A review of research on the effects of religion on adolescent tobacco use published between 1990 and 2003.

Weaver, Andrew J. ; Flannelly, Kevin J. ; Strock, Adrienne L. 等

Tobacco use is the chief preventable cause of premature disease and death in the United States (An, O'Malley, Schulenberg, Bachman, & Johnston, 1999; Centers for Disease Control and Prevention [CDC], 1994). Tobacco use causes almost 1 out of every 5 deaths (McGinnis & Foege, 1993) and the annual health care expenditure for treating smoking-related illnesses was estimated at $274 for each American adult in 1993 (Miller, Ernst, & Collins, 1999). Every year more than 400,000 Americans die from smoking-related illnesses and 2 million die in all developing countries combined (Peto, Lopez, Boreham, Thun, & Heath, 1994). Smoking kills more Americans annually than AIDS, automobile accidents, suicide, murder, fire, alcohol, and illegal drugs combined (CDC, 1994).

Early adolescence (eleven through fifteen years of age) is the crucial life stage for prevention of tobacco use, since it is uncommon for tobacco use to begin after high school (Johnston, O'Malley, & Bachman, 1995). Each day, nearly 3,000 American youths begin smoking (An et al., 1999). Smoking rates among teenagers in the United States have increased from 27.5% in 1991 to 36.4% in 1997 (Morbidity and Mortality Weekly, 1998) with a similar trend seen in Canada (Spurgeon, 1999). It is estimated that between one-third and one-half of adolescents who try cigarettes will become regular smokers, a process that takes an average of 2 to 3 years (Henningfield, Cohen, & Slade, 1991).

Tobacco is associated with the increased likelihood of using other addictive substances, acting for some as a "gateway drug" (Elders, Perry, Erikson, & Giovino, 1994). It is generally the first substance used by teens who later use alcohol and illicit drugs. The Surgeon General found that 12- to 17-year-olds that claimed to have smoked in the past 30 days were three times more likely to have used alcohol, eight times more likely to have smoked marijuana, and 22 times more likely to have used cocaine within the past 30 days compared to those teens who had not smoked (Elders et al., 1994).

The tobacco industry spends billions of dollars on advertising, product promotion, and promotional items, such as clothing and catalogue products that directly appeal to adolescents (Altman, Levine, Coeytano, Slade, & Jaffe, 1996). Research has shown that teens exposed to these promotions are more likely to be smokers (Altman et al., 1996).

In the United States, annual sales of tobacco products to minors total 950 million packs of cigarettes and 26 million containers of smokeless tobacco (Heishman, Kozlowski, & Henningfield, 1997). About one-half of minors who attempt to purchase tobacco products in stores report never being asked for proof of age (Centers for Disease Control and Prevention [CDC], 1996). Minors have even easier access to cigarettes via the Internet because many Internet vendors have weak or nonexistent age-verification procedures. In a recent study, minors successfully received cigarettes 93.6% of the times they attempted to purchase them with a credit card (Ribisl, Williams, & Kim, 2003) and Internet vendors sent 1,650 packs of cigarettes to these underage adolescents (Ribisl et al., 2003) without verifying their ages.

Greater religious involvement was found to be associated with lower risk of use of tobacco and other addictive substances in 26 separate studies (Koenig, McCullough, & Larson, 2001). Blyth and Leffert (1995) cite a number of studies specifically on teen drug use (including nicotine) that report an inverse relationship between drug use and religious involvement among teens and young adults.

Some examples of the nature of this effect are provided in the results of the National Study of Youth and Religion that the University of North Carolina began in 2001 (Smith & Faris, 2002). The survey of a national sample of 2,478 teens found that religious high school seniors were less likely to smoke, and those who do smoke started smoking at an older age than their less religious counterparts. Weekly religious service attendees, those who said religion was very important and those who have been involved in a religious youth group six or more years, were more likely to delay their first use of cigarettes when compared to nonattendees. Three in 10 teens who were not involved in religious activities smoked regularly, compared to 2 in 10 of all the teens in the study. Catholic, Mormon, Jewish, Baptist, and other Protestant students were all less likely to smoke than the nonreligious students. The inverse relationship between smoking and various measures of religiousness were statistically significant after controlling for race, age, sex, rural/urban residence, region, education of parents, number of siblings, whether the mother works, and the presence of a father/male guardian in the household.

Given these results and the public health concerns over the continuing use of tobacco by adolescents, we conducted electronic searches of the biomedical and psychological literature for recent research (1990-2030) on the relationship between religion and tobacco use in adolescents. Relevant articles were retrieved, and their methods and results were assessed to examine the nature and extent of the reported effects.

METHOD

An electronic search was conducted on the American Psychological Association's database (PsycINFO) and the National Library of Medicine's database (PubMed) for articles published between 1990 and 2003 in English-language journals. The search phrase we used was: (cigarette OR smoking OR tobacco or nicotine) AND (religio* OR spiritual*) AND (adolescent* OR youth OR teen*). The search produced an unduplicated count of 163 articles.

The abstract of each study was read by two judges who mutually agreed to select or reject an article for further examination based on the information contained in the abstract (Critchley, Jadad, Taniguchi, Woods, Stevens, Reyno, & Whelan, 1999). If either judge thought the abstract provided insufficient information for making a decision, the article itself was obtained.

Then, two judges read the retrieved articles and followed the same procedure to decide if an article should be included in the sample. In order to be included, the article had to be an original research study that analyzed the influence of some measure of religion on tobacco use in adolescents. The final selection of articles for the study was made by mutual consent after each article was read by two judges.

The articles were classified by type of journal, year of publication, and the number and kind of independent and dependent variables that were analyzed. Frequency data were analyzed by the chi-square ([chi-square]) test, and interval data were analyzed by analysis of variance (ANOVA) (Ferguson, 1966; Siegel, 1956). Correlation analysis was also conducted on some of the data as explained in the text. Sample size was transformed into logarithms for analysis because the distribution was extremely skewed.

RESULTS

Of the 163 articles found in the initial electronic search, only 29 specifically studied the influence of religion on tobacco use in adolescents. The other articles found in the search were eliminated from the sample because: (a) they were not research studies; (b) they studied adult samples; (c) they studied religion and smoking in relation to health outcomes but not each other; (d) they studied the relationship between religion and attitudes about cigarette smoking; (e) they measured attitudes about using tobacco but not its use; or (f) they studied substance abuse without analyzing the effects of religion on tobacco use per se.

The age of study participants ranged between 8 and 19 years, with five of the 29 studies including participants who were less than 12 years old. The sample size of the studies varied greatly--from 53 to over 17,000 participants--with the median being around 2,200. Some studies had larger samples, including one with nearly 190,000 participants, but the sample size we refer to here is the number of participants used in any given statistical analysis. The majority of studies (58.6%) were sample sizes between 1,000 and 5,000 participants, but 24.1% had sample sizes less than 500, and 17.2% had sample sizes over 10,000.

The number of studies on religion and tobacco use increased between the first half (1990-1996; n = 9) and second half (1997-2003; n = 201) of the 14-year time-period covered by the search, [chi square] = 4.17, p < .05. Indeed, over half of the 29 studies in the sample were published from 2000 to 2003 (n = 15).

The majority of the studies were published in medical journals (20.7%) or specialty journals in the field of addiction and substance abuse (31.0%). The remainder of the studies were published in nursing (6.9%), public health (6.9%), a range of health journals (20.7%), and other kinds of journals (13.8%). Three categories of journals were created and used in subsequent analyses of frequency ([chi square]) and interval data (ANOVA). The first category included all addiction and substance abuse journals (n = 9), the second category combined the medical, nursing, and public health journals (n = 10), and the third category consisted of the remaining journals (n = 10).

Six of the studies were published in journals that focus on adolescents. Two were published in the Journal of Child and Adolescent Psychiatric Nursing, and the others were published in Adolescence, Archives of Pediatric and Adolescent Medicine, the Journal of Adolescent Health, and the Journal of Child and Adolescent Substance Abuse.

Table 1 indicates the types of religious measures that were analyzed as dependent variables in the 29 studies. The majority of the studies measured religion with a single question or item. The items tended to fall into three categories or dimensions (Sherrill, Larson, & Greenwold; 1993; Weaver, Flannelly, Flannelly, Krause, Strock, & Costa, in press). Religious attendance primarily refers to attending church or religious services, although a few studies asked about participation in other kinds of church-related activities. While some studies looked at denominational differences in tobacco use, religious affiliation was sometimes measured as a dichotomous variable; e.g., Do you belong to any religion? One study specifically asked participants: "How effective do you think religious advice is in preventing youth from smoking?" Apart from religious denomination and affiliation, the responses to most questions were scored on an interval scale.

Six of the studies summed the responses from 2-4 questions to form a single composite measure of religiousness, whereas eight included a number of separate religious variables in their statistical analyses. Table I indicates the religious dimensions covered by these composite and multiple measures of religion. Although most studies measured one or more of the three religious dimensions that were noted above (importance/religiosity, attendance/participation, affiliation/denomination), a few studies measured other dimensions of religion, such as religious coping, religious beliefs, and private religious activities. No significant differences were found in the number of religious questions asked (ANOVA), the types of measures used (single item, composite, or multiple measures) ([chi square]), or the number of dimensions measured (ANOVA), either across time or among the types of journals.

Dependent Variables

Cigarette smoking was the sole dependent variable in 25 of the 30 studies, while 5 studies measured other types of tobacco use. Data on smoking and other uses of tobacco were collected on a nominal scale in 21 of the studies in which participants were categorized into groups on the basis of their responses. The other studies used interval scales or the frequency of cigarette smoking, but some of these grouped participants into dichotomous categories in the statistical analyses.

Three categories of tobacco users, or smokers, were commonly used: regular use, lifetime use, and occasional use. Statistical comparisons were typically made between one or more of these categories and nonsmokers. Lifetime use is a measure of whether individuals had ever used tobacco in their lives. Occasional use is sometimes referred to as experimental use in the research literature.

Table 2 lists the number of studies that analyzed different kinds of dependent measures of tobacco use. Regular use was the most common measure--used alone or in combination with other measures. The second most frequently used measure was lifetime use. Four studies analyzed tobacco use on an interval scale instead of analyzing it as a dichotomous category. These are labeled continuum of use in the table. No significant differences were found in the number (ANOVA) or the type ([chi square]) of dependent measures used across time or among the types of journals.

Analyses and Effects

Table 3 shows the number of instances in which various measures of religion were analyzed as dependent variables and the number of times they were found to have a statistically significant effect on tobacco use. The number of analyses exceeds the number of studies because several studies analyzed two or more dependent variables. Of the 43 analyses of religious effects conducted in the 29 studies, religion was found to be significantly related to reduced tobacco use in 33 analyses (see Table 3).

In all, 22 of the 29 studies in the sample found at least one significant effect of religion on tobacco use. Five of the seven studies that found no effect used small sample sizes (Ns between 53 and 441), and a biserial correlation (Guilford, 1956) showed that the failure to find a significant effect (no effect = 1, effect = 0) was directly related to the logarithm of the sample size, r (27) = .63, p < .001.

Religion was associated with significantly lower regular tobacco use in 12 of the 15 studies in which regular use was the only dependent variable, and significantly lower lifetime use in all seven studies in which lifetime use was the only dependent variable. Religion had a significant ameliorating effect on both of these variables in two of the three studies in which both of them were measured.

The results were somewhat more complicated among the four studies that analyzed the effects of religion on lifetime, occasional, and regular use. Kaufman et al. (2002) found that religious attendance had a significant effect only on current smoking. On the other hand, Ausem, Oman, Vesely, Aspy, and McLeroy (2003) found that nonsmokers were significantly more likely than occasional smokers to be religious, but they found no difference in religiosity between occasional and regular smokers. Swaim, Oetting, and Casaas (1996), who used structural equation modeling, reported that religion exerted an indirect influence on tobacco use through its modulating effect on school adjustment and achievement.

Nonnemaker, McNeeley, and Blum (2003) reported somewhat different effects for what they called public and private domains of religion. The public domain encompassed religious attendance and participation, whereas the private domain included the frequency of praying and the importance of religion. Both public and private religion were found to protect individuals from ever smoking cigarettes (i.e, lifetime use), but they were less consistent in their effects on occasional and regular use. Private religion had a significant effect only on occasional use, whereas public religion had a significant effect only on regular use.

Control of Extraneous Variables

Twenty-six of the studies used multivariate statistical techniques that controlled for the effects of a number of variables, in order to isolate the effects of religion on tobacco use. Table 4 lists the kinds of variables that were controlled for statistically, and the percentage of studies that included them in their analysis for the effects of religion on tobacco use. The variables are grouped into four categories for convenience.

Many studies used multiple measures of the types of variables listed in the table, and 31.0% of the studies controlled for 6 or more of the 14 variables in the table. The mean number of variables controlled for was 4.7 and the median was 5.0.

A majority of the studies included gender and ethnicity and some measure of education interest in their statistical analyses. Measures of family structure and parental involvement were included in less than half of the studies. The next most commonly used type of variable was a measure or measures of sibling and peer influences, such as smoking by siblings or peers.

DISCUSSION

Research has identified several risk and protective factors that appear to increase or decrease the likelihood of tobacco use among youth (Wills, Yaeger, & Sandy, 2003; Sperber, Peleg, Friger, & Shvartzman, 2001). The present review shows that studies published in various fields since 1990 have repeatedly found religion to be one of these protective factors. An inverse relationship between religious involvement and tobacco use was found in three-quarters of the 29 studies published in psychology and biomedical journals between 1990 and 2003. This effect was evident even after controlling for various demographic and other factors that influence tobacco use.

Self-reported church attendance (attending religious services) was one of the most widely used measures. It also yielded some of the most consistent effects, with significant effects of attendance reported in 11 of the 13 studies in which it was the only measure of religion. Participants' ratings of their own religiousness (i.e., religiosity) or the importance of religion in their lives was used most often, but their effects were less consistent.

The size of the effect that religion had on tobacco use was relatively small. Although effect size was not reported in most of the studies, the presentation of standardized beta values (i.e., partial correlations) and adjusted odds ratios in some studies indicate that it accounts for less than 10% of the variation in tobacco use. Two studies reported significant bivariate correlations between religion and tobacco use as low as .16, which suggests that religion accounts for only 2.5% of the variation in tobacco use in those studies. An effect of this size would require a sample size of nearly 800 participants in order to assure its detection. This helps explain why five studies with sample sizes of less than 500 participants did not find significant religious effects. The other two studies that did not find effects had sample sizes of approximately 1,300 and 2,300. The median sample size of the 29 studies was 2,200.

The results of a few of the studies are particularly enlightening, such as those from Brown, Schulenberg, Backman, O'Malley, and Johnston (2001), who examined a nationally representative sample of high school classes who graduated between 1976 and 1997 (N = 188,000). The researchers, who were interested in whether the correlates of tobacco use changed across time, surveyed a random sample of 15,000 and 19,000 high school seniors from approximately 135 high schools each year. The investigators found a highly consistent association between lower cigarette use and greater religious involvement across the 22-year study period.

One of the smaller studies in the sample presents some valuable findings about the factors contributing to tobacco use (Atkins, Oman, Vesely, Aspy, & McLeroy, 2002). Based on interviews of a randomly selected sample of 1,350 teens and their parents living in inner-city areas of two midsized Midwestern cities, the authors developed a logistic regression model to predict tobacco use. Significant predictors of lower tobacco use included: (a) nonparental adult role models, (b) peer role models, (c) family communication, (d) attending religious services and participating in religious activities, (e) participating in other organized activities, (f) good health practices, (g) community involvement, (h) future aspirations, and (i) responsible choices. The findings support the view that several resources including religious involvement protect youth from risk-taking behaviors, particularly tobacco use. Religious attendance and participation along with positive peer role models appeared to have the greatest protective effects. Adolescents with either of these two assets were approximately 2.5 times less likely to report tobacco use in the past 30 days compared to those without these resources. Even after controlling for important demographic factors including youth age and race, the assets remained significantly predictive of adolescent tobacco nonuse (Atkins et al., 2001).

A British study in our sample (Hope & Cook, 2001) examined what the authors called Christian commitment, which asked participants if they regularly attended church, prayed, and read the Bible. A fourth asked if they agreed or disagreed with the statement, "I have given my life to Jesus." All four items made significant contributions to a logistic model predicting lifetime tobacco use (i.e., never smoked) among 12-16-year-olds from England, Scotland, and Wales. The study, which also studied participants between 17 and 30 years of age, found the only items that significantly predicted smoking in this age group were Bible reading and giving one's life to Jesus.

To Hope and Cook (2001, p. 109) "The findings suggest[ed] that for church-affiliated young people it is initially the socialization of religion that acts as prohibitory against substance use, though, as age increases, a greater internalization of Christian commitment becomes more important." Whatever the case, the results of Ausem et al. (2003) and Nonnemaker et al. (2003) suggest that the primary effect of religion on smoking may be its effect on lifetime use. That is, its capacity to inhibit adolescents from trying cigarettes.

An inverse relationship between religion and tobacco use was reported in three quarters of the studies in our sample, after controlling for demographic and other variables that influence tobacco use. Religiosity and attending religious services were the two independent variables that were used most frequently. While the effects of these and other measures of religion on tobacco use are relatively small, they are very consistent.

There are probably several reasons why religious faith and practice might protect against smoking. It may be linked to positive influences found in religious peer groups. Adolescents who worship regularly socialize with people of similar beliefs and may avoid contact with peers who are more likely to smoke. In addition, religious faith may give adolescents emotional balance, which reduces the stress that leads to tobacco use. Religion may be related to perceived meaning and purpose in life (Pargament, 1997) as well as values and attitudes about substance use (Brody, Stoneman, & Flor, 1996), factors which could moderate the impact of negative life events. Positive parental religious beliefs may also mediate and neutralize other risk factors as well as enhance family resilience in the face of stressful events (Weaver, Revilla, & Koenig, 2002). Previous research has emphasized that factors related to tobacco use

can be moderated in several different ways (Atkins et al., 2001), so for future research it seems appropriate to examine multiple processes through which religion can affect behavior.

If involvement in faith communities has a positive effect on the attitudes and behavior of teens toward smoking cigarettes, then encouraging young people to be involved in religious life may be beneficial to those seeking to avoid tobacco. Teens who connect to a religious youth group may find it a helpful place to find peer support that can help them quit smoking. Tobacco use prevention programs for teens would be a valuable activity for faith communities, especially since the surveys indicate that about 1 in 10 teenagers who regularly attend religious services do smoke (Smith & Faris, 2002). Faith-based intervention programs need to address teens' abilities to recognize social and advertising pressures to use tobacco as well as develop skills to resist pressures (Bruvold, 1993). Increased self-reliance and self-esteem with decreased social alienation appear to be important factors in resisting the pressure to smoke (Bruvold, 1993). Faith-based tobacco-cessation programs focused on minority groups that have high levels of religious participation and suffer a disproportionately higher burden of tobacco-attributable illnesses and deaths may be of particular value (Spangler et al., 1998; Schorling et al., 1997).

The authors wish to thank The Henry Luce Foundation, The Clark Foundation, and The Fannie E. Rippel Foundation for their generous and long-time support of The Health Care Chaplaincy. The authors also thank the Research Department's Administrative Assistant John Barone for his help preparing and editing the manuscript.

REFERENCES

Altman, D. G., Levine, D. W., Coeytano, R., Slade, J., & Jaffe, R. (1996). Tobacco promotion and susceptibility to tobacco use among adolescents aged 12 through 17 years in a nationally representative sample. American Journal of Public Health, 86(11), 1590-1593.

An, L. C., O'Malley, P. M., Schulenberg, J., Bachman, J. G., & Johnston, L. D. (1999). Changes at the high end of risk in cigarette smoking among U.S. high school seniors. American Journal of Public Health, 89(5), 699-705.

Atkins, L. A., Oman, R. F., Vesely, S. K., Aspy, C. B., & McLeroy, K. (2002). Adolescent tobacco use: The protective effects of developmental assets. American Journal of Health Promotion, 16(4), 198-204.

Ausems, M., Mesters, I., Van Breukelen, G., & De Vries, H. (2003). Do Dutch 11-12-year-olds who never smoke, smoke experimentally or smoke regularly have different demographic backgrounds and perceptions of smoking? European Journal of Public Health, 13(2), 160-167.

Blyth, D. A., & Leffert, N. (1995). Communities as contexts for adolescent development: An empirical analysis. Journal of Adolescence Research, 10(1), 64-87.

Brody, G. H., Stoneman, Z., & Flor, D. (1996). Parental religiosity, family processes, and youth competence in rural, two-parent African-American families. Developmental Psychology, 32, 696-706.

Brown, T. N., Schulenberg, J., Backman, J. G., O'Malley, P. M., & Johnston, L. D. (2001). Are risk and protective factors for substance use consistent across historical time? National data from the high school classes of 1976 through 1997. Prevention Science, 2(1), 29-43.

Bruvold, W. H. (1993). A meta-analysis of adolescent smoking prevention programs. American Journal of Public Health, 83(6), 872-880.

Centers for Disease Control and Prevention. (1994). Cigarette smoking among adults-United States, 1993. Morbidity and Mortality Weekly Report, 43, 925-930.

Centers for Disease Control and Prevention. (1996). Tobaco use and usual source of cigarettes among high school students--United States, 1995. Morbidity and Mortality Weekly Report, 45, 413-418.

Critchley, P., Jadad, A. R., Taniguchi, A., Woods, A., Stevens, R., Reyno, L., & Whelan, T. J. (1999). Are some palliative care delivery systems more effective and efficient than others? Journal of Palliative Care, 15(4), 40-47.

Elders, M. J., Perry, C. L., Eriksen, M. P., & Giovino, G. A. (1994). The report of the surgeon general: Preventing tobacco use among young people. American Journal of Public Health, 84(4), 543-547.

Ferguson, G. A. (1966). Statistical analysis in psychology and education. New York: McGraw-Hill.

Guilford, J. P. (1956). Fundamental statistics in psychology and education: 3rd Edition. New York: McGraw-Hill.

Heishman, S. J., Kozlowski, L. T., & Henningfield, J. E. (1997). Nicotine addiction: Implications for public health policy. Journal of Social Issues, 53(1), 13-33.

Henningfield, J. E., Cohen, C., & Slade, J. D. (1991). Is nicotine more addictive than cocaine? British Journal of Addiction, 86, 565-569.

Hope, L. C., & Cook, C. C. H. (2001). The role of Christian commitment in predicting drug use amongst church-affiliated young people. Mental Health, Religion, & Culture, 4(2), 109-117.

Johnston, L D., O'Malley, P. M., & Bachman, J. G. (1995). National survey results on drug use from the monitoring the future study, 1975-1994 (NIH Publication No. 95-4026). Washington, DC: U. S. Government Printing Office.

Kaufman, N. J., Castrucci, B. C., Mowery, P. D., Gerlach, K. K., Emont, S., & Orleans, T. (2002). Predictors of change on the smoking uptake continuum among adolescents. Archives of Pediatric and Adolescent Medicine, 156, 581-587.

Koenig, H. G., McCulloch, M. E., & Larson, D. B. (2001). Handbook on religion and health. Oxford: Oxford University Press.

McGinnis, J. M., & Foege, W. H. (1993). Actual causes of death in the United States. Journal of the American Medical Association, 270, 2207-2212.

Miller, V. P., Ernst, C., & Collins, F. (1999). Smoking-attributable medical costs in the USA. Social Science and Medicine, 48, 375-391.

Morbidity and Mortality Weekly. (1998). Tobacco use among high school students--United States, 1997. Morbidity and Mortality Weekly, 229-233.

Nonnemaker, J. M., McNeely, C. A., & Blum, R. W. (2003). Public and private domains of religiosity and adolescent health risk behaviors: Evidence from the National Longitudinal study of adolescent health. Social Science & Medicine, 47, 2049-2054.

Pargament, K. I. (1997). The psychology of religion and coping: Theory, research, practice. New York: Guilford Press.

Peto, R., Lopez, A. D., Boreham, J., Thun, M., & Heath, C. (1994). Mortality from smoking in developed countries, 1950-2000. New York: Oxford University Press.

Ribisl, K. M., Williams, R. S., & Kim, A. E. (2003). Internet sales of cigarettes to minors. Journal of the American Medical Association, 290(10), 1356-1359.

Schorling, J. B., Roach, J., Siegel, M., Baturka, M., Hunt, D. E., Guterbock, T. M., & Stewart, H. L. (1997). A trial of church-based smoking cessation interventions for rural African-Americans. Alternative Medicine, 26, 92-101.

Sherrill, K. A., Larson, D. B., & Greenwold, M. (1993). Is religion taboo in gerontology? The American Journal of Geriatric Psychiatry 1(2), 109-117.

Siegel, S. (1956). Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill.

Smith, C., & Faris, R. (2002). Religion and American adolescent delinquency, risk behaviors and constructive social activities. Chapel Hill, NC: The National Study of Youth and Religion.

Spangler, J. G., Bell, R. A., Knick, S., Michielutte, R., Dignan, M. B., & Summerson, J. H. (1998). Church-related correlates of tobacco use among Lumbee Indians in North Carolina. Ethnicity and Disease, 8(1), 73-80.

Sperber, A. D., Peleg, A., Friger, M., & Shvartzman, P. (2001). Factors associated with daily smoking among Israeli adolescents: A prospective cohort study with a 3-year follow-up. Preventive Medicine, 33, 73-81.

Spurgeon, D. (1999). Studies reveal increased smoking among students in Canada. British Medical Journal, 319(7222), 1391.

Swaim, R. C., Oetting, E. R., & Casas, J. M. (1996). Cigarette use among migrant and nonmigrant Mexican American youth: A socialization latent-variable model. Health Psychology, 15(4), 269-281.

Weaver, A. J., Flannelly, L. T., Flannelly, K. J., Krause, N., Strock, A. L., & Costa, K. G. (submitted). A comparison of the quantity and quality of research on religion in three gerontology journals: 1985-1991 versus 1988-2002. Research on Aging.

Weaver, A. J., Revilla, L. A., & Koenig, H. G. (2002). Counseling families across the stages of life: A handbook for pastors and other helping professionals. Nashville, TN: Abingdon Press.

Wills, T. A., Yaeger, A. M., & Sandy, J. M. (2003). Buffering effect of religiosity for adolescent substance use. Psychology of Addictive Behaviors, 17(1), 24-31.

Andrew J. Weaver, Director of Research; Kevin J. Flannelly, Associate Director of Research; Adrienne L. Strock, Research Librarian, The HealthCare Chaplaincy.

Request for reprints should be sent to Kevin J. Flannelly, Ph.D., Research Dept., The HealthCare Chaplaincy, 307 E 60th St. New York, NY 10022. E-mail: kfiannelly@healthcarechaplaincy.org

Table 1
Number/Percentage of Studies Using Different Types
of Measures of Religion as Independent Variables

 Number of Percentage
Type of Measure Studies of Studies

Single-Item Measure 15 51.7
 Importance of Religion or Religiosity 7 24.1
 Religious Attendance or Participation 5 17.2
 Religious Affiliation or Denomination 2 6.9
 Does Religion Prevent Smoking 1 3.4

Composite Measure 6 20.7
 Importance of Religion or Religiosity 2 6.9
 Religious Attendance and Participation 1 3.4
 Religious Attendance and Religiosity 2 6.9
 Religious Affiliation and Religiosity 1 3.4

Multiple Measures 8 27.6
 Religious Attendance and Religiosity 1 3.4
 Religious Attendance and Affiliation 1 3.4
 Religious Attendance and Beliefs 1 3.4
 Religious Attendance, Affiliation, 2 6.9
 and Religiosity
 Religious Attendance, Religiosity, 1 3.4
 and Prayer
 Spiritual Advice, Comfort, and Support 1 3.4
 Religious Attendance, Commitment, 1 3.4
 and Private Religious Activities

Table 2
Types of Dependent Variables Measuring Tobacco Use

 Number Percentage
Dependent Variable of Studies of Studies

Regular Use 11 37.9
Lifetime Use 7 24.1
Lifetime and Regular Use 3 10.3
Lifetime, Occasional, and Regular Use 4 13.8
Continuum of Use 4 13.8

Table 3
Number of Times Each Independent Variable Was Analyzed and
Times It Was Found to Have a Significant Effect on Tobacco Use

 Number Number
 of Times of Times
Independent Variable Analyzed Significant

Importance of Religion or Religiosity 14 9
Attend Religious Services 13 11
Religious Affiliation or Denomination 5 4
Religious Affiliation and Religiosity 2 1
Religious Attendance and Religiosity 1 1
Other Religious Measures 5 5
Measures of Spirituality 3 0
Total 43 31

Table 4
Number/Percentage of Studies that Statistically Controlled
for Different Types of Variables

Types of Variables Number (1) Percentage

Demographic Variables
 Gender 19 65.5
 Ethnicity 17 58.6
 Socioeconomic Status 11 37.9
 Urban Density (e.g., Urban, Rural) 5 17.2

Parental Variables
 Family Structure/Composition 14 48.3
 Parental Involvement 13 44.8
 Parent Tobacco Use 9 31.0

Personal Variables
 Educational Interest, Performance, 16 55.2
 or Aspirations
 Self-Esteem 5 17.2
 Risky Behaviors/Activities 6 20.7
 Positive Behaviors/Activities 4 13.8
 Other Intrapersonal Variables 3 10.3

Other Variables
 Sibling or Peer Influence 12 41.4
 Stress or Arousal 4 13.8

(1) The number of studies exceeds the number of studies in the
sample (N = 29) because most studies controlled for several
variables.