A two-year study of patterns and predictors of substance use among Mexican American youth.

Zapata, Jesse T. ; Katims, David S. ; Yin, Zenong 等

A number of longitudinal studies concerning licit and illicit substance use among adolescent and preadolescent populations have been conducted (see, for example, Huizinga, Loeber, & Thornberry, 1993; Maddahian, Newcomb, & Bentler, 1985; Windle, 1990). Each of these studies investigated a particular combination of correlates, predictors, or risk factors associated with substance use. Of particular interest are those studies that focused on individual attributes, characteristics, situations, and environmental conditions that may increase the probability of substance use and abuse by school-age youth.

Specific risk factors for the initiation and use of substances by youngsters have been identified in both cross-sectional and longitudinal studies (Brook, Whiteman, Gordon, Nomura, & Brook, 1986; Newcomb, Chou, Bentler, & Huba, 1988; Wingard, Huba, & Bentler, 1979; Shedler & Block, 1990; Hawkins, Lishner, & Catalano, 1985; Rhodes & Jason, 1988). These risk factors include early antisocial behavior, depression, anxiety, socioeconomic status, hyperactivity, academic failure, lack of commitment to school, alienation, rebelliousness, and lack of social bonding.

In these studies, the number of risk factors to which youth are exposed has been linearly associated with substance use and abuse. These findings corroborate and substantially extend the preliminary work of Bry, McKeon, and Pandina (1982) and suggest that a risk-factor model of substance use is not unlike that of other epidemiological phenomena (e.g., heart disease, HIV) which follow patterns of vulnerability and susceptibility due to increased exposure to risk. Initiation and continued use of substances may be attributed to a risk factor or a combination of risk factors. The number of risk factors present directly increases the likelihood of drug use and abuse (Brook et al., 1986; Clayton, 1992; Newcomb, 1992).

Although these studies have shed light on the problem of substance use and associated risk factors among adolescents and preadolescents, only a few longitudinal investigations have included Hispanics as subjects. Those studies that have included Hispanics have traditionally focused on comparisons with whites. Within this context, some investigators suggest that risk factors associated with substance use may operate in much the same way for Hispanics as they do for non-Hispanics (Vega, Zimmerman, Warheit, Apospori, & Gil, 1992; De Barona & Simpson, 1984). Other researchers indicate that there are substantive differences between Hispanic and white youth in the patterns, frequency, and predictors of substance use (Chavez & Swaim, 1992; Fredlund, Spence, & Maxwell, 1989). Until recently, more white youth engaged in substance use than did Hispanic youth (United States Department of Health and Human Services, 1994; Rebach, 1991), contradicting traditional stereotypes about substance use among minorities.

Researchers have suggested that this comparative approach ignores intragroup variations among Hispanics. De La Rosa, Khalsa, and Rouse (1990) have concluded that there are substantial differences between Mexican Americans, Puerto Ricans, and Cubans in their substance use. In addition, Hispanic subgroups may be exposed differentially to risk factors (Vega et al., 1992). As Mexican Americans make up 63% of the Hispanic populations in the United States and are one of the fastest growing ethnic minorities in the country (Knouse, Rosenfeld, & Culbertson, 1992), it is important to understand substance use issues in the population.

The objectives of the present longitudinal study were to determine patterns of substance use from year one to year two and to examine the relationship between risk factors and the initiation of substance use across both years for a sample of low socioeconomic status Mexican American youth. Demographic, environmental, and psychological characteristics of elementary and middle school students and their use of nine substances were examined via a self-report survey.

METHOD

Subjects

The south Texas, low SES, working-class school district in which the study took place was selected because it consists primarily of Mexican American students whose characteristics are believed to be representative of the Mexican American population of the United States (see Knouse, Rosenfeld, & Culbertson, 1992, for a complete demographic and economic description of Hispanics in the United States). Subjects for the initial year of the study were 2,205 students from the district's fourteen elementary and four middle schools.

The survey was administered to all students for whom parental permission was obtained (a response rate of 89% was achieved). Eighty-three percent of the sample of usable surveys were completed by students who identified themselves as Mexican American. Students surveyed during the first year were tracked and asked to participate in the second year of the survey's administration. A total of 1,805 Mexican American students, who were now in grades 5, 6, and 7, participated in the second year of the study. The matched rate between years one and two was 56% (n = 1,236; 52% male and 48% female). The high attrition rate from year one to year two was primarily due to the mobile nature of the district's population (it was estimated that 26% of the student population transferred in or out of the school district or between school campuses within the district). Indicators of substance-use prevalence revealed no significant difference between the intact groups from years one and two and the matched cases in the study (see Table 1).

Procedure and Instrument

Data for both years of the study were collected via a self-report survey, which was administered in 50 minutes by trained university students majoring in the social and behavioral sciences (e.g., education, educational psychology, sociology). The entire survey was read aloud in English to ensure standardized administration and to avoid problems caused by the different reading levels of subjects. No school personnel were present during its administration.

The survey was designed to include demographic, environmental, and psychological domains. In year one, students were asked how many times they had ever used a substance in their lifetime and in the last year. In year two, the students were again asked how many times they had used a substance in the last year. Four minor substances (cigarettes, beer, wine/liquor, and marijuana) and five major substances (pills, cocaine, crack, inhalants, and hallucinogens) were included. The remainder of the survey consisted of the following:

Demographic Index. Students were asked to provide information on age, grade, sex, ethnicity, and occupations of parents.

Deviant Behavior Index. Students were asked to indicate the extent to which they had participated in the violation of social norms (e.g., gang association, carrying weapons in school or in the neighborhood, damaging property, staying out late at night without parental permission). A high score indicated a high level of "deviance." Cronbach's alpha was .93 for this sixteen-item index.

Family Conflict Index. Students were asked to indicate the quality and quantity of family conflict. A high score indicated a high level of family conflict. Cronbach's alpha was .81 for this four-item index.

Susceptibility to Peer Influence Index. Students were asked to indicate the degree to which their behavior is influenced by the behavior of close friends. A high score indicated a high level of susceptibility to peer influence. Cronbach's alpha was .68 for this seven-item index.

Peer Use of Substances Index. Students were asked to indicate if any of their close friends used one or more of the nine substances listed in the survey. A high score indicated a high level of peer use of substances. Cronbach's alpha was .91 for this six-item index.

Substances Offered Index. Students were asked to indicate if they were offered specific substances in school and in the neighborhood. A high score indicated a high level of availability of various substances in the school or neighborhood. Cronbach's alpha was .75 for the availability of minor substances and .76 for the availability of major substances.

Dysphoria Index. Students were asked to indicate how they felt most of the time over the past few weeks. A high score indicated a high level of dysphoric feelings. Cronbach's alpha was .76 for this six-item index.

Stressor Index. Students were asked to indicate if specific environmental and interpersonal stressors had occurred at any time during the school year. A high score indicated a high level of stress. Cronbach's alpha was .74 for this twenty-item index.

School Satisfaction Index. Students were asked to rate their personal satisfaction with general and specific aspects of both elementary and middle school environments. A low score indicated dissatisfaction with the school environment. Cronbach's alpha was .82 for this five-item index.

Data Analysis

In order to ascertain patterns of substance use in this population, two types of indicators were generated based on responses to the substance-use items at both years one and two. First, the lifetime prevalence of substance use was analyzed for all the students surveyed in year one and year two rather than the matched cases. Prevalence was separately assessed for minor and major substances by categorizing responses into three levels: never used, used one substance, and used at least two substances. Second, in order to identify the status of substance use, a classification index was created to reflect the patterns of minor and major substance use across the two years. Four nonoverlapping groups were identified: (1) students who reported never using substances (abstainers), (2) students who had used substances in the past but who did not use them during year two, (3) students who initiated use in year two (new users), and (4) students who used substances in the past and continued to use them in year two.

An incidence-of-substance-use variable was derived by including only those cases identified as abstainers and new users. This was used as the dependent variable in a logistic regression with nine risk-factor variables and two demographic variables (student gender and grade in school) as independent variables. Following the suggestion of Hosmer and Lemeshow (1989), all eleven independent variables were screened using stepwise logistic regression (SPSS, 1992), with alpha = .20 as the criterion for entry and alpha = .25 as the criterion for removal of a variable from the regression equation. Those independent variables that met the selection criteria were used in the final logistic regression. However, an independent variable was retained for inclusion in the final analysis if it did not meet the selection criteria but was considered theoretically plausible (Hosmer & Lemeshow, 1989; Mickey & Greenland, 1989). Unadjusted conditional odds ratios (ORs) and 95% confidence intervals were calculated for the independent variables included in the final models.

RESULTS

Table 1 shows the patterns of minor and major substance use in years one and two for intact and matched groups. Chi-square tests indicated that there were significant changes in the percentages of students who reported using both minor and major substances between years one and two.

In addition, Figure 1 illustrates the patterns of change in the use of both minor and major substances from year one to year two. Specifically, it shows the percentages of students who abstained from substance [TABULAR DATA FOR TABLE 1 OMITTED] use in their lifetime, students who used substances for year one only, students who initiated substance use in year two (new users), and students who used substances in both years.

Based on the results of the preliminary stepwise logistic regression analysis (stepwise selection results are available from the authors upon request), four significant independent variables and two nonsignificant but theoretically substantiated independent variables (gender and availability of minor substances) were reentered in a final multivariate logistic regression for the incidence of minor substance use (see Table 2). Odds ratios for the risk-factor variables ranged from 1.15 to 1.94 for each unit change. However, deviant behavior score (OR = 1.94 for reporting more deviant behaviors), grade attended in year one (OR = 1.39 for being in upper grades), and lack of serf-esteem (OR = 1.22 for reporting lower self-esteem) were the only significant predictors of the incidence of minor substance use over a period of one year.

Table 2

Multivariate Logistic Regression for the Incidence of Minor
Substance Use at Year 2

 Conditional odds ratio and
 95% confidence limits

Predictor variables OR Lower Upper

Student gender 1.29 0.90 1.86
Grade attended in year one 1.39 1.14 1.69
Deviant behavior 1.94(a) 1.46 2.57
Lack of self-esteem 1.22(a) 1.02 1.45
Susceptibility to peer influence 1.64(a) 0.90 2.50
Minor substances availability 1.15(a) 0.91 1.44

a OR was calculated using Z-score as unit of change.

Applying the same analytic approach to the prediction of major substance use, only four independent variables were selected based on the stepwise screening, which were reentered in a final multivariate logistic regression (see Table 3). Gender (OR = 1.64 for being female), grade attended in year one (OR = 1.34 for being in upper grades), deviant behavior score in year one (OR = 2.15 for reporting more deviant behaviors), and the use of minor substances (OR = 1.51 for having used at least one minor substance during the previous year) emerged as significant predictors of the incidence of major substance use in year two.

DISCUSSION

The results of the present investigation are relevant to the study of substance use among Mexican American school-age youth for a number of reasons. First, it is encouraging to note that nearly 48% of the students surveyed reported abstaining from the use of minor substances over their lifetime, and approximately 72% reported abstaining from the use of major substances. However, the dynamic nature of substance use was confirmed. For example, a number of students who used minor and major substances (7.4% and 6.0%, respectively) in year one did not use these substances in year two of the study. In addition, a substantial number of students initiated substance use during year two. Specifically, there was a 15% increase from year one to year two in the number of students using at least one minor substance and a 10% increase for major substance use. These findings substantiate the pattern observed by Shedler and Block (1990).

Table 3

Multivariate Logistic Regression for the Incidence of Major
Substance Use at Year 2

 Conditional odds ratio and
 95% confidence limits

Predictor variables OR Lower Upper

Student gender 1.64 1.12 2.38
Grade attended in year one 1.34 1.09 1.65
Deviant behavior 2.15(a) 1.71 2.70
Using minor substances in year one 1.51 1.01 2.24

a OR was calculated using Z-score as unit of change.

Employing logistic regression analysis, risk factors found to be predictive of the initiation of minor substance use were grade of the student in year one of the study, deviant behavior, and low self-esteem. Risk factors found to be predictive of the initiation of major substance use were gender, grade during year one, deviant behavior, and use of minor substances. Identification of these risk factors for a younger population than that traditionally addressed in the research literature suggests critical areas for early intervention and prevention efforts in a school setting. In addition, the findings emerge within the context of a low socioeconomic status Mexican American population that is believed to be representative of Mexican Americans in general.

Grade. School grade during year one was found to be a factor that predicted the initiation of the use of both minor and major substances in year two of the study. In other words, use of substances significantly increased from one year to the next. This common pattern of substance use increasing with age has been demonstrated in cross-sectional work on Mexican American children (Padilla, Padilla, Morales, & Olmedo, 1979; Katims & Zapata, 1993; Zapata & Katims, 1994). The results of the present study confirm and extend these findings over a period of two years and are significant in the context of early intervention and prevention efforts. Since this is a common pattern, the axiom "the earlier the better" seems appropriate for such prevention efforts.

Self-esteem. Students who scored low on self-esteem in year one tended to be at greater risk for the initiation of minor substance use in year two. This pattern has also been confirmed in cross-sectional and other longitudinal studies with adolescents and preadolescents (Bry, McKeon, & Pandina, 1982; Newcomb, Maddahian, & Bentler, 1986; Labouvie, Pandina, White, & Johnson, 1986). These studies have found that low self-esteem is important to the understanding of substance use and abuse; in fact, changing self-esteem is at the heart of many programs designed to prevent substance use, delinquency, or other problem behaviors. Intervention efforts could focus on helping Mexican American students improve self-esteem through the use of positive role models and culturally relevant and reinforcing group guidance and counseling activities (Zapata, 1995). This could lead to what is referred to in the literature as a "protective factor" against the initiation of substance use (Newcomb, 1992).

Gender. Gender predicted the initiation of major substance use but not minor substance use in the second year of the study. Most cross-sectional and longitudinal studies have found that males engage in more substance use than do females. The present study confirms this pattern, yet reveals a more complex relationship between gender and substance use. As the present study focused on the pre- and early adolescent period, the findings suggest that females are more vulnerable to the initiation of major substance use than are males during this stage of development, and perhaps need differential prevention approaches focusing on coping mechanisms that assist females as they progress through school.

Deviant behavior. Engaging in deviant behavior has acted as a predictor of substance use in several longitudinal studies (Brook, Whiteman, Gordon, Nomura, & Brook, 1986; Huizinga, Loeber, & Thornberry, 1993), and the present study confirms this relationship for the use of both minor and major substances. This relationship suggests that curtailing deviant behavior should be a major element of intervention efforts. School districts must learn to work cooperatively with local police and parents to enforce curfews and other community laws, and students must learn alternatives to deviant patterns of behavior.

Use of minor substances in year one. The use of minor substances in year one was a risk factor for the use of major substances in year two. This is not surprising, as the use of minor substances has been found to lead to "harder" substances among the general youth population (Kandel & Faust, 1975). Kandel and Faust suggest that involvement in substance use appears to occur in a typical sequence, or series of stages. They believe that youngsters typically go through the "gateway" of using alcohol and in many cases, cigarettes before progressing to stronger substances.

Two additional variables, availability of minor substances and susceptibility to peer influence, were included in the final analysis for minor substances because they met the screening criteria. Other research has also found these variables to be associated with substance use. Glantz and Pickens (1992) and Wagenaar et al. (1993) reported that availability of substances in the adolescent's environment is associated with use. Coombs, Paulson, and Richardson (1991) found that children who use substances tend to be more strongly influenced by their peers than by their parents. Van Roosmalen and McDaniel (1989) and Kandel (1985) found that peer groups are crucially important in the initiation and use of both minor and major substances among young adolescents. Although these two variables were not statistically significant in the final analysis, they appear to be of importance and should be examined further to ascertain the dynamics of their influence.

CONCLUSION

It seems clear that certain characteristics and patterns of behavior are risk factors in the initiation and continuation of substance use for Mexican American school-age youth. It is imperative to find ways to reduce these risk factors. The confirmation here of a specific set of risk factors over a two-year period should prove helpful for designing early intervention and prevention programs aimed at this population.

Support for this research was provided by National Institute on Drug Abuse Grant #5R24DA07234 to Jesse T. Zapata and David S. Katims. The authors wish to express their appreciation for the assistance provided by Jan Gallagher, Director of Research and Development Programs in the Harlandale Independent School District, San Antonio, Texas.

REFERENCES

Brook, J., Whiteman, M., Gordon, A., Nomura, C., & Brook, D. (1986). Onset of adolescent drinking: A longitudinal study of interpersonal and intrapersonal antecedents. Advances in Alcohol and Substance Abuse, 5, 91-110.

Bry, B. H., McKeon, R., & Pandina, R. J. (1982). Extent of drug use as a function of number of risk factors. Journal of Abnormal Psychology, 91, 273-279.

Chavez, E. L., & Swaim, R. C. (1992). An epidemiological comparison of Mexican American and white non-Hispanic 8th- and 12th-grade students' substance use. American Journal of Public Health, 82(3), 445-447.

Clayton, R. R. (1992). Transition to drug use: Risk and protective factors. In M. Glantz & R. Pickens (Eds.), Vulnerability to drug abuse (pp. 15-51). Washington, DC: American Psychological Association.

Coombs, R. H., Paulson, M. J., & Richardson, M. A. (1991). Peer vs. parental influence in substance use among Hispanic and Anglo children and adolescents. Journal of Youth and Adolescence, 20(1), 73-89.

De Barona, S., & Simpson, D. D. (1984). Inhalant users in drug abuse prevention programs. American Journal of Drug and Alcohol Abuse, 10(4), 503-518.

De La Rosa, M., Khalsa, J. H., & Rouse, B. A. (1990). Hispanics and illicit drug use: A review of recent findings. International Journal of Addictions, 6, 665-691.

Fredlund, E. V., Spencer, R. T., & Maxwell, J. C. (1989). Substance use among students in Texas secondary schools - 1988. Texas: Texas Commission on Alcohol and Drug Abuse.

Glantz, M. D., & Pickens, R. W. (1992). Vulnerability to drug abuse: Introduction and overview. In M. Glantz & R. Pickens (Eds.), Vulnerability to drug abuse. Washington, DC: American Psychological Association.

Hawkins, J. D., Lishner, D. M., & Catalano, R. F. (1985). Childhood predictors and the prevention of adolescent substance use. In C. L. Jones & R. J. Battjes (Eds.), Etiology of drug abuse: Implications for prevention. Washington, DC: National Institute on Drug Abuse.

Hosmer, D. W., & Lemeshow, S. (1989). Applied logistic regressions. New York: Wiley.

Huizinga, D., Loeber, R., & Thornberry, T. P. (1993). Longitudinal study of delinquency, drug use, sexual activity, and pregnancy among children and youth in three cities. Public Health Reports, 108(1), 90-96.

Kandel, D. B. (1985). On the processes of peer influences in adolescent drug use: A developmental perspective. Advances in Alcohol and Substance Abuse, 4, 139-163.

Kandel, D. B., & Faust, R. (1975). Sequence and stage in patterns of adolescent drug use: Archives of General Psychiatry, 32, 923-932.

Katims, D. S., & Zapata, J. T. (1993). Gender differences in substance use among Mexican American school-age children. Journal of School Health, 63(9), 397-401.

Knouse, S. B., Rosenfeld, P., & Culbertson, A. L. (1992). Hispanics in the workplace. Newbury Park, CA: Sage Publications.

Labouvie, E. W., Pandina, R. J., White, H. R., & Johnson, V. (1986). Risk factors of adolescent drug use: A cross-sequential study. Unpublished manuscript.

Maddahian, E., Newcomb, M. D., & Bentler, P. M. (1985). Single and multiple patterns of adolescent substance use: Longitudinal comparisons of four ethnic groups. Journal of Drug Education, 15(4), 311-326.

Mickey, J., & Greenland, S. (1989). A study of the impact of confounder-selection criteria on effect estimation. American Journal of Epidemiology, 129, 125-137.

Newcomb, M. D. (1992). Understanding the multidimensional nature of drug use and abuse: The role of consumption, risk factors, and protective factors. In M. Glantz & R. Pickens (Eds.), Vulnerability to drug abuse (pp. 255-297). Washington, DC: American Psychological Association.

Newcomb, M., Chou, C., Bentler, P., & Huba, G. (1988). Cognitive motivations for drug use among adolescents: Longitudinal tests of gender differences and predictors of change in drug use. Journal of Counseling Psychology, 35(4), 426-448.

Newcomb, M. D., Maddahian, E., & Bentler, P. M. (1986). Risk factors for drug use among adolescents: Concurrent and longitudinal analyses. American Journal of Public Health, 76(5), 525-531.

Padilla, E. R., Padilla, A. M., Morales, A., & Olmedo, E. L. (1979). Inhalant, marijuana, and alcohol abuse among barrio children and adolescents. International Journal of the Addictions, 14(7), 945-964.

Rebach, H. (1991). Alcohol and drug use among American minorities. In J. E. Trimble, C. S. Bolek, & J. Niemcry (Eds.), Ethnic and multicultural drug abuse (pp. 23-57). New York: Haworth.

Rhodes, J. E., & Jason, L. A. (1988). Preventing substance use among children and adolescents. Elmsford, NY: Pergamon Press.

Shedler, J., & Block, J. (1990). Adolescent drug use and psychological health: A longitudinal inquiry. American Psychologist, 45, 612-630.

SPSS, Inc. (1992). SPSS user's guide. New York: McGraw Hill.

United States Department of Health and Human Services. (1994). National household survey on drug abuse: Main findings, 1993. Rockville, MD: Public Health Service (publication 93-7980).

Van Roosmalen, E. H., & McDaniel, S. A. (1989). Peer group influence as a factor in smoking behavior of adolescents. Adolescence, 24(96), 801-816.

Vega, W. A., Zimmerman, G. I., Warheit, G. J., Apospori, E., & Gil, A. G. (1992). Risk factors for early adolescent drug use in four ethnic and racial groups. American Journal of Public Health, 83(2), 185-189.

Wagenaar, A. C., Finnegan, J. R., Wolfson, M., Anstine, P. S., Williams, C. L., & Perry, C. L. (1993). Where and how adolescents obtain alcoholic beverages. Public Health Report, 108, 459-464.

Windle, M. (1990). A longitudinal study of antisocial behaviors in early adolescence as predictors of late adolescent substance use: Gender and ethnic group differences. Journal of Abnormal Psychology, 99(1), 86-91.

Wingard, J., Huba, G., & Bentler, P. (1979). Beginning adolescent drug use and peer and adult interaction patterns. Journal of Consulting and Clinical Psychology, 47(2), 265-276.

Zapata, J. T. (1995). Counseling Hispanic children and youth. In C. Lee (Ed.), Counseling for diversity: A guide for school counselors and related professionals (pp. 85-108). Boston: Allyn and Bacon.

Zapata, J. T., & Katims, D. S. (1994). Antecedents of substance use among Mexican American school-age children. Journal of Drug Education, 24(3), 233-251.