A Prospective Longitudinal Study of High School Dropouts Examining Multiple Predictors Across Development

Download A Prospective Longitudinal Study of High School Dropouts Examining Multiple Predictors Across Development

Post on 02-Jul-2016




2 download


<ul><li><p>Journal of School Psychology, Vol. 38, No. 6, pp. 525549, 2000Copyright 2000 Society for the Study of School Psychology</p><p>Printed in the USA0022-4405/00 $see front matter</p><p>Pergamon</p><p>PII S0022-4405(00)00051-0</p><p>A Prospective Longitudinal Study ofHigh School Dropouts</p><p>Examining Multiple Predictors Across Development</p><p>Shane JimersonUniversity of California</p><p>Byron Egeland, L. Alan Sroufe, and Betty CarlsonUniversity of Minnesota</p><p>Prior studies report a variety of demographic, school, individual, and family charac-teristics that are related to high school drop out. This study utilizes data from a 19-year prospective longitudinal study of at-risk children to explore multiple pre-dictors of high school dropouts across development. The proposed model of drop-ping out emphasizes the importance of the early home environment and the qualityof early caregiving influencing subsequent development. The results of this studydemonstrate the association of the early home environment, the quality of earlycaregiving, socioeconomic status, IQ, behavior problems, academic achievement,peer relations, and parent involvement with dropping out of high school at age 19.These results are consistent with the view of dropping out as a dynamic develop-mental process that begins before children enter elementary school. Psychosocialvariables prior to school entry predicted dropping out with power equal to later IQand school achievement test scores. In our efforts to better understand processesinfluencing dropping out prior to high school graduation, early developmental fea-tures warrant further emphasis. 2000 Society for the Study of School Psychology.Published by Elsevier Science Ltd</p><p>Keywords: Dropouts, Development, Longitudinal study, Early caregiving, Home en-vironment, Parent involvement, Behavior problems, IQ, SES, Academic achieve-ment, Peer relations.</p><p>The seriousness of the drop out problem among American youth is welldocumented including both personal and societal implications (Cairns &amp;Cairns, 1994; Rumberger, Ghatak, Poulos, Ritter, &amp; Dornbusch, 1990). Theestimated 3.4 million nongraduating youth (National Center for EducationStatistics, 1994) are ill-equipped for the modern work force, thus ultimatelypaying less tax, adding costs to welfare programs, and being disproportion-ately represented in crime and incarceration statistics (Kirsch, Jungeblut,</p><p>Received August 23, 1999; accepted February 1, 2000.Address correspondence and reprint requests to Shane R. Jimerson, University of Califor-</p><p>nia, Graduate School of Education, 2208 Phelps Hall, Santa Barbara, CA 93106-9490. E-mail:Jimerson@education.ucsb.edu</p><p>525</p></li><li><p>526 Journal of School Psychology</p><p>Jenkins, &amp; Kolstad, 1993; Rumberger, 1987). Nearly a decade ago, the an-nual financial cost of the drop out problem was estimated at $240 billion(Dryfoos, 1990).</p><p>In some ways, dropping out is no longer mysterious. Five decades of re-search have uncovered numerous correlates of withdrawal from highschool. Prior research highlights various demographic status variables, indi-vidual characteristics, psychological and behavioral measures, and familyfactors associated with high school drop out (Rumberger, 1987, 1995;Rumberger et al., 1990). They are now well-known but not always useful.Demographic factors such as low socioeconomic status (SES), neighbor-hood-level variables, gender, ethnic minority status, and low parental edu-cation, for example, are consistently found to be related to school with-drawal (Cairns, Cairns, &amp; Neckerman, 1989; Ensminger, Lamkin, &amp;Jacobson, 1996; Fine, 1989; Oakland, 1992; Weis, Farrar, &amp; Petrie, 1989).However, such broad status variables leave considerable variance unex-plained and are not very informative with regard to the processes of drop-ping out. Of course, achievement problems and failing grades also arestrong correlates (Borus &amp; Carpenter, 1984; Ekstrom, Goertz, Pollack, &amp;Rock, 1986; Ensminger &amp; Slusarick, 1992; Garnier, Stein, &amp; Jacobs, 1997;Lloyd, 1974, 1978), but these may be viewed as early indicators of droppingout itself rather than as root causes.</p><p>In contrast to the above status variables, other studies have identifiedmore direct behavioral influences associated with drop out status such asmeasures of behavior problems, poor peer relationships, and certain familyvariables (Cairns &amp; Cairns, 1994; Cairns, Cairns, &amp; Neckerman, 1989; Ens-minger &amp; Slusarick, 1992; Feldhusen, Thurston, &amp; Benning, 1973; Garnier,Stein, &amp; Jacobs, 1997; Parker &amp; Asher, 1987). These measures (most ob-tained in middle school or late elementary school) have predicted laterdropping out quite well, often with some specificity. For example, Cairnsand Cairns (1994) found that association with others on the pathway todropping out increased the likelihood that a student would drop out. Stud-ies that include family factors have isolated variables such as parentalschool involvement, monitoring of the child, quality of parentchild inter-actions, and family lifestyles and values (Alpert &amp; Durham, 1986; Brooks-Gunn, Guo, &amp; Furstenburg, 1993; Dornbusch, Ritter, Leiderman, Rob-erts, &amp; Fraleigh, 1987; Garnier, Stein, &amp; Jacobs, 1997; Morris, Ehren, &amp;Lenz, 1991; Rumberger, 1995). However, most of these studies relied onquestionnaires or interview data, and with the exception of the Garnier etal. (1997) study, none of them began in the early years.</p><p>It is also the case that many of the factors predicting dropping out areinterrelated. Peer problems, behavior problems, and achievement prob-lems are strongly correlated with each other. Therefore, sorting out theircausal role in later behavior is challenging. Moreover, the predictors ofdropping out have known antecedents. For example, observed quality of</p></li><li><p>527Jimerson et al.</p><p>care in infancy and early childhood has well-demonstrated links to peerrelationships, behavior problems, and achievement in elementary schooland high school (Carlson et al., 1998; Elicker, Englund, &amp; Sroufe, 1992;Renken, Egeland, Marvinney, Mangelsdorf, &amp; Sroufe, 1989; Sroufe, Ege-land, &amp; Carlson, 1999; Teo, Sroufe, &amp; Carlson, 1996). Thus, what are cur-rently considered to be predictors may be midpoint markers of a lengthydevelopmental pathway to dropping out.</p><p>Rumberger (1987, 1995) provided a review of the literature regardingdropouts and concludes that although it is useful, the literature, in general,has several shortcomings. First, few studies have examined multiple influ-ences of dropping out in a comprehensive fashion (e.g., individual, family,and peer effects). Second, the longitudinal component often is very briefsuch as from an early high school period to a later high school period. Fi-nally, many of the factors (especially those related to family background)focus on structural characteristics rather than process (i.e., one-parent fam-ilies versus parentchild interactions). Rumberger (1987, 1995) arguedthat further research efforts should focus on developing multivariate, longi-tudinal, and comprehensive models of the causes of dropping out.</p><p>The present study was designed to further explore the developmentalprocesses and precursors that lead to dropping out of high school. The pur-pose of this study was to examine developmental adaptation and the factorsthat influence adaptation that may lead to a pathway toward dropping out.Lower achievement, poorer peer relations, behavior problems, and lessparent involvement during elementary school have all been associated withstudents who later drop out (relative to students who graduate). Prior re-search illustrates the influence of the family on each of these elementaryschool factors. However, most current literature examining early predictorsof dropping out of high school has been limited to information from ele-mentary school and high school. In an effort to better understand trajector-ies toward dropping out, developmental features prior to elementaryschool were examined. This study provides a conceptual model of drop-ping out as a dynamic developmental process that begins before childrenenter elementary school. From a developmental perspective, peer rela-tions, behavior problems, and academic achievement do not simply appearin elementary school; instead, development in each of these domains is re-lated to prior development. Dropping out is hypothesized to occur as a re-sult of current circumstances and prior development (i.e., experiences andadaptations).</p><p>We have been guided by a developmental transactional model (Samer-off, 1992; Sroufe, Egeland, &amp; Carlson, 1999). In this perspective, early de-velopmental history is given some priority, not because it ineluctably causeslater outcomes, but because what the child takes forward from these experi-ences, in part, frames subsequent transactions with the environment. Be-havior always is a product not only of current circumstances but also of de-velopmental history. There is a homeorhetic tendency in development</p></li><li><p>528 Journal of School Psychology</p><p>Figure 1. Variables examined in this study of dropping out as a developmental process.</p><p>(Sameroff &amp; Fiese, 1989) such that when a pathway is enjoined, numerousfactors conspire toward its continuation. Individuals engage, select, and in-terpret experiences within the previously established framework, so thatlater experiences often support prior experiences. As other researchers ofdrop out have proposed,</p><p>of special importance is the developmental nature of the patterns, so thatevents occur which then set other events in motion. This snowballing effect isone that will have to be mapped out in order to intervene at critical points intime before a trend or development has moved into a less malleable phase.(Evans &amp; DiBenedetto, 1990, p. 68)</p><p>Dropouts present a variety of profiles; however, certain early characteristicsmay increase the likelihood that a student will drop out. The hypothesis ofthis investigation is that ones early developmental history (e.g., quality ofcaregiving and home environment) will predict not only drop out anteced-ents (i.e., academic achievement, problem behaviors, peer relations, andparent involvement), but dropping out of high school itself (see Figure 1).Although there is converging evidence illustrating the role of the familyupon decisions to drop out, no longitudinal studies have illustrated the im-portance of the early quality of caregiving and home environment for sub-sequent school withdrawal. By focusing on earlier characteristics than priorstudies afforded, the power of family factors will be examined.</p></li><li><p>529Jimerson et al.</p><p>METHODS</p><p>Participants</p><p>The sample consisted of 177 children and their families from an originalsample of 267 first-time pregnant women in their last trimester of preg-nancy who were followed from the birth of their child through age 19 (Ege-land &amp; Abery, 1991; Egeland &amp; Brunnquell, 1979). The sample was at riskdue to poverty and associated risk factors such as age, education, and singleparenthood. All had received prenatal care through public assistance at theMaternal and Infant Care Clinic of the Minneapolis Health Department.To examine the antecedents of school drop out, data collected at age 1through age 19 was used. A primary concern with longitudinal research isparticipant attrition. For the period of 12 months of age to kindergarten,22 subjects left the study. Comparing the group of 22 to those who re-mained in the study did not yield any differences on race, mothers age, ed-ucation, or occupation. During the period first grade to age 19, 22 partici-pants left the study. Comparing the group of participants who left the studywith the remaining sample of 177 revealed no differences on the basic de-mographics including race, mothers age, education, or occupation at thetime of the childs birth.</p><p>The students were classified in terms of their high school graduation sta-tus at age nineteen. The classification was determined by reviewing formscompleted by the high school counselors, high school teachers, annual in-terviews with the students, and phone calls to verify academic enrollmentat age 19. In this study, students who were not enrolled in any form of aneducational program and were not making progress toward a high schooldiploma or GED were classified as Dropouts (n 5 43). Students who werecurrently enrolled in full-time attendance and making progress toward ahigh school diploma within a traditional setting or had graduated were clas-sified as Traditional students (n 5 100). Some students withdrew from atraditional educational program and entered alternative programs to con-tinue their education (n 5 34); because these students did not fit the crite-ria for either of the above groups they were not included in this study.</p><p>Within this subsample (n 5 143), mothers age at the time of birthranged from 1624 years (M 5 19.90; SD 5 2.08) for the Dropouts, and1534 years (M 5 21.01; SD 5 4.04) for the Traditional students. Educa-tional attainment of the mothers in both groups ranged from junior highschool to college. Whereas 28% of the mothers of Dropouts had not com-pleted 12 years of education at the time of the babys birth, 33% of themothers of Traditional students had not completed 12 years of educationat the time of the babys birth. For the Dropouts, 63% had Caucasian par-ents, 14% had African American parents, and 14% were of mixed race. Forthe Traditional students, 68% had Caucasian parents, 8% had African</p></li><li><p>530 Journal of School Psychology</p><p>American parents, and 21% were of mixed race. The only significant differ-ence regarding the composition of these two groups was that only 28% ofthe dropouts were females (i.e., 12 of the 43), whereas 52% of the Tradi-tional students were females.</p><p>Of 97 Caucasian students in this subsample, 28% dropped out, and of 29mixed-race children, 24% dropped out. In contrast, 46% of the AfricanAmerican children dropped out. The disproportionate percentage of Afri-can American students who dropped out relative to Caucasian studentsfound in this study is consistent with prior research (McMillen, Kaufmen,Hausken, &amp; Bradby, 1993).</p><p>Measures and Procedures: Family Factors</p><p>Early quality of caregiving composite. This composite included (a) a rat-ing of overall maternal sensitivity (an observation of play and two observa-tions of feeding) at age 6 months, (b) the quality of the infant and motherattachment relationship at age 12 and 18 months, and (c) the structure andlimits set by the mother on a series of tasks and the quality of the instructionprovided by the mother on these same tasks at age 42 months. This compos-ite is not only a measure of the caregiver, but also reflects the dyadic rela-tionship established between the caregiver and infant. Z-score transforma-tions were performed, and the three components were combined andaveraged. Each of the three initial variables is described below.</p><p>Overall maternal sensitivity. At age 6 months, mothers were observed intwo feeding situations and one play situation with their infants. Sensitivityratings were made for each session and an overall maternal sensitivit...</p></li></ul>


View more >