Review of Educational Research

Summer 1987, Vol. 57, No. 2, pp. 101-121

High School Dropouts:

A Review of Issues and Evidence
Russell W. Rumberger
University of California, Santa Barbara
The problem of high school dropouts has generated increased interest among
researchers, policymakers, and educators in recent years. This paper examines the
many issues involved in trying to understand and solve this complex social and
educational problem. The issues are grouped into four areas covering the incidence,
causes, consequences, and solutions to the problem. Within each area, the discussion
identifies the important issues involved, the current state of research on the issues,
and considerations for future research.

Dropping out of high school has long been viewed as a serious educational and
social problem. By leaving high school prior to completion, most dropouts have
serious educational deficiencies that severely limit their economic and social well-
being throughout their adult lives. The individual consequences lead to social costs
of billions of dollars.
Over the last 40 years, the proportion of young people who have failed to finish
high school has decreased substantially. In 1940, more than 60% of all persons 25
to 29 years old had not completed high school; by 1980, that proportion had
dropped to less than 16% (U.S. Bureau of the Census, 1985, Table 215).
Despite these long-term declines in dropout rates, interest in the dropout issue
among educators, policymakers, and researchers has increased substantially in
recent years. State and local education officials are currently devoting more time
and resources to measuring the extent of the problem, to examining its causes, and
setting up programs for dropout prevention and recovery. Policymakers are pro-
moting and supporting these efforts and passing legislation to fund them. More
research has appeared on the problem of dropouts in the last 2 years than in
perhaps the previous 15.
If the long-term incidence of dropping out is declining, why has the concern for
this problem increased of late? Several explanations could account for this increased
concern. First, although the long-term trend of dropping out has declined, the
short-term trend has remained steady and even increased, especially for some
groups. The proportion of white male dropouts, for example, increased from 14%
to 17% between 1968 and 1978, and then declined to 16% in 1984 (Table 1).
A second reason for the increased attention to dropouts is that minority popu-
lations, who have always had higher dropout rates than the white population, are
increasing in public schools. In 1982, racial and ethnic minorities represented the

An earlier version of this paper appeared in California Public Schools Forum, Fall 1986.1
would like to thank Jim Block, Edwin Bridges, Henry Levin, G. Alfred Hess, Michael Olivas,
four anonymous reviewers, and the editor for their helpful comments.

101
Russell W. Rumberger
TABLE 1
Dropout Rates by age, sex, race, and ethnicity: Selected years, 1968-1984 (percentages)
Cohort 1968 1978 1980 1982 1984
3 to 34-year-olds 18.3 12.9 12.7 12.7 12.6
white males 17.1 12.2 12.2 12.4 12.5
white females 17.3 12.4 11.9 11.9 11.7
black males 25.8 17.2 16.5 16.7 15.7
black females 25.6 16.2 16.2 14.9 15.0
Hispanic males — 28.1 28.3 26.9 27.0
Hispanic females — 29.0 27.3 27.3 26.7
18 to 19-year-olds 15.7 16.7 15.7 16.7 15.2
white males 14.3 16.3 16.1 16.6 15.8
white females 14.6 15.0 13.8 14.9 14.0
black males 23.8 25.8 22.7 26.4 19.7
black females 24.7 22.8 19.8 18.1 14.5
Hispanic males — 36.6 43.1 34.9 26.2
Hispanic females — 39.6 34.6 31.1 26.0
16 to 17-year-olds 7.8 8.8 8.8 7.3 6.8
white males 6.9 9.6 9.3 7.3 7.3
white females 7.6 8.7 9.2 8.0 6.9
black males 10.1 5.2 7.2 6.4 5.5
black females 14.2 9.4 6.6 5.5 4.9
Hispanic males — 15.6 18.1 12.2 13.6
Hispanic females — 12.2 15.0 15.9 12.7
Note. Dropout rates represent the percent of each cohort who are dropouts. Dropouts are
defined as persons of a given cohort who are not enrolled in school in October of the year
in question and have not received a high school diploma or an equivalent high school
certificate. Source: U.S. Department of the Census, School Enrollment, Current Population
Reports, Series P-20, various issues (Washington, D.C.: U.S. Government Printing Office,
various years).

majority of students enrolled in most large U.S. cities and more than 90% of all
students in such cities as Newark, Atlanta, and San Antonio (Plisko & Stern, 1985,
Table 1.5). And the proportions will increase in the future; this alone could drive
up high school dropout rates.
A third reason for increased concern about dropouts is that many states have
recently passed legislation to raise academic course requirements for high school
graduation. Although increasing academic demands could help to motivate some
students, others—especially those who already have a tenuous commitment to
school—might be more inclined to drop out (McDill, Natriello, & Pallas, 1985,
1986). Major efforts will be required to prevent more students in this high-risk
population from dropping out, efforts that have often not been included in the
initial round of educational reforms (Levin, 1986).
A fourth reason for the increased concern about dropouts is a widespread belief
that the educational requirements of work will increase in the future. Most of the
recent state reform efforts and the national reports on education that prompted
them have been predicated on a belief that the increased use of new technologies
and structural changes in the composition of jobs in the economy will require more

102
 High School Dropouts
educational skills (e.g., National Commission on Excellence in Education, 1983;
Task Force on Education for Economic Growth, 1983). While these visions of the
future have yet to be substantiated (Levin & Rumberger, in press), they do suggest
that dropouts will be even more disadvantaged in the future job market than they
have been in the past (National Academy of Sciences, 1984).
A final reason for increased concern about dropouts is political. Leading educa-
tion officials at the state and federal levels have recently initiated the idea of using
a series of "indicators" to judge the performance of the nation's and states' school
systems. These efforts began with the U.S. Department of Education's publishing
a series of education "wall charts" that compared state systems of education along
a series of dimensions, including the high school completion rates. Now the states
themselves, through the efforts of the Council of Chief State School Officers and
the U.S. Department of Education, are beginning to define and collect comparable
data that can be used for such comparisons. This requirement has prompted
increased attention not only to the problem of how to define and measure dropouts
(Wittebols, 1986) but also how to reduce the incidence of dropping out and improve
this indicator of school performance.
Whatever the reasons for this heightened concern, the dropout issue is likely to
command increased attention from researchers, policymakers, and educators at the
local, state, and national levels for some time to come. The purpose of this paper
is to begin to examine the many issues involved in trying to understand and solve
this complex social and educational problem.
In the remainder of the paper, I will briefly analyze four major facets of this
problem: its (a) incidence; (b) causes, including social, economic, psychological,
and educational factors; (c) individual and social consequences; and (d) solutions.
For each facet, I will identify the important issues involved, the current state of
research on the issues, and considerations for future research. I will also attempt to
identify the various academic disciplines—such as psychology, sociology, and
economics—that are useful in understanding the nature of this problem and its
solutions.
The Incidence of the Problem
Before this or any perceived social problem can be fully understood, it is
important to know the magnitude or incidence of the problem. What is the
incidence of dropping out of high school? Is the incidence increasing or decreasing?
How do the incidence of dropping out and trends in the dropout rate vary among
different social groups and different educational settings?
The Incidence of Dropping Out
The first question is the most important in understanding the nature of this
problem, yet it is also the most difficult to answer. In fact, no one knows what the
high school dropout rate really is in the United States. That is because there is no
consensus definition of a high school dropout, nor is there a standard method for
computing the dropout rate. At the national level, the two most widely cited
dropout statistics—the dropout rate computed from U.S. Census data and the high
school attrition rate computed from state-level school enrollment data—show
widely different dropout rates and probably represent lower and upper limits to the
true rate.

103
Russell W. Rumberger
The U.S. Census Bureau computes the dropout rate as the proportion of a given
age cohort that is not enrolled in school and has not completed high school. The
latest figures available are for October 1984. They show a dropout rate of 6.8% for
persons 16 and 17 years old and a dropout rate of 15.2% for persons 18 and 19
years old (Table 1).
The other widely cited national dropout statistic is based on attrition data. It
shows the proportion of a given entering high school class, usually the ninth grade,
that graduates 4 years later. The latest figures show an average attrition rate of
29.1% for high school class of 1984 in the U.S., with state-level attrition rates
varying from a low of 10.7% in Minnesota to a high of 43.3% in Louisiana (Table
2).
One reason these two statistics are so different is that they were designed to
answer different questions about dropouts. The Census statistic is designed to

TABLE 2
Attrition rates by state (percentages)
State 1972 1982 1984 State 1972 1982 1984
U.S. average 22.8 27.2 29.1 Missouri 22.5 24.6 33.8
Alabama 34.6 32.9 37.9 Montana 21.0 17.8 17.9
Alaska 20.7 29.0 25.3 Nebraska 14.1 16.4 13.7
Arizona 26.2 27.6 35.4 Nevada 25.0 24.7 33.5
Arkansas 31.1 25.3 24.8 New Hampshire 19.3 21.7 25.8
California 20.1 31.1 36.8 New Jersey 20.3 21.9 22.3
Colorado 15.2 23.7 24.6 New Mexico 23.1 28.4 29.0
Connecticut 16.6 28.8 20.9 New York 25.3 33.7 37.8
Delaware 22.0 18.2 28.9 North Carolina 31.4 31.6 30.7
District of Columbia 45.2 44.2 44.8 North Dakota 11.0 12.7 13.7
Florida 27.9 34.6 37.8 Ohio 19.7 22.5 20.0
Georgia 35.2 36.9 Oklahoma 20.7 22.4 26.9
Hawaii 10.9 15.8 26.8 Oregon 20.8 28.3 26.1
Idaho 15.3 23.1 24.2 Pennsylvania 15.0 21.2 22.8
Illinois 22.0 25.2 25.5 Rhode Island 18.9 27.1 31.3
Indiana 23.9 23.1 23.0 South Carolina 30.8 35.7 35.5
Iowa 10.5 14.2 14.0 South Dakota 9.5 16.1 14.5
Kansas 17.2 19.1 18.3 Tennessee 27.6 31.1 29.5
Kentucky 29.6 33.1 31.6 Texas 29.8 31.8 35.4
Louisiana 33.5 36.0 43.3 Utah 16.7 18.6 21.3
Maine 19.1 27.9 22.8 Vermont 29.1 22.3 16.9
Maryland 19.8 24.4 22.2 Virginia 23.6 25.0 25.3
Massachusetts 22.1 24.1 25.7 Washington 16.1 23.1 24.9
Michigan 19.0 27.3 27.8 West Virginia 28.1 25.2 26.9
Minnesota 8.5 10.8 10.7 Wisconsin 10.9 16.5 15.5
Mississippi 42.4 37.0 37.6 Wyoming 16.9 21.7 24.0
Note. Attrition rates were calculated by subtracting graduation rates from 100%. Graduation
rates were calculated by dividing the number of public high school graduates by the public
9th-grade enrollment four years earlier. 1984 data were adjusted for migration and unclas-
sified students. Source: U.S. Department of Education, State Education Statistics (Wash-
ington, DC: U.S. Department of Education, January 1984 and January 1986.

104
 High School Dropouts
determine the number and proportion of persons from a given demographic cohort
who are dropouts. State attrition data, which are constructed from enrollment and
graduation information, are designed to reveal how well the educational system is
doing in graduating students. But beyond differences in purpose, differences in
these statistics can be traced to six factors that must be considered in computing
any dropout rate: (a) choice of cohort, (b) initial membership in cohort, (c)
definition of dropout, (d) time for determining dropout status, (e) source of
information, and (f) level of determination.
The first factor is the choice of cohort to be used as a base for determining the
dropout rate. The Census figure is based on an age cohort, while the education
figure is based on a class cohort. Neither one is necessarily superior to the other,
although the latter creates more problems because it requires measuring a cohort's
status over an interval of time—first as a member in the cohort and then as a
dropout—while the former can be determined at one point in time. In either case
it is necessary to determine the particular age or class cohort to use in computing
a dropout rate. One problem with using the entering ninth grade class is that it
ignores those students who have dropped out in earlier grades.
The second factor is determining a person's initial status as a member of the
cohort. It is not difficult to determine one's status in an age cohort, but it is in a
class cohort. In the latter case it is necessary to determine which students should
be included as members of the cohort, such as students in special education or in
juvenile and mental institutions, as well as when membership in the cohort is
determined (Wittebols, 1986). At the school or district level this may be particularly
problematic because some students shop around for schools or are transients (such
as the children of migrant workers), which means they could be "enrolled" in more
than one school or district at any one time (California State Department of
Education, 1986, pp. 26-27). Students who are counted twice as members of a
given class cohort (including students held back a grade) would overstate enrollment
figures and therefore overstate the attrition rate.
The third, and most critical, factor in measuring dropout rates is determining
whether or not a person is a high school dropout. Currently, local and state
education agencies employ widely different methods for defining dropouts (Witte-
bols, 1986). "Dropout" is generally defined as a residual status, indicating someone
who has not graduated from, or is not currently enrolled in, a full-time, state-
approved education program. That is, most states recognize only persons who are
studying for, or who have received, regular high school diplomas. Persons who have
completed the General Educational Development (GED) examination or receive a
high school equivalency certificate from the state are often considered dropouts. So
too are persons still enrolled in school at the time dropout status is determined
(usually 4 years after entering the ninth grade).
The Census Bureau uses a much narrower definition of dropouts. Persons who
have a regular or equivalent high school certificate or who are still attending school
are not considered dropouts. This method also raises questions. First, is an alter-
native high school certificate equivalent to a regular high school diploma? Some
recent research suggests that GED holders do not do as well as regular high school
graduates in the labor market or further educational opportunities, which may
indicate that the two credentials are not equivalent (Fields, 1986; Quinn & Haber-
man, 1986).

105
Russell W. Rumberger
Second, how should subsequent enrollment status be determined? Some students
who are officially enrolled in school may not be attending regularly and at some
point should be considered dropouts rather than students. A more difficult problem
concerns students who transfer from one institution to another. Most states require
some official verification that a student has enrolled in another state-approved
institution, such as an official request for the student's transcript received within a
specified number of days (Wittebols, 1986). Surveys of individual school districts
show that many school leavers cannot be easily traced, and even those who reenter
may not have their transcripts requested (California State Department of Education,
1986; Hammack, 1986). Thus, this procedure not only could overstate the number
of dropouts but also puts schools and districts with large proportions of school
leavers and transient students at a disadvantage, in that they would have a greater
proportion of students to follow up.
A fourth factor that must be considered in computing any dropout rate is the
time interval for determining dropout status. Currently, state-level attrition data
are based on graduation rates of a ninth-grade class 4 years later. Dropouts include
not only persons still enrolled or holding high school certificates at that time but
also persons who may receive a regular or equivalent high school diploma at a later
point in time.
A recent study of 1980 high school sophomores who later dropped out of school
shows that 38% had received a regular or equivalent high school diploma by 1984
(Kolstad & Owings, 1986, Table 1). A 1985 survey found that almost 50% of young
Americans who did not complete the 12th grade had studied for the GED, and
40% of those had received it (Kirsch & Jungeblut, 1986, Figure 3). The current
30% attrition rate in California could be reduced to 20% if adjustments were made
for those persons who are likely to receive regular or equivalent high school
diplomas by the time they are 30 years of age (California State Department of
Education, 1986, pp. 30-31).
The issue of determining an appropriate time interval may become more
important in the future if more students opt to finish high school at a later time or
through alternative means. This, in fact, could be happening as more students
become aware of, and have, the opportunity to finish their high school diplomas at
community colleges, at night school programs, or through other means.
A fifth factor that must be considered in computing any dropout rate is the
source of information on enrollment and graduation status. The Census Bureau
uses population data collected in October of each year as part of the ongoing
Current Population Survey. Census enumerators obtain enrollment information
about each member of the sample household. This information, which can come
secondhand, could be inaccurate because parents, for example, might not know
the actual enrollment or graduation status of their children, or they may not want
to reveal it.
Dropout rates computed with official education data at a state or local level are
also subject to error because of current differences in measuring and collecting
dropout data. Some of these differences may be overcome by current efforts to
define and collect comparable data by states and localities. But it appears unlikely
that all these differences will be overcome because of the inherent difficulties in
collecting and sharing information among schools, districts, and state education
agencies over an extended period of time.

106
 High School Dropouts
The final factor to consider in measuring dropout rates is the level of determi-
nation. Currently, Census data are used to compute a dropout rate at a national
and regional level. Attrition rates are currently computed at a state level and
averaged across states to compute a nationalfigure.Dropout rates are also computed
for schools and districts in some states, with more likely to be computed in the
future.
Dropout statistics, computed at any level of the educational system, are clearly
appropriate for measuring the number of students who are failing to finish high
school and for measuring, in part, how well the educational system is performing.
However, the existence of these statistics, even if they were comparably defined
and measured, inevitably leads to inappropriate comparisons between schools,
districts, and states. Such comparisons are inappropriate because school systems
enroll different types of students who have very different propensities to drop out
of school. One remedy to this problem is to produce "adjusted" or disaggregated
statistics that reflect differences in the kinds of students attending different school
systems, although appropriate procedures have yet to be established (Toles, Schulz,
& Rice, 1986).
This discussion has pointed out the numerous factors that must be considered
in deriving suitable and accurate dropout statistics. Additional research can help to
illuminate these issues further, but ultimately policymakers and educators must
decide on an appropriate definition of dropping out before accurate data can be
collected. Such efforts are currently under way through cooperative work between
the U.S. Department of Education and the Council of Chief State School Officers
(Wittebols, 1986). But even with such cooperation, differences in official dropout
rates are likely to continue because different dropout measures are used to address
different questions and are computed by different organizations.

Trends in Dropout Rates

A second major question about the incidence of dropping out concerns trends:
Is the dropout rate getting better or worse? Since there is no consensus on an
appropriate dropout measure, it is difficult to assess trends in the dropout rate.
Census data, which have been computed on a yearly basis for more than 20
years, show no long-term increase in the aggregate dropout rate, only short-term
increases. The dropout rate for persons 16 and 17 years old increased from 7.8%
in 1968 to 8.8% in 1980 and then declined to 6.8% in 1984. The dropout rate for
persons 18 and 19 years old increased from 15.7% to 16.7% between 1968 and
1982, and then dropped to 15.2% in 1984 (Table 1).
State-level attrition data show that average dropout rates for the nation increased
from 22.8% in 1972 to 27.2% in 1982 and then remained steady between 1982
and 1984 (Table 2). These same data show that dropout rates increased in most
states between 1972 and 1984, but decreased quite substantially in some states,
such as Arkansas and Vermont.
Some of the observed increases in dropout rates could have come about from
changes in the school participation patterns. California enrollment data show that
the proportion of 9th-, 10th-, and llth-grade students failing to enroll in school
the following year actually went down between 1977 and 1983, while the proportion

107
Russell W. Rumberger
of l2th-grade students who failed to graduate went up dramatically (California
Assembly Office of Research, 1985, Figure 4). This shift could have come about
from an increasing number of high school seniors deciding to finish high school at
a later time, such as through night school programs or through community colleges.

Variations Among Social Groups and School Systems

A third major question about the incidence of dropping out concerns variations
among social groups and school systems. It is well-known that dropout rates vary
widely among social groups. Dropout rates are higher for members of racial, ethnic,
and language minorities, for men, and for persons from lower socioeconomic status.
A national survey of 1980 high school sophomores shows an overall dropout rate
of 13.6%. But dropout rates computed with these data varied from 12.2% for
whites to 18.7% for Hispanics, from 12.6% for women to 14.6% for men, and from
8.9% for students from the highest socioeconomic levels to 22.3% for students
from the lowest socioeconomic levels (Kolstad & Owings, 1986, Tables 1, 2, and
3). Census data show dropout rates for 18- and 19-year-olds varying from 14.0%
for white females to 26.2% for Hispanic males (Table 1). Dropout rates are also
higher for American Indians and particular Hispanic subgroups, particularly Cu-
bans and Mexican Americans (Brown, Rosen, Hill, & Olivas, 1980, Table 2.31;
LaFromboise & Rudes, 1983, Table 12.1).
Trends in dropout rates based on Census data also show variations among social
groups. Dropout rates for persons 18 and 19 years old declined from 23.8% in 1968
to 19.7% in 1984 for black males, from 24.7% to 14.5% for black females, and
from 14.6% to 14.0% for white females. For white males, the dropout rate increased
from 14.3% in 1968 to 17.9% in 1981, although it declined to 15.8% in 1984.
Dropout rates for Hispanics show little change over the last decade, except for
Hispanic males 18 to 19 years old, for whom the rate declined from 35.5% in 1972
to 26.2% in 1984 (Table 1).
These data suggest that some racial and ethnic minorities, particularly blacks,
who traditionally have always had much higher dropout rates than whites, have
shown marked improvements in their propensity to finish high school. The increas-
ing dropout rates for white males could be related to changing patterns of school
participation mentioned earlier and therefore may not indicate that the problem is
actually getting worse for this group.
Dropout rates vary widely among school systems as well as social groups. Not
only are there widespread variations in dropout rates among state educational
systems, as the earlier data clearly show, but there are also widespread variations
among school districts and even among schools within the same district.
In Chicago, for example, dropout rates for the 63 Chicago high schools range
from 10% to 62% (Toles, Schulz, & Rice, 1986, Table 1). Of course, some of these
differences are due to differences in the populations of students enrolled in the
educational systems. But even after controlling for differences in student popula-
tions, widespread differences among school systems remain, differences that more
clearly illustrate the ability of schools to educate and graduate their students. In the
Chicago schools, observed dropout rates were 50% higher to 50% lower than the
rate expected given the composition of the students in the schools (Toles, Schulz,
& Rice, Table 1).

108
 High School Dropouts

The Causes of the Problem

No one really knows what causes students to drop out of high school. Dropouts
themselves report a number of different reasons for leaving school, with marked
differences reported by different social groups (Table 3). Almost one half of all
dropouts and more than half of white and black males cite school-related reasons
for leaving school, such as disliking school or being expelled or suspended. Twenty
percent of all dropouts, but almost 40% of Hispanic males, cite economic reasons
for leaving school. A third of all female dropouts report personal reasons for leaving
school, such as pregnancy or marriage. While these reasons suggest that students
drop out of school for a variety of reasons, they are unable to reveal the underlying,
causal factors that lead to dropping out. Many reasons, such as getting a job or
getting pregnant, should probably be considered symptoms rather than causes of
dropping out.

Factors Associated With Dropping Out

A large body of empirical research has identified a wide range of factors that are
associated with dropping out. The factors can be grouped into several major
categories: demographic, family-related, peer, school-related, economic, and indi-
vidual. Within each of these categories there can be a large number of specific
factors. Some are well-known and widely documented in numerous studies; others
have not been well-explored in relation to this particular problem. Some of these
factors can be manipulated through policy interventions within and outside of the
schools; others cannot.
The demographic factors associated with dropping out are well-known and were

TABLE 3
Primary reason high school dropouts left school, by sex, race, and ethnicity: 1979
(percentage distribution)
Males Females
Reason Total
White Black Hispanic White Black Hispanic
School-related:
Poor performance 9 9 4 5 5 4 7
Disliked school 36 29 26 27 18 15 29
Expelled or suspended 9 18 6 2 5 1 7
School too dangerous 1 0 0 2 1 1 1
Economic:
Desired to work 15 12 16 5 4 7 10
Financial difficulties 3 7 9 3 3 9 4
Home responsibilities 4 4 13 6 8 8 6
Personal:
Pregnancy 0 0 0 14 41 15 17
Marriage 3 0 3 17 4 15 9
Other 20 21 23 19 11 25 19
TOTAL 100 100 100 100 100 100 100
Note. Data are for persons 14 to 21 years of age. Source: "Dropping out of high school: The
influence of race, sex, and family background" by R. W. Rumberger, 1983, American
Educational Research Journal, 20, p. 201.

109
Russell W. Rumberger
cited earlier. Members of racial and ethnic minorities are much more likely to drop
out of school than white, Anglo students, and males are somewhat more likely to
drop out of school than females.
A large number of factors associated with family background and structure have
also been identified in the research literature. Perhaps the most important is
socioeconomic status. Numerous studies have found that dropout rates are higher
for students from families of low socioeconomic status, no matter what particular
factors are used to measure socioeconomic status (e.g., Kolstad & Owings, 1986;
Rumberger, 1983a). Particular family-related factors associated with dropping out
include low educational and occupational attainment levels of parents, low family
income, speaking a language other than English in the home, single-parent families,
and the absence of learning materials and opportunities in the home (e.g., Ekstrom
et al., 1986; Rumberger, 1983a; Steinberg, Blinde, & Chan, 1984).
The influence of peers has not received much attention in previous research on
this problem, although it is a subject of considerable interest in other areas of
educational achievement (Bridge, Judd, & Moock, 1979, chapter 8). Many dropouts
have friends who are also dropouts, just as friends' educational aspirations and
expectations are related more generally. But it is not clear to what extent and in
what ways a student's friends and peers influence the decision to leave school.
School-related factors associated with dropping out have received considerable
attention, particularly because many of these factors are ones that can be manipu-
lated through practice and policy. It is fairly well-documented that poor academic
achievement in school, as measured by grades, test scores, and grade retention, is
associated with dropping out (Borus & Carpenter, 1984; Ekstrom et al., 1986;
Wehlage & Rutter, 1986). It is also known that behavioral problems in school are
also associated with dropping out, including absenteeism, truancy, and discipline
problems (Bachman, Green, & Wirtanen, 1971; Wehlage & Rutter).
Most research on school-related factors has focused on students' behaviors and
performance in school. Little attention has been given to the influences of schools
themselves—their organization, leadership, teachers—on students' decisions to
drop out. Yet many dropouts attend schools with very poor facilities and inadequate
teaching staffs, conditions that could affect their performance in school and
ultimately their decision to leave (Fine, 1986). School-level dropout rates vary
widely even controlling for differences in student populations; this further suggests
that school-related factors exert a powerful influence on students' decisions to leave
school (Toles, Schulz, & Rice, 1986).
Economic factors also influence students' decisions to leave school. About 20%
of dropouts report that they left school because they wanted to or felt they had to
work to help out their families (Table 3). But as with other reasons students report,
it is unclear to what extent work is initially seen as a desirable or necessary
alternative to school before a student drops out or whether students first decide to
leave school and then decide to find a job.
Finally, there are a host of individual factors associated with dropping out.
Dropouts have lower levels of self-esteem and less sense of control over their lives
than other students. They have poor attitudes about school and low educational
and occupational aspirations (Ekstrom et al., 1986; Wehlage & Rutter, 1986).
Finally, many dropouts report that they leave school to get married or because they
are pregnant (Ekstrom et al., 1986; Rumberger, 1983a).

110
 High School Dropouts
Developing a More Comprehensive Model
While previous research on the causes of dropping out has been helpful in
identifying the wide range of factors associated with this behavior, the empirical
literature is still lacking. Many studies have focused on only a few of the many
factors known to be associated with this problem, and many are based on correla-
tional models that simply identify the direct relationship between one factor and
dropout behavior, sometimes controlling for the influence of other factors.
What is needed is a more comprehensive, causal model of the dropout process.
Such a model should successfully identify the full range of proximal and distal
influences, the interrelationships among them, and their long-term, cumulative
effects. Developing such a model should be the goal of researchers working in this
area. In pursuit of this goal, researchers should consider the following factors.
First, they should attempt to uncover the processes that underlie and lead to this
problem. Many of the factors known to be associated with this problem are
structural in nature and reveal little of the underlying processes. For example, why
do Hispanics or children from socioeconomically poor families have higher dropout
rates? Some possible explanations include the use of language in the home, the
amount of time that parents spend with their children, and the type of "parenting"
style used in the home (Dornbush et al., in press; Liebowitz, 1977; Steinberg,
Blinde, & Chan, 1984).
Understanding school processes also deserves further attention (Natriello, Pallas,
& McDill, 1986). In fact, dropping out itself might better be viewed as a process of
disengagement from school, perhaps for either social or academic reasons (Catterall,
1986), that culminates in the final act of leaving. Identifying and understanding
the processes within homes and schools is an area where ethnographic studies can
make important contributions (e.g., Fine, 1986).
Second, future research efforts need to explore the interrelationships among the
various factors associated with dropping out. This is particularly important in trying
to separate actual causes of this problem from correlates such as attitudes and
behaviors. For example, while many female dropouts say they leave school because
they are pregnant, both getting pregnant and leaving school may be caused by a
number of other, related factors (Hofferth & Moore, 1979). Similarly, poor attitudes
and disruptive behavior in school might be better considered symptoms of under-
lying problems than actual causes of dropping out.
Third, researchers should attempt to measure the long-term, cumulative effects
of the various influences on dropping out. This is particularly important given the
influences of family background and early school achievement. Family background
can have a powerful, cumulative influence on school achievement through its
effects on such things as kinds of schools children attend, their attitudes about
school, and learning that takes place in the home. These influences affect a student's
achievement at an early age, which, in turn, influences subsequent attitudes and
performance in school (e.g., Stroup & Robins, 1972). Yet the cumulative influence
of family background and early school achievement has not been fully explored
either generally or with respect to the problem of dropping out.
Finally, a comprehensive model of dropout behavior should address the notion
that there are different types of dropouts who leave school for different reasons.

111
Russell W. Rumberger
That is, there is no "typical" dropout. A poor, urban black may drop out of school
because he is doing badly, his school is understaffed, and he believes his economic
prospects are poor whether or not he finishes school. A suburban, middle-class
white may drop out of school because he is bored although doing reasonably well
in school, he wants to spend some time with his friends, and he knows he can
finish school later on at the community college. The causes and the nature of
dropping out are very different for these two types of teenagers. Such differences
should be explored further and used to develop separate models of dropping out
for different types of students (Grant & Sleeter, 1986).
A more comprehensive model of dropout behavior should help identify the
actual causes of dropping out and their relative influence. The result would be of
immediate use to educators and policymakers. It would help identify potential
dropouts at an early age, when effective interventions could be designed and
implemented. It would also help identify the kinds of educational interventions,
both academic and psychological in nature, that could be most effective in address-
ing this problem. And it would help identify the mix of both educational and social
interventions that could be effective in helping potential and actual dropouts.

The Consequences of the Problem

Concern for dropouts is predicated on a belief that leaving high school before
graduation is bad for the individual and for society. Dropping out of high school is
generally viewed as a visible form of academic failure in the same way that high
school graduation is seen as a visible form of academic success at the secondary
level. Yet neither act may reveal much about educational .achievement. A recent
study of the Chicago school system found, for example, that 47% of students
enrolled in the ninth grade in 1980-81 graduated by 1984, but only 15% graduated
who could read at or above the national average (Designs for Change, 1985).
Similarly, not all dropouts are behind in school and have poor test scores (Fine,
1986).
In fact, dropping out could be beneficial for some kids as well as for the schools
they attend. Some students are not able or willing to get anything out of school;
others choose other alternatives over going to school, alternatives that in some
cases can be more fulfilling and rewarding. And some students who remain in
school can be very disruptive to those students who want to be there and to learn.
A recent study of high school graduates and dropouts found, for example, that
dropouts showed equal or greater improvements in self-esteem and a sense of
control than high school graduates (Wehlage & Rutter, 1986).
Overall, however, most evidence supports the notion that dropping out has
negative individual and social consequences. Individual dropouts suffer because
many have difficulty finding steady, well-paying jobs not just when they first leave
school but over their entire lifetimes. Society suffers as well because unemployment
and lost earnings lower tax revenues and increase demands on social services. But
the consequences of dropping out go beyond simple economic losses, no matter
how large they may be. They cover a wide range of individual and social outcomes
that need to be better understood.

112
 High School Dropouts
Individual Consequences
The most immediate individual consequence of dropping out of school is a low
level of academic skills. While graduating from high school does not ensure that a
person has sufficient academic skills for successful employment and further edu-
cation, failing to graduate usually ensures that a person does not. Recent studies
confirm that dropouts, on average, have lower academic skills than high school
graduates generally and even those graduates of similar personal characteristics
(e.g., Alexander, Natriello, & Pallas, 1985).
Because of their low levels of academic skills, many high school dropouts find it
difficult to secure steady employment and an adequate income. These economic
effects are sizable and are well-documented with published government statistics.
In the fall of 1982, for example, dropouts from the 1981-82 school year had
unemployment rates almost twice as high as 1982 high school graduates, 42%
versus 23% (Table 4). Even those dropouts who are able to secure year-round, full-
time employment still earn from 12% to 18% less than workers who complete high
school (Table 4).
Dropouts' lower level of educational achievement does not just have an imme-
diate economic consequence, it becomes an even bigger disadvantage over time
because dropouts have fewer opportunities to obtain additional education and
training needed to remain even relatively competitive in the job market. For
example, Census data reveal that the difference in expected lifetime earnings from
ages 18 to 64 between a male high school graduate and a male high school dropout
in 1979 was more than $250,000 (U.S. Bureau of the Census, 1983, Table 1).
Of course, the economic consequences of dropping out of school are not the
same for all social groups. The relative economic disadvantage of dropping out is
larger for whites than for Hispanics or blacks. In the fall of 1982, for example,
white dropouts had unemployment rates almost twice as large as white high school
graduates, while the unemployment rates for Hispanics and blacks were only one
fourth to one third as high (Table 4). Thus, there may be less economic incentive
for minorities to stay in school than whites, especially when black high school
graduates still experience an unemployment rate of over 50%.
While the educational and economic consequences of dropping out are generally
well-documented, much less attention has been focused on trying to identify and
measure other individual consequences. These would include effects on psycholog-
ical well-being and health. It is possible that dropping out of school leads to poorer
mental and physical health either directly or indirectly through its effects on
employment and income. One study found, for example, that increased unemploy-
ment was associated with increases in total mortality, suicides, and admissions to
state mental hospitals (Brenner, 1976). Since dropouts have higher rates of unem-
ployment than other persons, this study would suggest that they would also suffer
greater rates of mortality, suicide, and mental disorders. Yet the causal relationship
between dropping out of school and subsequent mental and physical health has
not been fully explored.
Social Consequences
Dropping out of high school affects not only those who leave school but also
society at large. Moreover, the social consequences go beyond the economic and

113
Russell W. Rumberger

TABLE 4
Unemployment rates (percentages) and annual earnings (in dollars) for high school dropouts
and high school graduates, by sex, race, and ethnicity
1976 1982

Cohort Ratio Ratio

Level Dropouts/ Level Dropouts/
Graduates Graduates
Unemployment3
Total 1.78 1.85
Dropouts 30.9 41.6
Graduates 17.4 22.5
Males 1.75 2.05
Dropouts 28.3 43.4
Graduates 16.2 21.2
Females 1.96 1.60
Dropouts 36.9 38.3
Graduates 18.8 23.9
Whites 1.90 1.89
Dropouts 27.5 36.0
Graduates 14.5 19.0
Blacks 1.24 1.35
Dropouts 56.9 71.4
Graduates 45.8 53.0
Hispanics 2.01 1.23
Dropouts 27.9 42.2
Graduates 13.9 34.3
Annual earnings15
Males .82 .84
Dropouts 7,351 10,964
Graduates 9,004 13,088
Females .87 .82
Dropouts 5,758 8,414
Graduates 6,620 10,235
a
Unemployment rates in October of each year for persons 16 to 24 years old who either
graduated or dropped out of high school in the preceding academic year.
b
Annual earnings for year-round, full-time workers who were 18 to 24 years old as of March
of the following year.
Sources: Unemployment data from A. M. Young, "Students, graduates and dropouts in the
labor market," Monthly Labor Review, 100 (July 1977), Table 3, and A. M. Young, "Youth
labor force marked turning point in 1982," Monthly Labor Review, 106 (August 1983),
Table 4. Earnings data from U.S. Bureau of the Census, Current Population Reports, Series
P-60, No. 114 (July 1978), Table 48, and No. 142 (February 1984), Table 48.

psychological impacts that befall individual high school dropouts. In the only
comprehensive study that has ever been done on this facet of the dropout problem,
Levin (1972) identified seven social consequences of inadequate education, which
he defined as the failure to complete high school:
1. Forgone national income;
2. Forgone tax revenues for the support of government services;

114
 High School Dropouts
3. Increased demand for social services;
4. Increased crime;
5. Reduced political participation;
6. Reduced intergenerational mobility; and
7. Poorer levels of health, (p. 10)
For each of these areas he examined the research literature and summarized what
was known about the relationship between education and that particular social
outcome. He then estimated the social costs associated with the first four outcomes.
Forgone income is perhaps the most often cited social consequence of dropping
out of high school. It is also the most widely documented. Levin estimated that the
forgone income from a cohort of males 25 to 34 in 1969 who failed to finish high
school amounted to $237 billion. This forgone income resulted in forgone govern-
ment revenues of $71 billion (Levin, 1972, p. ix). A more recent study estimated
the forgone income of both male and female dropouts from the national high
school class of 1981 at $228 billion and forgone government revenues at more than
$68 billion (Catterall, 1985, Table 2).
The social consequences and social costs of dropping out go beyond forgone
income and revenues, however. As Levin documented, high school dropouts are
more likely to require a wide range of social services, including welfare, medical
assistance, and unemployment assistance. They are also more likely to engage in
crime, have poorer health, have lower rates of intergenerational mobility, and lower
rates of political participation. Based on the research literature and cost data
available at that time, he estimated the social costs of providing social services and
fighting crime associated with dropping out at $6 billion per year (Levin, 1972, p.
ix). Today, of course, the figures would be much higher.
Levin's study is remarkably useful in helping to identify a wide range of social
consequences that are due to dropping out. However, the study and the literature
it reviews are more than 14 years old. Given the large amount of research that has
been undertaken in the last 14 years in the areas of crime, health, and welfare, it
would be useful to once again examine this large body of literature to try to
ascertain the many social consequences of dropping out.
It is likely, for example, that the social consequences are greater today than in
the past. That definitely appears to be the case with respect to earnings and
unemployment. The earnings differential between male high school dropouts and
male high school graduates increased from $73,000 in 1968 to $260,000 in 1979,
a much greater increase than the increase in consumer prices over the same period
(Levin, 1972, Table 8; U.S. Bureau of the Census, 1983, Table 1). Unemployment
rates for high school dropouts were 20% higher than overall unemployment rates
in 1950, but 100% higher in 1979 (Rumberger, 1983b, Table 5).
Yet the economic consequences of dropping out may not be changing similarly
for all racial and ethnic groups. In 1976, for example, Hispanic dropouts had
unemployment rates twice as high as Hispanic high school graduates, but by 1982
unemployment rates for Hispanic dropouts were 25% higher (Table 4). These
differences need to be better documented and linked to students' decisions to drop
out of school.
In the future, the relative economic disadvantage of dropping out of high school
could be even greater than today. The skill requirements of many jobs could be
altered in the future because of the increased use of new technologies. Without a

115
Russell W. Rumberger
sound, basic education, dropouts will be less able to learn new skills and adapt to
a changing work environment (National Academy of Sciences, 1984). Therefore
they could become even less employable.
The consequences of dropping out of high school deserve more attention from
researchers and policy analysts. An attempt should be made to document more
fully the wide range of impacts, both good and bad, resulting from dropping out of
high school, as well as differences in these impacts among racial and ethnic groups.
And more comprehensive estimates should be prepared on the total social costs of
this major problem. Such estimates could be valuable in rallying political support
to fund dropout prevention and recovery programs.

Solutions to the Problem

Not all students can or should be expected to finish high school. Yet many
students want to finish school and could be helped through effective policy
interventions. Without such interventions the dropout rate could easily increase
due to therisingproportion of minorities in the school-age population and increased
academic requirements for high school graduation. Whereas further research on
the causes of dropping out will help to better identify and measure the influence of
the many factors associated with this problem, the search for effective solutions
cannot wait.
Concerted efforts are currently being made in finding and promoting solutions
to the dropout problem throughout the United States. At least 15 state-level
commissions have been convened in the last couple of years to address the problems
of at-risk students (Smith & Hester, 1985). Many districts, especially those with
large at-risk populations, have enacted a variety of programs designed to reduce
the incidence of dropping out. And new studies have been undertaken to identify
and characterize successful dropout programs around the country (e.g., Bullis,
1986; National Foundation for the Improvement of Education, 1986; Stern et al.,
1985).
Toward Effective Interventions
Recent reviews of dropout programs and the vast literature on the causes of
dropping out, although far from complete, do suggest some of the elements needed
to develop a successful strategy of dropout prevention and recovery. Some of these
elements include: (a) different programs designed for different types of dropouts;
(b) an appropriate mix of educational and noneducational services in each program;
(c) accurate and timely identification of students with a high risk of dropping out;
and (d) programs designed for early prevention, late prevention, and recovery.
The first element follows from the discussion of the causes of dropping out.
Different kinds of students drop out for different reasons. Some are related to
problems in school, such as a lack of interest or poor performance; others are
related to factors outside of the school, such as the need to find work or having a
child. A comprehensive strategy will need to address all of these factors, providing
programs for different children with different needs. For example, recovery pro-
grams for teenage mothers will have to serve their needs for child care as well as
for an education; prevention programs for similar kinds of students should include
birth control information and services as well as educational services.

116
 High School Dropouts
Beyond these apparent needs, there are a host of other needs that effective
programs must address. They must first address the particular academic needs of
the students, by providing an appropriate type of curriculum, teaching staff,
instructional progress, and even schedule and location (Stern, 1986). For example,
reviews of dropout programs suggest that successful programs often mix academic
and vocational studies, provide more individualized instruction, and use a teaching
staff more sensitive and responsive to the needs of the students (Bullis, 1986; Olsen
& Edwards, 1982; Stern, 1986).
Besides these educational elements, successful programs need to address other
needs of students. Perhaps the most important is their psychological need for
someone to care about them individually, a need that is often met through the
provision of counseling (Bullis, 1986; Olsen & Edwards, 1982; Treadway, 1985).
The third element is crucial if dropout prevention programs are to be successful.
That is, the schools must be able to successfully identify those students who are
most likely to drop out of school if they hope to do something about it. A recent
study of California dropouts found that half of the dropouts interviewed did not
discuss their decision with anyone at school before they left (Olsen & Edwards,
1982, p. 32).
Timely identification is equally important. The earlier a student with a high risk
of dropping out is identified, the more likely it is that a sustained effort at dropout
prevention will be successful. Research has shown that some dropouts begin
showing signs of academic failure and disengagement in school in the early
elementary grades (Lloyd, 1978; Stroup & Robins, 1972). Successful identification
of high-risk students in elementary and junior high school would provide more
time to intervene and address the needs of these kids at an early age.
The final element follows from the third. If students with a high risk of dropping
out can be identified at an early age, prevention programs should be started at an
early age as well. Even if accurate, early identification is not possible, it still makes
sense to initiate early interventions for disadvantaged kids who generally have a
high probability of dropping out. A recent evaluation of one preschool program for
the disadvantaged found that it reduced the incidence of dropping out (Schweinhart
etal., 1985).
Many current efforts to address the problem of dropouts are premised on a belief
that dropouts constitute a relatively small percentage of the school population. In
this case, special, supplemental programs can be set up to target these at-risk
populations while schools continue to serve other students with regular programs.
But such a strategy is insufficient in school systems, such as those in New York,
Boston, and Chicago, where close to half of the students drop out. In this case,
where dropping out is often the rule rather than the exception, more fundamental
and systematic changes will be needed to address the problem (Hess, 1986).
The Need for Systematic Evaluations
While descriptive reviews of programs are helpful in helping to design effective
interventions, systematic evaluations are needed to determine both the effectiveness
of dropout prevention and recovery programs and their costs (Rossi, Freeman, &
Wright, 1979). Evaluations of program effectiveness determine the extent to which
programs are effective in producing their desired outcomes. In dropout prevention
programs, for example, desired outcomes may be a reduced incidence of dropping

117
Russell W. Rumberger
out among the target population or improvements in some of the known correlates
of dropping out, such as student involvement or academic performance. Evalua-
tions must be able to determine whether the outcome was actually caused by the
program or caused by something else.
Evaluations of costs involve determining all the resources used in the program,
not only so that the full cost of the program can be determined, but so that the
program can be properly implemented in another setting (Levin, 1983). Information
on program effects and costs can then be compared in two ways. Cost-benefit
studies determine whether a program's benefits exceed its costs; cost-effectiveness
studies determine whether one program is more effective for each dollar spent than
other, alternative programs.
Although some early cost-benefit studies of dropout programs were done in the
1960s (e.g., Weisbrod, 1965), to my knowledge there have been no recent studies
where both the effectiveness and the costs of dropout programs have been fully
evaluated (U.S. Government Accounting Office, 1986). Yet only by considering
both the costs and the effectiveness of dropout programs will it be possible to
produce the greatest improvement in the dropout problem at the lowest social cost.
One of the problems that plagues even modest efforts to identify effective programs
is a lack of suitable information on program features and outcomes (Stern et al.,
1985, p. 34).
There have been some attempts to measure the effectiveness of some regular
school programs in keeping students in school. Vocational education in high
schools, for example, is often said to make schooling more relevant for certain
kinds of kids and thus increase their likelihood of staying in school. Other kinds of
alternative educational programs, such as alternative or continuation schools or
employment training programs, are thought to have similar effects (Stern et al,
1985). Yet the effectiveness of such programs on reducing the likelihood of dropping
out has not been demonstrated (Catterall & Stern, 1986; U.S. Government Ac-
counting Office, 1986).
While no cost-effectiveness evaluations of dropout programs have been under-
taken, there have been attempts to compare the costs of dropout prevention to the
economic benefits associated with the completion of high school. In his earlier
study, Levin estimated that each dollar of social investment in dropout prevention
would produce $6 in national income and almost $2 in tax revenues over the
lifetime of the 25- to 34-year-old males he examined (Levin, 1972, p. 30). A more
recent study estimated that the costs of dropout prevention in Chicago would be
less than 1 % of the economic benefits derived from increased tax revenues, re-
duced welfare payments, and savings from the costs of crime (Hess & Labuer,
1985, p. 7).
Conclusions
Dropping out of high school is considered to be an important educational and
social problem. As such, it has commanded the attention of researchers, policy-
makers, and educators who are trying both to better understand the nature of the
problem and to do something about it. Part of the difficulty in pursuing these
efforts is that this problem, like many others, is complex and multifaceted. Each of
the four major facets discussed in this review—the incidence of the problem, its
causes, consequences, and remedies—requires attention. And each requires a broad,

118
 High School Dropouts
interdisciplinary approach that acknowledges not only the educational aspects of
the problem, but the social, economic, and psychological ones as well.
Because dropping out of school has been a long-standing problem, there is a
considerable research literature on this topic, especially from the 1960s (e.g.,
Bachman, Green, & Wirtanen, 1971). This earlier literature, although useful in
exploring the wide range of factors associated with this problem, in general suffers
from several shortcomings. First, many of the studies were largely correlational in
nature and are able to show only bivariate relationships between dropping out and
a host of antecedents or outcomes. Second, at best they explore the longitudinal
nature of the problem over a short period of time, such as from an early high
school to a later high school period. Third, many of the factors associated with this
problem, especially those related to family background, focus on structural char-
acteristics rather than processes.
Current and future research efforts need to move beyond these earlier efforts
while building on them. Much recent research on dropouts has simply replicated
the descriptive nature of earlier studies with more recent data. Such efforts are
necessary and useful as a first step, but to move beyond them will require
overcoming the limitations of these earlier studies. That is, new research efforts
should focus on developing multivariate, longitudinal, and comprehensive models
of the causes and consequences of dropping out. Additional research effort is also
needed in conducting systematic evaluations of dropout prevention and recovery
programs.
The dropout problem is unlikely ever to go away. But concerted and cooperative
efforts by educators, policymakers, and educational researchers can improve our
understanding of the problem and help reduce its incidence.
References
Alexander, K. L., Natriello, G., & Pallas, A. M. (1985). For whom the school bell tolls: The
impact of dropping out on cognitive performance. American Sociological Review, 50, 409-
420.
Bachman, J. G., Green, S., & Wirtanen, I. D. (1971). Dropping out: Problem or symptom?
Ann Arbor: Institute for Social Research, University of Michigan.
Borus, M. E,, & Carpenter, S. A. (1984). Choices in education. In M. E. Borus (Ed.), Youth
and the labor market (81-110). Kalamazoo, MI: W. E. Upjohn Institute for Employment
Research.
Brenner, M. H. (1976). Estimating the costs of national economic policy. Study prepared for
the Joint Economic Committee, U.S. Congress, 94th Congress, 2nd Session. Washington,
DC: U.S. Government Printing Office.
Bridge, R. G., Judd, C. M., & Moock, P. R. (1979). The determinants of educational outcomes.
Cambridge, MA: Ballinger.
Brown, G. H., Rosen, N. L., Hill, S. T., and Olivas, M. A. (1980). The condition of education
for Hispanic Americans. Washington, DC: U.S. Government Printing Office.
Bullis, B. M. (1986) Dropout prevention. Berkeley, CA: Policy Analysis for California Educa-
tion.
California Assembly Office of Research. (1985). Dropping out, losing out: The high cost for
California. Sacramento: Assembly Office of Research.
California State Department of Education. (1986). California dropouts: A status report.
Sacramento: California State Department of Education.
Catterall, J. S. (1985). On the social costs of dropping out of high school (Report 86-SEPI-3).
Stanford, CA: Stanford Education Policy Institute, Stanford University.

119
Russell W. Rumberger
Catterall, J. S. (1986). A process model of dropping out of school. Unpublished paper,
University of California, Los Angeles.
Catterall, J. S., & Stern, D. (1986). The effects of alternative school programs on high school
completion and labor market outcomes. Educational Evaluation and Policy Analysis, 8,
77-86.
Designs for Change. (1985). The bottom line: Chicago's failing schools and how to save them.
Chicago: Author.
Dornbusch, S. M., Ritter, P. L., Leiderman, P. H., Roberts, D. F., & Fraleigh, M. J. (in press).
The relation of parenting style to adolescent school performance. Child Development.
Ekstrom, R. B., Goertz, M. E., Pollack, J. ML, & Rock, D. A. (1986). Who drops out of high
school and why? Findings from a national study. Teachers College Record, 87, 356-373.
Fields, C. M. (1986, May 14). Value of high-school-equivalency certificates for dropouts is
questioned by Wisconsin Study. Education Week, p. 30.
Fine, M. (1986). Why urban adolescents drop into and out of public high school. Teachers
College Record, 87, 393-409.
Grant, C. A., & Sleeter, C. E. (1986). Race, class, and gender in education research: An
argument for integrative analysis. Review of Educational Research, 56, 195-211.
Hammack, F. M. (1986). Large school systems' dropout reports: An analysis of definitions,
procedures, and findings. Teachers College Record, 87, 324-341.
Hess, G. A., Jr. (1986). Educational triage in an urban school setting. Metropolitan Education,
2, 39-52.
Hess, G. A., Jr., & Lauber, D. (1985). Dropouts from the Chicago public schools. Chicago:
Chicago Panel on Public School Finances.
Hofferth, S. L., & Moore, K. A. (1979). Early childbearing and later economic well-being.
American Sociological Review, 44, 784-815.
Kirsch, I., & Jungeblut, A. (1986). Literacy: Profiles of America's young adults. Princeton,
NJ: Educational Testing Service.
Kolstad, A. J., & Owings, J. A. (1986, April). High school dropouts who change their minds
about school. Paper presented at the annual meeting of the American Educational Research
Association, San Francisco.
LaFromboise, T., & Rudes, B. (1983). Student attendance and retention. In Development
Associates (Ed.), The evaluation of the impact of the Part A entitlement program funded
under Title IV of the Indian Education Act. Washington, DC: U.S. Department of Educa-
tion.
Levin, H. M. (1972). The costs to the nation of inadequate education. Study prepared for the
Select Committee on Equal Educational Opportunity, U.S. Senate. Washington, DC: U.S.
Government Printing Office.
Levin, H. M. (1983). Cost-effectiveness: A primer. Beverly Hills, CA: Sage.
Levin, H. M. (1986). Education reform for disadvantaged students: An emerging crisis. West
Haven, CT: National Education Association.
Levin, H. M., & Rumberger, R. W. (in press). Educational requirements for new technologies:
Visions, possibilities, and current realities. Educational Policy.
Liebowitz, A. (1977). Parental inputs and children's achievement. Journal of Human Re-
sources, 12, 242-251.
Lloyd, D. N. (1978). Prediction of school failure from third-grade data. Educational and
Psychological Measurement, 38, 1193-1200.
McDill, E. L., Natriello, G., and Pallas, A. M. (1985). Raising standards and retaining
students: The impact of the reform recommendations on potential dropouts. Review of
Educational Research, 55, 415-433.
McDill, E. L., Natriello, G., & Pallas, A. M. (1986). A population at risk: Potential conse-
quences of tougher school standards for student dropouts. American Journal of Education,
94, 135-181.
National Academy of Sciences. (1984). High schools and the changing workplace: The
employers' view. Report of the Panel on Secondary School Education for the Changing
Workplace. Washington, DC: National Academy Press.

120
 High School Dropouts
National Commission on Excellence in Education. (1983). A nation at risk. Washington, DC:
U.S. Government Printing Office.
National Foundation for the Improvement of Education. (1986). Operation rescue: A blueprint
for success. Washington, DC: National Foundation for the Improvement of Education.
Natriello, G., Pallas, A. M, & McDill, E. L. (1986). Taking stock: Renewing our research
agenda on the causes and consequences of dropping out. Teachers College Record, 87,
430-440.
Olsen, L., & Edwards, R. (1982). Push out, step out: A report on California's public school
drop-outs. Oakland, CA: Citizens Policy Center.
Plisko, V. W., & Stern, J. D. (1985). The condition of education: 1985 edition. Washington,
DC: U.S. Government Printing Office.
Quinn, L., & Haberman, M. (1986). Are GED certificate holders ready for postsecondary
education? Metropolitan Education, 2, 72-82.
Rossi, P. H., Freeman, H. E., & Wright, S. R. (1979). Evaluation: A systematic approach.
Beverly Hills, CA: Sage.
Rumberger, R. W. (1983a). Dropping out of high school: The influence of race, sex, and
family background. American Educational Research Journal, 20, 199-220.
Rumberger, R. W. (1983b). Education, unemployment, and productivity (Project Report No.
83-A14). Stanford, CA: Institute for Research on Educational Finance and Governance,
Stanford University.
Schweinhart, L. J., Berrueta-Clement, J. R., Barnett, W. S., Epstein, A. S., & Weikart, D. P.
(1985). The promise of early childhood education. Phi Delta Kappan, 66, 548-553.
Smith, R. C, & Hester, E. L. (1985). Who's looking out for at-risk youth?New York: MDC.
Steinberg, L., Blinde, P. L., & Chan, K. S. (1984). Dropping out among language minority
youth. Review of Educational Research, 54, 113-132.
Stern, D. (1986). Dropout prevention and recovery in California. Unpublished paper, Univer-
sity of California, Berkeley.
Stern, D., Catterall, J., Alhabeff, C, & Ash, M. (1985). Reducing the high school dropout rate
in California. Report to the California Policy Seminar. Berkeley: University of California.
Stroup, A. L., & Robins, L. N. (1972). Elementary school predictors of high school dropout
among black males. Sociology of Education, 45, 212-222.
Task Force on Education for Economic Growth. (1983). Action for excellence. Denver, CO:
Education Commission of the States.
Toles, T., Schulz, E. M, & Rice, W. K., Jr. (1986). A study of variation in dropout rates
attributable to effects of high schools. Metropolitan Education, 2, 30-38.
Treadway, P. G. (1985, November). Beyond statistics: Doing something about dropping out
of school. Paper presented at the School Drop-out Prevention Conference, Aptos, CA.
U.S. Bureau of the Census. (1983). Lifetime earnings estimates for men and women in the
United States: 1979. Current Population Reports Series P-60, No. 139. Washington, DC:
U.S. Government Printing Office.
U.S. Bureau of the Census. (1985) Statistical abstract of the United States, 1986 (196th ed.).
Washington, DC: U.S. Government Printing Office.
U.S. Government Accounting Office. (1986). School dropouts: The extent and nature of the
problem. Washington, DC: U.S. Government Printing Office.
Wehlage, G. G., & Rutter, R. A. (1986). Dropping out: How much do schools contribute to
the problem? Teachers College Record, 87, 374-392.
Weisbrod, B. A. (1965). Preventing high school dropouts. In R. Dorfman (Ed.), Measuring
benefits of government investments (pp. 117-171). Washington, DC: The Brookings Insti-
tution.
Wittebols, J. H. (1986). Collecting national dropout statistics. Washington, DC: Council of
Chief State School Officers, State Education Assessment Center.
Author
RUSSELL W. RUMBERGER, Associate Professor of Education, Graduate School of Edu-
cation, University of California, Santa Barbara, CA 93106. Specializations: economics of
education, education policy.

121