eLetters to:

ARTICLE: James D. Sargent, Michael L. Beach, Anna M. Adachi-Mejia, Jennifer J. Gibson, Linda T. Titus-Ernstoff, Charles P. Carusi, Susan D. Swain, Todd F. Heatherton, and Madeline A. Dalton Exposure to Movie Smoking: Its Relation to Smoking Initiation Among US Adolescents Pediatrics 2005; 116: 1183-1191 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

eLetters published:

Alternative explanation for the association between movie-smoking exposure and smoking initiation

Joel M. Moskowitz, Ph.D.   (18 November 2005)

Response to Dr Moskowitz

James D. Sargent, Micheal Beach, Linda Titus-Ernstoff, Madeline Dalton   (6 April 2006)

Alternative explanation for the association between movie-smoking exposure and smoking initiation 18 November 2005
Joel M. Moskowitz, Ph.D., Director Center for Family and Community Health, School of Public Health, University of California, Berkeley Send letter to journal: Re: Alternative explanation for the association between movie-smoking exposure and smoking initiation jmm{at}berkeley.edu Joel M. Moskowitz, Ph.D.	In this paper and in an earlier study, the authors provide alternative explanations for the results of their research on the effect of exposure to smoking in movies on smoking initiation among children 10 to 14 years of age (1,2). A methodologic explanation for the results of this research may also be applicable. In the present study, telephone interviews were used to assess smoking among children in a cross-sectional survey (1). In an earlier cohort study, the authors used telephone interviews at followup to assess initiation of smoking (2). Our review of survey method studies suggests that children tend to under-report sensitive behaviors like cigarette smoking in telephone surveys (3). Of particular relevance to the authors' research, we found in a population-based, telephone survey of smoking among children 12 to 17 years of age that parental presence during the interview was common (3). In our conventional telephone interview condition, 59% of children reported at the end of the interview that a parent heard the interview (3). Even when most of the telephone interview was conducted by computer and students entered their responses using their telephone keypads (similar to what the authors did), we found that 42% of children reported that a parent heard all or part of the interview (3). In our research parental presence during an interview was strongly associated with whether children reported ever puffing/smoking a cigarette in their lifetime. Re-analysis of our data for children 12 to 14 years of age indicates that 57% reported that a parent heard them being interviewed. Among those who reported that a parent heard the entire interview, only 9% reported puffing/smoking a cigarette in their lifetime. In contrast, among those who reported that a parent did not hear the interview, 20% reported puffing/smoking in their lifetime, and among those who did not have a parent at home during the interview, 27% reported puffing/smoking a cigarette in their lifetime (p <.001; unpublished data). Although we did not experimentally manipulate parental presence during the interview, in conjunction with other data we did report, these results suggest that younger children are less honest about reporting cigarette smoking initiation when a parent is present during the interview (3). The interpretation of results from the authors' studies of exposure to smoking in movies is of particular concern because children who reported low exposure to smoking in movies (i.e., those whose parents did not allow them to see R-rated movies) may have more protective parents who are more likely to be present during their child's telephone interview compared to children who reported high exposure to smoking in movies. If this were true, and if parental presence during the interview caused these children to under-report smoking initiation, then the results of the authors' studies could be explained by this confounding. References 1. Sargent JD, Beach ML, Adachi-Mejia AM, Gibson JJ, Titus-Ernstoff LT, Carusi CP, Swain SD, Heatherton TF, Dalton MA. Exposure to movie smoking: its relation to smoking initiation among US adolescents. Pediatrics. 2005;116:1183-1191. 2. Sargent JD, Beach ML, Dalton MA, Ernstoff LT, Gibson JJ, Tickle JJ, Heatherton TF. Effect of parental R-rated movie restriction on adolescent smoking initiation: a prospective study. Pediatrics. 2004;114:149-156. 3. Moskowitz JM. Assessment of cigarette smoking and smoking susceptibility among youth: telephone computer-assisted self-interviews versus computer-assisted telephone interviews. Public Opinion Quarterly. 2004;68:565-587. Conflict of Interest: None declared
Response to Dr Moskowitz 6 April 2006
James D. Sargent, Professor of Pediatrics Dartmouth Medical School, Micheal Beach, Linda Titus-Ernstoff, Madeline Dalton Send letter to journal: Re: Response to Dr Moskowitz james.d.sargent{at}hitchcock.org James D. Sargent, et al.	Dr. Moskowitz offers a methodologic explanation for the association we report between exposure to movie smoking and adolescent smoking. His research involving adolescent telephone interviews indicated that adolescents reported lower smoking rates when their parents overheard their survey responses. He suggests that the adolescent children of protective parents will both underreport smoking, and see fewer R-rated movies, a phenomenon that might account for our study results. Interestingly, he does not raise the possibility that children of protective parents see fewer movies, have lower exposure to movie smoking, and consequently are less likely to smoke. This possibility seems equally likely and is fully consistent with our findings that higher exposure to movie smoking is associated with adolescent smoking. We point out that our studies control for several parenting characteristics in our multivariable model, including maternal responsiveness, maternal demandingness, parental oversight of smoking behavior and access to cigarettes in the household. Parental presence during the interview is likely correlated with age and the other parenting covariates, and should have been controlled for in the multivariate analysis. In addition, we conducted a sensitivity analysis which showed that, in order to account for our results, a potential confounder would need to be independent of all other variables assessed in our multivariable model (including parenting style) and would also need to be strongly associated with both exposure to movie smoking and adolescent smoking. In particular, the product of the z-statistics of the residuals of this covariate, when regressed on the other covariates in the model, would need to be greater than 64 in order to confound the results we report (note that a z-value of 1.96 is associated with a p-value of 0.05). Thus, it is highly unlikely that another potential confounder exists that would be both uncorrelated with any of these parenting characteristics and strongly associated with both exposure to movie smoking and adolescent smoking and why we consider confounding an unlikely explanation for our study findings. Finally, we are reassured by the consistency and robustness of our estimates of the movie smoking – adolescent smoking association across two adolescent studies (Table). 1, 2 In both studies, we determined exposure to movie smoking by randomly selecting movie titles from a pool of popular contemporary movies that had been content-coded for smoking, and assessed an extensive pool of covariates. Whereas the US National Study, recently published in Pediatrics,1 interviewed 6522 adolescents by telephone, the earlier Northern New England Study of 4919 middle school-aged adolescents2 surveyed them in school, capturing over 90% of students present on the day of the survey. In the school-based survey, smoking prevalence was 17%, higher than the 10% prevalence reported in the telephone based study1. Despite the difference in smoking prevalence, the adjusted odds ratios for the movie smoking – adolescent smoking association are similar across the two studies. Moreover, when the never smokers from the Northern New England cross-sectional survey were followed longitudinally, exposure to movie smoking at baseline predicted smoking initiation during the two-year follow-up period.3 The concordance of these three analyses, which employed two different study designs, survey environments, and study populations offers strong evidence for the validity of the association between exposure to movie smoking and adolescent smoking.   Table: Summary of Three Studies of the Association Between Exposure to Movie Smoking and Adolescent Smoking Initiation Characteristic Northern New England: Cross-Sectional (N=4919)2 Odds Ratio (95%CI) Crude (Adjusted) Northern New England: Longitudinal (N=2603)3 Relative Risk (95%CI) Crude (Adjusted) US National Sample: Cross-Sectional (N=6522)1 Odds Ratio (95%CI) Crude (Adjusted) Movie Smoking Exposure Quartile 1 Reference Reference Reference Quartile 2 3.1(1.9 [1.3, 2.7]) 3.0 (2.0 [1.3, 3.2]) 2.9 (1.7 [1.1, 2.6]) Quartile 3 5.5 (2.6 [1.8, 3.7]) 3.7 (2.2 [1.4, 3.4]) 5.5 (1.8 [1.2, 2.8]) Quartile 4 8.8 (2.5 [1.7, 3.5]) 5.3 (2.7 [1.7, 4.3]) 12.6 (2.6 [1.7, 4.1]) References 1) Sargent JD, Beach ML, Adachi-Mejia AM, et al. Exposure to movie smoking: its relation to smoking initiation among US adolescents. Pediatrics. 2005;116:1183-1191. Link 2) Sargent JD, Beach M, Dalton M, et al. Effect of seeing tobacco use in films on trying smoking among adolescents: cross sectional study. BMJ 2001;323:1394-1397. Link 3) Dalton MA, Sargent JD, Beach ML, et al. Effect of viewing smoking in movies on adolescent smoking initiation: a cohort study. Lancet 2003;362:281-285. Conflict of Interest: None declared