to participate often have done poorly
in school in the past and may be alienated from traditional
educational institutions and modes of instruction. Unlike
the voluntary, or traditional, students, however, students
connected to a welfare-to-work program may initially be motivated
to attend classes less by the desire to learn or to obtain
a credential than by the need to comply with welfare-to-work
program requirements in order to avoid reductions in their
welfare grant.
|
Definition of Key Terms
Many of the terms used in
this report are not reintroduced and redefined in each chapter.
In some cases, these terms are ambiguous or have a somewhat different
meaning in the context of welfare-to-work programs than is common
in the adult education field. In this box, we introduce some of
the most commonly used terms in this report.
Throughout
the report, adult education refers to any or
all of the following:
-
Adult basic education
(ABE) classes. These provide reading and math instruction
to students whose achievement levels are lower than is required
for high school completion or GED classes¾typically
at the eighth grade level or lower.
-
General
Educational Development (GED) classes. These
prepare students to take the GED tests that are
used by states to award certificates signifying
knowledge of basic high school subjects (social
studies, literature, science, math, and writing).
Students entering GED classes usually are expected
to have language and math skills at a ninth grade
level or higher so that they can use the instructional
materials.
-
High
school completion classes. These replicate a
high school curriculum in an adult school setting.
Students usually must have language and math skills
at a ninth grade level or higher to enter a high
school completion program. When students finish
the course of study, they receive a high school
diploma.
-
English
as a Second Language (ESL) classes. These provide
instruction in how to read and write English to
people who are not fluent English speakers.
The
term welfare encompasses both Temporary Assistance
for Needy Families (TANF) and its predecessor, Aid to
Families with Dependent Children (AFDC).
Program
group members are sample members who were randomly
assigned to a welfare-to-work program.
Control
group members are sample members who were randomly
assigned to a control group, were excused from a welfare-to-work
program mandate, and were ineligible to receive most
program services. However, control group members could
seek out similar services in the community on their
own.
Program
participants are sample members who participated
in a particular program activity. This could be a class,
a training program, a job club, or something similar.
Program participation is not limited to program group
members assigned to welfare-to-work programs, because
control group members could access services outside
welfare-to-work programs on their own. Depending on
the context, program participants are sometimes referred
to as “students,” as “enrollees,” or as “recipients”
of education or training services.
|
Finally,
in the context of welfare-to-work programs, adult education
is viewed as an intermediate goal, not as an end in itself.
Financial self-sufficiency of adult students and their families
is the ultimate goal of these programs.
These
factors have helped shape the analyses and interpretation
of results in this report. However, one additional factor—one
of importance to the adult education community—cannot be taken
into account in this examination. The prevalence of learning
disabilities among welfare recipients is estimated to be between
25 and 50 percent.
[7] As implemented during the study period and
probably continuing into current operations, most programs
did not assess welfare recipients for learning disabilities,
which could affect the programs’ ability to address these
disabilities, clients’ skill development in the programs,
and clients’ subsequent labor market success.
C. Findings in Brief
Although
the five chapters following this one use various analytical
techniques and samples, taken as a whole they support the
following broad conclusions about adult education in the NEWWS
welfare-to-work programs serving those without a high school
diploma or GED:
- In
providing services for welfare recipients, adult education
programs generally did not adapt their curricula or teaching
methods to fit the specific needs of this group of students.
- Even
when welfare recipients preferred not to enter adult education,
welfare-to-work programs substantially increased their
receipt of such education. There was no evidence that
those who were mandated to participate (most of whom did
not express a preference for adult education) benefited
any less from their participation in terms of educational
attainment and literacy or math gains than those who volunteered.
- On
the whole, assignment to education-focused programs did
not appear to have a substantial payoff for the welfare
recipients in our study in terms of their education outcomes.
Although the programs increased GED receipt, most participants
did not earn a GED, and few experienced significant increases
in their reading and math skills. Three-year impacts on
earnings and welfare receipt in HCD programs were smaller
than those experienced by welfare recipients in LFA programs.
|
Combining Experimental and Nonexperimental
Research Methods
The
random assignment research design (described in the
box on page 2) was used for certain analyses in this
report, but others were conducted with nonexperimental
methods, which go beyond the random assignment research
design. In general, the distinction between these two
methodologies depends on whether a question concerns
effects of the programs as a whole, which the
random assignment research design is well equipped to
address, or whether a question concerns the effects
of program components or program outcomes,
such as participation in adult education, GED receipt,
or participation in postsecondary services. Because
of the protection offered by the random assignment research
design, findings about the programs as a whole (concerning
questions like “By how much did the HCD programs increase
participation in adult education?”) are more reliable
and can be presented with greater confidence. Findings
about program components or outcomes (concerning questions
like “By how much does a GED increase subsequent enrollment
in postsecondary education or training?”) are not protected
by random assignment and therefore have greater uncertainty
surrounding them.
In
addition to the difference in research methods, the
two types of questions outlined here differ in their
substantive focus. The “program” questions are less
precise than the “component” or “outcome” questions,
describing how assignment to a broad program affected
outcomes, not how specific events and services did.
For example, many of those assigned to welfare-to-work
programs did not participate in program activities,
or participated for short periods. This limits the extent
to which the program could affect sample members’ education
outcomes. No such limitations exist for analyses involving
specific program components or outcomes because participation
is explicit in the definition of the measure studied.
An estimate of the effects of GED receipt on earnings
in principle (and on average) applies fully to everyone
who received such a credential, as does an estimate
of the payoff to an additional month of adult education.
Thus, in summary, the experimental findings presented
are robust and reliable but apply to programs that do
not always reach participants as intended, whereas the
nonexperimental findings are less reliable analytically,
but answer more concrete policy questions facing the
adult education community.
|
- Gains
in reading skills appeared to vary with the length of time spent in
the adult education programs. Stays shorter than a year (which the
majority of participants in adult education had) did not improve reading
skills measurably, whereas longer stays were associated with substantial
gains, comparable - for this sampleto those associated
with regular high school attendance.
- Improvements
in math skills were associated with shorter spells of adult education.
After six months of adult education, most participants’ math skills
no longer improved.
- GED
receipt also was associated with shorter spells of participation in
adult education. Additional participation beyond six months did not
increase GED receipt, possibly because most GED recipients were close
to being able to pass a GED test when they entered the programs.
- Higher
average levels of teachers’ experience and education in the adult
education programs appeared to enhance the payoff to participation
in adult education in terms of reading and math skills.
- The
welfare recipients who were most likely to get GED certificates and
receive postsecondary services were those who had higher initial reading
and math skills when they entered the welfare-to-work programs.
-
As students earned GEDs, increased basic skills, or subsequently participated
in postsecondary programs, they appeared to have substantial benefits
in terms of employment, earnings, and self-sufficiency. However,
relatively few adult education participants received a GED, increased
their basic skills, or entered postsecondary programs.
- Receipt
of a GED credential was an important predictor of subsequent enrollment
in postsecondary programs. Participants in basic education programs
who went on to postsecondary education or training programs appeared
to experience substantial benefits from them in terms of increased
earnings and self-sufficiency.
II.
Summary of Each Chapter’s Findings
This
report on adult education in 11 welfare-to-work programs addresses many
different aspects of the adult education experience of welfare recipients
in welfare-to-work programs. Specifically, the report traces the steps
outlined in Figure 1.1, [8] which describes the hypothesized effects of adult
education in welfare-to-work programs.
In
Chapter 2 we focus on the second and third boxes in Figure
1.1, addressing the questions of how education-focused welfare-to-work
programs affect participation in adult education, what adult education
that is provided to welfare recipients looks like, and how these welfare-to-work
programs affect education outcomes. Specifically, program effects are
presented on participation in adult education, basic skills, and GED receipt
in three education-focused programs (in Atlanta, Grand Rapids, and Riverside).
We explore how these effects vary across different subgroups and attempt
conceptually to link effects on one outcome to effects on other outcomes.
In
Chapter 3 this analysis is taken a step further, focusing more directly
on the third box in Figure 1.1, describing the
relationship between participation in adult education and improvements
in the skill and educational attainment of participants. For example,
we examine how an extra month of participation affects key education outcomes
and how this effect varies with (1) total time spent in adult education,
(2) individual student characteristics, and (3) program and staff characteristics.
In
Chapter 4 we look at the fourth box and link education outcomes
to employment outcomes and self-sufficiency, addressing questions
like “What is the GED worth to welfare recipients?” and “How
are additional reading and math skills valued in the labor
market?” Chapter 4 uses survey and administrative data across
all 11 welfare-to-work programs in each of the seven NEWWS
sites.
Chapter
5 revisits the experimental comparisons (of randomly created
program and control groups), comparing all 11 programs for
their effects on earnings and welfare receipt and attempting
to isolate factors that made some programs more successful
than others.
Last,
Chapter 6 completes our analyses by focusing on the box to the right in
Figure 1.1, addressing the important intermediate step of postsecondary
education and training, which is often believed to be an important intermediary
link between participation in adult education and longer-term improvements
in earnings and other employment outcomes. The chapter addresses the questions:
“What determines whether adult education participants enter postsecondary
programs?” and “Do participants benefit from these programs?”
A. Characteristics of Adult Education in Welfare-to-Work
Programs and the Effects of Education-Focused Welfare-to-Work Programs
on Educational Attainment and Achievement (Chapter 2)
Site
visits and surveys of education providers in the HCD programs
in Atlanta, Grand Rapids, and Riverside found that these welfare-to-work
programs used a wide variety of educational institutions to
provide adult education to the welfare recipients enrolled
in these programs. Research conducted at the adult education
provider sites concluded that the inclusion of welfare-to-work
program participants in the adult education classes usually
did not greatly affect the providers’ operations, curricula,
and teaching methods. In other words, welfare-to-work program
participants took classes together with non-welfare recipient
adult education students and generally did not receive services
specially tailored to their needs from adult education providers
or classroom teachers. In some cases, hours were expanded
to enable welfare recipients to participate for 20 hours a
week, as required by welfare-to-work program regulations.
In other cases, additional counseling or job-readiness instruction
were added for welfare recipients. Only in Riverside did the
welfare-to-work program negotiate contracts with adult education
providers and use welfare-to-work funds to pay providers serving
their clients. However, aside from these contractual differences,
there were few systematic differences in the adult education
provided across the three sites.
The
welfare-to-work programs substantially increased participation
in adult education for those who entered the study without
a high school diploma or GED (the sample analyzed throughout
the report). Without the programs, about one-fifth of sample
members sought out adult education programs on their own (as
evidenced by two-year participation rates in the control group).
The three programs studied in Chapter 2 (Atlanta, Grand Rapids,
and Riverside) more than doubled this rate of participation:
one-half of program group members participated in adult education.
When they enrolled in adult education, program group members
also stayed longer. Across all sample members (including those
who did not participate at all), the program more than tripled
the average number of hours of adult education, from 68 for
control group members to 244 for program group members. This
means that the average participant in program-provided
adult education classes was enrolled for about 488 hours.
Thus, the programs induced more individuals to participate
in adult education, and those who participated did so for
more hours.
As
part of the analysis, impacts on adult education participation
were estimated for 20 different subgroups, defined using individual
characteristics measured at program entry (hereafter referred
to as “baseline characteristics”). [9] Examples of such subgroups include persons with
young children, those who dropped out of school having completed
eighth grade or less, those expressing a lack of desire to
go back to school, and those with personal or family barriers
to participation. Without exception, the three HCD programs
increased participation in adult education for each of these
subgroups. This shows that mandatory welfare-to-work programs
can increase welfare recipients’ exposure to adult education
even among welfare recipients with barriers to participation.
The
three programs achieved modest impacts on GED receipt during
a two-year follow-up period. Whereas only 4 percent of control
group members received a high school diploma or GED during
the follow-up period, 11 percent of program group members
received such a credential. This impact more than doubled
the proportion with an education credential; however, fewer
than one in five participants in adult education earned a
credential. (Many sample members might not be expected to
attain such a credential during the two-year follow-up period,
because they entered the programs with low achievement levels
or limited English skills.)
Moreover,
the three programs did not increase scores on standardized
reading and math tests, conducted as part of the two-year
follow-up interview. As discussed more extensively in Chapter
2, the combination of modest increases in GED receipt and
a lack of gains in measured literacy and math skills has been
found in several previous studies. There are several possible
explanations for this apparent discrepancy. First, it is possible
that the GED test and the basic literacy and math tests administered
in the survey do not capture the same underlying skills. In
that case, someone might be able to pass a GED test without
showing concomitant gains in basic reading and math skills.
It also is possible that difficulties with the administration
of the literacy and math tests reduced the statistical reliability
of our findings. (These tests were administered as part of
a long interview in sample members’ homes—not the ideal environment
in which to concentrate on a skills test.) However, it is
reasonable to conclude that, on average, participants experienced
limited benefits in terms of increased skills and credentials
from their participation in adult education. Subsequent analyses
presented in Chapter 3 and 4 further explore this issue.
Impacts
on GED receipt and educational achievement also were estimated
separately for 20 subgroups. Researchers found that impacts
on GED receipt were strongest for those entering the welfare-to-work
programs with already high reading and math skill levels.
Those entering with high reading scores experienced an impact
(that is, an increase relative to the control group) of 16
percentage points. Conversely, those entering with low reading
scores experienced an impact of only 3 percentage points.
Those entering the program having left school below the ninth
grade experienced no impact on GED receipt. Thus, it seems
that the programs’ effects on GED receipt were closely tied
to program participants’ entry-level skills. Those who needed
little basic education to earn a GED were much more likely
to be successful in this regard. This finding is also consistent
with prior research, involving both mandatory and voluntary
programs for school dropouts.
[10]
Interestingly,
motivation to participate in adult education programs did
not affect program success in terms of GED receipt. Even though
some sample members indicated that they did not want to attend
school, those who made this claim as they entered the programs
and were required to participate anyway experienced substantial
increases in GED receipt, just like sample members who did
express a preference for adult education at program entry.
This shows not only that welfare-to-work programs can induce
individuals to do things they might not otherwise do but also
that participants in adult education can be successful even
if they prefer not to participate.
B. Individual Efforts and School Effects: The Payoff
to Participation in Adult Education (Chapter 3)
Nonexperimental
analyses in the three HCD programs discussed above (Atlanta,
Grand Rapids, and Riverside) suggest that the amount of time
that participants spent in adult education classes affected
their educational attainment and achievement outcomes. However,
these relationships were not straightforward. During participants’
first year of participation in basic education, additional
months of participation were not associated with higher literacy
test scores. However, after a year of participation, additional
months in adult education appeared to substantially increase
test scores (an increase of .55 of a standard deviation for
six additional months of participation). [11] In this sample,
this effect was comparable to the differences in baseline
literacy scores associated with having attended an additional
month of high school. These findings suggest that a threshold
level of participation of approximately one year is needed
to achieve meaningful literacy gains lasting until the test
administered in the two-year follow-up survey. (There could
be a significant lag between the end of participation and
this survey, which could make it difficult to reliably identify
more modest gains in basic skills.)
For
measured math skills, this relationship looked markedly different.
Increases in math skills were associated with additional months
of basic education during the first six months only. After
that, no further increases in these skills were found. This
suggests a “plateau” rather than a “threshold” type of relationship.
Such a pattern could reflect limitations in the math skills
being taught in adult education classes.
The
relationship between time in adult education and GED receipt
followed a similar pattern. Additional months of participation
increased the likelihood of GED receipt during the first six
months of participation but not thereafter.
When
these analyses were conducted separately for program group
members and control group members, it appeared that both groups
experienced similar patterns of gains in literacy and math
skills. This is interesting, because participation for program
group members was mandatory, while control group members sought
out adult education services on their own initiative and participated
in them voluntarily. One might expect the payoff to the control
group to be greater, but no such difference was found. A difference
was found for GED receipt, but it was the opposite: program
group members were more likely than controls to receive a
GED as a result of additional participation in adult education.
When
these analyses were conducted for different groups of sample
members defined using baseline characteristics, little variation
in the estimated effects of additional education on literacy
and math skills was found. Sample members who faced greater
barriers to participation or who were less motivated to participate
in adult education benefited from additional education in
similar ways as sample members who did not face these barriers.
However, the analyses also found that the relationship between
adult education participation and GED receipt varied across
the subgroups. The increased probability of earning a GED
from short-term participation, as mentioned above, did not
hold up for sample members entering the program having completed
less than ninth grade. Conversely, it was found that, for
those with low initial skill levels, the likelihood of GED
receipt continued to increase with additional participation
beyond six months. This is unsurprising, because one might
expect longer-term participation in adult education to help
those who started out with low skills more.
Last,
we analyzed how the relationship between time spent in adult
education and education outcomes was affected by differences
in provider characteristics, including measures such as class
size, teachers’ experience and education levels, individual
attention, the strength of the link between the education
provider and the welfare-to-work program, and program exit
standards. The reliability of this analysis was limited, because
we were unable to study the links between individual teachers
and individual students, instead having to correlate student
outcomes with school-level averages of teacher characteristics.
However, within these limitations, teachers’ experience and
teachers’ education each appeared to enhance significantly
the payoff to additional time spent in education classes in
terms of reading and math skills. No provider characteristics
were identified that affected programs’ effectiveness in increasing
GED receipt. (However, provider data were limited, and the
same caveats about our ability to match students to specific
teachers applies here.)
C. Does the Low-Wage Labor Market Value Basic Education?
Effects of GED Receipt and Literacy Gains on the Self-Sufficiency of Welfare
Recipients (Chapter 4)
Having
analyzed how participation in adult education programs affects
literacy, math scores, and GED receipt, the next step (in
Chapter 4) is to analyze the extent to which improvements
in these education outcomes affect welfare recipients’ employment
outcomes and welfare receipt. If those effects are strong,
improving welfare-to-work programs’ ability to improve education
outcomes would also improve the programs’ effectiveness in
terms of employment and welfare receipt. However, this would
not be the case if these education outcomes had little effect
on sample members’ employment and welfare receipt. In that
case, a focus on immediate employment might be more appropriate.
The
analyses presented in Chapter 4 also contribute to our knowledge
about the value of the GED and the importance of basic skills
in the low-wage labor market. Much of the GED-related research
has focused on comparisons of GED holders and high school
graduates, using national data sets. Those findings may have
limited relevance for programmatic choices made on behalf
of welfare recipients, whose needs and experiences may be
different from those of other school dropouts.
GED
receipt appeared to substantially increase earnings. The estimated
effect on annual earnings in the third year of follow-up was
approximately $771. GED recipients also received fewer welfare
benefits (an estimated reduction of $331). These estimated
effects remained largely unchanged when measures of time spent
in adult education or of reading test scores were introduced
as control variables into the analysis. This suggests that
our estimates of the value of this credential reflect the
effects of the credential itself, not the underlying basic
skills or participation in adult education programs. It also
was found that earning a GED had stronger estimated effects
for program group members than for control group members.
This suggests that the other aspects of the welfare-to-work
programs (ranging from the program’s message to services like
job search assistance, skills training, and college programs)
further enhanced the value of this credential by increasing
participants’ ability to make use of their newly acquired
credential in the workplace.
Like
all analyses of educational attainment, analyses of GED receipt
are potentially affected by selection bias. Such bias occurs
when recipients of GED credentials are different from nonrecipients
in ways that are not controlled in the analysis. In Chapter
4, effects of GED receipt were estimated in various ways to
assess the sensitivity of the findings to selection bias and
other problems. In general, the different estimates were consistent
with one another, and there was no evidence that uncontrolled
differences in motivation or ability explained the apparent
effects of GED receipt on earnings and self-sufficiency. However,
the findings presented in Chapter 4 could not be confirmed
with an advanced statistical method (an “instrumental variables”
estimator) because statistical precision was lacking.
Analyses
of the effects of greater reading skills on employment outcomes
and self-sufficiency found those effects to be substantial.
An increase of one standard deviation in reading scores was
associated with $355 in additional earnings during the year
following the test (the third year of follow-up). This effect
was independent of (that is, in addition to) any effect from
earning a GED credential. (Introducing math skills separately
did not show an independent effect separate from that associated
with greater reading skills.) All this suggests that HCD programs
could have more substantial effects on the economic outcomes
of welfare recipients if these programs managed to improve
their effects on mediating education outcomes. Our analyses
suggest that increased retention might be one way to achieve
this. Research on “best practices” in adult education for
welfare recipients has suggested that programs’ ability to
retain students and improve their skills is affected by many
program characteristics, including: [12]
-
developing
a well-defined mission,
-
providing
specially targeted classes to students who are welfare
recipients,
-
having
skilled, experienced teachers,
-
emphasizing
staff development,
-
adopting
varied instructional approaches, including small group
and computer activities,
-
communicating
frequently with welfare-to-work program staff,
-
stressing
regular attendance,
-
aggressively
following-up on absences,
-
adopting
relatively intensive class schedules, and
-
promoting
a high degree of teacher-student and student-student interaction.
Finally, a greater
emphasis on identifying and addressing learning disabilities,
which now remain largely undiagnosed, could greatly improve
programs’ ability to serve their students successfully.
D. Beyond Basic Education: The Benefits of Skills Training
and College (Chapter 6)
In
nonexperimental analyses, participation in postsecondary programs
was found to have substantial benefits in terms of greater
earnings and lower welfare receipt. These benefits did not
appear until after sample members completed their education
and training. Effects for postsecondary participants appeared
in the third year following their initial adult education
spell. In that year, their earnings were $1,542 (or 47 percent)
higher than those of sample members who received only adult
education, while their welfare benefits were $919 (or 32 percent)
lower. These estimated effects were not contingent on participants
completing their spell of postsecondary education or training
with a credential or certificate.
III.
Conclusions and Implications for Policy
A.
The Challenge of Making Adult Education Work for Welfare Recipients
This
study of adult education for welfare recipients who do not
have a high school credential uncovered several different
patterns of effects. Assignment to a Human Capital Development
program had substantial impacts on these welfare recipients’
participation in adult education, modest impacts on their
GED receipt, and no impacts on measured literacy and math
skills. The study also found that, within a three-year follow-up
period, the effects of HCD programs on earnings and welfare
receipt were positive but limited, especially compared with
labor force attachment programs.
A
second pattern of findings concerns the dynamics of participation,
learning, graduation, and skills acquisition that underlie
the experimental impact story. The exploration of this pattern
begins (in Chapter 3) with a discussion of one of the key
questions underlying education and learning: What is the value
of additional instruction? In addressing this question for
three education-focused programs in Atlanta, Grand Rapids,
and Riverside, Chapter 3 uncovers that additional months in
school (our only reliable measure of additional instruction)
matter; they increase literacy, math skills, and GED receipt.
However, long-term participation (longer than a year) is necessary
to achieve a measurable payoff in increased literacy skills,
while increases in math skills and GED receipt seem to be
limited to the first six months of instruction.
At
the same time, the average program group member in these three
programs received only about 244 hours (or about twelve 20-hour
weeks) of adult education, which is substantially less than
one year of high school—insufficient to make up the education
deficit with which most of these welfare recipients entered
the programs. This could explain why only 11 percent of program
group members earned a GED during the follow-up period and
why fewer than 15 percent of participants in adult education
went on to skills training and college programs.
Next,
using nonexperimental methods, the report shows that for those who did
reach these milestones, participating in adult education was beneficial.
Chapters 4 and 6 show that payoffs from GED receipt, increased literacy
skills, and postsecondary education and training were substantial. Thus,
in summary, the analyses presented in this report confirm the internal
logic underlying the human capital program model (as outlined earlier
in Figure 1.1). However, too few program group
members made it through the different steps to experience the anticipated
payoffs at the end. Assuming that other program participants would experience
comparable benefits, the challenge facing welfare-to-work program administrators
and adult education providers is to find a way to retain more students
long enough, to help them reach the immediate goal of earning a GED, and
to help them access postsecondary services that allow them to capitalize
on this credential. This could occur while people are receiving welfare
benefits or after they have left the welfare rolls. However, there is
no guarantee that simply increasing the duration of participation is sufficient
to reach educational goals. The quality of instruction, the appropriateness
of the material and technology for participants with low skills and possibly
with learning disabilities, and the sometimes limited motivation of program
participants are likely important as well, although our study did not
examine these factors.
B. Policy Implications
The
analyses presented in this report show that the long-term payoffs of an
education-focused approach for welfare recipients without a high school
diploma or GED can be substantial. However, the report also shows that
it is a challenge to improve the basic skills and educational attainment
of welfare recipients, even in programs that are directly focused on education
outcomes. For those entering with low skills and lacking years of high
school, several years of basic education and postsecondary education may
be needed to promote long-term success and self-sufficiency. In the current
welfare environment, such a long-term commitment carries some risk inasmuch
as long-term participation may exhaust welfare recipients’ limited allotted
time on welfare.
-
For
welfare recipients who are within easy reach of earning
a GED, pursuit of such a credential is a good program
option that produces substantial benefits, increasing
welfare recipients’ earnings and their access to postsecondary
education or training.
Success
is easier to achieve for those who are close to passing the GED test.
Our findings suggest that this credential is a worthwhile short-term program
goal, especially if it is combined with a targeted postsecondary activity.
Together, a GED and skills training greatly increased earnings and reduced
welfare receipt in the third year of follow-up. Education-focused welfare-to-work
programs may be most successful when they can combine GED preparation
and postsecondary services in a relatively short and intensive program
(an option that, however, will not work with the most educationally disadvantaged).
-
Too
few adult education students and GED recipients continue
on to postsecondary education or training. Links between
adult education programs and postsecondary programs could
be strengthened, and adult education students should be
made aware of the limitations of having just a GED credential
as a way to improve one’s employment outcomes.
Although
GED receipt and increased basic skills appear to have positive effects
on the earnings of welfare recipients, those effects appear to be much
stronger when spells of adult education and receipt of a GED are followed
by enrollment in postsecondary education or training programs. The orientation
of many of these programs toward specific jobs and career opportunities
may be a factor in explaining these programs’ apparent benefits. Participation
in postsecondary education or training carries a price in terms of lower
short-term earnings, but the longer-term effects of these services are
substantial. Especially after a long spell of participation in adult education,
it makes sense to cap off this investment with some college or vocational
skills training.
-
The
analyses presented in this report address important questions
about adult education provided to welfare recipients,
but many important questions remain, especially regarding
the quality of participation in adult education, the appropriateness
of adult education services provided to welfare recipients,
and the possible benefits of education and training for
those who do have an education credential when they enter
the programs.
In
this report, we present program effects on participation in
adult education and key educational outcomes. We relate these
effects to one another and to the employment and welfare outcomes
of those being targeted by the programs. However, although
we assessed the effects of participation in adult education,
we did not capture all the reasons for nonparticipation.
Although we measured the effects of earning a GED, we do not
know why so many participants never received this credential.
Answers to both of these questions, and others like it, may
provide a greater understanding of the “quality” of participation
(that is, the actual commitment to learning manifested by
students who were coerced to participate in adult education)
and by the quality of the instruction (that is, the appropriateness
of teaching materials and techniques for the welfare recipients
in these programs). The study data did not capture either
one of these “quality” measures accurately, and it is therefore
difficult to say whether simply increasing enrollment in adult
education programs beyond current levels would significantly
improve the outcomes of these programs. More detailed data
about the quality of students’ program experience must be
collected to address these questions, and more systematic
comparisons of the different types of adult education programs
that serve welfare recipients are needed.
The
analyses presented in this report are limited to welfare recipients
who did not have a high school diploma or GED and were considered
to be in need of basic education. No parallel study was conducted
to examine the effects of education and training on welfare
recipients who did have an education credential. Our findings
do not generalize to this group, and additional research may
be needed to assess whether and how additional education and
training benefits welfare recipients who are less disadvantaged
academically.
NOTES:
|
