| Study | Design | Measure of exposure | Categories of covariates used in adjustmenta | Outcome (prevalence) | Exposure comparison categories | Measure of association, association (95% CI)b | Comments |
|---|---|---|---|---|---|---|---|
| Cross-sectional | |||||||
| Distefan et al. 1999 | Multiethnic Aged 12–17 years Cross-sectional random-digit-dialing survey N = 6,252 (analysis performed on 3,510 never smokers) United States (California) 1996 | Identified favorite movie stars of ever smokers (vs. never smokers) | S, P, SCH, SI, M | Susceptibility to smoking among never smokers (42%) | Adolescent never smokers choosing a favorite star typical of ever smokers vs. choosing a favorite star typical of other never smokers | AOR 1.35 (1.12–1.62) | Favorite actors and actresses were defined by the nominations of the subjects; study examined commonly chosen actors/actresses; 52% of adolescents were excluded because they nominated a star chosen by fewer than 5 respondents |
| Sargent et al. 2001a, 2002, 2009a; Tickle et al. 2006 | White Aged 10–15 years Cross-sectional school-based survey N = 4,919 (3,766 never smokers) United States (Northeast) 1999 | Movie title recognition—Beach method 50 titles/survey 601 U.S. box office releases, 1989–1999 | S, P, SCH, PS, SI, M | Tried smoking (17%) | Quartile of exposure to movie smoking: | AOR | A cross-sectional structural equation model (Tickle et al. 2006) identified indirect paths from exposure to movie smoking to intentions to smoke through positive expectancies and identification as a smoker, but not through normative beliefs |
| 1 | Reference | ||||||
| 2 | 1.9 (1.3–2.7) | ||||||
| 3 | 2.6 (1.8–3.7) | ||||||
| 4 | 2.5 (1.7–3.5) | ||||||
| Lifetime smoking level among triers (n = 794): puffers (57%), 1–19 cigarettes (19%), 20–100 cigarettes (9.7%), >100 cigarettes (13.8%) | No association between exposure to movie smoking and higher levels of lifetime smoking | ||||||
| Among never smokers: susceptibility to smoking (20%) | Quartile of exposure to movie smoking: | AOR | |||||
| 1 | Reference | ||||||
| 2 | 1.2 (0.9–1.5) | ||||||
| 3 | 1.5 (1.1–1.9) | ||||||
| 4 | 1.6 (1.2–2.1) | ||||||
| Positive expectancies (61% endorsed no positive expectancies) | Quartile of exposure to movie smoking: | APOR | |||||
| 1 | Reference | ||||||
| 2 | 1.2 (1.0–1.5) | ||||||
| 3 | 1.3 (1.1–1.6) | ||||||
| 4 | 1.4 (1.1–1.7) | ||||||
| Views adult smoking as normative (55%) | Quartile of exposure to movie smoking: | AOR | |||||
| 1 | Reference | ||||||
| 2 | 1.2 (0.9–1.4) | ||||||
| 3 | 1.3 (1.1–1.6) | ||||||
| 4 | 1.4 (1.1–1.7) | ||||||
| Tickle et al. 2001 | White, low-income communities Aged 10–19 years Cross-sectional school-based survey N = 632 (281 never smokers) United States (New Hampshire, Vermont) | Movie character smoking status of favorite star averaged for films released up to 3 years before survey | S, SCH, SI, M | Smoking index: 0 = nonsusceptible never smoker (37%), 1 = susceptible never smoker (7%), 2 = 1–99 lifetime cigarettes smoked, but not a current (30 days) smoker (26%), 3 = 1–99 lifetime cigarettes smoked and a current smoker (9%), 4 = ≥100 cigarettes lifetime (20%) Susceptibility among never smokers (17%) | Character smoking by favorite star averaged over 3 years: | APOR | Study examined commonly chosen actors/actresses; 51% of adolescents were excluded because they nominated a star chosen by fewer than 5 respondents |
| None | Reference | ||||||
| 1 | 0.78 (NS) | ||||||
| 2 | 1.53 (1.01–2.32) | ||||||
| ≥3 | 3.09 (1.34–7.12) | ||||||
| AOR | |||||||
| None | Reference | ||||||
| 1 | 2.16 (0.86–5.45) | ||||||
| 2 | 4.78 (1.60–14.2) | ||||||
| ≥3 | 16.2 (2.33–112) | ||||||
| Goldberg and Baumgartner 2002 | Asian Aged 14–17 years Cross-sectional school-based N= 1,338 Thailand 1998 | Recall measure—how many American movies have you seen in the past 2 months in theater or on video (0–1 vs. 2–3 vs. ≥4)? | None | Intent to smoke in the future | 0–1 movies (15%), 2–3 (14%), ≥4 (15%) | NS | Results shown for exposure to American movies on video; results similar for exposure to American movies in theater |
| Tried smoking | 0–1 movies (24%), 2–3 (29%), ≥4 (32%) | p <0.05 | |||||
| Smoked at least 1 cigarette | 0–1 movies (19%), 2–3 (24%), ≥4 (27%) | p <0.05 | |||||
| Dixon 2003 | White Aged 12–18 years Cross-sectional school-based N = 2,610 participants, 1,858 experimental smokers Australia 1999 | Movie character smoking status of favorite male and female star (mean smoking scenes per movie) | S, SCH, SI | Smoking uptake index: 0 nonsmokers (67%), 1 occasional smoker (12%), 2 light smokers (8%), 3 heavy smokers (5%), 4 chain smokers (1%) Null findings for negative health effects of smoking, endorsement of smokers as more popular, intent to smoke in future | APOR male actors: 1.16, p = 0.04 APOR female actors: NS | Stronger evidence for association among girls than in boys; study examined commonly chosen actors/actresses; 31% of adolescents were excluded because they nominated a star chosen by fewer than 5 respondents | |
| Goldberg 2003 | Asian Aged 14–17 years Cross-sectional school-based N = 1,762 Hong Kong 1998 | Recall measure—how many American movies have you seen in the past 2 months (0–1 vs. 2–3 vs. ≥4)? | No covariate adjustment | Intent to smoke in the future (27%) | 0–1 movies (21%), 2–3 (26%), ≥4 (30%) | p<0.01 | |
| Tried smoking (40%) | 0–1 movies (34%), 2–3 (41%), ≥4 (47%) | p<0.01 | |||||
| Current (7 days) smoking (30%) | 0–1 movies (18%), 2–3 (21%), ≥4 (22%) | NS | |||||
| Henriksen et al. 2004b | Multiethnic 6th–8th grades Cross-sectional school-based N = 2,125 California 2003 | Recall measure—how often have you seen smoking in the movies or on television in the past week (never vs. sometimes/often)? | M, P, PS, S, SI, SCH | Tried smoking (prevalence not described, current [30 days] smoking 2.6–7.6%, depending on grade in school) | Past-week viewing of smoking in movies or television: | AOR | Unadjusted OR was statistically significant = 2.2 (95% CI = 1.7–2.8) |
| Reference | |||||||
| Never vs. sometimes/often | NS (OR estimate did not survive stepwise regression) | ||||||
| McCool et al. 2005 | Multiethnic Aged 12 or 16 years Cross-sectional school-based survey N = 3,041 New Zealand | Recall measure— 3 items (How often do you see a film at the cinema?), α = 0.65 Positive smoker stereotypes (smokers in films are stylish, smart, sexy, healthy, intelligent), α = 0.79 | S | Intent to smoke in the future Mediators Imagery pervasiveness ("smoking in films is common"), 3 items, α = 0.61 Nonchalance ("smoking in films is not important to me"), 3 items, α = 0.67 | Continuous structural equation model; the relation between exposure to smoking in movies mediated through image pervasiveness and nonchalance Positive smoker stereotypes had a direct relation with intent to smoke in the future but were not predicted by higher exposure | ||
| Sargent et al. 2005 | Multiethnic national sample N = 6,522 Aged 10–14 years Cross-sectional random-digit-dial ed survey United States 2003 | Movie title recognition—Beach method 50 titles/survey 532 U.S. box office hits released from 1998 to 2003 | S, P, SCH, PS, SI, SINC, ACH, EA | Tried smoking (10%) | Quartile of exposure to movie smoking: | AOR | |
| 1 | Reference | ||||||
| 2 | 1.7(1.1–2.6) | ||||||
| 3 | 1.8(1.2–2.8) | ||||||
| 4 | 2.6(1.7–4.1) | ||||||
| AAF | 0.38 (0.20–0.56) | ||||||
| Hanewinkel and Sargent 2007 | White Aged 10–17 years Cross-sectional school-based survey N = 5,586 Germany (Schleswig-Holstein) 2005 | Movie title recognition— Beach method 50 titles/survey 398 internationally distributed movies that were German box office hits and released from 1994 to 2004 | S, P, SCH, PS, SI, M | Tried smoking (41%) | Quartile of exposure to movie smoking: | AOR | |
| 1 | Reference | ||||||
| 2 | 1.7(1.4–2.1) | ||||||
| 3 | 1.8(1.5–2.3) | ||||||
| 4 | 2.2(1.8–2.8) | ||||||
| Current (30 days) smoking (12%) | Quartile of exposure to movie smoking: | AOR | |||||
| 1 | Reference | ||||||
| 2 | 1.4(0.9–2.2) | ||||||
| 3 | 1.7(1.1–2.6) | ||||||
| 4 | 2.0(1.3–3.1) | ||||||
| Laugesen et al. 2007 | Annual school-based surveys 10th graders N = 96,156 New Zealand 2002–2004 | How often do you watch R-rated movies? (3 venues: cinema, video, TV) Never <1/month Once/month 2–3/month ≥once/week | S (sensitivity analysis adjusted also for SI, SINC, and PS did not change the conclusion) | Tried smoking among not current smokers | ARR | ||
| Watched R-rated movies: | Reference | ||||||
| Never | 1.20(1.12–1.28) | ||||||
| 2–3 times/month | 1.67 (1.55–1.80) | ||||||
| Once/month | 2.04(1.90–2.18) | ||||||
| Weekly | 2.28(2.12–2.45) | ||||||
| Current (30 days) smoking | Watched R-rated movies: | Reference | |||||
| Never | 0.80 (0.73–0.88) | ||||||
| 2–3 times/month | 1.15 (1.05–1.26) | ||||||
| Once/month | 1.59 (1.44–1.75) | ||||||
| Weekly | 2.31 (2.10–2.54) | ||||||
| Song et al. 2007 | Multiethnic Aged 18–25 years Cross-sectional Web-based survey N = 1,528 United States | Movie title recognition— Beach method 60 titles/survey 500 top-grossing movies released from 2000 to 2004 | S, P, SRA, SI, M, PPS | Current (30 days) smoking (31%) | AOR with exposure to movie smoking divided into quartiles and entered as a continuous variable | 1.21 (1.05–1.38) for each quartile increase in exposure | For the established smoking analysis, a mediational model that showed significant paths from movie smoking to established smoking through friend smoking and positive expectancies |
| Established smoking (>100 cigarettes lifetime) (25%) | AOR, same analytic approach as above | 1.08 (0.93–1.25) | |||||
| Thrasher et al. 2008 | Hispanic Aged 10–14 years Cross-sectional school-based survey N = 3,874 Mexico (Cuernavaca and Zacatecas) 2005 | Movie title recognition— fixed list of 42 box office hits (2002–2006) with >1 minute of smoking, 15 Mexican, 23 U.S., 4 other foreign | S, P, SI, BOF | Current (30 days) smoking (12%) | Quartile of exposure to movie smoking: | AOR | Significant multivariate association not found for perceived prevalence among adults |
| 1 | Reference | ||||||
| 2 | 1.4(0.9–2.4) | ||||||
| 3 | 1.8(1.0–3.2) | ||||||
| 4 | 2.7(1.5–4.7) | ||||||
| Ever smoked (41%) | Quartile of exposure to movie smoking: | AOR | |||||
| 1 | Reference | ||||||
| 2 | 1.3(0.9–1.6) | ||||||
| 3 | 1.8(1.4–2.4) | ||||||
| 4 | 2.3(1.5–3.6) | ||||||
| Among never smokers susceptible to smoking (40%) | Quartile of exposure to movie smoking: | AOR | |||||
| 1 | Reference | ||||||
| 2 | 1.5(1.1–2.0) | ||||||
| 3 | 1.8(1.2–2.5) | ||||||
| 4 | 1.6(1.1–2.3) | ||||||
| Attitudes toward smoking (good or bad; pleasant or unpleasant; safe or dangerous) | Quartile of exposure to movie smoking: | UAβ | |||||
| 1 | Reference | ||||||
| 2 | 0.17 (0.03–0.31) | ||||||
| 3 | 0.18 (0.02–0.34) | ||||||
| 4 | 0.41 (0.23–0.57) | ||||||
| Perceived prevalence among adults and youth | Quartile of exposure to movie smoking: | UAβ | |||||
| 1 | Reference | ||||||
| 2 | 0.21(0.03–0.39) | ||||||
| 3 | 0.30(0.16–0.44) | ||||||
| 4 | 0.34(0.18–0.50) | ||||||
| Hunt et al. 2009 | White Aged 19 years Cross-sectional N = 948 Scotland (Glasgow) 2002–2004 | Movie title recognition— Beach method 50 titles/survey 532 U.S. box office hits released from 1998 to 2003 | S, P, SCH, SI | Ever smoked (63%) Current smoker (33%) Occasional social smoker + regular smoker vs. never smoker + trier + former smoker | No bivariate or multivariate association with movie smoking No bivariate or multivariate association with movie smoking | AOR Not: significant | None of the associations between exposure categories was significant |
| Longitudinal | |||||||
| Dalton et al. 2003,2009; Tickle et al. 2006; Wills et al. 2007; Adachi-Mejia et al. 2009; Sargent et al. 2009a | Longitudinal school-based survey with telephone follow-up, baseline = 1,999 N = 2,603 baseline never smokers followed up at 18 months, 1,791 at 7 years United States (New Hampshire, Vermont) Follow-up at 18 months, 5 years White Aged 10–14 years at baseline Baseline smoking status: never smoker | Movie title recognition— Beach method 50 titles/survey 601 U.S. box office releases, 1989–1999 | S, P, SCH, PS, SI, M | 18-month endpoint Incidence of tried smoking (10%) | Quartile of exposure to movie smoking: | ARR | Dalton et al. (2003) also found a significant moderation effect on parental smoking (higher movie effects among adolescents whose parents did not smoke); Tickle et al. (2006) found significant indirect paths to intentions to smoke through positive expectancies and identification as a smoker; there was also a pathway to smoking behavior at 18 months through smoking status of favorite star; Wills et al. (2007) found that change in friend smoking status from time 1 to time 2 partially mediated the effect of movie exposure on smoking at 18 months; Adachi-Mejia et al. (2009) found a moderation effect for the 7-year endpoint, with stronger effect for adolescent team sports participants |
| 1 | Reference | ||||||
| 2 | 2.02 (1.27–3.20) | ||||||
| 3 | 2.16 (1.38–3.40) | ||||||
| 4 | 2.71 (1.73–4.25) | ||||||
| 7-year endpoint | AAF 0.52 (0.30–0.67) | ||||||
| S, P, SCH, PS, SI, M | Established smoking incidence (≥100 cigarettes lifetime at survey point) (27.8%) | Quartile of exposure to movie smoking: | ARR | ||||
| 1 | Reference | ||||||
| 2 | 1.36 (0.95–1.94) | ||||||
| 3 | 1.68 (1.15–2.44) | ||||||
| 4 | 1.98 (1.35–2.90) | ||||||
| Distefan et al. 2004 | Longitudinal random -digit-dial survey N = 2,084 never smokers at baseline Follow-up 3 years Multiethnic Aged 12–15 years Baseline smoking status: never smoker United States (California) | Movie character smoking status of favorite star Nonsmoker star smoked in <2 movies in preceding 3 years Smoker star smoked in ≥2 movies in the preceding 3 years | S, SCH, PS, SI, M | Tried smoking (not given, approximately 30%) | Nonsmoker star | Reference | Significantly stronger effect was found for females, with no effect for males |
| Smoker star | 1.36 (1.02–1.82) | ||||||
| Among females | Nonsmoker star | Reference | |||||
| Smoker star | 1.86 (1.26–2.73) | ||||||
| Jackson et al. 2007 | Longitudinal school-based survey, 2001–2002 N = 735 Follow-up at 2 years White and Black Mean age 13.6 years Baseline smoking status: never smoker United States (North Carolina) | Title recognition measure—93 film titles released 2001–2002 7 (G-rated), 14 (PG-rated), 49 (PG-13 rated), 23 (R-rated) | S, SI, PS, SCH, P | Tried smoking (30%) | No movie effect for Black adolescents | Television in the bedroom also found to be related to smoking; after controlling for this variable, the AOR for tercile 3 among White adolescents = 2.69 (1.25–5.77) | |
| Among White adolescents, tercile of exposure to R-rated movies: | AOR | ||||||
| 1 | Reference | ||||||
| 2 | 1.57 (0.73–3.35) | ||||||
| 3 | 2.67 (1.07–6.55) | ||||||
| Sargent et al. 2007a; Wills et al. 2008; Tanski et al. 2009 | Longitudinal random -digit-dial survey N = 6,522 baseline (5,829 never smokers) National sample Follow-up at 8 months (5,503), 16 months (5,019), 24 months (4,574) Multiethnic Aged 10–14 years at baseline Baseline smoking status: never smoker for outcome of tried smoking, not established smoker for outcome of established smoking United States 2003 | Movie title recognition—Beach method 50 titles/survey Baseline pool: 532 U.S. box office hits released from 1998 to 2003 Follow-up pools: movies released to box office or DVD during interim periods (approximately 150 titles for each follow-up survey wave) | S, SI, P, PS, EA, SCH | Tried smoking (15.9% by 24 months) | Continuous measure windsorized and scaled so 0 = 5th percentile and 1 = 95th percentile, assessed by character type: | AHR | Interaction effect for negative character smoking: AHR = 2.55 (1.50–4.32) for adolescents low in sensation seeking; Wills et al. (2008) found that the relation of movie exposure and onset of smoking was partially mediated through positive expectancies and change in the smoking status of friends; interaction effect for established smoking: AHR = 12.7 (2.0–80.6) for adolescents low in sensation seeking |
| Mixed | 1.39 (1.04–1.85) | ||||||
| Negative | 1.46 (1.07–1.98) | ||||||
| Positive | 1.39 (0.99–1.96) | ||||||
| S, SI, P, PS, EA, SCH | Established smoking (≥ 100 cigarettes lifetime) | Continuous measure windsorized and scaled so 0 = 5th percentile and 1 = 95th percentile | AHR 2.04(1.01–4.12) | ||||
| Hanewinkel and Sargent 2008; Sargent and Hanewinkel 2009 | Longitudinal school-based survey N = 2,711 Follow-up at 1 year White Aged 10–16 years at baseline Baseline smoking status: never smoker Germany (Schleswig-Holstein) 2005 | Movie title recognition—Beach method 50 titles/survey 398 internationally distributed movies that were German box office hits and released from 1994 to 2004 | S, P, SCH, PS, SI, M | Tried smoking (19%) | Quartile of exposure to movie smoking: | ARR | Hanewinkel and Sargent (2008) also found a significant moderation effect on parental smoking (higher movie effects among adolescents whose parents did not smoke); this and the dose-response curve were similar to Dalton et al. (2003) |
| 1 | Reference | ||||||
| 2 | 1.37 (1.09–1.68) | ||||||
| 3 | 1.78 (1.39–2.29) | ||||||
| 4 | 1.96 (1.55–2.47) | ||||||
| Smoking index (composed of lifetime smoking and current smoking items, α = 0.87) | Continuous measure windsorized and scaled so 0 = 5th percentile and 1 = 95th percentile, assessed by character type | APOR among baseline never smokers: 2.85 (1.90–4.26) Among baseline ever smokers, the interaction term was 0.55 (0.34–0.92), indicating a significantly lower response in this category of baseline smoker | |||||
| Titus-Emstoff et al. 2008 | Longitudinal, school-based, elementary schools, telephone N = 2,627 (2,499 baseline never smokers) United States (New Hampshire, Vermont) 2002–2003 Follow-up at 1 year (2,354) and 2 years (2,255) White Aged 9–12 years at baseline Baseline smoking status: never smoker | Movie title recognition—Beach method 50 titles/survey 550 popular contemporary movies, top 100 releases for each of the 5.5 years preceding baseline survey Follow-up movie pools selected on rolling basis from top 100 box office hits plus top 100 video rentals for the 12 months preceding survey | Tried smoking (9.6% by 24 months) | Exposure entered as continuous measure, with each 1-point increase equivalent to a 1-decile increase in exposure: | ARR for trying smoking at 24 months | AAF = 0.35 (0.16–0.53); majority of movie smoking exposure was from youth-rated movies | |
| Baseline (B) exposure | 1.09 (1.03–1.15) | ||||||
| 12-month exposure | 1.09 (1.03–1.16) | ||||||
| 24-month exposure | 1.07 (1.00–1.14) | ||||||
| B + 12-month exposure | 1.11 (1.04–1.17) | ||||||
| B + 12-month + 24-month exposure | 1.09 (1.02–1.16) | ||||||
| Using <25th percentile as reference | AAF 0.35 (0.16–0.53) | ||||||
| Using <10th percentile as reference | AAF 0.46 (0.11–0.70) | ||||||
| Thrasher et al. 2009 | Longitudinal school-based survey N = 3,874 baseline (2,093 never smokers) Mexico (Cuernavaca and Zacatecas) 2005 Follow-up at 1 year (1,741) Hispanic Aged 10–14 years Baseline smoking status: never smoker | Movie title recognition— fixed list of 42 box office hits (2002–2006) with >1 minute of smoking, 15 Mexican, 23 U.S., 4 other foreign | BOF, M, P, PI, S, SI | Tried smoking (36%) | Quartile of exposure to movie smoking: | ARR | |
| 1 | Reference | ||||||
| 2 | 1.01 (0.64–1.60) | ||||||
| 3 | 1.54 (1.01–2.64) | ||||||
| 4 | 1.41 (0.95–2.10) | ||||||
| Current (30 days) smoking (8%) | Quartile of exposure to movie smoking: | ARR | |||||
| 1 | Reference | ||||||
| 2 | 1.22 (0.59–2.51) | ||||||
| 3 | 2.44 (1.31–4.55) | ||||||
| 4 | 2.23 (1.19–4.17) | ||||||
| Wilkinson et al. 2009 | Longitudinal household survey N = 1,328 Follow-up at 6,12, 18, and 24 months (1,286) Hispanic Aged 11–13 years Baseline smoking status: never smoker for new experimentation United States (Texas) | Movie title recognition—Beach method 50 titles/survey 250 popular contemporary movies, top 50 releases each year 1999–2004 | P, S, SCH, SI | Ever tried cigarettes (n = 1,286) | Continuous measure windsorized and scaled so 0 = 5th percentile and 1 = 95th percentile | AOR 1.27 (1.10–1.39) | Interaction effect found for country of birth, with Mexican- born adolescents having a stronger response to smoking in movies, AOR = 1.52 (1.14–2.05), than did U.S. born, AOR = 1.04 (0.86–1.27) |
| New experimentation with cigarettes (n = 1,129) | Continuous measure windsorized and scaled so 0 = 5th percentile and 1 = 95th percentile | AOR 1.19 (1.01–1.40) | |||||
Note: Multiple citations within one cell are for multiple reports on the same sample. U.S. = United States.
Covariates: ACH = access to cigarettes in household; BOF = reported seeing bogus title; EA = extracurricular activities; M = other media/advertising influences; P = personality characteristics; PPS = perceived prevalence of smoking; PS = parenting style/parental oversight of smoking behavior; S = sociodemographics; SCH = school attachment and function; SI = other social influences (friend and family smoking); SINC = weekly spendable income; SRA = smoking-related attitudes/cognitions.
Measures of association: AAF = adjusted attributable fraction; AHR = adjusted hazard ratio; AOR = adjusted odds ratio; APOR = adjusted proportional odds ratio; ARR = adjusted relative risk; CI = confidence interval; NS = not significant; OR = odds ratio; UAβ = unstandardized beta coefficient.
From: 5, The Tobacco Industry’s Influences on the Use of Tobacco Among Youth
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.