03064603/86 S3.00 + .OO Copyright e 1986 Pergamon Press Ltd
Addictive Behaviors, Vol. 11, pp. 67-70, 1986 Printed in the USA. All rights reserved.
BRIEF REPORT DOES TELEVISED DRINKING INFLUENCE CHILDREN’S ATTITUDES TOWARD ALCOHOL? JONATHAN
B. KOTCH, MARTHA L. COULTER,
and ANGELA LIPSITZ
University of North Carolina at Chapel Hill Abstract-Fifth and sixth grade children randomly assigned to experimental and control groups saw videotaped television programs in which the drinking of alcoholic beverages by the principal characters appeared in the experimental condition but not the control condition. Parents’ drinking and home television watching were similar for both groups. Immediately after viewing the television programs, the children completed alcohol attitude and subjective expected utility (SEU) questionnaires. There were no significant differences in attitudes or SEU scores with one exception. Among boys only, those who saw the film with drinking were significantly more likely to respond that the good things about alcohol are more important than the bad things. Further research into the potential effects on children of drinking on television serials is indicated.
Alcohol is the “drug of choice” (Kovar, 1981) among adolescents. Both the Surgeon General (Healthy People, 1979) and the Select Panel for the Promotion of Child Health (Select Panel, 1981) have called attention to drinking problems in adolescence. Nearly half of all deaths among 16 to 19 year olds is due to automobile accidents, and 30 to 50% of these accidents involve alcohol (Insurance Institute, 1981). Alcohol may also play a role in adolescent deaths due to fire, falls, drowning and suicide (Eisenberg, 1980). At the same time, television has been shown to influence on both children’s attitudes and behavior. Two successive scientific committees have documented the relationship between violence on television and aggressive behavior in children (Comstock, 1971; Television and Behavior, 1982). Television commercials for candy and sugared cereals can influence children’s dietary habits (Goldberg, 1978), while commercials for proprietary medicines have been shown to influence their perceptions and attitudes (Atkin, 1978). In the case of alcohol specifically, exposure to media advertising including television commercials is associated with adolescents’ drinking behavior (Atkin, Hocking, & Block, 1984). Given this evidence, and knowing that alcohol is the drug most frequently depicted on television, constituting over two-thirds of all coded substance acts involved alcohol in the 1979-1980 season (Greenberg, 1980), it is surprising that no experimental work examining the effect of televised drinking on children has been reported. Recently, however, a study examining the effect on young adults’ drinking behavior of televised advertisements for alcoholic beverages has appeared in the literature (Kohn & Smart, 1984). METHOD
To test the hypothesis that exposure to dramatizations in which main characters are shown drinking alcoholic beverages can influence children’s attitudes toward drinking, Supported by grants from R.J. Reynolds Industries and the North Carolina Alcoholism Research Authority. Requests for reprints should be sent to Jonathan B. Ketch, Department of Maternal and Child Health, School of Public Health, Rosenau Hall 201-H, University of North Carolina at Chapel Hill, Chapel Hill, N.C. 27514. 67
68
JONATHAN B. KOTCH et al.
the investigators selected a broadcast television program seen regularly on commercial television which had consistently high rates of drinking, was popular with early adolescents (10 to 12 year olds), and was acceptable to parents. After recording numerous episodes of the selected TV program, two separate videotapes were created by splicing together scenes from separate programs. This technique was chosen in order to increase the frequency of drinking in the experimental condition without having to produce a television program for this purpose. On the first videotape, scenes in which characters drank alcoholic beverages in social or ceremonial contexts without negative consequences were interspersed with scenes in which no drinking occurred. The second tape contained only scenes with no drinking. The net result was two videotapes which were similar in length (about 35 minutes each), number of scenes, and number of episodes represented, but which differed in the presence of drinking in one tape but not in the other. In the drinking tape, there were 13 scenes out of 32 in which drinking occurred. The total number of drinking incidents (the number of times any character brought an alcoholic beverage to his or her lips) was 35. To measure the attitude of the children toward drinking, a questionnaire incorporating two separate schedules was constructed. This first schedule was Bauman and Bryan’s Subjective Expected Utility (SEU) Scale (Bauman & Bryan, 1980). The second was an adaption of Gorsuch and Arno’s “How Wrong Is It” instrument (Gorsuch & Arno, 1979), which was renamed “Do You Approve or Disapprove.” The SEU schedule asks 45 matched questions about the salience and probability of specific outcomes which could occur if the respondent had a drink of beer or hard liquor. The “Do You Approve or Disapprove” instrument includes 24 items describing situations in which drinking occurs and asks the respondent to rank his or her approval/disapproval on a 5-point Likert-type scale, from - 2 (strongly disapprove) to +2 (strongly approve). After a pilot study involving 13 fifth and sixth graders from a private school demonstrated that the videotapes held the children’s attention and that the design was feasible, two items were added to the children’s questionnaire testing overall utility and overall importance of the effects of drinking. Letters inviting participation in the study were sent to the parents of 237 fifth and sixth graders of a public elementary school serving rural and suburban children. Of the 61 children whose parents returned consent forms, 43 children ultimately participated, 19 fifth graders and 24 sixth graders, 19 of whom were boys and 24 girls. All 237 fifth and sixth graders completed daily television diaries for a week, although only the lists of the 43 participating children were analyzed. During the succeeding eight week period, each child was seen individually and was randomly assigned to the experimental or control condition by blindly selecting a colored poker chip. The child then saw the corresponding videotape with drinking or without drinking while the accompanying parent filled out an alcohol attitude and drinking behavior questionnaire. Children’s memory for the tapes they had just seen as a tested first, and then the children were given the SEU and “Do You Approve or Disapprove” questionnaire to fill out. RESULTS
Differences between the experimental and control groups on a few demographic variables, pre-test TV viewing, parental attitudes toward drinking and reported drinking behavior, and recall of the content of the videotapes were statistically evaluated using t-tests and Chi Square tests. The experimental group had significantly more boys than girls, five versus 14, respectively (p = .03). In order to reduce any confounding in
Televised drinking and children’s attitudes
69
interpretation between sex and the treatment factors, all the analyses investigating the major hypotheses were conducted separately for boys and girls. In order to obtain more meaningful dependent measures, and in order to group the questions from the “Do You Approve or Disapprove” questionnaire empirically, a factor analysis was performed on the approval items employing the principal components method with Piomax rotation. Three distinct and interpretable factors emerged accounting for 12.48% of the total variance. The questions with factor loadings higher than .55 were used to define each of the three factors. The factors thus created, “social relaxation”, “to drink in response to problems”, and “drinking in excess”, were considered as three additional dependent variables in the analysis. The analysis was conducted using 41 of the 43 subjects. Two subjects had to be dropped from the analysis of variance due to missing questionnaire items. All differences were statistically tested for each sex separately using analysis of variance. No significant results were obtained for SEU scores or the three “Do You Approve or Disapprove” factors, although differences, where they occurred, in approval factor scores were in the direction of the hypotheses. The overall utility of drinking question, indicating respondents’ estimates of the probability of drinking’s doing more harm than good or more good than harm, showed no significant differences. Indeed, all the boys, experimental and control, gave exactly the same answer. The overall importance question showed significant differences for the boys 0) = .03), but not for the girls. The question read, “Think of all the good and bad things that might happen to you if you drank alcoholic beverages. Which are more important to you?” The three response options were: (a) The good things are more important than the bad things; (b) The good and bad things are equally important; and (c) The bad things are more important than the good things. A Fisher’s exact test was carried out by collapsing responses b and c into a single category in order to create four cells. All five experimental boys chose response a. Five control boys also chose response a, and the remaining seven control boys chose either response b or c. The difference between experimental and control boys was significant (p = .04). No significant results were apparent for girls. CONCLUSIONS
The analysis of these data do not show any significant short term impact of the treatment on the children, with one exception. While all boys felt that alcohol causes more harm than good, experimental boys felt after viewing the film with drinking that the good effects of alcohol are more important than the bad, while, on average, controls felt that the good and bad effects are equally important. Some possible reasons why the other results were not significant include (a) the decision to use scenes in which drinking was followed either by positive or neutral consequences of drinking rather than positive consequences only; (b) failure to include a third condition in which the drinking episodes were followed by a negative consequences; (c) the small sample; (d) the failure of randomization to generate sufficient experimental boys; (e) the parents’ more powerful influence on the children; (f) the selfselected nature of the sample; and (g) the possible lack of salience of the videotapes. While the experiment failed to demonstrate significant short term effects of televised drinking on children’s attitudes toward alcohol, to conclude that television has no effect may be premature. Unique characteristics of the sample or of the method of producing the videotapes may have precluded uncovering more definitive results. Both experimental and survey research will continue to be necessary before the role of televi-
70
JONATHAN B. KOTCH et al.
sion in the development of children’s attitude toward alcohol and children’s drinking behavior will be known. REFERENCES Atkin, C.K., (1978). Effects of drug commercials on young viewers. Journal of Communication, 28,71-79. Atkin, C., Hocking, J., &Block, N.. (1984). Teenage Drinking: Does Advertising Make a Difference? Journal of Communication, 34, 157- 167. Bauman, K.E., & Bryan, E.S., (1980). Subjective expected utility and children’s drinking. Journal of Studies on Alcohol, 41, 952-958. Comstock, GA., Rubinstein, E.A., & Murray, J.P. (Eds.). (1971). Television and Social Behavior. Reports and Papers, Vol. V. Television’s Effects: Further Explorations. Rockville, MD: DHEW/NIMH. Eisenberg, L., (1980). Adolescent suicide: on taking arms against a sea of troubles. Pediatrics, 66, 351-320. Goldberg, M.E., Gory, G.J., &Gibson, W., (1978). TV messages for snack and breakfast foods: do they influence children’s preferences? Journal of Consumer Research, 5, 73-81. Gorsuch, R.L., Arno, D.H., (1979). The relationship of children’s attitudes toward alcohol to their value development. Journal of Abnormal Child Psychology, 7, 287-295. Greenberg, B.S., (1980). Health issues on commercial television series. Paper presented at Institute of Medicine conference, Washington, DC. Healthy People. (1979). DHEW (PHS) Pub. No. 79-55071. Washington, DC: U.S. Government Printing Office. Insurance Institute for Highway Safety (1981). Teens and autos: A deadly combination. Status Report 16, l-11. Kohn, P.M., Smart, R.G., (1984). Impact of television advertising on alcohol consumption: an experiment. Journal of Studies on Alcohol, 45,295-301. Kovar, M.G., (1981). Better Health for our Children: A National Strategy. V. III. A Statistical Profile. DHHS (PHS) Pub. No. 79-55071. Washington, DC: U.S. Government Printing Office. Select Panel for the Promotion of Child Health, (1981). Better Health for our Children: A NationalStrategy. V. I. Major Recommendations and Findings. DHHS (PHS) Pub. No. 79-55071. Washington, DC: U.S. Government Printing Office. Television and Behavior: Ten Years of Scientific Progress and Implications for the Eighties. (1982). V. I. Summary Report. DHHS Pub. No. (ADM) 82-1195. Rockville, MD: National Institute of Mental Health.