Dropping Out of High School among Mexican-Origin Youths: Is Early Work Experience a Factor? 

ANANE N. OLATUNJI 
University of Florida, Gainesville 
aolatunji@coe.ufl.edu
olatunji4@yahoo.com

In this article, Anane Olatunji examines the effects of work experience on early high school attrition among Mexican-origin adolescents. He proposes a theoretical model that takes assimilation into account as a potential predictor of the consequences of work for this group. In order to estimate the effects of eighth-grade work experience on dropping out of school in 1990, he analyzed data from the National Education Longitudinal Survey of 1988. He begins this article with a brief review of literature on high school dropouts and youth employment. He argues that this topic is particularly salient for Mexican-origin youths because they not only comprise the majority of Hispanics, the nation  largest ethnic minority group, but also exhibit an alarmingly low high school completion rate. Overall, the results support conventional models that researchers have used to predict outcomes of teenage employment. Among Mexican-origin adolescents, however, girls were three and a half times more likely than boys to leave school, after controlling for work and other factors. He concludes the article with implications for future research, especially the role that gender plays in predicting early high school attrition among Mexican-origin youths. 

Since the mid-1980s, social scientists have addressed the concern that the U.S. economy’s need for part-time, low-skilled, minimum-wage labor might spur adolescents into employment prematurely, thereby undermining their emotional well-being as well as their investment in education (Carr, Wright, & Brody, 1996; D’Amico, 1984; Green, 1990; Marsh, 1991; Steel, 1991; Steinberg, Fegley, & Dornbusch, 1993). The National Research Council (1998) summarized the literature on this topic, reporting that 44 percent of 16- and 17-year-olds hold jobs, and two out of every five seventh and eighth graders work during the school year. The council also called attention to the potential risks associated with youth employment, reiterating the fact that the benefits parents and policymakers so often attribute to adolescent employment — better time management and work habits, improved self-esteem, greater responsibility, and an appreciation for the value of both education and work (Marsh, 1991; Phillips & Sandstrom, 1990; Steel, 1991) — may in fact be erroneous. 

If education is the gateway to employment opportunity, then a high school diploma is the sine qua non of upward social mobility. Consequently, it remains especially important that we understand the effects of working while in school on educational attainment. If early work experience enhances educational outcomes, then maybe adults should encourage teenagers to pursue employment. If, on the other hand, managing a job in addition to school reduces academic performance, then parents and policymakers should discourage youth employment. Despite considerable interest in the effects of adolescent employment, few researchers have paid attention to how this issue relates specifically to Latino youths. The paucity of such information raises concern, given that Latino youths appear to enter the labor force relatively early, compared to non-Latinos (Barton, Farkas, Kushner, & McCreary, 1985; Johnson & Herring, 1989). This study examines this salient topic as it relates to Mexican-origin youths for several reasons.  

In the first place, people of Mexican descent, who comprise the majority (58.5%) of the U.S. Hispanic population (U.S. Census, 2001, p. 3), exhibit not only the lowest level of education among the three major Latino subgroups (Mexicans, Puerto Ricans, and Cubans), but also the lowest of any major ethnic group in the nation as a whole (U.S. Bureau of the Census, 1998, p. 167). In 1997, only 54.7 percent of all Hispanics and 48.6 percent of Mexican-origin individuals over the age of twenty-five had completed four years of high school, compared to the national average of 82.1 percent (U.S. Bureau of the Census, 1998). Rumberger (1995), in his study on middle school dropouts, insisted that this disparity is cause for concern because those who do not finish high school “experience higher levels of unemployment, receive lower earnings, . . . are more likely to have health problems, engage in criminal activities, and become dependent on welfare ” (p. 584). Importantly, his study also revealed that half of the Hispanic males who drop out of high school do so by the ninth grade. Might adolescent work experience be an unrecognized culprit that encourages Mexican-origin youths to drop out of school?  

The mental health literature suggests that Mexicans who come to the United States in search of employment may suffer from a trade-off, exchanging psychological well-being for the jobs and wages they procure (Brindis, Wolfe, McCarter, Ball, & Starbuch-Morales, 1995; Diaz & Draguns, 1990; Guarnaccia, 1997). If this is so, then Mexican adolescents who work may be especially at risk of psychopathology, which in turn may undermine their educational achievement. For example, some researchers link early work experience among the general population to poor performance in school, depression, delinquency, and drug use (Bachman & Schulenberg, 1993; Mortimer, Finch, Shananhan, & Ryu, 1992; Shanahan, Finch, Mortimer, & Ryu, 1991; Steinberg & Dornbusch, 1991). 

Finally, Hispanics not only comprise the fastest growing minority group in the United States, but also have superseded African Americans as the nation’s  largest ethnic minority group.¹ If these trends stay on course, one in every five people living in the United States could be Hispanic by the middle of the 21st century (National Association of Hispanic Publications, 1995). Consequently, the extent to which early work experience constitutes a benefit or detriment to the well-being of youths in this rapidly expanding segment of the population could have broad social and economic repercussions.  

Does early work experience undermine or promote the likelihood of finishing high school among Mexican-origin youth? I posit that traditional cultural values may alter the effects of early employment on high school completion for Mexican-origin youth. To test this idea, I conduct two sets of analyses. The first compares all youths of Mexican descent to their non-Hispanic White counterparts. The second compares Mexican-origin youths who have retained their culture, as measured by their speaking Spanish at home, to those who have not, that is, those who usually speak English at home. Each set of analyses consists of six models: Models 1 and 2 examine the effects of background factors on dropping out of school by tenth grade; Models 3 and 4 explore the nature of the relationship between work experience and dropping out of school; and Models 5 and 6 confirm how work experience affects dropping out of school by tenth grade. I employ the logistic regression statistical methodology and report findings based on data supplied by the subsample of 12,700 participants in the National Educational Longitudinal Survey from 1988 to 1990. 

Theoretical Framework and Hypotheses 

Researchers have studied the question of whether early work experience promotes or impedes adolescent well-being principally from two theoretical orientations: the developmental and zero-sum models (Bachman & Schulenberg, 1993; Greenberger & Steinberg, 1986; Marsh, 1991; Steinberg et al., 1993). Developmental models view adolescent employment positively, as an opportunity for young people to learn the skills, attitudes, and habits necessary to participate successfully in the workplace. This perspective contends that young people will transfer the benefits they derive from work to other areas of their lives, including school. In this manner, work speeds their maturation into responsible adults. Most youth employment policy initiatives rest upon the developmental orientation (Steel, 1991). In contrast, zero-sum models view the issue of youth employment in terms of priorities, whereby other activities take precedence over academic interests. Accordingly, time invested in employment subtracts from time that could be devoted to school activities (Marsh, 1991). Therefore, one can expect the quantity of hours worked weekly (work intensity) to be inversely related to educational attainment. Articulating a zero-sum perspective, D’Amico (1984) suggested that the rise in high school employment accounts in part for the decline in educational performance of U.S. students. Note, however, that this perspective makes the rather questionable assumption that the time teenagers invest in jobs would actually be spent on school-related activities if they had no job. 

Both developmental and zero-sum orientations presume that schooling is the normative experience for adolescents. In other words, these models are built on the premise that schooling is congruent with the social and emotional development of the average American youth. This congruence between adolescence and schooling reflects the relatively advanced standard of living in our wealthy, postindustrial society, which increasingly demands that teenagers invest time in education as a way to acquire an economically secure future. However, in Mexico, a developing country, the pursuit of education often is a luxury, while work usually is an economic necessity (Martin, 1994; Parker & Pederzini, 2001), as evidenced by the current and unrelenting stream of migration to the United States. In light of such disparate social and economic contexts, neither the zero-sum nor the developmental perspective would seem to provide an appropriate framework upon which to base predictions about the effects of work experience on Mexican-origin youths.  

One may argue, however, that work experience appears to be the rule rather than the exception for most U.S. adolescents, as studies demonstrate that the majority of U.S. teenagers work at some point prior to adulthood (National Research Council, 1998). But in terms of the two theoretical perspectives that scholars rely on for predicting the effects of youth employment, schooling supersedes work as the primary responsibility of adolescents. For example, in discussing the school-work dilemma, Shanahan et al. assert, “Clearly, school, not work, is widely considered to be the central ‘business’ of the adolescent’s life” (1991, p. 300). Greenberger and Steinberg (1986) echo this sentiment, explaining that recent changes in the American social and economic context render the workplace deleterious to adolescent development. These suppositions stem from the American social context, so it is doubtful that they apply equally to youths of Mexican origin.  

An examination of the role work plays in Mexican culture suggests that holding a job may carry a substantially different meaning and thus have different consequences for youths of Mexican descent than for the average American adolescent (Carraquillo, 1991; Greenberger & Steinberg, 1986; Suárez-Orozco, 1995). It is well known that secondary education is not the predominant experience for Mexican youth, as it is in the United States (World Bank, 2000, p. 241). Mexican adolescents typically seek employment after completing primary schooling, hence, work — not school — is their normative experience.  

How then might work and its consequences differ for Mexican-origin youths? The traditional cultural values still intact in many Latino families dictate that both male and female children share the responsibility of supporting the family (Carraquillo, 1991; Suárez-Orozco, 1985; Zambrana, 1995). Traditionally, taking a job at an early age fulfills this cultural expectation of mutual support among family members. In fact, finding work to improve the family’s standard of living is one of the primary reasons that many Mexican youths accompany their parents to the United States (Massey, 1987). Therefore, through employment youths of Mexican descent may derive a special sense of accomplishment from fulfilling their familial responsibility. In contrast, non-Hispanic youths, according to Greenberger and Steinberg (1986), may view jobs merely as a way to earn money to obtain material possessions for no other reason than personal gratification. They argue, for example, that some teenagers use their earnings to buy drugs and alcohol, spawning delinquency rather than achievement.  

The literature on the effects of early work experience in general tells us that teenage employment may have a negative impact on performance inside as well as outside of the classroom, thwarting rather than developing positive attitudes and behaviors among adolescents (Pallas, 1993; Shanahan et al., 1991; Steinberg & Dornbusch, 1991). Interestingly, high levels of work intensity (more than 20 hours per week) measured in terms of hours worked per week appear to be inversely related to educational outcomes, while moderate levels (less than 15 hours per week) seem to enhance them (Marsh, 1991; Mortimer et al., 1991; Stern et al., 1990). However, Bachman and Schulenberg (1993) cautioned that a selection factor may confound the relationship between intensity and attainment, whereby youths who perform poorly in school opt to work longer hours as an alternative route for success. In other words, the observed correlation between teenage employment and poor performance in school is simply a function of the kind of students who decide to work long hours rather than an effect of work intensity itself. 

In a study of at-risk youths, Chaplin and Hannaway (1999) underscored both positive and negative employment consequences, reporting high school employment to be associated with higher dropout rates, but also with higher earnings in later years. Further, they asserted that at-risk status — that is, low educational achievement and low parental support during high school — account for differences in the effects of early work experience. In terms of earnings, Chaplin and Hannaway found that at-risk students benefited more from work experience as much as eleven years later and despite relatively high dropout rates. Consequently, their study suggests that working during high school may be more beneficial for at-risk students in the long run. McNeal (1997) found similar positive and negative effects on dropping out of high school, based on work intensity as well as employment type. 

Of particular interest to this study, Steel (1991) observed that ethnicity makes a difference in how the advantages and disadvantages of employment are dispensed: “Hispanic and Black youths, who encounter barriers to socioeconomic attainment in general, do not appear to benefit from early work experience to the same extent or in the same way as Whites — and may in fact suffer in terms of subsequent enrollment” (p. 444). Specifically, among youths aged seventeen to eighteen, employment status and the number of hours worked weekly both were positively associated with subsequent enrollment in school for Whites but not for Hispanics. Consistent with Steel’s observation, Carr et al. (1996) reported that adolescents holding jobs during high school tend to be White males from intact, middle-income families whose parents are well educated. It seems that the opportunity to work is yet another benefit of advantaged youths, as opposed to those who are most in need financially. 

Hypotheses 

This study, then, addresses several important questions: What are the consequences of early work experience for Mexican-origin youths in terms of high school completion? Do they differ from those of their White youth counterparts? And what about within-group differences? That is, are the effects of early employment the same for all youths of Mexican descent? If the meaning and motives of work differ for youths based on ethnicity and culture, then the patterns observed in the literature, which pertain to the consequences of work experience among adolescents in general, may vary for Mexican-origin youths, especially among those who are less assimilated into U.S. society. The retention of group cultural identity may in fact reduce the deleterious effects of assimilation (Partida, 1996; Portes & Rumbaut, 1996; Suárez-Orozco, 1985). Consequently, I argue that traditional Mexican cultural values instill different motivations for working, which in turn mitigate or protect adolescents from the negative consequences of employment observed among American adolescents in general. And the less assimilated these youths are, the greater the protective effect of such cultural values. Therefore, I expect that Mexican-origin adolescents who retain their traditional culture will be less likely to suffer the negative consequences of early work experience as described in the literature. Offering further support to this argument is the research linking immigrant status to positive educational outcomes for Hispanics and other groups (Driscoll, 1999; Wojtkiewicz & Donato, 1995). As a central hypothesis of this investigation, therefore, I propose that work experience will make youths of Mexican descent less likely to drop out of high school than Whites. As a corollary, work experience will reduce the odds of dropping out of high school for Mexican-origin adolescents who are less assimilated into the dominant American culture than other youths of Mexican descent.  

Data and Methodology 

To examine the consequences of youth employment on high school completion, I extract data from the National Education Longitudinal Study of 1988 (NELS:88). Sponsored by the National Center of Education Statistics for the U.S. Department of Education, this survey collected extensive information on the social and educational conditions of a nationally representative sample of 24,599 eighth-grade students in the United States during the spring of 1988. The original NELS cohort was subsampled and interviewed in 1990, 1992, 1994, and 2000. Participants were identified during the base year using a two-stage sampling procedure that randomly selected one thousand schools and then selected clusters of twenty-five students per school. The design also over-sampled Hispanics. For all these reasons, NELS remains one of the best sources of data available for studying the effects of early work experience in particular.  

Sample 

I extracted respondents who participated in the first two waves of NELS, resulting in an initial pool of 13,649 cases, including 1,412 (10.3%) youths who identified themselves as Mexican or Chicano, and 12,237 (89.7%) who categorized themselves as non-Hispanic Whites. After a listwise deletion of missing data, however, this initial sample was reduced by 7 percent to a final sample of 12,700 cases, of which 1,299 (10.2%) were of Mexican origin and the remainder (11,401, or 89.8%) were non-Hispanic Whites. A bias check suggested that missing data were not randomly distributed, but tended to be concentrated among respondents with lower socioeconomic status, as well as among those of Mexican origin. Consequently, the results of the study may underestimate differences between the two ethnic groups in question.  

Dependent Variable 

To determine the dropout status of individual respondents in 1990, I used a variable created by NELS:88 as the result of an intensive follow-up that NCES conducted to ascertain the enrollment status of respondents. Based on guidelines set by the U.S. Department of Education, a dropout was defined as anyone who missed at least four consecutive weeks of school in spring not due to illness or accident (National Center for Education Statistics, 1992, p. 76). In these data, there were four categories in which respondents fell on this variable: 0 = did not drop out; 1 = dropout, but returned; 2 = dropout, no return; and 3 = more than 1 episode of dropping out. I collapsed the last three categories to create a dichotomous dependent variable (0 = did not drop out, 1 = dropped out). Even in cases where respondents dropped out but returned to school, they presumably remained at risk of not finishing high school.  

Independent Variables 

I operationalized the concept of work experience using two original variables from the dataset: hours worked per week and type of work. The first variable measures work intensity based on self-reports of how many hours a respondent worked for pay per week on the current or most recent job as of the interview. I transformed the original ordinal level variable for 1988 into a ratio one by computing interval midpoints. The new variable acquired the following values: 0 = no hours (representing those who reported not working), 2.5, 7.5, 15.5, and 25.5 hours. The second variable indicates the type of work respondents did for pay in 1988. I created a dichotomous variable, work status, where zero (0) represented those who did not work, and one (1) represented those who worked, by collapsing all job categories. 

Control Variables 

Other factors must be controlled statistically that otherwise may produce spurious effects between the independent and dependent variables. Consequently, I selected a parsimonious set of variables known to be associated with high school completion and whose variance inflation factors, as indicated by an F-test, produced no evidence of multicollinearity (see Fox, 1991; McClendon, 1994). I controlled for several individual and background characteristics including Mexican culture, sex, prior academic performance (operationalized as self-reported grade point average), and socioeconomic status (operationalized as a standardized measure consisting of parental education and occupation, and family income as reported by parents). Using dummy variables, I also controlled for job types.  

Because in this study I argue that Mexican culture may influence the effects of work experience, I used the language spoken at home (self-reported) as a proxy for this key concept. Given that language is a primary aspect of culture, I argue that those respondents who have retained the Spanish language are also more likely to have retained other aspects of traditional culture, including the meaning of work as explained in this study. Hence, Mexican-origin respondents that reported “English dominant or English only” as their home language were coded as a dummy variable “English,” while those that reported “non-English only or non-English dominant” were coded as a dummy variable “Spanish,” and non-Hispanic Whites were the reference category. Implied in this coding scheme, then, is the idea that Mexican-origin youths whose home language is primarily Spanish presumably have assimilated less into the dominant Anglo culture than their primarily English-speaking Mexican-origin counterparts, who in turn, and despite their obvious assimilation, still may manifest characteristics of traditional Mexican culture and thus be culturally distinct from Whites. In other words, although all youngsters of Mexican descent may share a common ethnic background, it is the process of assimilation that ultimately determines how much of that traditional background they retain. Hence, operationalizing culture in this fashion enabled me to test for effects of assimilation, which in turn may influence Mexican-origin youths’ motivations for working and thus its consequences for educational attainment. Two other methods for testing this cultural hypothesis would be to divide the sample of Mexican-origin youths on the basis of nativity (foreign born versus U.S. born) or by generational status (based in part on nativity), comparing first, second, and third generations. Unfortunately, despite the oversamples of Hispanics, preliminary analyses revealed that using either of these measures would preclude a statistical analysis of the central hypothesis, as there are not enough data for these variables in relation to the measures for dropouts and work experience.  

With concern for the threat of selection effects, I employed two key measures. First, I controlled for personal educational expectations (the highest educational level a respondent expected to complete as of 1988). In addition, I applied a more robust indicator of at-risk status than did Chaplin and Hannaway (1999), who contended that at-risk status explains ethnic differences in work effects. Specifically, the NELS measure indexes six factors associated with dropping out of school: single-parent family, low parental education, sibling dropout, three or more hours at home alone, limited English proficiency, and low family income. Controlling for these factors enabled me to examine the effects of work experience, net of respondents’ educational aspirations and risk of dropping out, thus strengthening the ability to infer a causal link between the criterion and predictors.  

Analytic Strategy 

I divided the analyses into two sets in order to determine if work experience had a more positive effect on high school completion among youths of Mexican descent who retained their ancestral culture when compared to Whites and compared to other youths of Mexican descent. The first set compared the two home language groups (English and Spanish) of Mexican-origin adolescents to Whites, whereas the second set consisted only of Mexican-origin respondents and compared those who retain Spanish as their home language to those who use English. Because the second set of analyses compared the two subgroups of Mexican-origin respondents to each other rather than to Whites as in the first set of analyses, results from the former set may differ from the latter. 

Each set of analyses consisted of six logistic regression models, given the dichotomous dependent variable (see Menard, 1995). The first two models tested whether any difference in the log odds of dropping out of tenth grade among Mexican-origin and White youths can be explained by background factors. Models 3 and 4 tested 1) whether work experience predicted the log odds of dropping out of tenth grade, once background and certain selection factors were accounted for; and 2) whether the relationship between working and dropping out was linear. Finally, Models 5 and 6 tested the hypothesis that work experience promotes high school completion among Mexican-origin youths, after other important factors are taken into account. 

The methodology also adjusts statistical estimates for the survey design effects. The design effect is the ratio of the standard error of an estimate, calculated by taking into account the survey design compared to that for the estimate based on a simple random sample. A multistage sample design produces statistics that are more variable than they would have been if the design had been a simple random sample of the same size. Consequently, the standard errors used in calculating statistical significance of regression coefficients should be adjusted in light of this variability. To adjust for the conservative standard errors, I conducted the analyses using AM software, which was developed expressly for this purpose by the American Institutes for Research. In addition, I used the NELS-created panel weight for the base year and first survey follow-up, making the results of the study generalizable to those Anglo and Mexican-origin youths in 1990 who, in 1988, were eighth-grade students in the United States.  

Finally, because the methodology includes strategies to strengthen the case for making causal inferences — specifically, a longitudinal design and control for spurious factors correlated with the predictors and outcome measures — I interpret the results of the analyses using causal rather than correlational language. However, in social science one cannot know for certain that the analysis identifies all possible spurious factors. Consequently, the reader should be mindful of this limitation with respect to the interpretation of the findings.  

Findings 

Table 1 displays the weighted characteristics of the final sample. As one might expect, youths of Mexican origin differ significantly from non-Hispanic Whites in almost every respect. Particularly noteworthy is that a majority of adolescents in both groups — more than half (54.6%) of the Mexican-origin respondents and nearly three quarters (74.3%) of Whites — reported having worked for pay by the eighth grade. This finding suggests that work experience is just as prevalent among younger adolescents as it is among older ones. Also worthy of mention is that among Mexican-origin youths, 60 percent of those reporting work experience were boys, whereas among Anglos, work experience was divided equally between boys and girls. Respondents who reported having work experience typically worked roughly eight hours per week, on average. A majority of adolescents performed such traditional adolescent jobs as babysitting, lawn care, and delivering newspapers. Less popular jobs included working on farms or other manual labor, such as clerking in stores or offices and waiting on tables. Many other youths worked in a variety of unspecified types of jobs (not shown).  

The data also reveal notable differences in educational characteristics between Mexican-origin and White youths. Stunningly, by the eighth grade, 72 percent of Mexican-origin youths were at risk of becoming high school dropouts. Only half as many Whites (36%) reported having one or more of the six factors associated with high school attrition. On average, Mexican-origin youths also had significantly lower grade point averages and educational expectations than their White counterparts, although these differences did not appear to be large in absolute terms. The clear exception, however, was in the case of early high school dropouts, where the proportion of Mexican-origin youths who left school by the tenth grade was more than twice as large as that of Whites (11.7% versus 5%). In stark contrast to Rumberger’s (1995) finding regarding male dropouts among Hispanics in general, these data reveal that females accounted for almost 70 percent of the dropouts among Mexican-origin adolescents by the second year of high school. Among Whites, however, the proportion of dropouts for girls was about eight percentage points lower than that for boys (46% versus 54%). 

Does early work experience increase the likelihood of completing high school among Mexican-origin adolescents, compared to Anglo youths? The results from these analyses produced no evidence to support the hypothesis that working reduces the likelihood of dropping out of high school among Mexican-origin youths. To begin, Model 2 in Table 2 clearly demonstrates that the greater likelihood among Mexican-origin youths of dropping out of school by tenth grade, compared to Whites as displayed in Model 1, can be explained by prior academic performance, individual educational expectations, socioeconomic status, and other background factors associated with high school attrition. Consistent with previous research on older adolescents, the positive and statistically significant coefficient for hours worked per week in Model 3 indicates that among eighth graders, work intensity increases the log odds of dropping out of school by tenth grade, net of background factors. Specifically, for each additional hour worked per week, the log odds of dropping out of school by the tenth grade increase on average by a multiplicative factor of .03 (p < .005). Expressed alternatively as odds to demonstrate the size of this effect, in 1988, Anglo and Mexican-origin eighth graders who worked at the average level of intensity, about eight hours per week, increased their odds of becoming a dropout within the first two years of high school by nearly 27 percent, compared to their non-working counterparts. For those who worked twice that amount (about 15% of the sample), their odds of dropping out of school by tenth grade increased by about 60 percent. 

In combination with the negative but not statistically significant coefficient for work status, the work intensity coefficient suggests the possibility that the relationship between working and dropping out of school may be nonlinear, as mentioned in the literature. Adding a squared term to the model for hours worked per week tested for such a relationship. The outcome of the squared term (p < .40) in Model 4 indicates insufficient evidence that the relationship between the criterion and work intensity is nonlinear. As an additional check, I tested for nonlinearity using dummy variables for levels of work intensity (analysis not shown), but the outcome was similar to that for the squared term. 

Model 5 in Table 2 includes two product terms to test whether the effect of hours worked per week varies for either one of the two home-language groups of Mexican-origin respondents (Spanish/English home language times hours worked per week). That is, Model 5 directly tested the hypothesis that work experience decreases the likelihood of dropping out of high school among Mexican-origin adolescents, compared to Whites. As neither of the product terms was statistically significant, I conclude that among Mexican-origin adolescents, the effect of work intensity on high school attrition does not differ from that for White youths, and that effect is positive. That is, work intensity increases the likelihood of dropping out of high school regardless of ethnicity. 

Finally, Model 6 controls for job type, which previous studies have linked to outcomes of work experience. Work intensity retained its effect demonstrated earlier, while one of the job categories, delivering newspapers, actually exhibited a rather sizeable and negative effect on dropping out of school within two years (b = –.97, p  < .05). Although working more hours per week generally increased the likelihood of dropping out of school, eighth graders who worked delivering newspapers were, on average, nearly 40 percent less likely than their non-working peers to drop out of school within two years. This outcome would appear to be consistent with the developmental perspective on youth employment, which might argue that the relatively unsupervised nature of this type of work fosters independence and responsibility, which helps youngsters become more successful students. Subsequent findings indicated that this effect did not apply to Mexican-origin youths. 

Are the effects of work experience on high school completion more positive for Mexican-origin youths who are less assimilated into the dominant culture than for other adolescents of Mexican origin? These data did not produce any evidence that assimilation mediates the effects of early work experience on dropping out of school among adolescents of Mexican descent in the United States. The findings from the second set of analyses are generally consistent with those in the previous set that included Whites, except for one important and unexpected distinction: The results indicate that among eighth graders of Mexican descent, girls are far more likely than boys to drop out of school by the tenth grade (see Table 3).  

Model 1 in Table 3 demonstrates that there is no difference in the log odds of dropping out of school by the tenth grade, based on home-language background. In contrast, Model 2 illustrates that several other background factors predict dropout status. Particularly noteworthy is that female status emerged as a major predictor of dropout status among Mexican-origin youths, but it was not statistically significant in the first set of analyses that included White respondents. Turning attention to work experience, in Model 3, we observe that hours worked per week in 1988 had a positive and significant effect on dropping out of school within two years (p < .05), and that the coefficient for work status is negative but not statistically significant. Despite this hint of curvilinearity, however, Model 4 produces similar results as in the previous set of analyses, indicating insufficient evidence of a nonlinear relationship between hours worked per week and the dependent variable. Consequently, Model 5 omits the squared term and includes product terms to test for interaction effects between hours worked per week and both Spanish home language background and female status, but neither of these terms enhanced the predictive power generated by Model 3.  

Finally, after controlling for job type, Model 6 exhibited three notable findings. First, work intensity retained the size and direction of its effect on the criterion as demonstrated in Model 3. Net of all else, each additional hour per week that Mexican-origin adolescents worked in 1988 increased the log odds of dropping out of school by tenth grade by a multiplicative factor of .05, on average. Therefore, the typical eighth-grade youngster (of Mexican descent) who worked about eight hours per week in 1988 increased his or her odds of dropping out of school within the next two years by about 50 percent over their non-working peers. Second, working as a babysitter actually reduced the likelihood of dropping out of school. Specifically, youngsters who worked as babysitters reduced the log odds of dropping out of school by a factor of –1.15, on average, compared to their non-working peers (p < .01). Alternatively, babysitters reduced the odds of dropping out of school by tenth grade by about 68 percent, compared to their jobless peers. Perhaps performing this type of work for pay fosters the development of maturity and responsibility that translates into success at school for Mexican-origin youths because of the traditional importance they may attach to taking care of younger children and family. Third, net of all else, female status produced the largest effect on the likelihood of leaving school prematurely. After taking all other factors into consideration, the odds of dropping out of school by tenth grade were three and a half times greater for Mexican-origin girls than for boys (p < .0005). This finding may be the single most important outcome of this investigation.  

Conclusion 

Out of concern for the persistent problem of dropping out of high school among Latino youths in general, this study examined the effects of early work experience among Mexican-origin adolescents in particular. In contrast to the typical American teenager who presumably works for money to purchase commodities for personal gratification (Carraquillo, 1991; Greenberger & Steinberg, 1986; Suárez-Orozco, 1995), I’ve argued that traditional Mexican cultural values, which encourage teenagers to assume familial responsibilities, might alter the meaning and thus the outcome of teenage employment for adolescents of Mexican descent in the United States. This line of reasoning led to formulation of the hypothesis that Mexican-origin youths, unlike their Anglo counterparts, would benefit from working as it relates to high school completion. By the same token, Mexican-origin youths who are less assimilated into mainstream U.S. culture would profit more from work experience than other youths of Mexican origin.  

The findings from this investigation, however, produced little evidence to support the research hypotheses. Specifically, among Mexican-origin youths but not Whites, babysitting was the only type of work that reduced the likelihood of dropping out of school. All other jobs had no effect. The number of hours worked per week, on the other hand, increased the likelihood of dropping out for both Mexican-origin and Anglo youths. Overall, then, consistent with research on older adolescents, the results of this study indicate that, among eighth-grade students, work intensity, measured as hours worked per week, increases the odds of dropping out of school. Moreover, traditional Mexican cultural values or assimilation into mainstream U.S. culture, at least when measured in terms of home-language background, do not appear to alter this effect. Consequently, for non-Hispanic Whites and for youths of Mexican descent, it was estimated that in 1988 an average work load of eight hours per week increased the odds of dropping out of school within two years by 27 percent, and those who worked at least twice that amount — approximately 15 percent of all eighth graders in these two groups — increased their odds of leaving high school prematurely by 60 percent over youths without jobs.  

In sum, although cultural orientation does not appear to mediate the effects of some types of early work experience on high school completion for Mexican-origin youths, this investigation provides clear evidence that early work intensity is a consistent and negative factor with respect to high school completion among these youngsters. Unlike most studies on youth employment, however, this research examined the effects of work on dropping out by the tenth rather than twelfth grade. Given that adolescent development and work experience may change substantially during this period, future research should examine the effects of employment among older Mexican-origin youths in order to evaluate the effects of working compared to other youths in the U.S. Still, the evidence presented in this study clearly points to work intensity as a factor that increases the probability of dropping out during the early years of high school for youths of Mexican descent, as well as for Whites. Fortunately, how much a youngster works is something over which individuals may exercise control. As a logical outcome of these findings, one recommendation is that early adolescents limit the number of hours they work per week to the extent that their personal circumstances allow. 

The results of this study also seem to support both of the two major theoretical perspectives that social scientists have used to predict outcomes of teenage employment. In favor of the zero-sum model, the effect of work intensity on dropping out of school was positive and significant, even after controlling for personal and background factors, as well as the type of work that students performed. On the other hand, supporting the developmental perspective of youth employment, work delivering newspapers or babysitting reduced the odds of dropping out of school by tenth grade. From the developmental perspective, both of these jobs typically require that youngsters work unsupervised and therefore may foster the type of maturity, independence, and responsibility that enhance their persistence in school.  

Finally, this study has illuminated the fact that the early high school dropout problem among Mexican-origin youths is a female one, in contrast to the apparently male-dominant dropout problem suggested by Rumberger (1995) among Latinos in general. Specifically, we observed that among Mexican-origin youngsters, the percentage of females dropping out early was more than twice that of males, whereas among Whites the percentage of females was lower than that of males. Moreover, after controlling for confounding factors, female status was the most powerful predictor of tenth-grade dropout status among Mexican- origin adolescents. By the time they should have been in the tenth grade, girls comprised 70 percent of the dropouts among Mexican-origin youths who were eight graders in 1988. And after taking into consideration all other factors, the odds of early high school attrition were three and a half times greater for Mexican-origin females than males. While beyond the scope of this investigation, future studies clearly must give closer attention to this phenomenon in an effort to determine its causes as well as potential solutions.  

Note 

1.    The U.S. Hispanic population grew by 57.9 percent, from 22.3 million to 35.3 million, between 1990 and 2000. Over the same period, the Asian and Pacific Islander population grew by about 50.3 percent, from 7 million to almost 10.5 million. In comparison, the African American population grew by 16.2 percent, to 33.9 million, while the non-Hispanic White population rose by only 3.4 percent, reaching 194.5 million (Census 2000 PHC-T-1, 2001). 

References 

Bachman, J. G., & Schulenberg, J. (1993). How part-time work intensity relates to drug use, problem behavior, time use, and satisfaction among high school seniors: Are these consequences or merely correlates? Developmental Psychology, 29, 220–235. 

Barton, M., Farkas, G., Kushner, K., & McCreary, L. (1985). White, Black, and Hispanic male youths in Central City labor markets. Social Science Research, 14, 266–286. 

Brindis, C., Wolfe, A. L., McCarter, V., Ball, S., & Starbuch-Morales, S. (1995). The association between immigrant status and risk-behavior patterns in Latino adolescents. Journal of Adolescent Health, 17, 99–105. 

Carr, R., Wright, J. D., & Brody, C. J. (1996). Effects of high school work experience a decade later: Evidence from the National Longitudinal Study. Sociology of Education, 69, 66–81.  

Carraquillo, A. L. (1991). Hispanic children and youth in the United States: A resource guide. New York: Garland. 

Census 2000 PHC-T-1: Population by race and Hispanic or Latino origin for the United States: 1990 and 2000 [Online]. (2001). Available at PDF/Excel/ASCII: http://www.census.gov/population/www/cen2000/phc-t1.html File: Table 4. 

Chaplin, D., & Hannaway, J. (1999). High school employment: Meaningful connections for at risk youth. Washington, DC: Urban Institute. 

D’Amico, R. (1984). Does employment during high school impair academic progress? Sociology of Education, 57, 152–164. 

Diaz, J. O. P., & Draguns, J. G. (1990). Mental health of two-way migrants: From Puerto Rico to the United States and return. Migration World Magazine, 18, 34–37. 

Driscoll, A. K. (1999). Risk of high school dropout among immigrant and native Hispanic youth. International Migration Review, 33, 857–875. 

Fox, J. (1991). Regression diagnostics. Newbury Park, CA: Sage. 

Green, D. L. (1990). High school student employment in social context: Adolescents’ perceptions of the role of part-time work. Adolescence, 25, 425–434. 

Greenberger, E., & Steinberg, L. (1986). When teenagers work: The psychological and social costs of adolescent employment. New York: Basic Books. 

Guarnaccia, P. J. (1997). Social stress and psychological distress among Latinos in the United States. In I. Al Issa & M. Tousignant (Eds.), Ethnicity, immigration, and psychopathology (pp. 71–94). New York: Plenum Press. 

Johnson, R. J., & Herring, C. (1989). Labor market participation among young adults: An event history analysis. Youth and Society, 21, 3–31. 

Marsh, H. W. (1991). Employment during high school: Character building or subversion of academic goals. Sociology of Education, 64, 172–189. 

Martin, C. J. (1994). Schooling in Mexico: Staying in or dropping out. Brookfield, VT: Ashgate. 

Massey, D. (1987). Understanding Mexican migration to the United States. American Journal of Sociology, 92, 1372–1403. 

McClendon, M. J. (1994). Multiple regression and causal analysis. Itasca, IL: F. E.Peacock. 

McNeal, R. B. (1997). Are students being pulled out of high school? The effect of adolescent employment on dropping out. Sociology of Education, 70, 206–220. 

Menard, S. (1995). Applied logistic regression analysis. Newbury Park, CA: Sage. 

Mortimer, J. T., Finch, M., Shanahan, M., & Ryu, S. (1992). Adolescent work history and behavioral adjustment. Journal of Research on Adolescence, 21, 59–80. 

Mortimer, J. T., Shanahan, M., & Ryu, S. (1991). The effects of adolescent employment on school-related orientation and behavior. In R. K. Silbereisen & E. Todt (Eds.), Adolescence in context: The interplay of family, school, peers, and work in adjustment (pp. 304–326). New York: Springer-Verlag. 

National Association of Hispanic Publications. (1995). Hispanics-Latinos: Diverse people in a multicultural society. A special report. Washington, DC: Author. 

National Center for Education Statistics. (1992). National education longitudinal study of 1988, second follow-up: Student component data file user’s manual. Washington, DC: U.S. Department of Education. 

National Research Council. (1998). Protecting youth at work: Health, safety, and development of working children and adolescents in the United States. Washington, DC: National Academy Press. 

Pallas, A. M. (1993). Schooling in the course of human lives: The social context of education and the transition to adulthood in industrial society. Review of Educational Research, 63, 409–447. 

Parker, S. W., & Pederzini, C. (2001). Gender differences in education in Mexico. In E. G. Katz & M. C. Correia (Eds.), The economics of gender in Mexico: Work, family, state, and market (pp. 9–45). Washington, DC: World Bank. 

Partida, J. (1996). The effects of immigration on children in the Mexican-American community. Child and Adolescent Social Work Journal, 13, 241–254. 

Phillips, S., & Sandstrom, K. L. (1990). Parental attitudes toward youth work. Youth and Society, 22, 160–183. 

Portes, A., & Rumbaut, R. G. (1996). Immigrant America: A portrait. Berkeley: University of California Press. 

Profiles of general demographic characteristics: U.S. Census, 2000 of population and housing [online]. (2001). Available at FTP/PDF: http://www.census.gov/Press-Release/www/2001/demoprofile.html File: United States: 2000.  

Rumberger, R. W. (1995). Dropping out of middle school: A multilevel analysis of students and schools. American Educational Research Journal, 32, 583–625. 

Shanahan, M. J., Finch, M., Mortimer, J. T., & Ryu, S. (1991). Adolescent work experience and depressive affect. Social Psychology Quarterly, 54, 299–317. 

Steel, L. (1991). Early work experience among White and non-White youths. Youth and Society, 22, 419–447. 

Steinberg, L., Fegley, S., & Dornbusch, S. M. (1993). Negative impact of part-time work on adolescent adjustment: Evidence from a longitudinal study. Developmental Psychology, 29, 171–180. 

Steinberg, L., & Dornbusch, S. M. (1991). Negative correlates of part-time employment during adolescence: Replication and elaboration. Developmental Psychology, 27, 304–313. 

Stern, D. S., Hopkins, C., & McMillion, M. (1990). Quality of students’ work experience and orientation toward work. Youth and Society, 22, 263–282. 

Suárez-Orozco, C., & Suárez-Orozco, M. (1985). Transformations: Immigration, family life, and achievement motivation among Latino adolescents. Stanford, CA: Stanford University Press. 

U.S. Bureau of the Census. (1998). Statistical abstract of the United States, 1998. Washington, DC: U.S. Government Printing Office. 

Wojtkiewicz, R. A., & Donato, K. M. (1995). Hispanic educational attainment: The effects of family background and nativity. Social Forces, 74, 559–574. 

World Bank. (2000). World development report, 1999/2000. New York: Oxford University Press. 

Zambrana, R. E. (1995). Understanding Latino families: Scholarship, policy, and practice. Thousand Oaks, CA: Sage.