C 2004) Journal of Behavioral Medicine, Vol. 27, No. 3, June 2004 (° Understanding Soft Drink Consumption Among Male Adolescents Using the Theory of Planned Behavior Nada O. Kassem1,2 and Jerry W. Lee1 Accepted for publication: September 17, 2003 This study identified factors that influence regular soda consumption among 564 male students, aged 13–18 years, attending North Los Angeles County public high schools. Participants completed a group-administered Theory of Planned Behavior based questionnaire. Almost all of the participants, 96.5%, reported that they currently drink soda, 60.2% reported drinking two glasses of soda or more per day during the past year. Students reported drinking regular soda more than diet soda and reported drinking phosphoric acid containing soda (cola) more than nonphosphoric acid containing soda (noncola). Attitude, subjective norm, and perceived behavioral control were significant predictors of intention to drink regular soda and together explained 61% of its variance. Our results suggest that parents, teachers/coaches, and health professionals should encourage the perception that there are other healthier drinks that quench thirst better than soft drinks and taste good, and that soda should not be excessively available at home. KEY WORDS: adolescents; soda; soft drink; carbonated beverage; cola. Soft drink consumption among children and adolescents has been associated with health risks (Mazariegos-Ramos et al., 1995; Wyshak, 2000). This is alarming since, in 2000, Americans spent over $60 billion on soft drinks, consuming more than 565 twelve-ounce cans per person per year (National Soft Drink Association, 2002). The recent increase in soft drink consumption 1Department of Health Promotion and Education, School of Public Health, Loma Linda University, Loma Linda, California. 2To whom correspondence should be addressed at 42030 Thornbush avenue, Lancaster, California 93536; e-mail: nadakassem@hotmail.com. 273 C 0160-7715/04/0600-0273/0 ° 2004 Plenum Publishing Corporation 274 Kassem and Lee is apparent in male adolescents. Since 1977–1978, soft drink intake has nearly tripled among adolescent boys (United States Department of Agriculture (USDA), 1994–1995). In 1996, adolescent boys age 12–19 years consumed an average of 1.78 twelve-ounce cans (606 g) of carbonated soft drinks per day (USDA, 1996). Adolescence is a crucial time for bone development. Approximately 50% of bone mass is accrued during the adolescent years (Anderson, 1995). In males, about 85–90% of peak bone mass is acquired by about age 20 years (Heaney, 2000) with maximal bone density gain occurring between the ages of 13 and 17 years (Kröger et al., 1993; Theintz et al., 1992). Substituting soft drinks for milk may adversely impact bone acquisition during adolescence predisposing to fracture risk (Heaney and Rafferty, 2001; Wyshak, 2000). In 1994–1995, adolescent boys age 12–19 years consumed twice as much soft drink as milk, 1.25 eight-ounce cups (316 g) of fluid milk versus 2.66 eight-ounce cups (609 g) of carbonated soft drinks per day (USDA, 1994–1995). Analyzing dietary data collected by the USDA in the Nationwide Food Consumption Survey (NFCS), 1977–1978, Guenther (1986) found that soft drink intake was negatively correlated with intake of milk, and the nutrients calcium, magnesium, riboflavin, vitamin A, and ascorbic acid among adolescents between the ages of 13 and 18 years. Using dietary data collected by USDA’s 1994 Continuing Survey of Food Intakes by Individuals (CSFII), Harnack et al. (1999) found that adolescents (13–18 years) consuming 13–25.9 oz (approximately 1–2 twelve-ounce cans) of soft drink per day were about 3.53 times more likely to consume less than 8 oz (1 cup) of milk per day than nonconsumers of soft drinks. Moreover, analyzing dietary data collected by USDA’s 1994–1996 CSFII, Ballew et al. (2000) found that carbonated soda consumption was negatively correlated with both milk and juice consumption among adolescent boys aged 12–17 years. These findings are of particular relevance because studies indicate that carbonated cola beverage consumption is associated with risk of bone fractures in children (Mazariegos-Ramos et al., 1995) and adolescents (Petridou et al., 1997; Wyshak, 2000; Wyshak and Frisch, 1994). The rising prevalence of obesity among adolescents is of public health concern. Obesity is associated with dyslipidemia, hyperinsulineia, hypertension, and cardiovascular disease (Freedman et al., 1999). Srinivasan et al. (1996) reported that obesity in adolescence persists into adulthood and it adversely impacts cardiovascular risk factors. According to the Third National Health and Nutrition Examination Survey (NHANES III) 1988–1994 data, over half of American adults are obese and 14% of adolescents aged 12–19 years are overweight (Johnson and Frary, 2001). Ludwig et al. (2001) found that consumption of sugar-sweetened drinks was associated with obesity in school children (age 11.7 years, SD 0.8) and for each additional serving of sugar-sweetened drink consumed, the odds of becoming obese increased Soda Consumption Among Male Adolescents 275 by 60%. Using the USDA’s 1994–1996 CSFII data, researchers found that the largest source of added sugars in the American diet was carbonated soft drinks accounting for one third (33%) of total intake (Guthrie and Morton, 2000; Krebs-Smith, 2001). Adolescent boys aged 12–19 years consumed an average of 280 calories per day from nondiet soda alone (Nestle, 2000). Additionally, using the NHANES III 1988–1994 data, Troiano et al. (2000) found that soft drinks contributed a higher proportion of energy intake (10.3%) for overweight than for nonoverweight adolescent males (7.6%) ages 12– 19 years. Frequent consumption of soft drinks, in particular cola-containing carbonated beverages, was associated with dental caries (Watt et al., 2000), gastric mucosal damage (Kapicioglu et al., 1998), decreased esophageal pH (Rubinstein et al., 1993), and increased periods of duodenal acidification predisposing to duodenal ulceration (Mccloy et al., 1984). Cola consumption was also positively associated with urinary stone disease (Shuster et al., 1985) and caused unfavorable changes in the risk factors associated with calcium oxalate kidney stone formation (Rodgers, 1999). Ellison et al. (1995) reported that soft drinks are the single highest source of caffeine in children’s diets. Caffeinated soft drinks may contain from 32 to 65 mg of caffeine per 12-oz can (Page, 1987) and some contain up to 100 mg [Jolt contains 75– 100 mg caffeine (Sizer and Whitney, 2000)], which is equivalent to a 6-oz cup of brewed coffee (Hughes and Hale, 1998). In 1994–1996, 1998 adolescent boys aged 12–19 years consumed an average of 85.5 mg of caffeine per day (USDA, 1994–1996, 1998). Griffiths and Mumford (1995, as cited in Hughes and Hale, 1998) reported that chronic caffeine consumption of 3.0 mg/(kg/day) may induce physical dependence in humans. Effects of chronic caffeine consumption among children and adolescents need to be further investigated to guide public health professionals in disease prevention. In summary, increased soft drink consumption among adolescents may adversely affect their health by increasing their sugar and caffeine intakes, and by displacing consumption of nutritious foods, thereby possibly increasing the risk of tooth decay, obesity, and bone fractures. Yet there is little literature on the factors influencing the increased soft drink consumption among male adolescents. The main purpose of this study was to use the Theory of Planned Behavior (TPB) (Ajzen, 1989) as a framework to identify those factors which influence regular soft drink consumption among male adolescents. THEORETICAL MODEL This study utilized Ajzen’s Theory of Planned Behavior to predict intention to drink regular soft drinks. The predictive power of this theoretical model has been established in many social and health behavior studies 276 Kassem and Lee (Godin and Kok, 1996; Libbus, 1995; Millstein, 1996). The TPB, a modified version of The Theory of Reasoned Action (Ajzen and Fishbein, 1980), is an expectancy-value model that addresses the problem of incomplete volitional control (Blue, 1995). Ajzen (1989) proposed that behavior is best predicted by intention. Intention is in turn determined by attitude toward the behavior (favorable or unfavorable), subjective norm (perception of social pressures to perform or not perform the behavior), and perceived behavioral control (perception of ease or difficulty of performing the behavior). Perceived behavioral control is also assumed to have a direct link to behavior. Persons’ beliefs about the outcomes of the behavior (salient behavioral beliefs) and their evaluation of those outcomes (outcome evaluations) influence their attitude. Persons’ beliefs about what others who are important to them want them to do (normative beliefs) and the motivation to comply with what those others want, influence their subjective norm. Persons’ beliefs about the availability of resources and opportunities necessary to achieve the behavior (control beliefs) and the power of those resources and opportunities to facilitate the behavior (perceived power) influence their perceived behavioral control. In summary, Ajzen proposed that “individuals will intend to perform a behavior when they evaluate it positively, believe that important others think they should perform it, and perceive it to be under their own control” (Courneya and McAuley, 1995, p. 501). We chose to use this theoretical perspective over others available (e.g., social cognitive theory; Bandura, 1986, 1997) because the TPB offers a finer grained examination of the belief structure underlying behavior. For example, social cognitive theory incorporates the concept of outcome expectancy which is roughly equivalent to outcome beliefs in the TPB. However, in social cognitive theory, outcome expectancy is treated in a monolithic way and as secondary to self-efficacy. The concepts of how specific salient outcome beliefs (and the values attached to those beliefs) are related to attitude and ultimately to intention and behavior are absent from social cognitive theory as are the concepts of attitudes, salient control beliefs and their power, and subjective norm. By examining this finer-grained belief structure, we hoped to gain insight into the specific beliefs that might be most crucial if one were to attempt to change the behavior of soda drinking. METHOD Participants There are a total of six public high schools in The Antelope Valley Union High School District located in North Los Angeles County. Piloting of the Soda Consumption Among Male Adolescents 277 final questionnaire was conducted in one high school located approximately 20 miles east of the five remaining high schools in the district. The final questionnaire was administered in those five schools. No data were collected from private high schools in this district. A total of 606 male adolescents, aged 13–18 years, participated in this study between February 1999 and March 1999. A total of 564 participants (93.1%) provided complete data on the TPB model variables and their data are included in this report. Data were also collected from female students and their data were the subject of a previous manuscript (Kassem et al., 2003). Instrument A questionnaire was developed using a two-stage process guided by the TPB (Ajzen and Fishbein, 1980). The first stage consisted of developing an initial open-ended questionnaire to obtain the salient beliefs underlying soft drink consumption. Utilizing 18 open-ended questions, the participants were asked to list behavioral, normative, and control beliefs related to regular soda and diet soda consumption. This questionnaire was administered to students with characteristics similar to the target population until saturation was reached (Strauss and Corbin, 1990). Saturation occurred when additional pretests resulted in no new information regarding the questions asked. A total of 31 public high school male students from four grade levels participated in the initial phase of the questionnaire development process. A content analysis was conducted and the results were utilized to develop the final questionnaire. The final instrument was designed to determine how salient beliefs relate to high soft drink consumption. The final questionnaire consisted of 80 items and was divided into three sections: (1) demographic characteristics of the survey respondents: age, gender, school grade level, and ethnicity (four questions); (2) Soda intake (12 questions); (3) behavioral intentions, attitudes, subjective norms, and perceived control related to daily regular soda consumption (64 questions). Two types of scales were used for section three of the questionnaire: 1) Semantic differential scales with bipolar adjective pairs (e.g., good–bad), and 2) Likert-like scales (e.g., not at all–very much). We asked students about their consumption of all types of soda but we based the questions that were relevant to the TPB (predictors of consumption) on regular soda only, because that type of soda is the one most commonly used by adolescents (USDA, 1996). A statement that the TPB questions were based only on the consumption of regular soda was clearly written in the instruction section of the questionnaire in several places. In addition, when the questionnaire was passed out, students were verbally 278 Kassem and Lee informed that the TPB section of the survey referred to regular soda rather than diet soda consumption. The time provided by the school district did not permit the assessment of regular soda and diet soda consumption. We chose regular soda consumption as the dependent variable based on two observations. First, results from a USDA 1996 survey, showed that out of the males in the 12–19 age group who reported drinking soft drinks, 95.7% drank regular carbonated soft drinks, and 7.5% drank diet (low calorie) carbonated soft drinks. Second, results of a survey we conducted during the pretesting phase showed that 91.15% of adolescent participants reported that they drank regular soda and 8.85% reported that they drank diet soda. The 80-item final questionnaire was piloted on 20 male students to detect ambiguous items and to determine the mean time of questionnaire completion. Students were also asked to record the time when they started and finished the survey and also to write comments or recommendations regarding the survey content or construction directly in the space provided on the survey instrument. The mean time of questionnaire completion was approximately 15 min. Procedure The Antelope Valley Union High School District and the Loma Linda University Institutional Review Board approved the study protocol. This study used an active parental consent protocol but a passive student assent protocol. Parents or guardians gave written approval for participation of their adolescent children, but the students were provided with a letter accompanying the final questionnaire that explained that they had the option not to answer the survey questions. The final questionnaire was group administered by the principal investigator to classes selected from among biology, chemistry, and healthful living classes in four grade levels by teachers who chose to participate in the study. A total of 20 teachers, three to five from each of the five public high schools selected 8–12 classes in which to conduct the study. Choice of classes was based on having no time overlap of classes so that the principal investigator would be able to be present in all classes to administer the questionnaire rendering the sample to be a convenience sample. To assure representation from the four grade levels, two to three classes per grade level per public high school were selected. A total of 50 classes participated in this study. One 55-min class period was devoted to completing the questionnaire. The survey was conducted on a voluntary and anonymous basis. This survey was a part of a doctoral dissertation which was originally designed to study only female behaviors since some diseases mentioned which were the focus of the dissertation (e.g., osteoporosis) affect females disproportionately; this study, part of doctoral dissertation, was originally designed to study only female behaviors. However, data were collected from both males and females Soda Consumption Among Male Adolescents 279 during all phases of the study to avoid the disruption that would be inherent in separating the males from the females in a school setting. After completing the first manuscript based on the findings of the female participants, we found that many of the adverse effects of soda consumption mentioned in our literature review section such as obesity, dental caries, and gastric mucosal damage may affect men as well; therefore, we decided to write this manuscript based on the findings of the male participants. To reduce inconsistencies of questionnaire administration, the principal investigator was present at all times during data collection using a standardized administration protocol. Upon completion of the survey, students were asked to place the response sheets into an envelope, seal it, and hand it in to the principal investigator. Students who completed their questionnaire early were asked to work on attached puzzles and not distract other students. Additionally, those students who did not wish to participate and those who were not given parental permission to participate as well as those who had filled out the survey in previous classes were asked to work on the puzzles as an alternative activity. Measures Behavior Soda consumption questions included items regarding average regular soda consumption. Participants indicated how frequently they drank different types of regular soda during the past 12 months (never/less than 1 glass, bottle or can per month, 1 glass per week or less, 2–6 glasses per week, 1 glass per day, 2 glasses per day, 3 glasses per day, more than 3 glasses per day). Behavioral Intention Participants rated on an 8-point scale (0 = strongly disagree, 7 = strongly agree; 0 = very unlikely, 7 = very likely) three items: “I intend to drink regular soda daily”; “How likely is it that you will drink regular soda daily?”; and “If everything goes as I plan, I will drink regular soda daily.” The reliability coefficient (alpha) for intention was 0.92. Attitude Participants rated on three 8-point bipolar adjective scales the following item: “When you think about drinking regular soda daily, how do you feel?” (0 = very bad, 7 = very good; 0 = very worthless, 7 = very valuable; 0 = very 280 Kassem and Lee unpleasant, 7 = very pleasant). The reliability coefficient (alpha) for attitude was 0.92. Subjective Norm Participants rated on an 8-point scale (0 = strongly disagree, 7 = strongly agree; 0 = none at all, 7 = a great deal) three items: “Most people who are important to me think I should drink regular soda daily,” “Important people in my life say I ought to drink regular soda daily,” and “How much pressure do you feel from other people to drink regular soda daily?” The reliability coefficient (alpha) for subjective norms was 0.83. Perceived Behavioral Control Participants rated on an 8-point scale (0 = very little, 7 = complete control; 0 = very difficult, 7 = very easy; 0 = strongly disagree, 7 = strongly agree) each of these three items: “How much control do you have over drinking regular soda daily?”; “For me drinking regular soda daily would be:”; and “If I chose to I would be able to drink regular soda daily.” The reliability coefficient (alpha) for perceived behavioral control was 0.73. The factors underlying attitudes, subjective norm, and perceived behavioral control were drawn from the initial elicitation and were measured on 8-point scales (0–7). The labels for the scales are shown in the table footnotes. During analysis scales on normative beliefs and control beliefs were recalibrated (–3.5 to +3.5). Then, each pair of underlying factors (behavioral beliefs/outcome evaluation, normative beliefs/motivation to comply, control beliefs/perceived power) was multiplied to obtain product scores of the strength of beliefs underlying attitudes, subjective norms, and perceived behavioral control. Data Analysis and Power Data was entered into the Statistical Package for the Social Sciences (1998) using an IBM compatible computer. Simple descriptive statistics were used to analyze the demographic data. Central tendency measurements and 95% confidence intervals were utilized to summarize the distribution of variables and their variability. Chi-square tests of independence were conducted to evaluate the differences between cases with complete data and cases with missing data. A paired-samples t test was conducted to Soda Consumption Among Male Adolescents 281 determine the mean difference in consumption of phosphoric acid containing soda (colas) and nonphosphoric acid containing soda (noncolas). Pearson correlation was used to examine the associations among the variables of the theoretical model. A five-step hierarchical multiple regression analysis was performed to determine the predictors of participants’ intention to consume regular soft drink. Behavior was regressed on intention and on the perceived behavioral control. Intention was regressed on attitudes, subjective norms, and perceived behavioral control. Attitude was regressed on the product of behavioral beliefs and outcome evaluations. Subjective norm was regressed on the product of normative beliefs and motivation to comply. Lastly, perceived behavioral control was regressed on the product of control beliefs and perceived power. A p value of 0.05 was considered significant for all statistical tests conducted. With our sample size there was 80% power to detect an effect size of 0.03 using 10 independent variables, which was the maximum number actually used in a regression analysis in this study. RESULTS Characteristics of Respondents Of the 606 students who participated in this study, a total of 564 participants (93.1%) provided complete data on the model variables. Results for this study relevant to the TPB variables were based only on the 564 cases with the complete data. Missing data analysis was conducted to evaluate whether there were consistent differences between the cases with complete data and those cases with missing data. The results of the chi-square test indicated that there were no significant differences among the ethnic backgrounds in amount of missing data [χ 2 (5, 600) = 4.483, p = 0.482]. Furthermore, no significant differences were found among the ages of the participants in the amount of missing data [χ 2 (5, 605) = 4.151, p = 0.528]. The response rate for this study was high. While exact counts were not kept of those students who failed to return the parental reply cards, there were, however, very few. The typical classes ranged in size from approximately 25–30 students, and within a class no more than two or three students failed to return the reply cards. In many classes all students returned the reply cards. The high response rate may have been due to the “Got milk” posters that were promised as incentives for students to return the parental reply cards. Students had verbally expressed strong desires to have those posters. The participants ranged in age from 13 to 18 years. The 15-year-olds 28.1% (n = 170) and 16-year-olds 25.2% (n = 153) were the most represented in this study followed by the 17-year-olds 23.6% (n = 143) and 282 Kassem and Lee the 14-year-olds 15.8% (n = 96), whereas, the 13-year-olds 0.2% (n = 1) and 18-year-olds 6.9% (n = 42) were the least represented. The majority of participants were White/Anglo Americans 53.8% (n = 323) followed by Hispanics or Latinos 18.5% (n = 111), Other 9.7% (n = 58), Asian/Pacific Islanders 8.8% (n = 53), Black/African Americans 7.8% (n = 47), and Native Americans 1.3% (n = 8). Compared to the ethnic mixture of The Antelope Valley Union High School District (AVUHSD), this study sample had fewer Hispanics or Latinos and fewer Black/African Americans. The ethnic mixture of the AVUHSD male students for the year 1999 was as follows: White/Anglo Americans 50.0% (n = 4491) followed by Hispanics or Latinos 30.3% (n = 2697), Black/African Americans 14.7% (n = 1324), Asians/Pacific Islanders 2.3% (n = 208), Native Americans 0.8% (n = 68), and Other 2.2% (n = 197). Information on the ethnic mixture of the AVUHSD was given to us by the district administration department. Compared to ethnic mixture of California public high schools (California Department of Education, 1999), this study sample had fewer Hispanic/Latinos and Fewer Asian/Pacific islanders but more other minorities. Patterns of Soda Consumption Almost all the participants, 96.5% (n = 544), reported that they currently drink soda, while only 3.0% (n = 17) reported that they do not. In response to an item asking students to indicate which type of soda they usually drink, 64.7% (n = 365) chose regular Coke, Pepsi, or other cola, and 24.8% (n = 140) chose regular noncola, for example, 7-up or Sprite. Additionally, 4.8% (n = 27) chose diet Coke, diet Pepsi, or other diet cola, and 2.3% (n = 13) chose diet noncola beverages. Table I shows patterns of various soda types consumed during the past year among study participants. Students in this study reported drinking regular soda more than diet soda and reported drinking phosphoric acid containing soda (colas) more than nonphosphoric acid containing soda (noncolas). Results of a paired-samples t test indicated that the mean intake for phosphoric acid containing soda (M = 1.67 glasses per day; 95% CI of 0.15) was significantly greater than the mean intake for nonphosphoric acid containing soda (M = 0.94 glasses per day; 95% CI of 0.12, t(558) = 10.2, p = 0.000). We did not directly ask students about their total soda consumption. To obtain an estimate of total soda consumption per day, we coded 1 or less glasses per week as 1/7th glass per day; 2–6 sodas per week as 0.5 glasses per day; 1, 2, and 3 glasses per day as 1, 2, or 3 glasses per day; and more than 3 glasses per day as 4 glasses per day. Then we summed Phosphoric acid containing soda Nonphosphoric acid containing Soda Regular Coke, Regular Diet Coke, diet Regular noncola, Regular Diet noncola, e.g. Pepsi, or other Dr. Pepper Pepsi, or other diet e.g. 7-up or Sprite Mountain Dew diet 7-up or diet cola (N = 564) (N = 564) cola (N = 564) (N = 564) (N = 564) Sprite (N = 564) Glasses of soda consumed N % N % n % n % n % N % Never/less than 1 glass per month 1 glass per week or less 2–6 glasses per week or less 1 glass per day 2 glasses per day 3 glasses per day More than 3 glasses per day Missing 65 68 153 104 81 37 51 5 11.5 12.1 27.1 18.4 14.4 6.6 9.0 0.9 218 140 103 41 27 13 16 6 38.7 24.8 18.3 7.3 4.8 2.3 2.8 1.1 402 82 31 16 12 7 8 6 71.3 14.5 5.5 2.8 2.1 1.2 1.4 1.1 107 188 138 55 36 13 21 6 19.0 33.3 24.5 9.8 6.4 2.3 3.7 1.1 262 165 69 29 15 2 15 7 46.5 29.3 12.2 5.1 2.7 0.4 2.7 1.2 452 57 23 10 6 4 6 6 80.1 10.1 4.1 1.8 1.1 0.7 1.1 1.1 Soda Consumption Among Male Adolescents Table I. Patterns of Soda Consumption During the Past Year Among Public High School Male Students in The Antelope Valley Union High School District, 13–18 years old, 1999 283 284 Kassem and Lee these values across all six types of soda listed in Table I. When these figures were used, 60.2% of the students reported that they drank two or more glasses of soda per day during the past year. Moreover, 40% reported that they drank three glasses of soda or more per day during the past year. Constructs of the Theory of Planned Behavior Figure 1 shows a path diagram representing the results of the five multiple regression analyses for predicting regular soda consumption. The diagram presents associations (simple correlations) of predictors with dependent variables (r s) and the independent association of each predictor with the dependent variables when other variables were held constant (βs). Both intention and perceived behavioral control were positively associated with the behavior. However, intention alone was the only significant independent predictor of the behavior. These two predictors together (R2 ) explained 15% of the variance in regular soda consumption. Intention as a Function of Attitude, Subjective Norm, and Perceived Behavioral Control Attitude, subjective norm, and perceived behavioral control were each significant predictors of intention to drink regular soda and together (R2 ) explained 61% of its variance. The strongest predictor was attitude, followed by perceived behavioral control and subjective norm. All three components also had moderate to high statistically significant positive associations with intention and these remained when the associations were assessed independently in a multiple regression. Underlying Factors of Attitude, Subjective Norm, and Perceived Behavioral Control Attitude-Outcome Beliefs and Evaluation A regression (results shown in Fig. 1) showed that the outcome products (R2 ) explained 38% of the variance in attitude toward regular soda consumption. While most of the outcome variables measured were significantly associated with attitude, there were only three variables that had independent positive predictive power. Those predictors were whether students enjoyed Soda Consumption Among Male Adolescents 285 Fig. 1. Path diagram for prediction of regular soda consumption based on the Theory of Planned Behavior. Numbers in rotated boxes are Rs from multiple regression. Numbers in parentheses next to arrows are correlations; numbers not in parentheses are beta coefficients. N represents cases with complete data. Note. ∗ p ≤ 0.05, ∗∗ p ≤ 0.001, ∗∗∗ p ≤ 0.0001. 286 Kassem and Lee the taste, it quenched their thirst, and whether they thought drinking regular soda would make them feel healthy. Table II shows which beliefs were most strongly and least strongly accepted. Students most strongly believed that they enjoy the taste of regular soda. Additionally, they strongly believed that drinking regular soda leads to tooth decay and getting too much caffeine. The outcomes that were rated most important were feeling healthy followed by developing tooth decay or cavities, face breaking out, and stomachache or gas. Subjective Norm-Normative Beliefs and Motivation to Comply A regression (see Fig. 1) showed that the referent products (R2 ) explained 36% of the variance in subjective norm. While all the normative variables measured were positively associated with the subjective norm, there were six that had independent predictive power. The primary predictor was parent. The weaker predictors were every one who drinks regular soda, friends, teacher, and/or coach, doctor, and fast food restaurant owners. Table III shows which beliefs were most strongly and least strongly accepted. It seems that students most strongly believed that soda companies, fast food restaurant owners, and famous people in regular soda ads wanted them to drink regular soda daily while doctors, teachers/coach, and parents did not want them to. Participants reported that they were motivated to comply the most with what their doctors and parents wanted them to do. Table II. Means and 95% Confidence Intervals for Regular Soda Consumption Behavioral Beliefs and Outcome Evaluations Behavioral beliefs (N = 564) Outcome evaluations (N = 564) Outcome Mean 95% CI Mean 95% CI Develop tooth decay or cavities Enjoy taste Too much caffeine Become hyper or have a sugar rush Quench thirst Feel addicted Gain weight Feel healthy Face break out Stomachache or gas 5.05 4.87, 5.23 6.00 5.86, 6.14 5.37 4.36 4.09 5.21, 5.53 4.16, 4.56 3.87, 4.31 5.01 3.76 2.38 4.84, 5.18 3.56, 3.96 2.18, 2.58 3.77 3.73 3.29 2.97 2.82 2.63 3.57, 3.97 3.51, 3.95 3.10, 3.48 2.79, 3.15 2.62, 3.02 2.43, 2.83 3.01 4.20 4.12 6.18 5.81 5.23 2.81, 3.21 4.00, 4.40 3.90, 4.34 6.05, 6.31 5.66, 5.96 5.05, 5.41 Note. The behavioral beliefs scale ranged from 0 (very unlikely) to 7 (very likely). The outcome evaluations scale ranged from 0 (not important) to 7 (extremely important). Soda Consumption Among Male Adolescents 287 Table III. Means and 95% Confidence Intervals for Regular Soda Consumption Normative Beliefs and Motivation to Comply Normative beliefs (N = 564) Motivation to comply (N = 564) Referent Mean 95% CI Mean 95% CI Soda companies Fast food restaurant owners Famous people in regular soda ads Every one who drinks regular soda Friends Sister or brother Parent Teacher and/or coach Doctor 6.07 5.59 5.92, 6.22 5.42, 5.76 1.91 1.89 1.72, 2.10 1.70, 2.08 5.04 4.82, 5.26 1.80 1.62, 1.98 3.84 3.66, 4.00 1.98 1.81, 2.15 2.31 1.99 1.85 1.69 1.41 2.12, 2.50 1.81, 2.17 1.68, 2.02 1.52, 1.86 1.24, 1.58 2.83 2.38 4.10 3.07 4.40 2.65, 3.01 2.20, 2.56 3.91, 4.29 2.88, 3.26 4.21, 4.59 Note. The beliefs about whether these people wanted them to drink regular soda (normative beliefs scale) ranged from 0 (not at all) to 7 (very much). The motivation to comply scale ranged from 0 (not at all) to 7 (very much). Perceived Behavioral Control–Control Beliefs and Perceived Power A regression (see Fig. 1) showed that the resource products (R2 ) explained 22% of the variance in perceived behavioral control. While all the resources measured were positively associated with perceived behavioral control, there were only two that had independent predictive power. The primary predictor was availability of regular soda at home. The weaker predictor was seeing advertisement to encourage drinking it. Table IV shows which resources students believed were most and least available. Students most strongly believed that availability of soda at school, access to vending machines containing regular soda and having enough money to buy regular soda at school help facilitate its consumption daily. The resources that were rated most important were having enough money to buy it at school, having access to vending machines containing regular soda and knowledge about the health risks of regular soda. DISCUSSION Studies indicated that adolescents who substitute soft drinks for milk and fruit juices compromise their calcium intake as well as other nutrients (Ballew et al., 2000; Harnack et al., 1999). In this study we found that almost all participants (96.5%) reported that they currently drink soda. Sixty 288 Kassem and Lee Table IV. Means and 95% Confidence Intervals for Regular Soda Consumption Control Beliefs and Perceived Power Control beliefs (N = 564) Perceived power (N = 564) Resource Mean 95% CI Mean 95% CI (1) Regular soda available at school (2) Access to vending machines containing regular soda (3) Enough money to buy Regular soda at school (4) Access to foods that make you thirsty (5) See advertisement to encourage to drink it (6) Regular soda available at home (7) Knowledge about the health risks of Regular soda 6.26 6.12, 6.40 4.72 4.54, 4.90 6.26 6.13, 6.39 5.09 4.93, 5.25 5.60 5.43, 5.77 5.20 5.03, 5.37 5.54 5.39, 5.69 3.91 3.72, 4.10 5.24 5.08, 5.40 2.50 2.30, 2.70 4.94 4.77, 5.11 4.86 4.69, 5.03 4.13 3.95, 4.31 5.04 4.87, 5.21 Note. The degree to which this resource was believed to be available for items (1 & 6) ranged from 0 (not available) to 7 (very much available); for items (2 & 5) ranged from 0 (none at all) to 7 (very much access); for item (4) ranged from 0 (not often) to 7 (very often); for items (3 & 7) ranged from 0 (not enough) to 7 (more than enough). The perceived power scale ranged from 0 (not important) to 7 (extremely important). percent of the respondents reported that they consumed two or more glasses of soda per day during the past year. Moreover, analysis of our data showed that cola-containing soda consumption among participants was significantly greater than non cola-containing soda consumption. This is of concern to public health professionals since a strong association was found between carbonated cola beverage consumption and fractures in children (Petridou et al., 1997) and teenagers (Wayshak and Frisch, 1994). The threat of developing weak bones and increased risk for bone fractures becomes greater when milk is replaced by soda during the teenage years when approximately 50% of bone mass is developed (Anderson, 1995; Wyshak, 2000). Prediction of Consumption Intention to drink regular soda was found to predict its consumption. Perceived behavioral control did not contribute independently to regular soda consumption; however, it was significantly and positively correlated with it. Ajzen argued that perceived behavioral control would predict behavior independently to the extent that it was an accurate predictor of actual control (Ajzen, 1991, p. 184). The lack of an independent contribution of Soda Consumption Among Male Adolescents 289 perceived control to prediction of behavior thus raises the interesting question of whether adolescents accurately perceive their degree of control over their soda-drinking behavior. The relatively low predictive power of intentions for behavior may result in part because the behavior being predicted was current use of soft drinks while the intention being measured was intention to engage in the behavior in the future. Prediction of Intention Attitudes, subjective norm, and behavioral control predicted 61% of the variance in intention substantially supporting the TPB. The strongest predictor was attitude followed by perceived behavioral control and subjective norm. This predictive pattern was consistent with other studies that utilized the TPB to research eating behaviors in adults (Godin and Kok, 1996) and also consistent in that subjective norm was found to be the weakest predictor of intention in 15 out of 19 study reviews by Ajzen (1991). Prediction of Attitude, Subjective Norm, and Perceived Control Regarding attitude, subjective norm, and perceived control each one could be predicted from the underlying belief components. However, the strength of the prediction was not as strong as was the case for intention with only 38, 36, and 22% of the variance being explained respectively by the three components of the theory. Possible reasons for this are that the individual belief measures were not measured as reliably as attitude, subjective norm, and intention. Each individual belief in the analysis was represented by the product of two conceptually different though related concepts (e.g., outcome belief and outcome value for outcomes) whereas attitude, subjective norm, perceived control, and intention were each the sum of multiple item scales with each item intended to measure the same thing. Another possibility is that though we used Ajzen’s qualitative methodology for determining the salient beliefs to include and selected the beliefs most frequently mentioned for inclusion, it is probably that the modal salient beliefs do not match the individual salient beliefs. That is, some individuals may have beliefs that are salient for them, but which are not salient for the typical adolescent. This would limit the overall ability of the beliefs used to predict individual attitude, subjective norm, and perceived control. Nevertheless, examining the individual beliefs that most strongly predicted attitude, subjective norm, and perceived control provides useful insight into what beliefs might be important in any campaign to change soda-consumption behavior. 290 Kassem and Lee Predicting Attitude Taste enjoyment of regular soda was one of the most predictive expected outcome beliefs of regular soda consumption among the male participants. The strong association between the reported consumption of food and taste enjoyment was also reported in earlier studies (Bobroff and Kissileff, 1986; Tuorila and Pangborn, 1988). Moreover, Anon (1986) (as cited in Tuorila et al., 1990) reported that almost all nationally surveyed American teenagers who drank regular sodas said that they drank it because of its taste. Therefore, taste enjoyment should be incorporated in health education programs aimed at encouraging teenagers to replace soda with healthier drinks that taste good as well. Quenching of thirst was the second most important predictor of attitude toward drinking soda. Moreover, in response to an item asking students to indicate which type of drink would most likely quench their thirst, approximately one third (31.6%) of respondents chose soda. This is alarming since soda was found to further dehydration in thirsty people (Brouns et al., 1998). This belief among adolescents that sodas quench thirst needs to be addressed when planning health and wellness programs. Feeling healthy was the third predictor of attitude toward drinking regular soda. This, despite the fact that the typical teen felt that drinking soda was unlikely to make them feel healthy. Reasons why some adolescents feel that sodas are not unhealthy drinks need to be explored in future studies. Predicting Subjective Norm While all referents were positively associated with the subjective norm, the most important independent predictor of subjective norm was parent. Other studies have also reported that adolescents’ choices of food are influenced by parents (De-Bourdeaudhuij and Van Oost, 1998; Doyle and Feldman, 1997; Neumark-Sztainer et al., 1999; Story et al., 2002). The influence of parents on adolescents’ food consumption behavior is important because parents provide a home environment for their teenage children (Kirk and Gillespie, 1990) (as cited in Frank, 1997) and they act as role models (Feunekes et al., 1998). Parents set the cooking and eating behaviors at home, for example, the purchase of foods and snack meals and the frequency of eating-out opportunities (Frank, 1997). Therefore, the involvement of parents in adopting healthy eating behavior among adolescents is important. Additionally our data indicated that others who drink soda, friends, teachers and/or coaches, doctors, and fast food restaurant owners did contribute to a lesser extent to prediction of subjective norm. Earlier studies indicated that dietary habits of adolescents are influenced by their friends (Barr, 1994) Soda Consumption Among Male Adolescents 291 and teachers, and/or coaches (Tatar, 1998), although in our study, parents had considerably more influence. Miller et al. (2001) reported that as adolescents become more independent, they tend to eat away from home more often and their food choices are influenced by their friends. According to Neumark-Sztainer et al. (1999), teenagers in his study reported that fast food restaurants tend to advertise meal deals with sodas rather than milk or juice. Moreover, French et al. (2001) reported that frequency of fast food restaurant use was positively associated with daily servings of sodas and was inversely associated with daily servings of fruits, vegetables, and milk. As the trend towards eating in fast food restaurant increases, it is important to develop strategies to assist adolescents to make healthful food choices (Miller et al., 2001). Predicting Perceived Control Availability of regular soda at home was the strongest predictor of perceived behavioral control. Male adolescents were more likely to drink regular soda if it was available at home. Participants believed that regular soda was very much available at home and they believed that its availability at home made it easy for them to drink it. In response to an item asking students to indicate whether they would drink something healthier if they ran out of soda at home, two thirds of the students (75.2%) reported that they strongly would do that. Messages to reduce regular soda consumption should be targeted at parents as well as the teens themselves. Lastly, seeing advertisements to encourage drinking soda was the second predictor of perceived behavioral control. Students believed that they have much access to advertisements that encourage them to drink it. Strasburger and Donnerstein (2000) reported that American children and adolescents spend an average of 3–5 h per day with the media such as television, radio, videos, video games, and the Internet. According to Nestle (2000), soft drink companies use various marketing strategies to compete for market share. Some of the marketing methods used by soft drink companies to promote their products to children at schools include contracts with school districts for exclusive use of particular brands of soft drinks, sponsorship of school sports, logos on vending machines and supplies, hallway advertising, advertisements on school buses, sports uniforms, scoreboards, free samples, and coupons for fast food (Nestle, 2000). Some of the marketing methods used outside of school include television, Internet, and magazine advertising, toys and clothing, telephone cards, celebrity product endorsements, product placements in movies, supermarket, and fast food chains (Nestle, 2000). As part of an effort to counteract the increased consumption of soft drink among adolescents, greater media advocacy on the 292 Kassem and Lee part of health professionals should be encouraged (Wallack and Dorfman, 1992). Limitations The limitations of this study include the use of self-report from the respondents that may limit the generality of the study. Retrospective selfreporting may be affected by memory consolidation or poor recall (Marx et al., 1996). Self-reports of dietary intakes and lifestyle behaviors may be influenced through researcher expectancies and social desirability (Cook and Campbell, 1979). Anonymity should have reduced this bias because the students were told the researcher and the teacher would not know what the students wrote; therefore, the students should have had no particular reason to try to impress the experimenter or teacher. Another limitation of the study is potential selection bias. Teachers chose which classes would participate in the study. Moreover, the high response rate might also be explained, in part, by the fact that teachers selected classes to participate; teachers may have selected classes they expected would be willing to participate. Generalizability of the study is limited by the sample of students who participated in the research, including their age, type of class they are enrolled in, ethnic mixture and region of the country. Additionally, the small number of participants who reported not drinking any type of soda might have slightly biased the associations between the constructs of the TPB model. Additionally, examining the association between intention to perform the behavior at some future time and current behavior is problematic. However, in the absence of compelling reason to change current behavior is likely to be closely related to future behavior. Thus, when examining the intention–behavior relationship, current behavior may serve as a rough marker for that future behavior when we are trying to understand how perceived control and intention jointly relate to behavior. However, to the extent that current stated behavior is an unreliable marker for future behavior, we would most likely be underestimating the intention–behavior relationship. Finally, the small number of participants who reported not drinking any type of soda might have biased the associations between behavior and intention. When examining soda consumption, the reasons why adolescents drink regular or diet (low calorie) soda may differ. However, the time provided by the school district did not permit the assessment of the full TPB model for both types of soda. Thus, we cannot determine the factors that influence adolescents’ intention to consume diet soda. This makes it more difficult to identify points of intervention to plan effective and comprehensive Soda Consumption Among Male Adolescents 293 educational programs to improve the health of adolescents. Future research should investigate both variables. Implications The results of the study provide general support for the TPB and suggest that it may be useful in providing insight into the nature of beliefs underlying a specific behavior of interest to health educators, in this case, soda consumption in male adolescents. For example, the results suggest that parents, teachers/coaches, and health professionals should encourage the perception that there are other healthier drinks that quench thirst better than soft drinks and taste good, that soda should not be excessively available at home, and that it is more important to target attitude toward soda consumption than subjective norms. However, such findings are only the first step. Further research is needed to determine whether programs developed to target specific beliefs found to predict TPB elements will have stronger impact on behavior than programs developed in a more generic fashion. ACKNOWLEDGMENTS The authors thank all the participating students, teachers, principals, and, in particular, Mikie Loughridge, EdD, the director of curriculum and instruction of The Antelope Valley Union High School District. The authors thank the following agencies that donated educational materials and “GOT MILK” posters to all the participating individuals: the Dairy Council of California, MilkPEP, and the National Institute of Child Health and Human Development. REFERENCES Ajzen, I. (1989). Attitude structure and behavior. In Pratkanis, A. R., Breckler, S. J., and Greenwald, A. G. (Eds.), Attitude Structure and Function, Lawrence Erlbaum, Hillsdale, NJ, pp. 241–274. Ajzen, I. (1991). The theory of planned behavior. Organ. Behav. Human Decis. Process 50: 179–211. Ajzen, I., and Fishbein, M. (1980). Understanding Attitudes and Predicting Social Behavior, Prentice-Hall, Englewood Cliffs, NJ. Anderson, J. J. (1995). Development and maintenance of bone mass through the life cycle. In Anderson, J. J., and Garner, S. C. (Eds.), Calcium and Phosphorus in Health and Disease, CRC press, Boca Raton, FL, pp. 265–288. 294 Kassem and Lee Anon. (1986). 2nd Annual National Survey. Teenage food survey. Food Prod. Dev. 47: 32– 41. Ballew, C., Kuester, S., and Gillespie, C. (2000). Beverage choices affect adequacy of children’s nutrient intakes. Arch. Pediatr. Adolesc. Med. 154: 1148–1152. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory, PrenticeHall, Englewood Cliffs, NJ. Bandura, A. (1997). Self-Efficacy: The Exercise of Control, W.H. Freeman, New York. Barr, S. I. (1994). Associations of social and demographic variables with calcium intakes of high school students. J. Am. Diet Assoc. 94: 260–267. Blue, C. L. (1995). The predictive capacity of the Theory of Reasoned Action and the Theory of Planned Behavior in exercise research: An integrated literature review. Res. Nurs. Health 18: 105–121. Bobroff, E. M., and Kissileff, H. R. (1986). Effects of changes in palatability on food intake and the cumulative food intake curve in man. Appetite 7: 85–96. Brouns, F., Kovacs, E. M., and Senden, J. M. (1998). The effect of different rehydration drinks on post-exercise electrolyte excretion in trained athletes. Int. J. Sports Med. 19: 56–60. California Department of Education, Educational Demographics Unit (1999). Statewide Enrollment in California Public Schools by Gender, Ethnic Designation and Grade, 1998–99 [Online]. Available at http://data1.cde.ca.gov/dataquest/StEnrAll.asp Cook, T. D., and Campbell, D. T. (1979). Quasi-Experimental Design and Analysis Issues for Field Setting, Houghton Mifflin Company, Boston, p. 67. Courneya, K. S., and McAuley, E. (1995). Cognitive mediators of the social influence-exercise adherence relationship: A test of the theory of planned behavior. J. Behav. Med. 18: 499– 515. De-Bourdeaudhuij, I., and Van Oost, P. (1998). Family members’ influence on decision making about food: Differences in perception and relationship with healthy eating. Am. J. Health Prom. 13: 73–81. Doyle, E. I., and Feldman, R. H. (1997). Factors affecting nutrition behavior among middle-class adolescents in urban area of Northern region of Brazil. Rev. Saude Public. 31: 342–350. Ellison, R. C., Singer, M. R., Moore, L. L., Nguyen, U. S., Garrahie, E. J., and Marmor, J. K. (1995). Current caffeine intake of young children: Amount and sources. J. Am. Diet Assoc. 95: 802–804. Feunekes, G. I., de Graaf, C., Meyboom, S., and van Staveren, W. A. (1998). Food choice and fat intake of adolescents and adults: Associations of intakes within social networks. Prev. Med. 27: 645–656. Frank, G. C. (1997). Methodological issues regarding eating behavior of high-risk adolescents. Ann. N. Y. Acad. Sci. 817: 66–82. Freedman, D. S., Dietz, W. H., Srinivasan, S. R., and Berenson, G. S. (1999). The relation of overweight to cardiovascular risk factors among children and adolescents: The Bogalusa Heart Study. Pediatrics 103: 1175–1182. French, S. A., Story, M., Neumark-Sztainer, D., Fulkerson, J. A., and Hannan, P. (2001). Fast food restaurant use among adolescents: Associations with nutrient intake, food choices and behavioral and psychosocial variables. Int. J. Obes. Relat. Metab. Disord. 25: 1823–1833. Godin, G., and Kok, G. (1996). The Theory of Planned Behavior: A review of its application to health-related behaviors. Am. J. Health Promot. 11: 87–98. Griffiths, R. R., and Mumford, G. K. (1995). Caffeine–A drug of abuse? In Bloom, F. E., and Kupfer, D. J. (Eds.), Psychopharmacology: The Fourth Generation of Progress, Raven Press, New York, pp. 535–580. Guenther, P. M. (1986). Beverages in diets of American teenagers. J. Am. Diet Assoc. 86: 493– 499. Guthrie, J. F., and Morton, J. F. (2000). Food sources of added sweeteners in the diets of Americans. J. Am. Diet Assoc. 100: 43–51. Harnack, L., Stang, J., and Story, M. (1999). Soft drink consumption among US children and adolescents: Nutritional consequences. J. Am. Diet Assoc. 99: 436–441. Soda Consumption Among Male Adolescents 295 Heaney, R. P., and Rafferty, K. (2001). Carbonated beverages and urinary calcium excretion. Am. J. Clin. Nutr. 74: 343–347. Heaney, R. P. (2000). Calcium, dairy products and osteoporosis. J. Am. Coll. Nutr. 19: 83S–99S. Hughes, J. R., and Hale, K. L. (1998). Behavioral effects of caffeine and other methylxanthines on children. Exp. Clin. Psychopharmacol. 6: 87–95. Johnson, R. K., and Frary, C. (2001). Choose beverages and foods to moderate your intake of sugars: The 2000 dietary guidelines for Americans–What’s all the fuss about? J . Nutr. 131: 2766S–2771S. Kapicioglu, S., Baki, A., Tekelioglu, Y. Arslan, M., Sari, M., and Ovali, E. (1998). The inhibiting effect of cola on gastric mucosal cell cycle proliferation in humans. Scand. J. Gastroenterol. 33: 701–703. Kassem, N. O., Lee, J. W., Modeste, N. N., and Johnston, P. K. (2003). Understanding soft drink consumption among female adolescents using the Theory of Planned Behavior. Health Educ. Res. 18: 278–291. Krebs-Smith, S. M. (2001). Choose beverages and foods to moderate your intake of sugars: Measurement requires quantification. J. Nutr. 131: 527S–535S. Kröger, H., Kotaniemi, A., Kröger, L., and Alhava, E. (1993). Development of bone mass and bone density of the spine and femoral neck—a prospective study of 65 children and adolescents. Bone Miner. 23: 171–182. Libbus, K. (1995). Women’s beliefs concerning condom acquisition and use. Public Health Nurs. 12: 341–347. Ludwig, D. S., Peterson, K. E., and Gortmaker, S. L. (2001). Relation between consumption of sugar-sweetened drinks and childhood obesity: A prospective, observational analysis. Lancet 357(9255): 505–508. Marx, B. P., Wie, V. V., and Gross, A. M. (1996). Date rape risk factors: A review and methodological critique of the literature. Aggress. Viol. Behav. 1: 27–45. Mazariegos-Ramos, E., Guerrero-Romero, F., Rodriguez-Moran, M., Lazcano-Burciaga, G., Paniagua, R., and Amato, D. (1995). Consumption of soft drinks with phosphoric acid as a risk factor for the development of hypocalcemia in children: A case-control study. J. Pediatr. 126: 940–942. Mccloy, R. F., Greenberg, G. R., and Baron, J. H. (1984). Duodenal pH in health and duodenal ulcer disease: Effect of a meal, Coca-Cola, smoking, and cimetidine. Gut 25: 386–392. Miller, G. D., Jarvis, J. K., and McBean, L. D. (2001). The importance of meeting calcium needs with foods. J. Am. Coll. Nutr. 20: 168S–185S. Millstein, B. (1996). Utility of the theories of reasoned action and planned behavior for predicting physician behavior: A prospective analysis. Health Psychol. 15: 398–402. National Soft Drink Association (2002). Soft Drink Facts [Online]. Available at http://www.nsda.org/softdrinks/History/funfacts.html Nestle, M. (2000). Soft drink “Pouring Rights”: Marketing empty calories. Public Health Rep. 115: 308–319. Neumark-Sztainer, D., Story, M., Perry, C., and Casey, M. A. (1999). Factors influencing food choices of adolescents: Findings from focus-group discussions with adolescents. J. Am. Diet Assoc. 99: 929–937. Page, R. M. (1987). Perceived consequences of drinking caffeinated beverages. Percept. Mot. Skills 65: 765–766. Petridou, E., Karpathios, T., Dessypris, N., Simou, E., and Trichopoulos, D. (1997). The role of dairy products and non-alcoholic beverages in bone fractures among school age children. Scand J. Soc. Med. 25: 119–125. Rodgers, A. (1999). Effect of cola consumption on urinary biochemical and physicochemical risk factors associated with calcium oxalate urolithiasis. Urol. Res. 27: 77–81. Rubinstein, E., Hauge, C., Sommer, P., and Mortensen, T. (1993). Oesophageal and gastric potential difference and pH in healthy volunteers following intake of coca-cola, red wine, and alcohol. Pharmacol. Toxicol. 72: 61–65. Shuster, J., Finlayson, B., Scheaffer, R. L., Sierakowski, R., Zoltek, J., and Dzegede, S. (1985). Primary liquid intake and urinary stone disease. J. Chron. Dis. 38: 907–914. 296 Kassem and Lee Sizer, F., and Whitney, E. (2000). Nutrition: Concepts and Controversies, Wadsworth/Thomson Learning, Belmont, CA, p. 466. Srinivasan, S. R., Bao, W., Wattigney, W. A., and Berenson, G. S. (1996). Adolescent overweight is associated with adult overweight and related multiple cardiovascular risk factors: The Bogalusa Heart Study. Metabolism 2: 235–240. Statistical Package for the Social Sciences (versions 7.5.1 and 8.0) [Computer software] (1998). SPSS Inc., Chicago, IL. Strasburger, V. C., and Donnerstein, E. (2000). Children, adolescents, and the media in the 21st century. Adolesc. Med. 11: 51–68. Story, M., Neumark-Sztainer, D., and French, S. (2002). Individual and environmental influences on adolescent eating behaviors. J. Am. Diet Assoc. 102: S40–S51. Strauss, A., and Corbin, J. (1990). Basics of Qualitative Research: Grounded Theory Procedures and Techniques, Sage, Newbury, CA. Tatar, M. (1998). Significant individuals in adolescence: Adolescent and adult perspectives. J. Adolesc. 21: 691–702. Theintz, B., Buch, B., Rizzoli, R., Slosman, D., Clavien, H., Sizonenko, P. C., and Bonjour, J.-P. H. (1992). Longitudinal monitoring of bone mass accumulation in healthy adolescents: Evidence for a marked reduction after 16 years of age at the levels of lumbar spine and femoral neck in female subjects. J. Clin. Endocrinol. Metab. 75: 1060–1065. Troiano, R. P., Briefel, R. R., Carroll, M. D., and Bialostosky, K. (2000). Energy and fat intakes of children and adolescents in the United States: Data from the National Health and Nutrition Examination Surveys. Am. J. Clin. Nutr. 72: 1343S–1353S. Tuorila, H., and Pangborn, R. M. (1988). Prediction of reported consumption of selected fatcontaining foods. Appetite 11: 81–95. Tuorila, H., Pangborn, R. M., and Schutz, H. G. (1990). Choosing a beverage: Comparison of preferences and beliefs related to the reported consumption of regular vs. diet sodas. Appetite 14: 1–8. United States Department of Agriculture (1994–1995). Continuing Survey of Food Intakes by Individuals [On-line]. Available at http://www.barc.usda.gov/bhnrc/foodsurvey/home.htm United States Department of Agriculture (1994–1996, 1998). Food and Nutrient Intakes by Children [On-line]. Available at http://www.barc.usda.gov/bhnrc/foodsurvey/home.htm United States Department of Agriculture (1996). Continuing Survey of Food Intakes by Individuals [On-line]. Available at http://www.barc.usda.gov/bhnrc/foodsurvey/home.htm Wallack, L., and Dorfman, L. (1992). Health messages on television commercials. Am. J. Health Promot. 6: 190–196. Wardlaw, G. M., and Insel, P. M. (1996). Perspectives in Nutrition, McGraw-Hill, Boston, p. 20. Watt, R. G., Dykes, J., and Sheiham, A. (2000). Preschool children’s consumption of drinks: Implications for dental health. Commun. Dent. Health 17: 8–13. Wyshak, G. (2000). Teenaged girls, carbonated beverage consumption, and bone fractures. Arch. of Pediatr. Adolesc. Med. 154: 610–613. Wyshak, G., and Frisch, R. E. (1994). Carbonated beverages, dietary calcium, the dietary calcium/phosphorus ratio, and bone fractures girls and boys. J. Adolesc. Health 15: 210–215.