Predicting Risky Behavior
Predicting Future Risky Behavior Among Those "Too Young" to Drink
as the Result of Advertising Desirability
Predicting risky behavior
Erica Weintraub Austin
Edward R. Murrow School of Communication
P.O. Box 642520
Washington State University
Pullman, WA 99164-2520
E-mail: [log in to unmask]
FAX: (509) 335-1555
M.A., Washington State University, 1997
Manuscript submitted to
The Communication Theory & Methodology Division of
The Association for Education in Journalism & Mass Communication
for 1998 presentation, Baltimore, Maryland
Predicting Risky Behavior
A convenience sample of 273 children in Washington state investigated the
validity of a predrinking behavior index as a behavioral outcome to assess media
effects on precursors to drinking among children for whom alcohol consumption is
not yet occurring. It also examined age trends in relevant beliefs and
behaviors. Perceptions of advertising desirability increased steadily from
third to ninth grade, whereas identification with portrayals leveled off after
sixth grade. Expectancies also increased with age, particularly between sixth
and ninth grade. Desirability predicted identification, and both predicted
expectancies, consistent with media decision-making theory. Expectancies
predicted alcohol predrinking behavior and risky behavior, which related
Predicting Future Risky Behavior Among Those "Too Young" to Drink
as the Result of Advertising Desirability
Research suggests that the media can make children more vulnerable to alcohol
experimentation (Aitken, Eadie, Leathar, McNeill & Scott, 1988; Atkin, Hocking &
Block, 1984; Austin & Johnson, 1997b; Grube & Wallack, 1994). Causation has
been difficult to demonstrate, however, because many cognitions that predict
drinking develop prior to first use (Austin & Johnson, 1997b; Miller, Smith &
Goldman, 1990; Wallack, Cassady & Grube, 1990) and most data are correlational
(Austin & Nach-Ferguson, 1994; Grube & Wallack, 1994; Wallack, et al., 1990).
This study, therefore, examines relationships among children's age, their
beliefs about advertising messages, their expectancies for drinking, behavioral
precursors to drinking, and self reported risky behavior.
A good deal of research suggests that children begin making decisions about
alcohol at a young age, with precursors to drinking in place, though not
unchangeable, by the elementary school years (e.g., Austin & Johnson, 1997b;
Miller, Smith & Goldman, 1990; Wallack, Cassady & Grube, 1990). Research also
has suggested that children's decisions about media messages have important
implications for the ways they make decisions about drinking alcoholic beverages
(e.g., Austin & Meili, 1994; Grube & Wallack, 1994). In particular, it appears
that the media can make children more vulnerable to future alcohol
experimentation because children do not develop adult-level comprehension skills
for media messages until about eighth grade (Collins, 1982), which is after many
children have begun to drink alcohol.
Several recent research projects have shown that children as young as seven
already are well engaged in decision making about alcohol and point to the role
of televised alcohol advertising in this decision-making process (Austin &
Nach-Ferguson, 1995; Austin & Johnson, in press; Wallack et al., 1990; Miller et
al., 1990) . Miller, et al. (1990) argue that because children by around third
grade have acquired better communication skills just as they are becoming more
attuned to social norms and peer influence, they are better able to understand
messages, such as those in the media, but simultaneously less able to resist
them. Exposure and attention to beer commercials correlates with brand
knowledge and positive attitudes toward alcohol (Aitken, Leathar & Scott, 1988;
Wallack, Cassady & Grube, 1990; Austin & Nach-Ferguson, 1995). Perceptions of
and admiration for alcohol advertising have been found to predict children's
intentions to drink earlier and engage in problem behaviors such as binge
drinking (Austin & Meili, 1994).
Despite this plethora of findings, research in this area has been weakened by
the lack of strong behavioral measures that would link exposure to media
portrayals of alcohol, such as advertising, to drinking behavior among children.
Prevention research would benefit from a behavioral measure that demonstrates an
unambiguous link between children's exposure to alcohol advertising and their
later drinking behavior. Measuring the precursors to drinking behavior would
provide a way to determine whether advertising effects on attitudes and beliefs
at a young age represent any real risk to children as they grow older and more
likely to engage in actual drinking behavior. A valid precursor, however, has
not been established.
This is not an unresolvable situation if researchers in the venue of substance
abuse follow the lead of researchers who have tackled similar challenges in the
study of the media and aggression. In the study of television violence effects,
for example, ethical concerns would prohibit a researcher from giving children
knives and guns in order to see whether viewing televised violence makes them
more likely to knife or shoot their playmates. Similarly, we cannot offer third
graders shots of whiskey or cans of beer to see if they might be more likely to
have a drink after watching televised beer advertisements. Nevertheless, by
paralleling the procedures used to study aggressive behavior in the media
effects literature (Liebert & Sprafkin, 1988), however, we can be somewhat
creative. For example, to study aggression researchers have measured activities
such as popping of balloons, aggressive play with toys, and choice of toy, with
more aggressive toys considered a measure of more aggressive tendencies in the
Similarly, play behavior can be measured in children to assess their affinity
for alcohol. Marketers have been producing a number of products emblazoned with
alcohol logos which could appeal to children. These items include beach towels,
piggy banks, toy trucks, salt and pepper shakers, hats, t-shirts, kazoos, model
airplanes, beach balls and basketballs. Children's desire to own such items
could indicate an affinity for alcohol products that later may be translated
into actual use of alcohol products. An alternative behavioral measure based on
children's desire to own products with alcohol logos has been developed and
tested successfully to represent children's receptivity to alcohol (Austin &
Johnson, 1997a, 1997b), but to date it has not been tested for its predictive
validity relative to actual drinking behavior. This study, therefore, pursues
this possibility in the context of a theoretical model that has been tested with
children of predrinking age to assess their decision-making strategies applied
to media message and alcohol.
The theoretical model that has been used, called the Message Interpretation
Process Model (Austin & Johnson, 1997b), suggests that children actively process
messages with strategies based partially on logic and partially on emotion or
wishful thinking. The theory, based on social cognitive theory (Bandura, 1986)
and decision-making theory (Beyth-Marom, Fischhoff, Jacobs, Quadrel and Furby,
1991) holds that a progression of logic-based decisions gradually leads to
conclusions about expectancies and behavioral intentions. It also suggests that
desirability, the perceived attractiveness of a portrayal, can overwhelm more
logically based decision-making criteria by directly influencing the degree to
which a child wants to emulate a portrayal, which then leads to expectancies,
which then lead to behavior (Austin & Johnson, 1997b; Austin & Nach-Ferguson,
1995; Austin & Johnson, 1997a; Austin & Meili, 1994). Advertising constructed
to maximize product desirability thus poses particular risks for the child whose
decision-making processes have not yet fully matured (Austin & Johnson, 1997a,
1997b; Miller, et al., 1990).
Critical viewing of mass media messages requires the ability to pay attention
selectively, to understand implicit as well as explicit information, to
understand the perspective and intentions of programmers and characters, and to
understand the meaning behind production techniques (Dorr, 1980). Researchers
suggest that children do not develop adult-level comprehension skills until
about eighth grade (Collins, 1982), with middle childhood representing a
critical period for decision making about a variety of topics, including alcohol
(Elias, Branden-Muller & Sayette, 1991; Miller, et al., 1990). In middle
childhood, children are developing better communication skills just as they
also are becoming more concerned with social norms, making them better able to
understand but less able to resist persuasive messages.
Media effects research has established that children nevertheless are not
persuaded by every media message. Mediators of effects include perceptions of
realism (Reeves, 1978; Hawkins & Pingree, 1982), similarity, and identification
with characters (Reeves & Garramone, 1989; Austin, Roberts & Nass, 1990).
Imitation tends to increase with frequent exposure to consistent messages
(Gerbner, Gross, Morgan & Signorielli, 1986), and with messages reinforced in
real life (DeFleur & Ball-Rokeach, 1982; Austin et al., 1990) and by significant
others (Austin et al., 1990; Austin & Meili, 1994). Television also becomes a
more influential source for topics when children get little information from
other sources (Rosengren & Windahl, 1972; Miller & Reese, 1982).
Key variables in the message interpretation process, therefore, include
perceptions of message desirability and the degree to which viewers identify
with portrayals. The Message Interpretation Process (MIP) model, for example,
holds that internalization of a television portrayal occurs at a number of
increasingly rigorous levels. Once identification has been established,
expectancies develop which in turn guide future behavior (Austin & Johnson,
1997a, 1997b; Austin & Meili, 1994). Desirability of a portrayal can be a
powerful influence on identification.
The majority of children do not begin drinking alcohol until the preteen or
teen years, but precursors of substance use behavior, such as beliefs about
media portrayals' desirability, are already well under development by the third
grade. In fact, attitudes towards drinkers exist by age six (Spiegler, 1983).
In addition, several recent research projects have shown that children as young
as seven already are well engaged in decision making about substance use and
point to the role of televised advertising in this decision making process
(Austin & Nach-Ferguson, 1995; Austin & Johnson, 1997a, 1997b; Wallack et al.,
1990; Miller et al., 1989). Finally, identification with portrayals and
expectancy development are well under way by third grade, although few children
have begun experimenting with entry-level drugs such as alcohol or tobacco. The
MIP model, however, theorizes that decisions that influence expectancy
development, such as determinations of desirability, identification and
expectancies, eventually will lead to decisions about actual behavior once
opportunity and desire converge.
It is reasonable, therefore, to expect that as children grow older, and
statistically more likely to engage in drinking behavior, their beliefs about
alcohol will become more positive as well. Exposure and attention to
commercials correlates with brand knowledge and positive attitudes toward
alcohol (Aitken, Leathar & Scott, 1988; Wallack, Cassady & Grube, 1990; Austin &
Nach-Ferguson, 1995). In addition, perceptions of and admiration for alcohol
advertising predicts children's intentions to drink earlier and engage in
problem behaviors such as binging (Austin & Meili, 1994).
H1: Children's beliefs regarding alcohol will be more favorable as grade level
increases, as measured by desirability, identification and expectancies.
H2: Children will exhibit more behavior representative of alcohol use as grade
level increases, as measured by their toy preference and self-reported frequency
of risky behavior.
Consistent with the MIP model and expectancy theory, it is appropriate to
expect that children's expectancies will be positively predicted by desirability
and identification. Specifically, the model holds that desirability should
predict identification, which should predict expectancies, which should predict
behavior. A survey of 154 at-risk adolescents, for example, found that
identification predicted behavioral intentions for alcohol use, with perceived
desirability of messages predicting identification (Austin & Meili, 1994).
H3: Children's expectancies towards alcohol will be positively predicted by
desirability and identification.
Finally, the study will test the proposition that, as with the study of media
effects on children's aggression, children's preferences for products displaying
alcohol logos will be representative of children's actual drinking behavior,
among those old enough to report actual drinking activity related to alcohol.
H4: Alcohol predrinking behavior will be positively predicted by children's
expectancies towards alcohol.
H5: Risky behavior will be positively predicted by children's expectancies
H6: Risky behavior will be positively relate to alcohol predrinking behavior.
A convenience sample of 273 third, sixth, and ninth graders in two Washington
state communities included 48.7% boys and 50.2% girls. A parent survey
indicated that respondents' income was higher than the median income for the
state (63% made more than $50,000/year), while the majority were white (90%),
consistent with the population of the state (U.S. Bureau of the Census, 1995).
The use of multiple classrooms ensured that results were not due to
teacher-specific effects. The survey was a pencil and paper questionnaire. The
questions were read aloud to the third graders.
EXPECTANCIES--Respondents' expectancies for alcohol, beliefs of positive
outcomes associated with drinking, were measured by a six-item index (Austin &
Johnson, 1997b; Austin & Meili, 1994), with acceptable reliability (alpha=.77)
on a scale ranging from yes, always (5) to no, never (1). Statements included
drinking beer helps you seem more grown up, helps you have fun, helps make you
happy, helps you make friends, helps make sports activities fun, and helps you
fit in. Table 1 displays the descriptive statistics for all variables used in
Table 1 About Here
IDENTIFICATION--Identification with, or the desire to be like, media portrayals
was assessed by three items (Austin & Johnson, 1997b; Austin & Meili, 1994), on
a scale ranging from yes, always (5) to no, never (1), with acceptable
reliability (alpha=.70), including I wish I could be like people on TV, be like
people in TV ads, and live my life like people in ads.
DESIRABILITY--An index (Austin & Johnson, 1997b; Austin & Meili, 1994), with
acceptable if somewhat low reliability (alpha=.65) assessed the degree to which
people in beer and wine ads appear to be positive role models, on a scale
ranging from yes, always (5) to no, never (1). Children reported whether they
thought people in beer and wine are seemed "popular" and "smart," whether women
in beer and wine ads are good looking, and whether men in beer and wine ads are
RISKY BEHAVIOR--Self-reported risky behavior, confirmed as a sufficiently
reliable measure (Johnson, O'Malley & Bachman, 1994), was assessed using three
measures of use for legal substances, which together provided a more reliable
estimate of risk taking than any single measure would among those whose behavior
patterns are still developing: cigarette smoking, use of chewing tobacco, and
drinking alcohol (alpha=.63). These data were collected only from sixth and
ninth graders on a scale ranging from never used (1) to use about every day (6).
ALCOHOL PREDRINKING BEHAVIOR--An alternative behavior measure has been
developed for use among children not yet making drinking decisions, based on the
research literature on the effects of televised violence on aggression in
children. The measures have been previously tested and peer reviewed (Austin &
Johnson, 1997a, 1997b). Preferences for products exhibiting beer or soda pop
logos were determined among third and sixth graders based on the presentation of
six items which respondents rated on a scale of 1-5 (not wanting it at all to
wanting it a lot). Each beer item had a corresponding item representing a soda
pop logo. Items included balls, toy trucks, motorized "dancing" cans, shirts,
piggy banks, towels, hats, and salt- and pepper-shakers, representing a variety
of brands. An improvement to the previous use of these measures was the
increase in number of items and the use of several balanced orders to improve
validity. A total of 18 items were randomly assigned to classrooms, with three
of each theme (soda pop or alcoholic beverage) represented in each order, to
avoid item-specific or order-specific effects. Items in each order crossed
gender stereotypes; for example, salt- and pepper-shakers countered athletic
Table 2 shows that as grade increased, so did desirability of portrayals
(p<.001), identification (p<.01), expectancies (p< .001) and risky behavior (p<
.0001), but not predrinking behavior (p< .10). Thus, hypothesis 1 was
supported, in that children's beliefs
Table 2 About Here
regarding alcohol were more favorable as grade level increased. Hypothesis 2,
however, which predicted that children would exhibit more behavior
representative of alcohol use as grade level increases, as measured by their toy
preference and self-reported frequency of risky behavior, received only partial
Relationship of media perceptions to alcohol beliefs and behavior
Table 3 displays the results of multiple regression analysis of relationships
between desirability, identification, expectancies and the two measures of
behavior predicted by the progressive decision-making model (e.g., Austin &
Meili, 1994; Austin & Johnson, 1997a, 1997b). Both desirability (p < .000) and
identification (p < .001) positively related to expectancies, as predicted by
hypothesis 3. Desirability related to identification (p < .000),
Table 3 About Here
suggesting that desirability predicts identification while both predict
expectancies. Expectancies predicted both predrinking behavior (p < .01) and
risky behavior (p < .0001), as predicted by hypothesis 4 and hypothesis 5.
Predrinking behavior related to risky behavior (r=.20, p<.01), as predicted by
hypothesis 6. Thus, a consistent pattern of relationships emerged for
predrinking behavior and risky behavior, in support of the MIP model's
prediction that desirability of messages and identification with portrayals of
alcohol use should positively predict expectancies and, therefore, actual use or
preuse of alcohol products.
This study has endeavored to establish the validity of a predrinking behavior
index as a relevant behavioral outcome when measuring media effects in order to
assess precursors to drinking among children for whom alcohol consumption is not
yet occurring. It also assessed the relationship of behavior to beliefs about
alcohol and of alcohol portrayals in advertising.
Desirability increased steadily from third to ninth grade, whereas
identification leveled off after sixth grade. That beliefs became more positive
between the third and sixth grades suggests that attempts to counter the appeal
of alcohol advertising need to occur well before sixth grade. If identification
has leveled off by sixth grade, which cannot be firmly established by this
small, cross-sectional study, this would suggest that for those likely to drink
alcohol, little may remain to motivate them other than opportunity. This would
suggest that prevention campaigns targeted to middle-school and
junior-high-school aged children take place far later than ideal. This would be
consistent with decision-making theory and developmental theories, which
indicate that middle childhood is a time of major decision making (Elias et al.,
1991; Miller et al., 1990).
Expectancies and risky behavior jumped between sixth and ninth grade, providing
cross-sectional support for the view that perceptions established by sixth grade
have the potential to feed into later substance use. Indeed, identification
and desirability predicted expectancies, which in turn predicted drinking and
predrinking behaviors. Identification and desirability did not have direct
effects on the behavioral outcomes, providing further support for the
theoretical model that suggests decision making progresses through a number of
steps over time to produce a cumulative effect on behavior.
The study also provided some support for the predrinking behavior measure as a
valid measure of behavior likely to evolve into drinking of alcoholic beverages.
The predrinking and risky behavior indices correlated moderately. This suggests
that products which appear designed to appeal to children, such as toy trucks,
basketballs, beach balls, piggy banks, and toy cans that dance when a switch is
flipped, do have the potential to prime them for future drinking. It is
important to note that the relationship between expectancies and predrinking
behavior was weaker than for risky behavior, suggesting that the desire for
alcohol-themed products, not surprisingly, is a less perfect measure of drinking
affinity than actual drinking behavior. In addition, the moderate correlation
between the two behavioral measures suggests that the desire for alcohol-themed
products will not necessarily lead to drinking behavior.
It should be noted that the moderate correlation represented only the
association between the two measures for sixth graders. No risky behavior data
was gathered from third graders, and no predrinking behavior data existed for
ninth graders. Thus, the only age at which both data existed was sixth grade.
Thus, it remains to be determined whether the relationship between the
predrinking behavior measure and actual risk taking strengthens, weakens or
remains the same as children grow into teenagers. That the associations between
expectancies and the behavior measures were consistent for the two behavior
measures across all age groups does provide additional confirmation that the
measures are related.
The limitations of the data and the moderate correlation between predrinking
and risky behavior should not diminish the worrisome implications of the
significant relationships that did emerge. The findings suggested that 1)
desirable images in alcohol advertising positively relate to the desire to be
like the images, even among third graders; 2) the desire to be like the images
levels off at sixth grade, but the belief that drinking will bring rewards is
predicted by this desire and continues to increase; 3) the belief that drinking
brings rewards predicts both the desire for alcohol-themed products and, among
older children, large amounts of variance in the frequency of risky substance
use; and 4) the desire for alcohol-themed products positively relates to actual
use of the product.
In short, this study adds to the increasing body of evidence that indicates
that alcoholic beverages, intentionally or unintentionally, are quite
effectively marketed to children. On its own, this study suffers from many
weaknesses common to modestly funded studies, such as a small, nonrandom sample
and a cross-sectional design. Its findings, however, are consistent with
existing theories and extend those theories with additional empirical evidence.
For example, this study used a convenience sample of children less at risk than
many others viewing alcohol advertisements, which limits the study's
generalizability but which should have made predicted relationships less likely
to emerge (Grant & Moore, 1995; Sillars, 1995; Young, 1993). That even these
data demonstrated a likely predictive link between beliefs about alcohol
portrayals, product appeal, and later consumption should cause concern.
People in ads are popular
People in ads are smart
People in ads are good looking
Men in ads are strong
I wish I could live like ads
I want to be like TV people
I wish I could be like ads
Drinking makes you happy
Drinking helps you fit in
Drinking helps make friends
Drinking helps you have fun
Drinking makes sports fun
Drinking makes you seem grownup
How often do you use cigarettes
How often do you use tobacco
How often do you drink alcohol
ANOVAs Assessing Grade-Level Differences
Independent df F P Means S.D. N
Desirability 2,265 36.02 .0001 11.15 3.55 268
third 9.02 3.23 87
sixth 11.71 3.21 143
ninth 13.87 2.79 38
Identification 2,267 4.62 .01 7.17 2.96 270
third 6.38 2.71 86
sixth 7.56 2.98 146
ninth 7.44 3.11 39
Expectancies 2,266 18.37 .001 9.06 3.70 269
third 8.03 2.93 86
sixth 8.87 3.37 144
ninth 12.05 4.77 39
Predrinking Beh. 1,231 2.76 .10 -3.54 4.94 233
third -4.23 5.35 87
sixth -3.12 4.65 146
Risky Behavior 1,176 34.36 .0001 4.01 1.93 178
sixth 3.59 1.32 138
ninth 5.45 2.84 40
*High means indicate higher levels of desirability, identification,
expectancies, and more engagement in risky behavior. A positive mean for
predrinking behavior indicates alcohol-item preferences outweigh soda pop
preferences, whereas a negative mean indicates more soda pop-related
Results of stepwise regressions predicting decisionmaking variables and
Variables: r2 chg. df F b P
Desirability .06 1,264 16.73 .24 .001
Desirability .18 1,260 58.05 .43 .0001
Identification .04 2,259 35.57 .38 .0001
Expectancies .02 1,260 4.96 .14 .027
Expectancies .31 1,173 80.42 .56 .0001
*Variables entered using forward-stepwise regression with models reflecting the
cumulative order of progression predicted by the MIP model. The model predicts
that desirability should predict identification, which should predict
expectancies, which should predict behavior.
Aitken, P.P., Eadie. D.R., Leathar. D.S., McNeill, R.E.J., & Scott, A.C.
(1988). Television advertisements for alcoholic drinks do reinforce under-age
drinking. British Journal of Addiction, 83, 1399-1419.
Atkin, C.K., Hocking, J., & Block, M (1984). Teenage drinking: Does
advertising make a difference? Journal of Communication, 34, 157-167.
Austin, E.W., & Johnson, K.K.(1997a). Immediate and delayed effects of media
literacy training on third graders' decision making for alcohol. Health
Communication, 9, 323-350.
Austin, E.W., & Johnson, K.K. (1997b). Effects of general and alcohol-specific
media literacy training on children's decision making about alcohol. Journal of
Health Communication, 2, 17-42.
Austin, E.W., & Meili, H.K (1994). Effects of interpretations of televised
alcohol portrayals on children's alcohol beliefs. Journal of Broadcasting &
Electronic Media, 38, 417- 435.
Austin, E.W., & Nach-Ferguson, B (1995). Sources and influences of young
school-age children's general and brand-specific knowledge about alcohol.
Health Communication, 7, 1-20.
Bandura, Albert (1986). Social cognitive theory. Hillsdale, NJ: Erlbaum.
Beyth-Marom, R., Fischhoff, B., Quadrel, M. J., & Furby, L. (1991). Teaching
decision making to adolescents: A critical review. In J. Baron & R. V. Brown
(eds.), Teaching Decision- Making to Adolescents, pp. 19-60 . Hillsdale, NJ:
Collins, W. A. (1982). Cognitive processing in television viewing. , 2. In
D. Pearl, L.
DeFleur, M. L., & Ball-Rokeach, S. (1982). Theories of mass communication
(4th ed.). NY: David McKay.
Dorr, A. (1980). When I was a child, I thought as a child. In S. Withey, &
R. Abeles, (Ed.), Television and Social Behavior: Beyond Violence and
Children. Hillsdale, NY: Erlbaum.
Elias, M. J., Branden-Muller, L. R., & Sayette, M. A. (1991). Teaching the
foundations of social decision making and problem solving in the elementary
school. In J. Baron & R.B. Brown (Eds.), Teaching decision making to
Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1986). Living with
television: The dynamics of the cultivation process. In J. Byrant, & D.
Zillmann, (Ed.), Perspectives on media effects, pp. 17-40. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Grant, A.E., & Moore, S. (1995). Minorities and television: A theoretical
perspective. Paper presented to the Mass Communication Division of the
International Communication Association: Albuquerque, NM, August.
Grube, J.W., & Wallack, L (1994). Television beer advertising and drinking
knowledge, beliefs, and intentions among schoolchildren. American Journal of
Public Health, 84, 254-259.
Hawkins, R., & Pingree, S. (1982). Television's influence on social reality.
I. Bouthilet, & J. Lazar, (Ed.), Television and behavior: Ten years of
scientific progress and implications for the eighties, pp. 9-23. Washington,
DC: U.S. Government Printing Office.
Johnson, L.D., O'Malley, P.M., & Bachman, J.G.(1994). National survey results
on drug use from the monitoring the future study, 1975-1993. NIDA: Rockville,
Liebert, R. M., & Sprafkin, J. (1988). The Early Window: Effects of Television
on Children and Youth. New York: Pergamon Press.
Miller, M. M., & Reese, S. D. (1982). Media dependency as interaction:
Effects of exposure and reliance on political activity and efficacy.
Communication research, 9, 227-248.
Miller, P.M., Smith, G.T., & Goldman, M.S.. Emergence of alcohol expectancies
in childhood: A possible critical period. Journal of Studies on Alcohol, 51,
Reeves, B. (1978). Perceived reality as a predictor of children's social
behavior. Journalism Quarterly, 55, 682-689, 695.
Reeves, B., & Garramone, G. (1982). Children's person perception: The
generalization from television people to real people. Human Communication
Research, 8, 4, 317-326.
Rosengren, K. E., & Windahl, S. (1972). Mass media as a functional
alternative. In D. McQuail, (Ed.), Sociology of Mass Communications, pp.
166-194. Middlesex, England: Penguin.
Sillars, A.L. (1995). Communication and family culture. In M. A. Fitzpatrick &
A. L. Vangelisti (Eds), Explaining Family Interactions, pp. 375-399. Thousand
Oaks, CA: Sage.
Spiegler, D. (1983). Children's attitudes toward alcohol. Journal of studies
on alcohol, 44, 545-552.
U.S. Bureau of the Census (1995). Statistical abstract of the United States
1993 (115th Edition). Washington, DC: Department of Commerce, Economic and
Wallack, L., Cassady, D., & Grube, J.W. (1990). TV beer commercials and
children: Exposure, attention, beliefs and expectations about drinking as an
adult. Washington, DC: AAA Foundation for Traffic Safety.
Young, T.J. (1993). Alcoholism prevention among Native-American youth. Child
Psychiatry and Human Development, 24, 41-47.
 Note: This study was conducted with assistance from Marilyn Bayona, Yin Ju
Chen, Wen-Wan Chen, Patrick DeLay, Yuki Fujioka, Michael Hall, Kristine Kay
Fortman, Chan-Ki Kim, Monakan Kiatikajornthada, Steven Liu, Eric Moreau, and
Tammie Wyers, all graduate students from the Edward R. Murrow School of
Communication, Washington State University.