Categories
American Journal of Drug and Alcohol Abuse

Determinants of Youth Tobacco Use in West Virginia: A Comparison of Smoking and Smokeless Tobacco Use

Determinants of Youth Tobacco Use in West Virginia: A Comparison of Smoking and Smokeless Tobacco Use – Statistical Data Included

Kimberly A. Horn

Since the 1964 U.S. Surgeon General’s Report first revealed the health hazards of tobacco, smoking and smokeless tobacco (ST) use have declined substantially in the upper and middle income groups. This has not been the case, however, among lower income groups, such as those in West Virginia. Unfortunately, this trend transcends age, and as such, youths who live in rural areas are at greater risk for using tobacco products than their nonrural counterparts (1, 2).

There are approximately 8000 West Virginians who start smoking or using other forms of tobacco each year, most of them are under age 18 (3, 4). Moreover, the 1997 Youth Risk Behavior Survey (YRBS) (3) found West Virginia youths had one of the highest rates of both current and frequent smoking in the United States. Specifically, 41.9% of West Virginia high school youths reported using tobacco products on 1 or more of the 30 days preceding the survey (current use), and 24.1% had smoked cigarettes on 20 or more of the 30 days preceding the survey (frequent use) (3). Further, West Virginia ranked as one of the highest in ST use nationally with 15.8% of its youths currently using compared to a U.S. average of 11.5%.

Both smoking and ST use are associated with other high-risk behaviors during adolescence. For example, adolescent smoking is associated with alcohol and illicit drug use, violence, suicide, stress, depression, and unprotected sex (5). The early onset of ST use has also been associated with alcohol consumption, marijuana use, and physical risk taking (6). Most youths who use tobacco products will continue beyond adolescence (5, 7, 8), increasing their risks for premature disease and death in adulthood. More specifically, youths who smoke increase their risk for eventually developing cardiovascular disease and various cancers, especially lung cancer (7). Continued ST use has serious health consequences as well. Research shows that ST use is related to gum recession, oral cancer, oral leukoplakia (9-11), adverse cardiovascular effects (12), and gastrointestinal cancer (11). Given these adverse health effects, preventing and intervening with tobacco use have serious implications for national health care. To illustrate, the heavy consumption of tobacco use in West Virginia, along with early initiation, clearly has an impact on the state’s health care costs. West Virginia has the fourth highest death rate from tobacco-related illnesses in the United States (4).

It is widely recognized that the most effective way to reduce tobacco-related morbidity and mortality is to prevent the initiation of tobacco use among youths, particularly those at high risk. In doing so, however, it is imperative to understand the reasons that youths start and maintain tobacco use. Empirical studies have revealed several primary determinants or risk factors for tobacco use. Some of these include favorable attitudes toward tobacco use, inadequate knowledge about tobacco and health, family tobacco use, tobacco-using friends, family problems, and school problems (13, 14).

An extensive review of the literature on youth tobacco use revealed limited comparisons of the risk factors for smoking and ST use, and fewer studies on the determinants of conjoint smoking and ST use. Information on the determinants of tobacco use among rural youths is even scarcer.

The current study examined the risk factors associated with four exclusive tobacco use categories (i.e., smoking only, ST use only, conjoint smoking and ST use, and non-tobacco use) among youths in rural West Virginia. Many past studies on tobacco use examined ST use and smoking as nonexclusive categories (e.g., 2, 13). To date, few studies have compared youths who use these two types of tobacco exclusively, as well as those who use both, given a single sample. The purpose of this article is twofold. The first is to identify and compare the determinants of smoking only, ST only, and the conjoint occurrence of smoking and ST use among a sample of rural youths. The second is to discuss the differences among youths in these exclusive tobacco use categories and implications for tobacco cessation interventions.

Participants

METHOD

Participants (n = 883) included adolescents from the class of 2000 (ninth graders in 1996) in rural West Virginia. This sample was drawn from ninth grade students from nine high schools, representing four West Virginia counties.

Setting

West Virginia is a predominantly white (95%), rural state. It is the only state in the nation that is fully encompassed by Appalachia and has a markedly low per capita income ($15,598), with almost one-fourth living below the poverty line. West Virginia is comprised of 55 counties, and each county is largely comprised of small communities (15). The small rural communities from which the school sites were drawn reflected the demographics of West Virginia. Specifically, almost all of the schools were located in communities with less than 3000 residents, except for one community that had a population of 15,000. Approximately 95% of all residents in these communities are white. Schoolwide populations ranged between 178 and 1045 students, and the percentage of ninth graders per school was between 10% and 29%.

Measurement

After an extensive review of the literature, we determined that there were no available instruments that examined all of the factors of interest, particularly with the target population. In the best interest of West Virginia and the study, we developed a new instrument. As such, data were gathered through a self-administrated, teacher-supervised survey entitled the Year 2000 Youth Tobacco Assessment (Y2YTA). The Y2YTA was a 77-item survey that contained questions about tobacco use attitudes, knowledge, and behavior, as well as a section on pyschosocial and demographic variables.

Content validity. The content validity of the instrument used in this study was assessed via expert review and pilot testing. Experts in tobacco and health services research reviewed the instrument and determined that the questions were relevant to conceptual domains and existing literature on youth tobacco use. Based on the expert feedback, changes were made in the wording and ordering of some items. Following the expert review, a pilot test was conducted with 6 adolescents of varying ages and reading levels.

Reliability. Although there are several types of reliability (e.g., stability, equivalence, inter- and intrarater reliability), internal consistency (homogeneity) is particularly important for multi-item scales (16). In this study, the internal consistency of the multi-item scales (knowledge and attitude) was confirmed by calculating the coefficient alpha and item-to-total correlations when appropriate.

Survey content. The survey contained four parts: (A) Tobacco Use, (B) Knowledge About Tobacco and Health, (C) Attitudes Toward Tobacco, and (D) Psychosocial Profile.

Part A included 26 heterogeneous items about tobacco use (smoking and ST) among youths and their friends and family. The items in this part included both dichotomous and continuous response options (e.g., “Does your mother smoke?” “How many of your close friends smoke?”). Similar questions were formatted for ST use. For the purpose of this study, tobacco use status was assessed by two specific questions in part A. This classification process is described below.

Part B (Knowledge About Tobacco and Health) included 10 true/false statements (e.g., “Lung cancer is the only cancer caused by smoking,” “Smokeless tobacco is a safe alternative to smoking cigarettes”). These 10 items constituted a multi-item scale (alpha = .51). Although the reliability of the knowledge scale was relatively low on administration, no item with low item-total correlations was deleted as all items were considered relevant measures of knowledge. The scale range was between 0 and 10.

Part C (Attitudes Toward Tobacco) included 16 items using a 5-point Likert scale, with 1 being “strongly agree” and 5 being “strongly disagree” (e.g., “Smoking is a good way to keep your weight down,” “Using smokeless tobacco helps you look cool”). These items were also combined as a multi-item scale (alpha = .71). The potential score range of the scale was between 16 and 80.

Part D included 10 items relating to self-perception and school and family problems (e.g., “I like school most of the time,” “I like myself most of the time”). Again, response options were defined at the poles from 1 (“strongly agree”) to 5 (“strongly disagree”). Fifteen additional items pertained to demographics such as place of residence, family composition, age, race, and gender.

Tobacco use classification. Because we were interested in exploring the knowledge, attitudes, and behaviors of current tobacco users, student respondents were assigned to one of two groups: current users and current nonusers. For instance, using items from part A, smoking status was determined by asking: “Do you smoke cigarettes?” Current nonsmokers included those who never smoked or those who quit. Youths who smoked at least occasionally (i.e., anything other than everyday use) or every day were categorized as current smokers. Students also were asked to report their ST use. ST status was determined by asking: “Do you use smokeless tobacco?” Current users and current nonusers of ST were defined using the same procedures described for current smokers and nonsmokers.

Four groups resulted: current smokers, current nonsmokers, current ST users, and current nonusers. These groups were nonexclusive (i.e., current smokers could also be current ST users). To obtain a better understanding of youth tobacco use, a new variable was created. The values of this variable represented four exclusive categories: nontobacco users, smokers only, ST users only, and conjoint smokers and ST users. Using this classification, participant responses could be in one and only one category.

Knowledge, attitude, and psychosocial factors. Factors of knowledge and attitude were assessed using a summated score from the Y2YTA knowledge and attitude scale (parts B and C), respectively. High scores reflected low levels of knowledge and favorable attitudes toward tobacco use.

Psychosocial factors were assessed using specific items from Y2YTA part A and part D. These items related to questions about the following: smoking and ST use among mother, father, and siblings; overall perception of how well youths get along with their families; smoking among close friends; ST use among close friends; and overall perception of how well youths get along at school. Higher scores reflected more smoking and ST use among family members and close friends and negative perceptions of getting along with family and school. The demographic factors included in this study were gender and age. Race was not included since 95% of the target population was white. In sum, this study examined the relationship between tobacco status (nontobacco users, smokers only, ST users only, conjoint smokers and ST users) and 14 separate factors (Fig. 1).

[Figure 1 ILLUSTRATION OMITTED]

Procedure

Prior to survey administration, participating schools obtained approval from the Department of Education and written parental consent. Although youths who did not want to participate were excused, few students refused. The overall response rate of all eligible students was 86.1%.

Participants were administered surveys in their school classrooms. To increase reliability, procedures were age appropriate and standardized. Teachers read aloud from scripted instructions that explained the general purpose of the survey. The format of the survey was reviewed, and students were allowed to ask questions. In addition, students were informed that neither teachers nor other school personnel would view their answers. After hearing explanations about the survey and matters of confidentiality, participants were asked to respond truthfully and privately to all questions. Students read and responded to the survey on their own; a teacher was present to answer questions. To ensure confidentiality, participants were asked to deposit their completed surveys into an envelope that was sealed and subsequently put into a “drop box.” The researchers collected the sealed envelopes later.

RESULTS

Sample Characteristics

Among 883 participants, there were approximately equal portions of males (49.2%) and females (48.8%). All participants were ninth graders from rural West Virginia public schools, ranging between 13 and 19 years old (M = 14.6 years). Consistent with the racial and ethnic composition of West Virginia, 95% were white, 0.9% were African-American, 1.6% were Native American, 0.6% were Asian-American, and 1.9% were of other racial/ethnic origin.

Overall Tobacco Use

Among sampled youths, 31.8% were current smokers, and 16.1% were current ST users. However, these two categories are nonexclusive. For instance, among students who were current ST users, 63.2% were also current smokers. On classification of youths into the four exclusive categories of current tobacco use, 20% were smokers only, 6% were ST users only, 10% were conjoint smokers and ST users, and 64% were nonusers of tobacco.

Table 1 illustrates a general comparison of means by multivariate analysis of variance (MANOVA) for all 14 tobacco use determinants using the four exclusive categories of tobacco use. The differences among the exclusive categories of tobacco use were significant for all factors: gender, age, attitude toward tobacco use, knowledge about tobacco use and health, smoking and ST use among parents and siblings, smoking and ST use among friends, and getting along at home and school.

Table 1. General Comparisons of Students in Different Tobacco Use Groups

Risk No tobacco Smoking

indicators use(a) only(b)

Gender

Male 46% (212) 35% (48)

Female 54% (248) 65% (89)

M [+ or -] SD M [+ or -] SD

Age 14.41 [+ or -] 0.61 14.68 [+ or -] 0.79

Attitude 18.74 [+ or -] 6.62 26.02 [+ or -] 7.54

Sibling smoking 1.33 [+ or -] 0.59 1.88 [+ or -] 0.78

Sibling ST use 1.18 [+ or -] 0.43 1.30 [+ or -] 0.54

Mother smoking 1.81 [+ or -] 0.90 2.19 [+ or -] 0.91

Mother ST use 1.05 [+ or -] 0.29 1.05 [+ or -] 0.28

Father smoking 1.92 [+ or -] 0.87 2.17 [+ or -] 0.86

Father ST use 1.69 [+ or -] 0.86 1.86 [+ or -] 0.91

Family problems 1.78 [+ or -] 0.71 2.53 [+ or -] 1.26

Friends’ smoking 3.02 [+ or -] 1.69 4.88 [+ or -] 0.49

Friends’ ST use 2.68 [+ or -] 1.75 3.30 [+ or -] 1.75

Knowledge 3.69 [+ or -] 1.53 3.77 [+ or -] 1.50

School problems 1.71 [+ or -] 0.59 1.88 [+ or -] 0.74

Risk Using ST Smoking and

indicators only(c) using ST(d)

Gender

Male 95% (39) 90% (65)

Female 5% (2) 10% (7)

M [+ or -] SD M [+ or -] SD

Age 14.81 [+ or -] 0.81 14.91 [+ or -] 0.88

Attitude 25.86 [+ or -] 5.38 30.84 [+ or -] 9.44

Sibling smoking 1.32 [+ or -] 0.59 1.80 [+ or -] 0.83

Sibling ST use 1.57 [+ or -] 0.73 1.66 [+ or -] 0.74

Mother smoking 1.92 [+ or -] 0.87 2.20 [+ or -] 0.90

Mother ST use 1.16 [+ or -] 0.49 1.19 [+ or -] 0.56

Father smoking 2.03 [+ or -] 0.87 2.27 [+ or -] 0.91

Father ST use 2.00 [+ or -] 0.91 2.22 [+ or -] 0.84

Family problems 1.76 [+ or -] 0.83 2.21 [+ or -] 0.97

Friends’ smoking 4.19 [+ or -] 1.40 4.53 [+ or -] 0.91

Friends’ ST use 4.29 [+ or -] 1.31 4.60 [+ or -] 0.90

Knowledge 3.90 [+ or -] 1.81 4.74 [+ or -] 2.23

School problems 1.90 [+ or -] 0.77 2.14 [+ or -] 0.74

The number of students in each group may change due to missing values of the variable.

The p values for all variables are less than .001.

(a) N = 462 (65%).

(b) N = 141 (20%).

(c) N = 42 (6%).

(d) N = 72 (10%).

Determinants of Different Types of Tobacco Use

To determine the extent to which the 14 factors predicted tobacco use, separate multiple logistic regression models were used for the exclusive tobacco use categories. Nonusers of tobacco constituted the comparison group. In each model, forward stepwise selection was used. Adding and removing variables was based on the likelihood ratio statistic using conditional parameter estimates. The model coefficients for independent factors; were transformed to odds ratios (ORs). This approach provided a comparison of the unique contribution of each factor to the model. Comparison of the data in the correlation matrices revealed no evidence of a problematic level of multicolinearity among predictors.

Smoking only. Multiple logistic regression analysis for the smoking-only group revealed that smoking among friends, favorable attitudes toward tobacco use, and family problems were significant determinants (see Table 2). As Table 2 indicates, significant predictors were smoking among close friends (OR = 2.75), smoking among siblings (OR = 1.96), having family problems (OR = 1.70), and favorable attitudes toward tobacco (OR = 1.11). Furthermore, the stepwise model indicated that the same variables were significant after adjusting for all other variables. Coefficients, ORs, and related statistics are shown in Table 2. The overall percentage accuracy classification (PAC) was 85.68%. The goodness of fit was 401.77, and the log likelihood (LL) was -141.77. The chi square was significant ([chi square] = 171.06, p [is less than] .001).

Table 2. Results of Multiple Logistic Regression Analysis Among Students in the Smoking-Only Group

Risk indicators B P OR(a) 95% CI(b)

Friends’ smoking 1.01 <.0001 2.75 1.87-4.03

Sibling smoking 0.68 .0011 1.96 1.31-2.95

Family problems 0.53 .0003 1.70 1.27-2.27

Attitude .10 <.0001 1.11 1.06-1.16

(a) OR = odds ratio using coefficients obtained with logistic regression.

(b) 95% CI = 95% confidence intervals.

ST use only. A multiple logistic regression analysis for ST users only was conducted. Most (95%) ST users were male. Therefore, gender was not included because it is an obvious determinant and because it represented the overwhelming majority of the ST sample. Excluding gender, results showed that ST use among siblings (OR = 4.28) and close friends (OR = .71) and favorable attitudes toward tobacco use (OR = 1.12) were significant factors. Coefficients, ORs, and related statistics are shown in Table 3. The overall PAC was 92.84%. The goodness of fit was 242.54, and the LL was -64.24. The chi square of the model was significant ([chi square] = 46.32, p [is less than] .001).

Table 3. Results of Multiple Logistic Regression Analysis Among Students in the ST-Only Group(a)

Risk indicators B P OR(b) 95% CI(c)

Sibling ST use 1.45 .0001 4.28 2.07-8.84

Friends’ ST use 0.53 .0016 1.71 1.23-2.38

Attitude .11 .0003 1.12 1.05-1.19

(a) Gender factor was excluded from the analysis.

(b) OR = odds ratio using coefficients obtained with logistic regression.

(c) 95% CI = 95% confidence intervals.

Conjoint smoking and ST use. Finally, a multiple logistic regression was conducted for conjoint users. Analysis revealed that 6 of 14 factors explored in this study were significant for this group. These included gender (OR = 8.62); sibling ST use (OR = 3.09); ST use among close friends (OR = .13); knowledge (OR = 1.39); family problems (OR = 2.41); and favorable attitudes toward tobacco use (OR = 1.15). Coefficients, ORs, and related statistics are shown in Table 4. The overall PAC was 92.31%. The goodness of fit was 409.91, and the LL was -61.37. The chi square of the model was significant ([chi square] = 138.33, p [is less than] 001).

Table 4. Results of Multiple Logistic Regression Analysis Among Students in the Conjoint Use Group

Risk indicators B P OR(a) 95% CI(b)

Gender 2.15 .0020 8.62 2.20-33.66

Sibling ST use 1.13 .0087 3.09 1.33-7.18

Family problems 0.88 .0053 2.41 1.30-4.47

Friends’ ST use .76 .0006 2.13 1.38-3.27

Knowledge .33 .0199 1.39 1.05-1.85

Attitude .14 <.0001 1.15 1.08-1.22

(a) OR = odds ratio using coefficients obtained with logistic regression.

(b) 95% CI = 95% confidence intervals.

DISCUSSION

A variety of psychosocial factors have been documented extensively in the literature as risk factors or determinants for youth tobacco use: favorable attitudes toward tobacco use; smoking among parents, siblings, and close friends; ST use among parents, siblings, and close friends; not getting along with family or at school; and lack of knowledge about tobacco and health. Age and gender have been examined as well. This study explored the extent to which these factors had an impact on exclusive tobacco use categories (i.e., smoking only, ST use only, conjoint smoking and ST use) among a single sample of rural youths.

As expected, a majority of youths (64%) in this sample did not smoke or use ST. Among tobacco users (36%), the selected factors varied in terms of their importance for predicting the three different types of tobacco use. Among smokers only, having friends and siblings who smoke, having favorable attitudes toward tobacco, and not getting along with family were all significant predictors.

Among ST users in this study, gender was the most powerful determinant. Similar to the pattern found with smokers, significant determinants of ST use included the two variables related to peer use of the specific tobacco product (i.e., having friends and siblings who use ST) and having favorable attitudes toward tobacco use. Different from smokers only, having family problems was not a significant predictor of ST use.

Of the youths in this sample, 10% smoked and used ST conjointly. There were more risk factors or determinants significantly associated with conjoint users than the other two groups (smoking only, ST use only). Like ST-only users, conjoint users were much more likely to be male than to be female. The determinants of conjoint use included all the determinants of ST-only use (sibling ST use, friend ST use, and attitudes toward tobacco) and two of the four determinants of smoking only (family problems and attitudes). Smoking among friends and siblings were not determinants of conjoint use, although they were determinants of smoking only. One additional factor that served as a predictor of conjoint use was knowledge about tobacco and health.

These findings show several interesting patterns. First, attitude was a significant, although weak, predictor for all three tobacco use categories. Second, for smoking only and ST use only, the two factors involving peer use (i.e., friends and siblings) of the particular tobacco product (i.e., cigarettes or ST) were also significant predictors. It is important to note that parental use of tobacco products was not a significant predictor for any of the tobacco use categories. Third, family problems were a determinant for smoking only and conjoint use, but not for ST only. Finally, conjoint use was predicted by a combination of all the predictors of ST only and two of the predictors of smoking only plus the knowledge factor. Further research is needed to explain why the risk factors for conjoint use are more similar to those for ST only rather than smoking only.

Consistent with previous studies on tobacco use risk factors, this study supports the importance of addressing attitude, knowledge, peer tobacco use, and family problems in comprehensive tobacco prevention efforts (e.g., 13, 14). Although tobacco prevention is extremely important, prevention does not target the needs of youths who already use tobacco products and who may want to quit.

There is a growing trend to focus on cessation efforts with youths. Findings from this study have important implications for tobacco use cessation with youths. Consistent with previous research (8), 60% of the smoking-only group reported trying to quit at least once. In addition, half of the ST-only users in this sample reported trying to quit at least once. Clearly, since these youths are still using tobacco products, their attempts met with failure–highlighting the need for effective cessation interventions with smokers and ST users.

In addition, the findings from this study have implications for the content of cessation interventions. For instance, adolescent smokers and ST users are highly likely to have close friends who smoke or use ST, respectively. To illustrate, ST users are likely to receive offers of cigarettes and ST from family members and to use it with them (17, 18). As such, tobacco use cessation programs should include strategies for helping youths identify new peer groups and/or stand up to the inevitable peer pressure of their smoking and/or ST-using friends (16).

Sometimes, cessation interventions may need to emphasize different factors or issues for different tobacco use categories. For instance, interventions involving smokers and conjoint users may need to include stress management techniques (for dealing with family problems), whereas this may not be as critical for ST users. Finally, for ST users and conjoint users, the primary target group is likely to be male. Thus, recruitment and retention strategies for ST cessation programs need to address issues that are important and meaningful to males.

Findings also suggest that lack of knowledge among smokers only and ST users only about the hazards of tobacco use was not an important determinant of tobacco use. Generally, these youths scored high on knowledge and were aware of potential health consequences of tobacco use. This finding underlines the need for public health approaches for tobacco control that go beyond providing knowledge on the general health effects.

However, different from smokers only and ST users only, conjoint users had significantly less knowledge about tobacco and health. This implies that interventions for conjoint users should aim to increase knowledge about tobacco and health. More specifically, among youths who smoke and use ST conjointly, effective cessation interventions may result from comprehensive approaches that highlight knowledge, as well as attitude and behavior change. For the purpose of intervention, it is also important to consider that conjoint users may be a more addicted and recalcitrant group than their other tobacco-using peers. Youths who increase smoking or maintain a high level of smoking are more likely to use ST (6).

While the findings of this study have merit, there are limitations. The study sample was rural with little variation in race, parental education levels, and household income. Therefore, caution must be exercised in generalizing the findings to populations of nonrural youths, particularly beyond the Appalachian region, and for non-white youths.

Second, the classification of tobacco use status used in this study (i.e., current tobacco user vs. current nonuser) might pose some limitations on the interpretation of the results. Inquiry about smoking status provided five response options for participants:

1. I never have (or only tried it).

2. I did, but quit.

3. I smoke occasionally, but not every day.

4. I smoke every day (less than a pack).

5. I smoke every day (a pack or more).

Participants who responded as never smoking or had quit smoking (either 1 or 2) were categorized as current nonsmokers, and participants who smoked every day (either 4 or 5) or occasionally (3) were categorized as current smokers. We did not specifically separate never users from quitters or occasional from every day users. We acknowledge that there may be important differences among never smokers, quitters, occasional smokers, and regular smokers.

Although we did not explore such differences in this study, understanding the behavior of youths who are nonsmokers/ST users, quitters, and experimenters is important for developing effective interventions. The field could benefit from future studies that are focused more explicitly on the positive health behaviors of adolescents who do not smoke or those who have quit.

Overall, the results of this study support the emphasis of specific risk factors for different types of tobacco intervention. Findings may be useful in helping educators and researchers better understand the factors that are associated with rural youths who smoke and/or use ST; the aim would be eventually to develop tailored and effective intervention strategies for different types of tobacco users in rural areas.

REFERENCES

(1.) Noland, M. P., Kryscio, R. J., Riggs, R. S., et al., Use of snuff, chewing tobacco, and cigarettes among adolescents in a tobacco-producing area, Addict. Behav. 15(6):517-530 (1990).

(2.) Noland, M. P., Kryscio, R. J., Hinkle, J., et al., Relationship of personal tobacco-raising, parental smoking, and other factors of tobacco use among adolescents living in a tobacco-producing region, Addict. Behav. 21(3):349-361 (1996).

(3.) West Virginia Department of Education, West Virginia Youth Risk Behavior Survey, Health Office of Healthy Schools, Charleston, West Virginia, 1997.

(4.) Centers for Disease Control and Prevention, Youth risk behavior surveillance–United States, 1995, MMWR Morb. Mortal. Wkly. Rep. 45:1-84 (1996).

(5.) McDermott, R. J., Clark-Alexander, B., and Westhoff, W. W., Tobacco use initiation: future intentions to use, and attitudes towards use in sample of Florida fifth graders, J. Florida Med. Assoc. 83(2):93-95 (1996).

(6.) Dent, C. W., Sussman, S., Johnson, C. A., et al. Adolescent smokeless tobacco incidence: relations with other drugs and psychosocial variables, Prev. Med. 16:422-431 (1987).

(7.) U.S. Department of Health and Human Services (USDHHS), Preventing Tobacco Use Among Young People: A Report of the Surgeon General, DHHS Publication no. S/N 017-00100491-0, Public Health Service, Washington, DC, 1994.

(8.) Institute of Medicine, Growing Up Tobacco Free: Preventing Nicotine Addiction in Children and Youths (B. S. Lynch and R. J. Bonnie, Eds.), National Academy Press, Washington, DC, 1994.

(9.) Winn, D. M., Blot, W. J., Shy, C. M., et al., Snuff dipping and oral cancer among women in the southern United States, N. Engl. J. Med. 304:745-749 (1981).

(10.) Frithiof, L., Anneroth, G., Lasson, U., et al., The snuff-induced lesion: a clinical and morphological study of a Swedish material, Acta Odontol. Scand. 1:53-64 (1983).

(11.) U.S. Department of Health and Human Services (USDHHS), The Health Consequences of Using Smokeless Tobacco: A Report of the Advisory. Committee to the Surgeon General, U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, National Cancer Institute, DHHS Publication no. (NIH) 86-2874, 1986.

(12.) Benowitz, N. L., Porchet, H., Sheiner, L., et al., Nicotine absorption and cardiovascular effects with smokeless tobacco use: comparison with cigarettes and nicotine gum, Clin. Pharmacol. Ther. 44:23-28 (1988).

(13.) Chassin, L., Presson, C., Sherman, S. J., et al., Psychosocial correlates of adolescent smokeless tobacco use, Addict. Behav. 10:431-435 (1985).

(14.) U.S. Department of Health and Human Services (USDHHS), Reducing Tobacco Use Among Youth: Community-Based Approaches. A Guideline. Prevention Enhancement Protocols System, Substance Abuse and Mental Health Services Administration Center for Substance Abuse Prevention, DHHS Publication no. (SMA) 97-3146, 1997.

(15.) McGrath, M., and Price, L., West Virginia Statistical Abstracts (1995-1996), Bureau of Business Research, College of Business and Economics, West Virginia University, Morgantown, West Virginia, 1996.

(16.) Litwin, M. S., How to Measure Survey Reliability and Validity. The Survey Kit, Sage, 1995.

(17.) Dino, G. A., Horn, K. A., Zedosky, L., et al., A positive response to teen smoking: Why N-O-T? NASSP Bull. 82(601):46-58 (1998).

(18.) Ary, D. V., Lichtenstein, E., Severson, H., et al., An in-depth analysis of male adolescent smokeless tobacco users: interviews with users and their fathers, J. Behav. Med. 12(5):449-467 (1989).

Kimberly A. Horn, Ed.D.(*)

West Virginia University Department of Community Medicine and Prevention Research Center

Xin Gao, M.S.

West Virginia University School of Pharmacy

Geri A. Dino, Ph.D.

West Virginia University Department of Community Medicine and Prevention Research Center

Sachin Kamal-Bahl

West Virginia University School of Pharmacy

(*) To whom correspondence should be addressed.

COPYRIGHT 2000 Marcel Dekker, Inc.

COPYRIGHT 2001 Gale Group