ERIC Identifier: ED339092
Publication Date: 1991-00-00
Author: Gaustad, Joan
Source: ERIC Clearinghouse on Educational Management Eugene OR.

Identifying Potential Dropouts. ERIC Digest.

The percentage of Americans completing high school has increased steadily over the past fifty years. Not only have dropout rates been falling, but many dropouts return to high school or obtain an equivalency certificate within a few years after dropping out. According to U.S. Census Bureau figures, in 1940 less than 40 percent of all persons aged 25 to 29 had completed high school; by 1980, over 85 percent had done so (Finn 1987). Frase (1989) notes that there was an increase in the annual dropout rate between 1968 and 1978, but the rate has been declining since then.

Nonetheless, concern about the dropout issue has increased among educators, policymakers, and the public. Dropout rates remain disturbingly high in certain areas, particularly major cities, and among certain populations, such as Hispanics. Moreover, as changes in the nation's economy eliminate jobs for unskilled workers, dropouts will increasingly suffer in the job market.

Recent legislation to raise high school graduation requirements has provoked concern that students unable to meet these requirements will drop out, says Rumberger (1987). State and federal efforts to evaluate and improve school performance have stimulated attempts to more accurately measure dropout rates, to identify potential dropouts, and to develop programs to help them.


Most studies agree that the main factors associated with dropping out include students' socioeconomic status, location, school behavior, and academic achievement.

"Dropout rates are higher for students coming from low socioeconomic backgrounds, from single-parent families, and from non-English language family backgrounds," stated Frase in the first annual report by the National Center for Education Statistics. This nationwide study also found higher dropout rates for students living in cities than in suburbs or rural areas, and in the South and West rather than in the Northeast. Students who marry or have children, or who have had problems with the law or school authorities, are also at greater risk.

Academic factors are clearly related to dropping out. Students who received poor grades, who had repeated a grade, who were overage for their class, and who had poor attendance for reasons other than illness were more likely to drop out. "A powerful predictor...was the attendance record during the first four months of tenth grade," Frase reported.

Barrington and Hendricks (1989) found that dropouts in a Wisconsin community showed clear indications of academic problems by the third grade. Their achievement test scores were significantly lower than those of their classmates and also below their ability as measured by intelligence tests; teacher comments alone identified potential dropouts with 63 percent accuracy. Poor attendance, failing grades, and low overall GPA marked these students' high school careers.

Unfortunately, uncontrolled variables complicate the process of accurately identifying dropouts, as we shall see below.


Definitions vary among states, districts, and even among schools within the same district. In addition, the criteria used to define dropout are sometimes questionable, resulting in statistics that don't accurately reflect the problem.

For example, some institutions count as dropouts students who transfer to other schools, are hospitalized, take longer than four years to graduate, or are admitted early to college. "One district treated a student who had died as a dropout," note Barber and McClellan (1987). Some don't count students who leave to get married, who attend for four full years, or who exceed the compulsory attendance age but haven't fulfilled graduation requirements.

LeCompte and Goebel (1987) describe how recordkeeping problems can contribute to inaccurate totals. In one pilot study, 25 percent of a group of "dropouts" turned out to have transferred; the transfer requests in their folders just hadn't been entered into the computer. And if "acts of dropping out" are measured, a student who drops in and out repeatedly will be counted more than once. On the other hand, if students who have dropped out remain on the rolls, the error benefits schools whose funding depends on enrollment--a situation that doesn't encourage vigorous attempts at accuracy.

When the count is made also affects the totals. If dropouts are recorded only during the academic year, students who don't return after the summer--which may constitute one-third of all dropouts--are overlooked. And collecting statistics only from grades 9 through 12 misses kids who dropped out earlier (LeCompte and Goebel).


Although a number of factors are correlated with dropping out, this does not mean that these factors "cause" individuals to drop out (Finn). In addition, the interaction of multiple, complex variables makes it hard to determine which ones are significant.

For example, overall dropout rates are higher for African-Americans than for whites, but race turns out not to be the crucial factor. When social background is factored in, reports Frase, "dropout rates for blacks are not higher, and in some cases may be lower, than those for whites."

On the other hand, Fernandez and Shu (1988) found that Hispanic students in a national sample dropped out at significantly higher rates than other students, even when factors such as family income, academic achievement, and parents' educational level were considered.

Different risk factors are important in different communities. Even within the same school, students drop out for different reasons. Finally, most dropouts are simply unexplained. The majority of students with any particular risk factor do not drop out, and the majority of dropouts are not in the at-risk groups.

For example, as Frase reports, "Of the dropouts from the 1980 sophomore class: 66 percent were white, 86 percent had an English language home background, 68 percent came from two-parent families, 42 percent had neither children nor spouses, and 71 percent had never repeated a grade."


The margin for error in current prediction models clearly limits their value. For example, in 1987 the Texas legislature required school districts to identify students who were at risk of dropping out. Four criteria were mandated: being overage for grade, being two or more years below grade level in mathematics or reading skills, failing two or more courses during a semester, or failing any section of the state minimum skills tests.

Evaluators at the Houston Independent School District found these criteria had serious limitations (Bowman and others 1991). To begin with, 40.6 percent of the district's secondary school students fell into at least one of the at-risk categories, over 50 percent on some campuses, too large a target for meaningful intervention.

In terms of predictive accuracy, only 13.9 percent of Houston's at-risk students actually dropped out, and almost half the dropouts were not predicted. A study of nine other Texas districts found even lower accuracy: 61.7 percent of dropouts were not predicted (Parsons, Saye, and McNamara 1990).

Bowman and his colleagues found that using combinations of the four criteria increased accuracy to 32.3 percent at best--an improvement, but still disappointing.


Before we can predict who will drop out, we must know who is dropping out. Experts agree that educators and policymakers must set a standard definition of dropping out before accurate, comparable data can be collected. LeCompte and Goebel say only a federal mandate can ensure uniformity in definitions as well as in basic reporting procedures among states and districts.

To date, most studies have covered only the high school years; more data are needed on younger dropouts. The National Education Longitudinal Study of 1988, which follows a representative nationwide sample of eighth-graders, will ultimately provide more data than its predecessor, High School and Beyond, which began with tenth-graders (Frase). Several researchers urge that statistics be collected starting with elementary school. It is hoped such studies will identify characteristics that permit earlier intervention.

More accurate recordkeeping is essential. Barber and McClellan believe the necessary technology and personnel are already available. However, LeCompte and Goebel express concern about the ability of smaller and poorer districts to finance up-to-date, computerized recordkeeping systems.

A systematic means of tracking transfers would be desirable. LeCompte and Goebel suggest a nationwide system modeled after the Migrant Student Record Transfer system but acknowledge the cost may be prohibitive. As a low-tech alternative, they suggest districts be required to request transcripts for all transfer students and record requests in a uniform manner.

Research may eventually produce more accurate models for predicting dropouts. In the meantime, although educators may want to pay special attention to students with the risk factors mentioned above, they must not overlook the majority of potential dropouts who are not obviously "at risk."


Barber, Larry W., and Mary C. McClellan. "Looking at America's Dropouts: Who Are They?" PHI DELTA KAPPAN 69,4 (December 1987): 264-67. EJ 363 376.

Barrington, Byron L., and Bryan Hendricks. "Differentiating Characteristics of High School Graduates, Dropouts, and Nongraduates." THE JOURNAL OF EDUCATIONAL RESEARCH 82,6 (July/August 1989): 309-19. EJ 398 453.

Bowman, Clay, and others. "Making the Best of State-Mandated At-Risk Criteria." Houston Independent School District Department of Research and Evaluation. Paper presented at the Annual Meeting of the American Educational Research Association, Chicago, Illinois, April, 1991.

Fernandez, Ricardo R., and Gangjian Shu. "School Dropouts: New Approaches to an Enduring Problem." EDUCATION AND URBAN SOCIETY 20,4 (August 1988): 363-86. EJ 378 368.

Finn, Chester E., Jr. "The High School Dropout Puzzle." THE PUBLIC INTEREST 87 (Spring 1987): 3-22. EJ 355 110.

Frase, Mary J. DROPOUT RATES IN THE UNITED STATES: 1988. Washington, DC: National Center for Education Statistics, 1989. 114 pages. ED 313 947.

LeCompte, Margaret D., and Stephen D. Goebel. "Can Bad Data Produce Good Program Planning? An Analysis of Record-Keeping on School Dropouts." EDUCATION AND URBAN SOCIETY 19,3 (May 1987): 250-68. EJ 356 409.

Parsons, James L.; Elaine Saye; and James F. McNamara. "1988-89 Student Dropout Research Project: An Aggregate Report of Nine Collaborative School Districts." Texas A&M School/University Research Collaborative. Paper presented at the Fourth Annual Collaborative Research Conference, January 12, 1990.

Rumberger, Russell W. "High School Dropouts: A Review of Issues and Evidence." Review of EDUCATIONAL RESEARCH 57,2 (Summer 1987): 101-21. EJ 369 707.

Library Reference Search

Please note that this site is privately owned and is in no way related to any Federal agency or ERIC unit.  Further, this site is using a privately owned and located server. This is NOT a government sponsored or government sanctioned site. ERIC is a Service Mark of the U.S. Government. This site exists to provide the text of the public domain ERIC Documents previously produced by ERIC.  No new content will ever appear here that would in any way challenge the ERIC Service Mark of the U.S. Government.

Popular Pages

More Info