ANNOTATED BIBLIOGRAPHY

Substance Use and Academic Performance among African American High School Students

James Herbert Williams, Larry E. Davis, Sharon D. Johnson, Trina R. Williams. Jeanne A. Saunders, and Von E. Nebbitt >

Academic performance among African American students continues to be a concern. • Adolescent developmental research has identified numerous factors that affect academic performance. School-based intervention programs have focused on substance use prevention

to imprave academic performance. This study investigated to what extent family fmancial concerns, individual and parental norms, and substance use are associated with academic

performance among a sample of 212 African American students attending an urban high school located in the midwestern United States. The student body was 99% African American, with

approximately equal numbers of boys and girls.The results indicated that marijuana use. parental substance use norms, and family fmancial concerns were associated with students" academic

intentions. Grade point average was associated with both marijuana use and parental substance use norms.These results highhght the importance of family-related correlates in addition to substance use when investigating academic performance among African American students.

KEY WORDS: academic performatue; adolescence; African Americans; substarKe use; urban education

T he lack of consistent academic success for African American youths is having dam- aging consequences on African American

families and communities (Williams. Davis, Miller- Cribbs, Saunders. & Williams, 2()()2).The character- istics that defme academic success emerge through complex interactions of school, individual, family. and community contextual factors.

Also, academic success for African American youths has significant implications for ftiture earning potential and life options. A high school diploma is required for postsecondary education and is a determinant of future life outcomes for employ- ment and economic and family stability (Smith, Brooks-Gunn. & Klebanov, 1997). Although aca- demic success and high school completion do not guarantee employment, they enhance the probability of securing employment and having a higher level of lifetime earning. Some research indicates that academic success can influence marriage patterns, family formation, and social relationships (Tucker &Mitchell-Kernan. 1995).

The purpose of this study was to assess the ef- fects of subsunce use on the academic performance (grade point average [GPA]) and positive intentions of a cohort of African American high school stu-

dents. This study used data from a sample of high school students attending a predominantly African American high school (99%) in the midwestern United States. Positive intentions to complete the school year are undergirded by the theory ot planned behavior (TPB), which presupposes that behavioral intentions are affected by personal atti- tudes, social norms, and perceived personal control (Ajzen, 1991).TPB hypothesizes that the primary predictor of behavior is the intention to execute the behavior. As previously stated, the intention to engage in the behavior is predicated on attitudes, norms, and perceived control. Attitvides are related to potential consequences for executing the behavior, norms are the influence of significant others (for example, family members, friends, community) as well as opmions about the behavior, and perceived control addresses the individual's ability to carry out the behavior (Ajzen). Overall GPA was used as the proxy for academic performance.

i LITERATURE REVIEW Alcohol and Marijuana Use Although alcohol use has remained consistent among adolescents over the past decade, research- ers have invariably identified some important

CCC Code: 1070-5309/07 $3.00 O2007 National Association of Social Workers 151

trends across racial groups, and their findings have substantiated significant differences for initiation, current use, dependency, and lifetime use (Johnston, O'Malley, & Bachman, 2000; Williams, Ayers, Ab- bott, Hawkim,& Catalano, 1999). Overall, African American youths report lower rates of alcohol use. Historically, African American youths have the low- est rates of lifetime alcohol use compared with other racial and ethnic groups (Johnston et al.).

Results from comparative studies investigat- ing marijuana use among racial groups have been mixed (Williams et al., 1999), Many studies have also identified racial differences in initiation and current use for marijuana (Ellickson, Bui, Bell, & McGuigan, 1998). At grades 10,11,and 12,African American youths were significantly less likely than white youths to have initiated marijuana use before age 13 and had lower lifetime, annual, and 30-day prevalence rates (Johnston et al., 2000).

Other studies have identified dissimilar rates of prevalence (Hill, Etchison. Hawkins. & Catalano. 2003;Wimams et al.. 1999). Using data fh)m the Se- attle Social Development Project, studies have iden- tified higher prevalence rates for African American youths compared with white youths for marijuana use (Hill et aI,;Williams et al.. 1999). It is clear from various studies that alcohol and marijuana use are ongoing concerns for positive youth development. and these rates may vary by race.

Across both racial groups, adolescents who drop out of school are more likely to be involved in sub- stance use than are those who complete {Guagliardo, Huang. Hicks, & D'Angelo, 1998). Hawkins and colleagues (1992), using truancy rates, amount of time spent on homework, and students' perception of the relevance of course work as indicators, found a negative relationship between academic achieve- ment and substance use. Likewise, Hallfors andVan Dorn (2001) reported data from a recent survey that indicated that students with lower grades were five times more likely to report marijuana use. In a study of 681 African American adolescents, Zim- merman and Schmeelk-Cone (2003) identified a strong interrelationship between substance use and school motivation and completion.

Although there is considerable epidemiological research on substance use among African American and white youths (Ensminger & Juon, 1998), our knowledge regarding the relationship between sub- stance use and academic performance for African American students specifically remains limited. In-

creasing our knowledge of how substance use affects African American students' academic performance is important for developing more relevant program- ming for prevention and intervention.

Many ot these studies used substance use as the primary antecedent for understanding academic performance (Guagliardo et al.. 1998). Other re- searchers have also identified relationships between various correlates and low academic achievement in addition to substance use (Dornbusch. Erickson, Laird, & Wong, 2001) .These studies have investigated these relationships primarily with samples that have included either low percentages of or no African American students. Few studies have focused on the relationship between academic performance and substance use specifically among African American students (Zimmerman & Schmeelk-Cone, 2003).

Social Correlates Research has indicated that various social influ- ences are among the most powerful predictors of adolescent behaviors. Social influences such as social modeling, perceived norms, and peer norms have a positive association with low academic achieve- ment (Peterson. Hawkins,Abbott, & Catalano, 1995; Wood, Read, & Palfai, 2001). Individual and peer norms about alcohol and marijuana use are impor- tant predictors of academic performance (Ellickson et al., 1998). Adolescents with norms supporting antisocial behavior have low school involvement and low academic achievement and are more likely to drop out (Ellickson et al.). Zimmerman and Schmeelk-Cone (2003) have documented a strong causal relationship between peer norms and academic performance. Building on these fmdings. it would be important to investigate the relationship between substance use norms and academic perfor- mance among African American students.

Parental norms and parental behavior are also considered to be strong predictors of school en- gagement and academic performance for youths (Ellickson et al., 1998). In their research, Peterson and her colleagues (1995) identified race differences in parental norms regarding alcohol use. Among African American families, parental norms regarding drinking and parenting styles are sigiiificant predic- tors of adolescent problem behaviors (Peterson et al.). Another study of African American girls from inner-city, high-poverty neighborhoods found par- enting styles and norms to be significantly related to academic achievement (Pittman & Chase-Lansdale,

152 SocialWirkResearch VOLUME 31, NUMBER 3 SEPTEMBER 2007

2001).These studies provide an excellent founda- tional basis for investigating the links of substance use-related correlates with academic performance for African American students.

In addition to parental and peer norms, African American youths growing up in poverty and in households with consistent financial concerns are more likely to have poor educational outcomes and low levels of school engagement (McLoyd, 1998). Research suggests that family income is the highest single correlate of academic performance (McLoyd).

It is important to recognize that even with the challenging environment faced byAfrican American youths, there is tremendous heterogeneity in the academic performance of these youths, with many of them managing to do quite well (Furstenberg, Cook, Eccles, Elder, & Sameroff. 1999). Many manage to remain in school,graduate,and avoid significant life- compromising situations, such as substance use and involvement with the criminal justice system, and become healthy adults (Smith, Lizotte,Thornberry, & Krohn. 1995).

The present study was conducted to further advance knowledge in this area by examining how these correlates may impede intentions to com- plete high school and academic success for African American students.The two research questions for this study are (1) to what degree are family finan- cial concerns, individual and parental norms, and substance use associated with positive intentions to complete the school year and (2) to what degree are family fmancial concerns,individual and parental norms, and substance use associated with school performance (GPA)?

METHOD

Research Design We used cross-sectional data from a study of school completion among African American youths. The student participants attended an urban high school located in the midwestern United States. The student body was 99% African American with ap- proximately equal numbers of boys and girls.There was little income variation among the sample. Ap- proximately 58% ofthe sample qualified for free or reduced-cost lunch, and many ofthe "nonquaUfy- ing" students either refused to apply or had income levels only marginally above the qualifying cut off. Thirty-two percent ofthe sample hved with both biological parents. A cohort of high school juniors

was queried with the assistance of a school guidance counselor who served as the primary Uaison among the students, parents, and members ofthe research team. This sample included 212 African American high school juniors {80% ofthe junior class). The focus of the study was explained in detail during the students' homeroom class. Participants were surveyed in groups of 15 to 40 during the school day. A team of African American graduate research assistants administered the survey, which took about an hour to complete. To accommodate the wide variability in reading levels among the students, one research assistant read the items to the group while students followed along with their questionnaire. The second research assistant was available to assist any students who had questions or experienced dif- ficulty in completing the items. Students were given $15 for participating. Before the larger study was conducted, a small pilot study ofthe questionnaire was conducted with 10 students from the target school. The pilot study was conducted to gauge the effectiveness ofthe methodology and the abil- ity ofthe students to understand and complete the questionnaire. No additional modifications were made to the instrument or the research design as an outcome of the pilot study.

Outcome Measures Positive Intentions. Students' intention to complete the current school year was measured by taking the mean score ofthe responses to five items (Cronbach's alpha = .82): (1) "I intend to complete the current school year," (2) '*I will try to compiete the current school year," (3) "I expect to complete the current school year," (4) "I am determined to complete the current school year," and (5) "I might not complete the current school year." Each item was measured on a seven-point Likert scale with I ~ disagree very much , 2 = disagree somewhat, 3 = disagree slighdy, 4 = neither, 5 = agree slightly, 6 = agree somewhat, and 7 = agree very much. These items followed the guidelines established by Ajzen (1991) for measuring this construct.

Grade Point Average. Overall GPA was used as a proxy for academic performance. These data were obtained from students' official school record at the end ofthe school year.The possible range for this measure was 0.0 to 4.0. GPA was dichotomized into two groups: students (n = 91) academically achieving below average with a GPA of less than " C " (< 2.0) and students (« = 88) achieving a GPA of " C " or

W I L L I A M S E T AL. / Substance Use and Academic Perfirmance among AfUcan American High School Students 153

better {> 2.0). Mean GPA was 1.97 (SD = .879). More than 50% of the students had a GPA of less than " C " (< 2.0), 34% had a GPA of "C," and 16% had a GPA of "B" or better {> 3.0). .

Control Mea5ure5 The models did not include the standard demo- graphic control variables.There was little variation in income, age, race, or ethnicity in the sample, so these variables were not included in the model. Although earlier research on academic performance found gender differences (Saunders, Davis, Williams, & Williams, 2004), a comparison between boys and girls in this study showed no significant gender differences; therefore, gender was not included in the model.

Social Correlates Substance use norms held by the student regarding substance use were measured by taking the mean score of the following three items (a = .85): How wrong do you think it is for someone your age to (1) "Drink beer, wine, or hard liquor (for example, vodka, whiskey, or gin) regularly?" (2) "Smoke mari- juana?" and (3) "Use LSD, cocaine, amphetamines, or another illegal drug?" Each item was measured on a four-point scale with 1 ~ very wrong, 2 = wrong, 3 = a little bit wrong, and 4 = not wrong at all (Arthur, Hawkins, PoUard, Catalano, & Baglioni, 2002). Items were reverse scored for scale develop- ment and analyses.

Social acceptability of substance use by peers was measured by the mean score of a three-item scale {a = .86) consisting of the following items: What are the chances you would be seen as cool if you (1) "Smoked marijuana?" (2) "Began drinking alcoholic beverages regularly, that is, at least once or twice a month?"and (3) "Smoked cigarettes?"Each item was measured on a five-point scale with 1 = no or very litde chance, 2 = iittle chance, 3 = some chance, 4 = pretty good chance, and 5 = very good chance (Arthur etal.. 2002).

Paretital .•^ttbstance use norms held by parents were measured by using the mean score of a three-item scale (a = .82), using the following items: How wrong do your parents feel it would be for you to (1) "Drink beer, wine, or hard liquor (for example, vodka, whiskey, or gin) regularly?" (2) "Smoke ciga- rettes?" and (3) "Smoke marijuana?" Each item was measured on a four-point scale with 1 = very wrong, 2 = wrong, 3 = a litde bit wrong, and 4 = not wrong

at all (Arthur et al., 2002).These items were reverse scored for scale development and analyses.

Family jitiandal concern was measured using a single item asking, "How often does your family have problems paying for basic necessities like food, cloth- ing, and rent/mortgage?" This item was measured on a five-point scale with 1 = never, 2 = rarely, 3 = occasionally, 4 = frequendy, and 5 = almost always (McLoyd,Jayaratne, Ceballo, & Borquez, 1994).

Adult substance use nonm held by adults known to the student were measured by using the mean score of a three-item scale (a = .89) created using the following items: How wrong would most adults in your neighborhood think it was for youths your age to (1) "Use marijuana?" (2) "Drink alcohol?" and (3) "Smoke cigarettes?" Each item was mea- sured on a four-point scale with 1 = very wrong, 2 = wrong, 3 = a little bit wrong, and 4 = not wrong at all (Arthur et al., 2002).These items were reverse scored for scale development and analyses.

Adult antisocial behavior by adults known to the student was measured by calculating the mean score of a four-item scale (a = .83) created from the following items: About how many adults have you known personally who in the past year (1) "Have used marijuana, crack, cocaine, or other illicit drugs?" (2) "Sold or dealt drugs?" (3) "Have done other things that could get them in trouble with the police, like stealing, selling stolen goods, mugging or assaulting others, etc?" and (4) "Got- ten drunk or high?" Each item was measured on a five-point scale with 1 = none, 2 = 1 adult, 3 = 2 adults, 4 = 3 or 4 adults, and 5 = 5 or more adults (Arthur etal., 2002).

Alcohol use was measured by the mean score of three items indicating various levels of alcohol use (a = .78): On how many occasions (if any) have you (1) "Had beer, wine, or hard liquor to drink in the past 30 days?" (2) "Had beer, wine, or hard liquor to drink in your lifetime?" and (3) "Think back over the last two weeks. How many times have you had five or more alcoholic drinks in a row?" Each item was measured on a seven-point Likert scale with 1 = 0, 2 = 1-2, 3 = 3-5, 4 = 6-9, 5 = 10-19, 6 = 20-39, and 7 = 40 or more occasions (Arthur etal., 2002).

Marijuana use was measured by the mean score of two items indicating various levels of marijuana use (a = .88): On how many occiisions (if any) have you (1) "Used marijuana during the past 30 days?" and (2) "Used marijuana in your lifetime?"

154 Social Work Research VOLUME 31, NUMBER 3 SEPTEMBER 1007

Each item was measured on a seven-point Likert scale with 1 = 0, 2 = 1-2. 3 = 3-5, 4 = &-9, 5 = 10-19, 6 = 20-39, and 7 = 40 or more occasions (Arthur etal.. 2002).

Analysis Strategy A series of statistical analyses were conducted to examine the study's primary research questions to determine the extent to which substance use and social correlates affect academic performance and intentions of African American high school juniors. The analyses were conducted in multiple stages. First, we conducted an attrition analysis to investigate differential attrition by social correlates. Analysis of variance (ANOVA) was conducted to compare students with a reported GPA to those without. We then conducted a descriptive analysis to assess means, standard deviations, ranges, and correlation coefficients for the outcome measure and social correlates.

To answer the primary research questions, we conducted a hierarchical linear regression model to estimate the degree of association between substance use and social correlates and intentions (Tabachnick & Fidell. 2000). The hierarchical linear regression model was analyzed by regressing the outcome measure—positive intentions—on the eight corre- lates. Building on earlier research, individual norms were entered into the model first, parental norms and family financial concerns were entered next. and community adults' norms and behaviors were entered in the third block, with substance-using behaviors entered in the final block. In the second analysis, a hierarchical logistic regression model was analyzed using overall GPA as a frinction of the eight correlates using the same order of entry as the linear regression model.

RESULTS i

Attrition Analysis Using overall GPA as an outcome measure decreased the sample size by 33. Decrease in sample size was a result of the high school not having current academic records. To investigate whether the students with GPA data (« = 179) were systematically different from those students with no GPA data (M ~ 33), an attrition analysis was conducted using AN OVA.The results indicated no differential attrition by positive intentions, social acceptability of substance use, adult antisocial behavior, adult substance use norms, family financial concerns, and alcohol use. These analyses

indicated some differential attrition for three of the correlates. Students with no GPA data had lower levels of substance use norms [F(l, 210) = 5.56, p < .05], lower perceptions of parental norms [F(l, 210) = 19.12,p < .001], and higher prevalence of marijuana use [F(l,2tO) = 7.68,p < .01] than stu- dents with GPA data.The external validity of results reported that are related to academic achievement may be constrained by these results.

Descriptive Analysis Twent>'-thrce percent (« = 49) of the students re- ported some alcohol use, and 49.5% (M = 105) of the students reported no marijuana use. Fewer than 4% (3.8%. fl = 8) of the students reported alcohol use 10 or more times over the past 30 days, whereas 13.2% (n = 28) of the students reported marijuana use 10 or more times over the past 30 days. Many of the youths in this study did not support norms against alcohol or marijuana use.The students also reported their parents and other adults to hold norms in support of their alcohol and marijuana use. The youths reported exposure to numerous adults involved in antisocial behavior and indicated some family financial concerns. (The means, standard deviations, and ranges of the outcome and correlates are presented in Table 1.)

The pattern of coefficients in the matrix indicated an overall positive association among many of the correlates to support the multivariate analysis (Table 2). All correlates were negatively associated with intentions. Many of the coefficients were in the low to moderate range, with the stronger coefficients between the substance use norms and marijuana use.

Table 1: Outcome Measure and Correlates of Measures Predicting

Academic Performance among African American Youths (Af = 212)

pweasure r . . . i r

.Siiii-JuiiKi' [iM- norm.s

Social acceptability of substance use

Parental substance use norms

Family financial concerns

Adul[ substance use norms

Adult antisocial behavior

Alcohol use

Marijuana use

ilff

6.45

2.19

1.71

1.43

1.90

2.03

3.53

2.19

2.47

5 0 0.97

0.93

0.98

0.69

1.08

0.96

1.23

1.26

1.96

Range

i - T

1-5

1-5

1-5

1-5

1-5

1-7

1-7

WtLLiAMS ET AL. / Suhstance Use and Academic Performance among African American High School Students 155

Table 2: Zero-order Intercorrelation Coefficients for Measures of Academic Performance of African American Youths (A/ = 212)

Factor

2. Substance unc norms

3. Social acccpiabiticy of substance

-.271** —

use 4. Parental .substance use norms

5. Family financial concerns

6. Adult substance use nornis

7. Adult antisocial behavior

8. Alcohol use

9. Marijuana use

-.187**

-.355***

-.193**

-.274**

-.142*

-.259**

-.339***

.235**

.573***

.054

.276**

.559**"

.641***

—

.285**

.121

.076

.094

.184**

.183**

.115

.490***

.21.1**

.435***

.501***

—

.020

-.002

.019

-.021

.257**

.495***

.482***

.310*** ~

.422*** .694*** 05. "p<.01.*"p<,001.

The majority of the coefficients were significant at t h e p = .01 or .001 level.

Social Correlates Associated with Positive Intentions In the first modei,substance use norms [F= 13.461, p < .001] and social acceptability of substance use (F = 3.592, ;i < .07] explained 9% of the variance [FC2,208) = I0.75.p<.001], with both correlates having a negative association with positive intentions (Table 3).When adding parental norms [F= 11.814, p < .001] and family financial concerns [F = 4.907, p < .05], 16% of the overall variance was explained

[F(4,206)= 10.365.p<.001].The unique variance explained by these two additional correlates was 7% (p < .001). After adding the family variable to the model,student's substance use norms and social ac- ceptability were no longer significant.When adding community adult substance use norms and antisocial behaviors to the model. 18% of the overall variance was explained |F(6,204) = 7.475.p < .001 ].These variables did not add to the overall model. As with the previous model, parental norms [F= 7.508,/) < .01] and family financial concerns |F = 5.244,/J < .05] continued to be significant in model 3. Alcohol andmarijuanause[F=5.040,jp< .05] vrere entered

Table 3: Model Statistics of Change for Correlates on Positive

Measurtt Imt-rtcpi

Substance use norms

Social acceptability of substance use

Patental substance use norms

Family financial concerns

Adult substance use norms

Adult antisocial behavior

Alcohol use

Marijuana use

'0<.10.'p*.OS. " p < . 0 1 . •••;><.001,

Intentions of African American Youths (Af = 212) ModeM

B SE F

:.2.K; .iti2

-.261 .071 13.461***

-.128 .067 3.592'

R=.m R-' = .094

Adjusted M' = .085

/^2,208) = 10.750***

Model 2

B SE F

-.107 .081

-.068 .066

-.383 .111 11.814**'

-.129 .058 4.907*

R= .409

^ = .168

Adjusted/f = .151

R'-. = .074*** dungc

fl;4.206) = 10.365***

Model 3

B S£ F

7.771 .248

-.056 .086

-.079 .067

' -.320 .117 7.508**

-.133 .058 5.244*

-.124 .078

-.031 .053

R=A25

R'-.ISO

Adjusted/?'= .156

us

F{6.204) = 7.475***

Model a

B

.038

-.076

-.275

-.147

-.097

.009

.018

-.114

ft

iP

SE F^ —HB

,095

.066

.117 5.482*

.058 6.473''*

.080

.055

.070

.051 5.040*

= .451

= .203

Adjusted ^ ' = . 172

1= 6.456***

156 Social Work Research VOLUME 31, NUMBER 3 SEPTEMBER 2007

in the final model. The unique variance explained by entering these variables was 2% (jt < .05). In the final model,parental substance use norms [F = 5.48, p < .05], family fmancial concerns [F = 6.473,p < .01], and marijuana use (F = 5.040, p < .05) were signifuarit correlates for the overall model. In the final model, the combination of correlates accounted for 20% of the variance in positive intentions [F (8, 202) =6.436,p<.001].

Social Correlates Associated with GPA The overall model was significant after substance use norms and social acceptability [x'(2, N = 179) = 25.487,p < .001] were added to the model.Students' substance use norms {B - -.849,p < .001,odds ratio [OR] = .42) accounted for the significant portion of the decrease in GPA (Table 4). Adding parental substance use norms and family financial concerns into the model indicated a significant improvement [X-(2,N= 179) = \\.02>2,p< .01]. Students'sub- sunce use norms (B = -.574, p < .01, O R - .56) and parental substance use norms (B = -1.415,p < .01, O R = .24) had significant associations with low GPA. Entering community adult substance use norms and antisocial behavior did not significantly improve the overall fit of the model. Parental sub- stance use norms (B =-1.257, p < . 0 1 , O R = .28) continued to be significant, witb students'substance use norms approaching significance (6 = -,413,p < .07, OR = .70). Entering alcohol and marijuana use in the final step significantly improved the overall fit of the model [x-{2,N= 179) = 8.661,p < .01]. In the final model, parental substance use norms (B = -1.294, p <. 01, O R = .27) and marijuana use (B = -.355, p < .05, O R = .70) were tbe only significant correlates.

DISCUSSION The descriptive results indicate that marijuana use is more prevalent among youths in this sample than alcohol use. Although the overall frequency of use for both marijuana and alcohol was unremarkable, the results are consistent with earlier studies show- ing that African American students consistently had lower rates of marijuana and alcohol use and that Afi-ican American students ages 12 to 17 reported rates of substance use 50% lower than other groups (Johnston etal.,2000).

Two research questions underpin this study. The first question investigated the associations between family fmancial concerns, individual and parental

norms, and substance use and positive intentions for Afi-ican American students. The second ques- tion investigated associations using the same set of social correlates, substance use, and academic performance. Parental substance use norms, family fmancial concerns, and marijuana use were found to negatively affect positive intentions. This find- ing supports the overall theoretical constructs of the study (for example, social norms and perceived personal control). It is apparent tbat positive inten- tions are influenced by family and individual factors and not by outside adults. Many students perceived their parents as having permissive norms and atti- tudes regarding alcohol and marijuana use. Similar findings were obtained when investigating students' school performance, witb the exclusion of family fmancial concerns. These findings are similar to those reported in eaHier studies investigating these relationships (Guagliardo et al., 1998;Hallfors &Van Dorn, 2001; Peterson et al., 1995; Williams et al., 1999).The negative relationship between parental norms and both outcome measures would suggest that the messages students receive from parents have a significant impact on their behaviors and academic performance. Tbe results also suggest a strong rela- tionship between students' marijuana use and both outcome measures. These relationships provide a basis for intervention and prevention programs to address the negative effects of these correlates on positive intentions and academic performance.

Given tbat tbe findings from tbis study identify negative relationships between family and individual factors and academic performance, there are some limitations to the findings.The use of cross-sectional data prohibits time-ordering examination for the regression models when investigating causal relation- ships between the various correlates and outcome nieasures.The use of cross-sectional data was limited by the scope of the research project.Tbe sample for this study represents just 80% of the juniors, with limited information on tbe 20% who chose not to participate. The lack of comparison data limits the study's ability to investigate possible selection bias in the sample.

This study used GPA as the sole proxy for aca- demic achievement. Although GPA is a reasonable measure of academic achievement, other measures (standardized test scores, assessment test scores) would strengthen such a measure.Tbe results from the attrition analyses also indicated some differential attrition on GPA for marijuana users and perceived

W I L L I A M S ET AL. / Suhstance Use and Academic Perfrrmance among African American High School Snidena 157

I

^ u

.7 86

.0 96

.2 74

5. 79 3

-1 .2 94 **

77 5 -

1, 62 3

0 0

ir^

.1 47

67 0

1. 44 3

.5 65

.9 03

.1 83

-. 10 2

14 6

1. 21 6

.6 82

.9 11

.4 03

-. 09 4

04 6

1. 23 6

.5 43

,8 19

.9 02

-. 19 9

.9 49

0 0

.7 01

5. 29 1

-. 35 5*

1 / ^

— fN

~ -^ r-i

oo 00 a\ "*t "* ,9 u S >A ^ (N IN M

H H

- V

levels of parental norms. Substance use data for this study were limited to alcohol and marijuana use. Prevalence of other substances (for example, tobacco, cocaine, opiates,and steroids) were not assessed.This study has a limited scope of overall substance use by this sample. The accuracy of substance use data is limited by the accuracy and social desirability to which respondents self-reported their involvement in substance-using behaviors. Mensch and Kandel (1988) reported tbat African American youths were more likely than white youths to underreport mari- juana use. On the other hand, a body of literature supports tbe validity of self-report data (O'Malley, Bachman, & Johnston, 1983). Despite these limita- tions, the findings from this study suggest important relationships between family and individual factors and academic performance for African American students.

Educational statistics report that 48.9% of all African American students entering ninth grade do not graduate on schedule or drop out (Paulson, 2006). These findings suggest that high school is one potential setting to introduce evidence-based intervention and prevention programming. Tbe impact of marijuana use and parental substance use norms on academic performance for African American youths suggests the need for better tools to assess marijuana use and improved service mechanisms to prevent marijuana use and better engage parents. For many youths, higb school is the concluding institution for implementing pro- grams to address problem behaviors. High school provides an excellent opportunity to further explore alcohol and marijuana use with youths, address real consequences, and provide substantive intervention. Schoolwide prevention programming or activities can be provided in a classroom setting (for example, health, physical education, and homeroom).The use of schoolwide strategies as one method to address alcohol and marijuana use may be sahent for Afri- can American stvidents given their low percentage of use.

The research literature contends that school-based programs should be more strategic in their efforts, broaden the focus beyond the exclusive scboolwide programming,and target or match interventions to- ward specific risk factors among subgroups (Gorman, 1996). A more targeted or matched intervention approach may be more promising for the limited number of students currently using alcohol and drugs. In an effort to provide more effective services

to tbe students who are currently using marijuana or alcobol,collaborative partnerships with community- based programs may increase the overall effectiveness of both sets of programs. School partnership witb various community-based youth treaonent programs expands the range of services available to youths.To address the service needs of alcohol and marijuana users, it is imperative to develop school-based pro- grams witb strategies to support decision making, address intrapersonal and interpersonal risk factors, and promote desistance of use (Gorman). These programs should promote school engagement and achievement. Another alternative approach to tbe risk-based approach is resihence education (Brown, 2001). Resilience education provides youths with a fijller sense of school as a community, higber levels of internal loci of control, and better conflict resolution skills (Watson, Battistich, & Soloman, 1997). Although theoretically and conceptually promising, more institutional research is needed to determine tbe effectiveness of the resilience educa- tion approach.

It is important to note that the effectiveness of both school-based and community-based drug prevention programs bas been mixed. There is a strong body of evaluation literature providing a critical review of tbe overall effectiveness of drug prevention and intervention programs (Beck, 1998; Gorman, 1996). School-based programs are among the most highly used methods of prevention and interventions. Social influence models (for example, social skills training and resistance skills) and drug education are tbe predominant strategy for drug use prevention and intervention. In a review of 12 prominent social skills training programs, Gorman found little or no effect on participants' alcohol use or related behavior. These 12 studies focused exclusively on alcohol use, with no mention of marijuana use, race, gender, or ethnic differences in outcomes. Other researchers have shown that both peer-led and adult-led strategies of preven- tion and interventions produce positive outcomes for students (Cuijpers, 2002;Toler et al., 2000). In a nieta-analytic study,Toler and colleagues found that peer-facilitated drug prevention programs are more effective tban adult-facilitated programs. Cuijpers found no difference between either type of scbool prevention programs.

in addition to scbool-based prevention program- ming, community-based initiatives have emerged to address drug use. In a review of community-based

W I L L I A M S E T AL. / Suhstance Use and Academic Perfi>rmance among African American Hi^ School Students 1S9

programs,Gorman (1996) found that these programs had minimal program effects. Similar to the school- based programs, alcohol was the primary focus of the community-based programs, with limited foci on urban settings. The current limitations of these studies are the lack of strong samples of numerous African American youths, and the primary focus of the intervention was alcohol use and related behaviors. Overall, the mixed results on the effec- tiveness of schooi-based and community-based drug intervention and prevention programming speak to challenges for intervention and prevention.

Over the past decade intervention research has made tremendous strides in developing and testing effective programs to address substance use behav- ion among youths. The best of these practices and promising programs have strong empirical evidence and positive outcomes.There are four such programs appropriate for high school students. Effective in- tervention programs such as Project Towards No Drug Abuse (Project TND) and the Behavioral Monitoring and Reinforcement Program arc model school-based programs that focus on substance use prevention for older middle school and high school students (Center for the Study and Prevention of Violence, 2006). Brief Strategic Family Therapy and Project Northland are both promising programs that address substance use and problem behaviors within the family and communitywide contexts (Center for the Study and Prevention of Violence).

Finally, this study highlights the effects of sub- stance use on positive intentions and academic performance of African American high school students. To promote the success of these students, schools must have the resources and personnel to implement programs and help African American students overcome the barriers to academic success. Economic and educational research indicates that African American high school students remain at a disadvantage in spite ofthe successful completion of high school. Without interventions, correlates for school failure may continue to diminish positive outcomes as these students enter adulthood.

REFERENCES Ajzen. I. (1991). The theory of planned behavior.

Oi^anizational Behavior and Human Decision Processes, 50,179-211.

Arthur, M. W.. Hawkins,J. D., Po!laTd,J. A., Catalano, R. F., & Baglioni,A.J.,Jr. (2002). Measuring risk and protective factors for substance use, delin- quency, and other adolescent problem behaviors: The Communities That Care Youth Survey. Evaluation Rii', 2(5,575-601.

Beck.J. (1998). lOO years of'just say n o " versus "just say know": Reevaluaiing drug education ^oals for the coming century. Fviiluation Revieti', 22, 15-45.

Brown,J.H.(2001).Youth, drugs, and resilience education, Journal of Drug Education. 31, 83-122.

Center for the Study and Prevention of Violence. (2006). Blueprints for violence prevention ovcrvieu'. Boulder: Univenity of Colorado, Institute of Behavioral Science. Retrieved June 25,2006, from www. colorado.edu/cspv/blueprints

Cuijpers, P. (2002). Petr-Ied and adult-led school drug prevention: A meta-analytic companion.Journal oJ' Drug Education, 32, 107-119.

Dornbusch,S. M., Erickson, K.. G., Laird.J.,& Wong, C.A. (2001).The relation of family and school at- tachment to adolescent deviance in diverse groups and communities. ̂ tiHf"'!/ of Adolescent Research, 16, 396-422.

EUickson. P. Bui, K.. Bell, R., & McGuigan, K. A. (1998). Docs early drug use increase the risk of dropping out of high school? /onrwfl/ ofDru^ Issues, 28,357-380.

Ensminger, M.E.,&: j u o n , H . S . (1998).Transition to adult- hood among high-risk youth. In R. jessor (Ed.), Neu> perspectiivs on adolescent risk beluwior (pp. 365-391). New York: Cambridge University Press.

Furstenberg, E E. Cook,T. D., Eccles.J., Elder, G. H . J r . , & Sameroff, A. (1999). Managing to make it: Urban families and adolescent success. Chicago: University of Chicago Press.

Gorman. I I H. (1996). Do school-based social skills train- ing programs prevent alcohol use among young people? r̂fdiffioH Research, 4{2). 191-210.

tiuagliardo, M. E, Huang, Z., Hicks,J.,& D'Angelo, L. (1998). Increased drug use among old-for-grade and dropout urban adolescents. American Journal of Preventiw Medicine, 15, 42—48.

Hallfors, D., & Van Dorn, R. A. (2001). Strengthening the role of two key institutions in the prevention of ado- lescent substance 3.buse.Journal of Adolescent Health, 30, 17-28.

Hawkins,J. D., Catalano, R. E, & Mi!ler,J.Y. (1992). Risk and protective factors for ali ohol and other drug problems in adolescence and early adult- hood: Implications for substance abuse prevention. Psychological Bulletin, 112. 64-105.

Hill, K. G., Etchison, K., Hawkins,J. D., & Catalano, R. E (2003, November). Ethnic dijferences in the criminal consequences of adolescent sub.-!tance u.<e. Paper presented at the American Society of Criminology, Denver.

Johnston, L. D., O'Malley. P. M , & Bachman.J. G. (2000, December 14). Ecstasy use rises sharply among teens in 2000: Use of many other drugs steady, but signifi- cant declines are reported for some [National Press Release], i^niversity of Michigan News and Informiition Services. Retrieved December 28, 20(11, from www. monitoringthefuture.org

McLoyd,V. C. (199H). Socioeconomic disadvantage and child development. American Psychologist, 53, 185-204.

McLoyd, V. C,Jayaratne, T. E., Cebailo, R , & Borquez, J. (1994). Unemployment and work interruption among African American single mothers: Effects on parenting and adolescent socio emotional tlinctioning. Child Development, 65. 562-589.

Mensch, B. S., & Kandel, D. B. (1988). Underreporting of substance use in a nation.^! longitudinal youth cohort. Public Opinion Quarterly, 52, 100-124.

O'Malley. P. M., Bachman.J. G., & Johnston. L. D. (1983). Reliability and con.sistency in self reports of drug use. International Journal of Addictions, 18. 805-824.

Paulson, A. (2006, March 3). Dropout rates high, but fixes under way. Christian Science Monitor. Retrieved June

160 Social Work Research VOLUME 31, NUMBER 3 SEPTEMBER 2007

25.2(M)6. from www,csmonitor.com/2006/0303/ p(tlsO2-legn.htm!

Peterson,P. L,. Hawkins.J, D.,Abbott.R, D.&; Catalano. R. F (1995), Disentangling the effects of parental drinking, family management, and parental akohol norms on current drinking by black and white ado- lcsi ents. In G. M. Boyd.J. Howard, & R. A. Zucker (Eds.), Alcohol problems among adolescents: Ctirrent direc- tions in preivntion research (pp. 33-59). Hillsdale. NJ: Lawrence Erlbaiini,

Pittman, L. 1>. ̂ Chase-Lansdale. P. L. (2001), African American adolescent girls in impovt-rished com- munities: Parenting style and adolescent outcomes.

Journal of Research on Adolescence, I t , 1 9 9 - 2 2 4 , Saunders.J,. Davis. L,,Williams,T. & Williams.J. H. (2004).

Gender diftVrences in self-perceptions and academic outcomes; A study of African American high school studcnu. Journal of Youth and Adolescence, .i3, 81-90.

Smith, C,,Lizotte,A,J.,Thornberry,T, I'.. & Krohn. M, D. (1995), Resilient youth: Identifying factors that pre- vent high-risk youth fTOni engaging in delinquency and drug use. In J. Hagen (Ed.), Delinquency and disrepute in the life course (pp. 217-247). Greenwich. CTiJAl Press.

Smith.J, R.. Bmoks-Gunn.J,, & Klebanov, P. K. (1997). Consequences of living in poverty for young chil- dren's cognitive and verbal ability and early school achievement. In G.J. Duncan & J. Brooks-Gunn (Eds.). Consetjucnces of growing up poor (pp. 132—189). New York: Russell Sage Foundation.

Tabachnick, B. C, & Fidell. L, S. (2000). Using multimriate statistia (4th ed,). Boston: Allyn & Bacon.

Toler. N. S.. Roona. M. R.. Ochshorn, P., Marshall, D, G., Streke. A.V. & Stackpolc. K, M. (2000). School-based adolescent drug prevention programs: 1998 meta- iinalysis. journal oJ Inimary Prevention, 20, 275—336.

Tucker. M. U.".& MitchelJ-Kernan,C. (1995).Trends in African-American family formation: A theoreti- cal and statistical overview. In M. B.Tucker & C, Mitchelt-Kernan (Eds.). Tlie decline in marriage among African-Americans: Causes, consequences, and policy implications (pp. 3-26), New York: Russell Sage Foundation.

Watson. M.. Battistich,V., & Soloman. D. (1997), Enhancmg students' social and ethical develop- ment in schools: An intervention program and its efiecK, liitcmatiorial Journal of Educational Research, 27, 571-586,

Williams.J, H.. Ayers. C D , Abbott. R, D,. Hawkins.J. D., & Catalano, R. F. (1999). Race differences in risk factors for delinquency and substance use among adolescents. Social IVork Research, 23,241-256,

Witlianis.T, R., Davis. L. E.. Miller-Cribbs.J,. Saunders, J.. & Williams.J. H. (2002). Friends, family, and neighborhood: Understanding academic outcomes of African American youth. Urban Education, 37, 408-^31,

Wood, M. D., Read,J. I*. & Palfai.T R. (2(W1). Social infiuence processes and college student dHnking:The mediational role of alcohol outcome expectancies. Journal of Studies on Alcohol, 62, 32—46.

Zimmerman. M A . & Schmeelk-Cone, K, H, (2003). A longitudinal analysis of adolescent substance use and school motivation among African American youth. Journal of Research on Adolescence, 13, 185—210.

James Herbert Williams, PhD, is dean and professor of sodal work. Graduate School of Social Work, Uniivrsity of Denver, 214H South High Street, Drntvr, CO 80208: e-mail: jatnes. [email protected]. Larry E. Davis, PhD, is dean and Donald

M. Henderson Professor, School of Social IVork, Cathedral of

Learttitig, University of Pittshtirgh. Sharon D.Johttson, PhD, is associate professor, Social M'ork Department, I'-niversity of

Missouri-Saint Louis. Tirina R. Williams, PhD, is assiS' tant professor. School of Social Work, Utiiversity of Michigan,

Ami Arbor. Jeanne A, Saunders, PhD, is assistant professor, School of Social Work. University of Iowa. Iowa City. Von E. Nebbilt, PhD, is assistant professor, School of Social Work, Howard University, Waslmgtoit, DC. Address correspottdcnce

to James Herbert Williams, PhD. TItis study, Choices of Life

for Adolescent Success (CLASS) MCJ-290644, was funded

by the National histitutes of Health, Maternal and Child

Health Division.

Original manuscript rtceived October 18, 3005 Finai revision received August 21, 2006 Accepted January 5, 2007

WILLIAMS ET AL. / Substance Use aneiAcatiemic Performance among African American Hi^ School Students 161