Tobacco-refusal skills and tobacco use among high-risk adolescents

Journal of Behavioral Medicine, VoL 16, No. 6, 1993

Tobacco-Refusal Skills and Tobacco Use Among High-Risk Adolescents

John P. Eider, 1 James F. Sallis, 2 Susan I. Woodruff , 1 and Mar ianne B. Wiidey 1

Accepted for publication: February 9, 1993

Psychosocial tobacco use prevention programs are based on the assumption that refusal skills training will have a suppressive effect on the onset of use by enabling non-using adolescents to refuse offers of cigarettes and smokeless tobacco. The present study investigated this assumption with 389 high-risk junior high-school students involved in a prevention program during their seventh, eighth, and ninth-grade years. Direct behavioral measures of refusal skills were taken by having subjects respond to audiotaped offers of tobacco and then rating the quality of their responses. These ratings were then linked to tobacco use measures obtained at the end of each of the 3 study years. Results showed that the comprehensive prevention program produced a favorable trend in delaying or preventing the onset of tobacco use. However, the refusal skills training, which was carried out throughout the 3-year intervention period, produced significant differences in overall refusal skill quality only at the seventh grade. Moreover, refusal skill quality was not related to overall tobacco use or cigarette use at any grade.

KEY WORDS: tobacco refusal skills; tobacco use; adolescents; prevention programs.

This work was supported by Grant R01-CA44921 from the National Cancer Institute. 1Graduate School of Public Health, San Diego State University, San Diego, California 92182. ZDepartment of Psychology, San Diego State University, San Diego, California 92182.

629

0160-7715/93/1200-0629507.00/0 �9 1993 Plenum Publishing Corporation

630 Elder, Sallis, Woodruff, and Wiidey

INTRODUCTION

Cigarette smoking is often cited as the single most preventable cause of death and disability in the United States, responsible for more than one of every six American deaths [Cummings et aL, 1989; Department of Health and Human Services (USDHHS), 1989]. Use of smokeless tobacco also is associated with significantly elevated risks of death and disease (USDHHS, 1986). Although cigarette smoking among Americans as a whole has de- creased since the publication of the first Surgeon General's report linking smoking to illness and disease (USPHS, 1964), the rate of decline still is not as steep as might be desired, especially in certain subgroups of the population. For example, the rate of young people initiating the smoking habit has been relatively resistant to change, with the equivalent of about 3000 new young persons becoming regular smokers each day in 1985 (Pierce et al., 1989). In addition, some evidence indicates that smoking on- set may be increasing among certain ethnic groups (Marcus and Crane, 1985).

Much research has been conducted to assess the effectiveness of dif- ferent techniques and programs to promote smoking cessation. However, because of the limited effectiveness of cessation efforts, attention has in- creasingly focused on prevention programs. These programs are largely targeted at the adolescent age group, because it is during this developmen- tal period that the smoking habit is usually initiated (Schinke et al., 1986).

Since the mid-1970s (Evans et al., 1976) the most promising preven- tion programs have been comprehensive psychosocial interventions designed to train adolescents in interpersonal skills thought to be effective in resisting social influences to use tobacco. These programs are based on empirical evidence linking smoking onset and peer usage, and the assump- tion that adolescents skilled in refusing tobacco will be less likely to try tobacco (or will be less likely to continue using tobacco if they have already begun experimenting). It has become almost "common knowledge" that psychosocial interventions which include tobacco refusal skills training will be effective in promoting the nonuse of tobacco. This assumes, however, that (a) such skills are appropriate and effective for resisting pressures and (b) participants actually learn and use the skills.

Several studies have evaluated the efficacy of tobacco-refusal skills training in meeting immediate and long-term prevention objectives. For ex- ample, adolescents tested for their ability to refuse tobacco after undergoing skills training demonstrated greater competency in refusing to- bacco than did controls (Hops et al., 1984; Sallis et al., 1990). Considering the effect of skills training on tobacco use behavior, Schinke and colleagues (1985) are among those reporting positive results, with intervention subjects

Tobacco-Refusal Skills in Adolescents 631

reporting lower past-week smoking than education-only and no-treatment controls at 6-, 12-, and 24-month follow-ups.

Most investigators in the area agree that psychosocial interventions have a positive effect on the onset of tobacco use and that refusal skills training specifically designed to counteract pressure to smoke enhances program effectiveness (Schinke et al., 1985; Flay, 1985). However, questions remain about how and why prevention programs work (McCaul and Glas- gow, 1985) and attempts have been made to identify specific mediators of program effects (MacKinnon et al., 1991). Sussman and colleagues (under review), for example, contrasted the relative effectiveness of three common components of psychosocial interventions, including (a) refusal skills train- ing, (b) general social skills training, and (c) physical consequences information. In addition, a comprehensive program was included as a com- parison, as was a "standard care" control group. Results showed that, while refusal skills training was effective for prevention of smokeless tobacco use, it was less effective than other programs (e.g., a comprehensive curriculum) for preventing the onset of cigarette smoking.

Information is also lacking about the conditions under which skills training is effective and for whom it is effective. For example, high-risk adolescents have the highest tobacco-use rates, yet may be less likely to respond to prevention efforts (Glynn et al., 1991). Furthermore, the degree to which the effectiveness of a skills training approach generalizes to vari- ous sociocultural groups is not known with certainty (Flay, 1985).

The present study assessed the effectiveness of a comprehensive to- bacco use prevention program with a refusal skills component for an ethnically diverse, high-risk adolescent group. The impact of behavioral skills training on ability to refuse tobacco was assessed by examining the quality of refusal skills for an intervention and control group over time. The long-term effect of skills training on tobacco-related behavior was evaluated by exploring the association between tobacco refusal skills with concurrent and subsequent tobacco use.

METHOD

Procedures

In 1988, 22 San Diego County high schools agreed to participate in a longitudinal study evaluating the effectiveness of a school-based tobacco prevention program, Project SHOUT (Students Helping Others Under- stand Tobacco). Approximately 1800 seventh-grade students at 11 of the schools were selected to serve as control subjects; a similar number at re-

632 Elder, Sallis, Woodruff, and Wildey

maining schools was assigned to the SHOUT intervention condition. Stu- dents and their parents provided consent for students to participate in a longitudinal study of adolescent tobacco and other substance use. Prior to the intervention, all participants were administered a self-report question- naire to collect baseline tobacco use, tobacco-related information, and demographic characteristics. The questionnaire was administered under bo- gus pipeline conditions in which saliva samples were obtained to increase the validity of subjective reports of tobacco use (Evans et al., 1977). Fol- lowing the SHOUT intervention and toward the end of the school year, all seventh-grade participants completed a post-intervention questionnaire identical to that used at baseline.

Early the next school year when participants had entered the eighth grade, the intervention group again received the SHOUT program, and toward the end of the year, all subjects provided questionnaire data and saliva samples. Due to logistical problems caused by school changes as par- ticipants entered the ninth grade, the SHOUT intervention evolved into a home-based, telephone "booster" program. As in the seventh and eighth grades, all participants provided questionnaire data and saliva samples, in classroom settings, toward the end of the school year.

SHOUT Intervention/Skills Training

Project SHOUT was a school-based comprehensive psychosocial in- tervention that combined various educational, activist, and behavioral strategies aimed at reducing the onset of regular tobacco use [see Elder et al. (1993) for a detailed description of Project SHOUT's content and struc- ture]. Project SHOUT was implemented over the 3 junior high-school years, with eight 50-min classroom sessions in both the seventh and the eighth grade followed by personalized phone calls and newsletters during the ninth grade. SHOUT classroom sessions were usually conducted during social studies and science classes, although this varied depending upon the individual school. A skills training component of the classroom sessions included learning about the role of social influences on tobacco use, re- hearsing methods of resisting pressure to use tobacco, practicing decision making, performing and watching tobacco-refusal skits, and earning prizes for practicing refusal skills. Skills training activities comprised approxi- mately 30% of the total classroom time.

When SHOUT evolved into a home-based "booster" program during the ninth grade, a total of five newsletters and four telephone calls was delivered to intervention students. Each four- to six-page newsletter in-

Tobacco-Refusal Skills in Adolescents 633

cluded information regarding local tobacco control events, the latest to- bacco-related research, articles on tobacco industry tactics, state and national tobacco legislation, cessation tips, and a question-and-answer col- umn. Telephone calls conducted by trained counselors would shortly follow each newsletter, and usually began with a discussion of one or two specific articles in the newsletter. The calls were designed to be as interactive as possible and covered a variety of topics such as opinion polls, normative education instruction, tobacco news and cessation, and refusal skills train- ing. Approximately 15% of the home-based program was devoted to practicing and review of tobacco refusal skills.

High-Risk Subsample

Subjects selected for specific testing of their refusal skills were a sub- group of the entire seventh-grade SHOUT sample that were considered at risk for tobacco use. This subgroup was chosen for skills testing for several reasons. First, behavioral skills testing is resource-intensive; therefore, test- ing of all SHOUT subjects was not feasible. Second, high-risk adolescents are an important group to consider in light of the fact that they comprise a disproportionate number of new smokers. Finally, a high-risk group would more likely provide variation in tobacco use and, consequently, sufficient statistical power to examine the relation between refusal skills and tobacco use. The subgroup was classified as high-risk based on several previously identified characteristics (Glynn et al., 1991) that were measured on the baseline questionnaire. Two types of individuals were included: (1) "experi- menters," who reported trying tobacco for the first time in the preceding 2 years but, at baseline, used tobacco no more than once a month, and (2) "never users," with a high predicted risk-of-use score based on peer and family tobacco use and school performance. More specifically, risk of use among never-users was estimated using a logistic function which was created by regressing tobacco use status (experimenter = 1, never-user = 0) on friends' and parents' tobacco use and school grades. A never-users predicted risk of use was estimated by substituting his or her reported val- ues on the independent variables into the logistic model, and students with the highest risk scores were selected for this study.

The resulting sample of 389 high-risk students (approximately 10% of the original SHOUT sample) was balanced by gender, SHOUT experi- mental condition, and risk status (experimenter versus never-user). The average age was 12.2 years (SD = 0.65), with a range from 11 to 15 years. Whites comprised about half of the sample, Hispanics represented 26%,

634 Elder, Sallis, Woodruff, and Wildey

and other racial/ethnic groups comprised 23%. Intervention-control group comparisons revealed that the two groups were not significantly different with regard to age, gender, or ethnic composition.

Because of resource limitations and because research objectives after the seventh-grade skills assessment were focused on examining the main-

tenance of refusal skills, it was decided that a 35% sample of the high-risk group (n = 135) would be sufficient for further skills testing. This sample was randomly selected and balanced by gender, experimental condition, and risk status. At the ninth grade, 30 students were unavailable for testing; the result was a group of 105 individuals tested at the final skills assessment.

Skills Assessments

Behavioral assessments of refusal skills were conducted at the sev- enth, eighth, and ninth grades following the SHOUT intervention. The test was based on that developed by Hops et al. (1986), and it was modi- fied by Sallis et al. (1990). The refusal test consisted of 23 realistic audio- taped scenarios in which peers offered subjects cigarettes (15 scenarios) and smokeless tobacco (eight scenarios) [see Sallis et aL (1990) for a de- tailed description of the skills assessment tool]. Subjects were instructed to refuse tobacco offers, and their verbal refusal responses to the sce- narios were recorded. Each recorded response was coded for content by trained coders in a manner indicating whether or not it was characterized by each of seven refusal categories: (1) simple direct refusal (e.g., "No thanks"), (2) reason (e.g., "It causes cancer"), (3) suggestions for alter- native activities accompanying a refusal (e.g., "No, let's go on the rides instead"), (4) direct "I" statement (e.g., "I don't want one"), (5) suppor- tive statement (e.g., "Not for me, but you go ahead"), (6) ignore/with- drawal (e.g., "Come on, let's go dance"), and (7) aggressive (e.g., "You're an idiot for smoking"). A response could potentially contain several of these categories. For example, the response, "No thanks. It makes your teeth yellow," was coded as containing both a simple direct refusal and a reason.

To minimize coder drift and ensure reliability, coders were retrained after one-fourth, one-half, and three-fourths of the responses had been coded. Coders' reliability was tested on approximately 10% of the coded cases. Results showed that reliability coefficients ranged from .78 (ig- nore/withdrawal) to .97 (simple direct refusal), with a mean of .89 across refusal categories.

Tobacco-Refusal Skills in Adolescents 635

Measures

Quality of Refusal Skills. Two scores indicating quality of tobacco re- fusai were computed for each subject at each grade: (1) an overall refusal quality score that reflected responses to both smokeless tobacco and ciga- rette offers and (2) a cigarette refusal quality score that reflected responses to cigarette offers only. In each case, responses to each scenario were first scored as poor, fair, or good (coded as 1 to 3, respectively). These quality ratings were based on formative research conducted with 12-year-old sub- jects (Sallis et al., 1990) and reflected the goals of the intervention. Responses operationalized as most appropriate and effective were coded as "good." For example, a simple direct refusal followed by a reason, a response combination prompted during intervention activities, was consid- ered effective and socially appropriate. Examples of "fair" responses were ones consisting of either a simple direct refusal or a reason. Responses were labeled "poor" if they were either nonassertive refusals (e.g., ignoring the offer or withdrawing from the situation) or aggressive (e.g., personally insulting).

For the overall refusal quality measure, a mean of the 23 quality rat- ings was computed. The internal consistency of this scale (Cronbach's alpha) averaged over the three years was .80. To construct the cigarette refusal quality score, a mean of the 15 cigarette-related quality ratings was computed. Cronbach's alpha for this measure averaged over the three years was .72.

Tobacco Use Measures. Smoking information was based on an item from the questionnaires that asked about the number of times the student smoked cigarettes in the last 30 days. Response options included 1 (none), 2 (1-3 times), 3 (4-6 times), 4 (7-9 times), and 5 (10 or more times). In the present study, the smoking variable was employed as a score ranging from 1 to 5, and it also was used to compute rates of smoking. For com- putation of rates, students indicating that they smoked cigarettes at least once during the last 30 days (a score of 2 or greater) were categorized as smokers.

A measure of generic tobacco use (i.e., use of either cigarettes or smokeless tobacco) was constructed from the smoking item and an identical one asking about smokeless tobacco use. These two items were combined into a mean generic tobacco use score for some analyses. Rates of generic tobacco use were also computed, with students reporting that they smoked or used tobacco at least once during the last 30 days categorized as generic

tobacco users.

636 Elder, Sallis, Woodruff, and Wildey

Analyses

Trends in tobacco use rates over four assessment periods (i.e., base- line, seventh, eighth, and ninth grade) were examined for the high-risk sample, as were 3-year longitudinal changes in overall tobacco refusal qual- ity and cigaret te refusal quality. Chi-square analysis and t tes ts for independent samples were conducted within each assessment period to test for differences between intervention and control groups on tobacco use and refusal skills variables.

A series of multiple regression procedures was used to investigate the relationship between cigarette and generic tobacco use with refusal skills and demographic variables. More specifically, generic tobacco use was the dependent variable, with prior- and concurrent-grade overall refusal qual- ity, ethnicity, gender, and experimental condition serving as independent variables. Similarly, cigarette use was predicted from prior- and concur- rent-grade cigarette refusal quality, ethnicity, gender, and experimental condition.

30

,- 20

O E

C

u 10 [1.

Baseline 7th 8th 9th Grade Grade Grade

�9 Control

[ ] Intervention

A s s e s s m e n t

Fig. 1. Past-month cigarette use by assessment period. Note: The number of subjects providing tobacco use information decreased over time, from 325 at baseline to 247 at the ninth-grade assessment.

Tobacco-Refnsal Skills in Adolescents 6 3 7

q~

Q.

RESULTS

30 ~ [] Control

[ ] Intervention

20

10

Baseline 7th 8th 9th Grade Grade Grade

Assessment

Fig. 2. Past-month generic tobacco use by assessment period. Note: The number of subjects providing tobacco use information decreased over time, from 325 at baseline to 247 at the ninth-grade assessment.

Figures 1 and 2 present rates of cigarette and generic tobacco use by experimental condition for the high-risk sample at baseline and three postintervention assessment periods. In general, tobacco use increased more sharply for control subjects than among intervention participants. At the final postassessment at the ninth grade, the groups reached widest di- vergence, with approximately 24% of control subjects reporting cigarette use compared to 16% of intervention subjects. With regard to generic to- bacco use at the end of the ninth grade, 27% of controls reported such use versus 19% of the intervention group. Although the trends suggest a tendency for improvement in the intervention group relative to high-risk controls, chi-square analyses conducted within each assessment period failed to show a statistical difference between the groups in the percentage using cigarettes or generic tobacco. Nevertheless, the observed pattern of differences seen over time paralleled an overall between-group significant difference found for the entire SHOUT sample (Elder et al., 1993).

Figure 3 presents mean overall refusal quality scores by group for skills tests conducted late in the seventh, eighth, and ninth grades. Only at the seventh-grade assessment were reliable group differences indicated,

638 Elder, Sallis, Woodruff, and Wildey

>, 3.00 . i , . ,

- i

O 1.75

m 1.50

> 1.25 0 r -

1.00

�9 Control r-- c~ [ ] Intervention

7th 8th 9th Grade* Grade Grade

Assessment

Fig. 3. Overall refusal quality by assessment period. Note. The number of subjects tested for quality of refusal skills decreased over time, from 389 at the seventh-grade assess- ment, to 135 at the eighth-grade assessment, to 105 at the final ninth-grade assessment. *Groups different at the .05 level of significance.

" 3.oo "1 m -1

O

-~ 1.75

" 1

rr 1.50

Q ;

m cn 1.25 i5 e-

l .00

�9 Control �9 Intervention

7th 8th 9th Grade Grade Grade

Assessment

Fig. 4. Cigarette refusal quality by assessment period. Note: The number of subject tested for quality of refusal skills decreased over time, from 389 at the seventh-grade assessment, to 135 at the eighth-grade assessment, to 105 at the final ninth-grade assessment.

with the intervention students having a significantly higher mean refusal quality score [t(387) = -1.96, p < .05]. Despite the initial significant group difference, by ninth grade the refusal scores for the groups had converged.

Tobacco-Refusal Skills in Adolescents 639

Figure 4 presents a similar description of cigarette refusal quality. No group differences reached statistical significance, although the spread between group means and the pattern of change were similar to that seen for the overall refusal quality measure.

Multiple regression analyses were conducted at each assessment period to determine the association between the two tobacco use scores with con- current and past-grade refusal skills level. In addition to skills level, experi- mental condition, gender, and ethnicity were included as predictor variables.

Regression results indicated that generic tobacco use was not related to experimental condition, ethnicity, or seventh grade overall refusal skill level. Gender of the subject, on the other hand, made a small but significant contribution, accounting for 2% of the variance in generic tobacco use in the seventh grade [F(1, 320) = 6.16, p _< .05]. The direction of the relation- ship indicated that seventh-grade girls were more likely to be using some form of tobacco than were boys. The prediction of cigarette smoking at the seventh grade was similar to that seen for generic tobacco use. Quality of cigarette refusal, experimental condition, and ethnicity made no contribution to the prediction of seventh grade smoking, and gender accounted for a significant yet small 1% of the variance [F(1, 317) = 5.55, p < .05].

At the eighth grade, generic tobacco use was weakly associated with gender [R e = .05; F(1, 103) = 6.04, p < .05], yet no significant associations were found with prior seventh-grade overall skill level, concurrent eighth- grade skill level, ethnicity, or condition. Regression results for cigarette smoking at the eighth grade closely paralleled that seen for eighth-grade generic tobacco use. Cigarette smoking was related to gender [R 2 = .06; F(1, 102) = 7.63, p < .01] but not to the other variables.

Neither generic tobacco nor cigarette use at the ninth grade was sig- nificantly associated with condition, ethnicity, refusal skill level at the two previous grades, or concurrent ninth-grade skill level. Further, the modest association between tobacco use and gender seen at the seventh and eighth grades was not found at the ninth grade. However, for this ninth-grade analysis, data were required for all three skills assessments. Because the number of students providing ninth-grade questionnaire smoking data and three skills assessments was small (n = 59), there may have been inade- quate power to detect significant relationships.

DISCUSSION

The present study showed that among adolescents at high risk for tobacco use, a psychosocial prevention program implemented over the 3 junior-high years produced favorable (albeit not statistically significant)

640 Elder, Sallis, Woodruff, and Wildey

trends in delaying or preventing the onset of tobacco use. Like other state- of-the-art prevention curricula, Project SHOUT places a heavy emphasis on interpersonal skills development, especially those related to resisting peer pressure. Analogue assessment of these skills indicated that the first year of intervention produced a salutary change in overall refusal skill qual- ity. However, this effect was no longer significant at the end of the second year and was even less apparent after the third and final year of interven- tion. This diminishing effect occurred in spite of resistance skill boosters delivered through the latter 2 intervention years. Moreover, resistance skill quality was not significantly related to generic tobacco use or cigarette use at any grade level. Only the gender of the subject proved to be significantly predictive, with girls reporting higher tobacco use than boys.

The present results run counter to the current prevention zeitgeist in two ways: skills training did not result in long-term skill acquisition, and skill acquisition did not suppress tobacco use. It is likely that the skill train- ing format (and measu remen t p ro toco l ) p red ica ted on behaviors appropriate for 12 year olds became less appropriate over time, thereby leading the subjects to forsake these skills as they progressed through this rapid period of development. Consistent with operant learning theory, moreover, performance deficits are not necessarily rooted in skill deficits. Powerful contingencies of reinforcement in the natural environment can override behaviors learned in an educational setting, no matter how well trained these behaviors are. Students who are eager to please their peers by accepting offers of tobacco may conceivably be the same students who are eager to impress their teachers with skillful performance on an ana- logue test of their refusal skills.

In spite of the results among this high-risk subgroup, Project SHOUT had an overall substantial prevention effect when considering the entire cohort. Intervention and control group differences actually widened during the final year when the intervention format was changed from classroom lessons to telephone calls and newsletters (Elder et al., 1993). It is felt that while the skills training component did not produce a significant inde- pendent effect, it comprised an important component of a comprehensive package that utilized antiindustry exercises, lotteries, skilled undergraduate facilitators, and personalized as well as group interactions.

One of the most notable findings from the present study is that the prevention program did not produce long-term changes in refusal skills, and those skills did not predict tobacco use. These results call into question the central theoretical role of peer pressure resistance in smoking preven- tion (Evans and Raines, 1982). The results also suggest that current programs may not be effective in maintaining tobacco refusal skills, at least among those at-risk for tobacco use. The theoretical basis of tobacco pre-

Tobacco-Refusal Skills in Adolescents 641

vention programs may need to be reconsidered. It is possible that the role of refusal skills training has been overemphasized.

One of the limitations of the present study is related to the use of a high-risk study group. These youth were selected because they were already experimenting with tobacco or because they had risk factors (e.g., peer and family use) for tobacco use. Thus, the refusal skills component may be less effective for this group whose skills are severely and frequently tested. Gen- eralizability of these findings to the general adolescent population should be evaluated and not assumed.

Future studies should continue to examine what types of prevention protocols are most effect for what types of participants, especially in terms of their gender, age, and risk status. Refinements in the skill training as- pects of both intervention and measurement protocols should be made with an eye toward the developmental appropriateness of the target skills. Meas- urement should continue to include direct observations of refusal skills in order to determine whether, at a minimum, short-term intervention goals are being met. To the greatest extent possible, however, interventions must do a better job at remediating performance deficits as well as skills deficits by altering the social and interpersonal environments in such a way as to promote the nonuse of tobacco.

REFERENCES

Cummings, S. R., Rubin, S. M., and Oster, G. (1989). The cost-effectiveness of counseling smokers to quit. JAMA 261: 74-79.

Elder, J. P., Wildey, M., de Moor, C., Sallis, J. F., Eckhardt, L., and Edwards, C. (1993). The long-term prevention of tobacco use among junior high school students: Classroom and telephone interventions. Am. J. PubL Health. 83: 1239-1244.

Evans, R. I., and Raines, B. E. (1982). Control and prevention of smoking in adolescents: A psychological perspective. In Coates, T. J., Peterson, A. C., and Perry, C. (eds.), Promoting Adolescent Health: A Dialogue on Research and Practice, Academic Press, New York, pp. 101-136.

Evans, R., Hansen, W., and Mittelmark, M. (1977). Increasing the validity of self-reports of smoking behavior in children. J. Appl. Psychol. 62: 521-523.

Flay, B. R. (1985). Psychosocial approaches to smoking prevention: A review of the findings. Health Psychol. 4: 449-488.

Glynn, T. J., Anderson, D. M., and Schwartz, L. (1991). Tobacco-use reduction among high-risk youth: Recommendations of a National Cancer Institute expert advisory panel. Prevent. Med. 20: 279-291.

Hops, H., Weissman, W., Biglan, A., Thompson, R., Failer, C., and Severson, H. H. (1986). A taped situation test of cigarette refusal skills among adolescents. Behav. Assess. 8: 145-154.

MacKinnon, D. P., Johnson, C. A., Pentz, M. A., and Dwyer, J. H. (1991). Mediating mechanisms in a school-based drug prevention program: First-year effects of the Midwestern Prevention Project. Health Psychol. 10: 164-172.

Marcus, A. C., and Crane, L. A. (1985). Smoking behavior among Hispanics: A preliminary report. World Smoking and Health, Spring.

642 Elder, Sallis, Woodruff, and Wildey

McCaul, K. D., and Glasgow, R. E. (1985). Preventing adolescent smoking: What have we learned about treatment construct validity? Health Psychol. 4: 361-387.

Pierce, J. P., Fiore, M. C., Novotny, T. E., Hatziandreau, E. J., and Davis, R. M. (1989). Trends in cigarette smoking in the United States: Projections to the year 2000. JAMA 261: 61-65.

SaUis, J. F., Elder, J. P., Wildey, M. B., de Moor, C., Young, R. L., Shulkin, J. J., and Helme, J. M. (1990). Assessing skills for refusing cigarettes and smokeless tobacco. J. Behav. Med. 13: 489-503.

Schinke, S. P., Gilchrist, L. D., and Snow, W. H. (1985). Skills intervention to prevent cigarette smoking among adolescents. Am. J. Publ. Health 74: 665-667.

Schinke, S. P., Gilchrist, L. D., Schilling, R. F., Snow, W. H., and Bobo, J. K. (1986). Skills methods to prevent smoking. Health Educ Q. 13: 23-27.

Sussman, S., Dent, C. W., Stacy, A. W., Sun, P., Craig, S., and Simon, T. R. (1993). Project Towards No Tobacco Use one-year behavior outcomes (under review).

U.S. Department of Health and Human Services (1986). The health consequences of using smokeless tobacco. A report of the advisory committee to the surgeon general, 1986. USDHHS Public Health, Service NIH Publication No. 86-2874.

U.S. Department of Health and Human Services (1989). Reducing the health consequences of smoking: 25 years of progress. A report of the surgeon general. USDHHS Publication No. (CDC) 89-8411.

U. S. Public Health Service (1964). Smoking and health. Report of the Advisory Committee to the Surgeon General of the Public Health Service. USDHHS Public Health Service, PHS Publication No. 1103.