Abstract
There is evidence for a negative influence of exercise/weight-loss television on explicit attitudes towards exercise. However, it is unclear whether these findings are specific to viewing intense exercise or the overweight/obese status of the viewed exercisers. Additionally, there is evidence that exposure to exercise cues can induce people to eat more, an effect hypothesized to reflect compensatory eating in response to cues of exercise rather than actual exercise. In this study, we examined the relative influence of viewing overweight/obese versus normal weight exercisers on eating behaviors (calories consumed) and attitudes towards exercise.
We randomized 102 college students to view one of three video clips manipulating viewing of vigorous exercise as well as the weight status of exercise participants: 1) overweight/obese individuals engaging in vigorous exercise; 2) normal weight individuals engaging in vigorous exercise; or 3) no-exercise video with participants of various weight statuses (control condition). Participants subsequently completed a taste test assessing calories consumed; a computerized, implicit attitudes-towards-exercise task; and an explicit attitudes-towards-exercise questionnaire.
We found that participants with higher BMIs and those viewing normal weight exercisers (vs. overweight/obese exercisers) ate significantly more. No significant effects were found for the interaction between BMI and video condition or for outcomes of explicit or implicit attitudes towards exercise.
This study extends findings of the impact of viewing vigorous exercise to eating behaviors. If replicated under more naturalistic conditions, these findings have implications for health promotion initiatives particularly identifying cues for overeating that may be impacted by pop cultural depictions as well as overweight status of the exercisers and viewers.
Keywords: eating, exercise, physical activity, health behavior, body mass index
Adults in the United States watch a mean of 2.77 hours of television per day (Krantz-Kent, 2018), making television a potentially powerful context for influencing health behaviors. Berry and colleagues (2013)_S1_Reference10 found that depictions of intense exercise on the television show, Biggest Loser, negatively biased individuals’ explicit (but not implicit) attitudes towards exercise relative to viewing a control video (American Idol). It is not clear whether this effect was due to the exercise depicted, the apparent suffering of individuals exercising, or the obese status of the exercisers. Indeed, it has been suggested that media depictions of vigorous exercise, with an emphasis on discomfort experienced by participants, may negatively bias attitudes towards exercise (Maibach, 2007). Similarly, weight-loss reality shows may increase anti-fat attitudes and dislike for overweight individuals (Domoff et al., 2012).
Further, exposure to exercise cues may induce increased eating, an effect hypothesized to reflect compensatory eating in response to cues of exercise rather than actual exercise (Albarracin, Wang, & Leeper, 2009). Compensatory eating (increased food intake after exercise) is prevalent (Moshier et al., 2016) and limits weight loss from exercise (King et al., 2007; Melanson, Keadle, Donnelly, Braun, & King, 2013). Accordingly, viewing depictions of exercise may have two additive negative effects: negatively biasing attitudes toward exercise while increasing calorie intake.
The present study was designed to evaluate these dual effects, while better understanding the nature of viewed exercise influences. We introduced a novel control video (fit exercisers) to be able to parse the influence of the weight status of the exercisers (overweight/obese) versus intense exercise itself. Hence, our design allowed for the comparison of the effects of viewing normal weight versus overweight/obese individuals vigorously exercising with a no-exercise control on the implicit and explicit attitudes towards exercise (evaluated by Berry et al. (2013)) and adds to the design an evaluation of calories consumed. We hypothesized that those viewing individuals exercising vigorously would consume more calories in a taste test paradigm, thus exhibiting compensatory eating to viewed exercise (as opposed to actual exercise). We also hypothesized that those viewing overweight/obese exercisers would demonstrate greater negative implicit and explicit attitudes towards exercise due to watching vigorous exercise specific to overweight participants. Finally, we hypothesized that participant body mass index (BMI) would moderate the associations such that overweight/obese participants would consume more calories and have more negative implicit/explicit attitudes towards exercise than normal weight participants.
Materials and Methods
Participants and Procedures
Young adults, ages 18+ (n=102), were recruited from a large university. Demographic characteristics are in Table 1.
Table 1.
Sample characteristics and outcome values.
| Mean ±SD/ %(n) | Biggest Loser | Challenge | American Idol | p value | |
|---|---|---|---|---|---|
| Sample Characteristics | |||||
| Age | 19.2 ± 1.4 (range 18–25) | 19.3 ± 1.5 | 19.3 ± 1.5 | 19.1 ± 1.2 | 0.74 |
| Sex (% female) | 56% (57) | 51% (18) | 59% (20) | 58% (19) | 0.81 |
| Race | |||||
| Ethnicity | |||||
| BMI | |||||
| Outcome Measures | |||||
| Total Calories Consumed | 190.1 ± 93.8 | 153.6 ± 86.0 | 217.0 ± 92.1 | 201.1 ± 94.0 | 0.01 |
| Implicit Attitudes Towards Exercise | 0.30 ± 0.35 | 0.29 ± 0.36 | 0.29 ± 0.37 | 0.32 ± 0.35 | 0.94 |
| Explicit Attitudes Towards Exercise | 13.0 ± 4.9 | 12.5 ± 5.9 | 13.9 ± 5.0 | 12.4 ± 3.3 | 0.35 |
Procedures were approved by our university Institutional Review Board. Following written informed consent, participants were randomized to view one of three 5-minute videos: Biggest Loser (overweight/obese individuals vigorously exercising; n=35), Challenge (normal weight individuals vigorously exercising; n=34), or American Idol (no-exercise control; n=33). Videos were chosen to balance demographic characteristics, exercise type (stationary biking), and intensity of and distress related to exercise. The control included individuals of both normal and overweight/obese status. After watching the video, participants were asked to record as many details as they could remember, and rated the familiarity and enjoyment of the depicted program. They then completed the taste test procedure, the implicit association test, and, finally, the rating of explicit attitudes. Height and weight were collected to inform BMI. Participants were subsequently debriefed and provided course credit for participation.
Calories consumed.
Participants completed a taste test, reporting taste ratings for snack foods (chips, M&Ms, cookies; 1–1.5 ounces of each snack was provided) after being informed that they could consume as much food as they would like, with requirements to eat at least one bite of each food. Participants were given 7-minutes for tasting, which was conducted in private to reduce social desirability bias. The outcome measure (calories consumed) was computed from the weight of snacks consumed over the tasting period.
Implicit attitudes.
A computerized Implicit Association Test-Physical Activity (IAT-PA; Conroy, Hyde, Doerksen, & Ribeiro, 2010), similar to that used in Berry et al. (2013), was used to assess negative attitudes towards exercise by observing latencies in responses to categorization. Participants rapidly categorized words under one of two headings (e.g., “good/physical activity” or “bad”) and then the location of the “physical activity” prompt was switched (e.g., “good” or “bad/physical activity”). Faster pairings are interpreted as more strongly associated in memory; thus, a stronger “bad/physical activity” pairing would indicate more negative attitudes towards exercise.
Explicit attitudes.
As per Berry et al. (2013), participants completed an Explicit Attitudes toward Physical Activity (Hyde, Doerksen, Ribeiro, & Conroy, 2010) self-report measure, with items rated on a 7-point Likert scale, with lower scores indicating more positive ratings about exercise.
Results
Preliminary Analyses
Participants did not differ significantly on baseline demographics across conditions (all p>0.41). Mean BMI was 23.3 (±3.5; range 17–39) with 26% of the sample in the overweight/obese category and no significant differences in BMI by condition (F(2,99)=0.81, p=0.45). Participants were significantly more familiar with American Idol as opposed to either other video (p<0.01); however, participants did not rate any program as more enjoyable than another (F(2,67)=1.70, p=0.19). Outcome values are provided in Table 1.
Effects on Outcome Variables
Separate univariate 2×3 ANOVAs, including BMI (low <25 and high ≥25;(World Health Organization, 2000) and the three video conditions, were conducted assessing effects on three outcomes: calories consumed, implicit attitudes, and explicit attitudes towards exercise. The main effects of BMI (F(1,96)=4.77, p=0.03, d=0.44) and video condition (F(2,96)=5.46, p=0.006, d=0.67) significantly predicted total calories consumed. As hypothesized, those with higher BMIs consumed more calories. Follow-up contrasts indicated participants watching normal weight individuals exercising consumed more calories than those viewing overweight/obese individuals exercising (p=0.01, d=0.72). Follow-up contrasts also indicated a trend (p=0.08, d=0.53) for individuals in the no-exercise control to consume more calories than those viewing overweight/obese individuals exercising. Contrary to hypotheses, the interaction between BMI and condition was not significant (p=0.41, d=0.27).
ANOVAs with the outcomes of explicit (all p>0.17, all d<0.33) and implicit (all p>0.16, all d<0.08) attitudes towards exercise did not any yield significant effects.
Discussion
This study expanded upon previous research by manipulating the weight status of depicted exercisers and including assessment of eating behaviors. We were unable to replicate Berry and colleagues’ (2013) findings of significantly more negative explicit attitudes towards exercise for individuals viewing overweight/obese exercisers versus a no-exercise control (reflecting a small effect size, d=0.17 in the Berry et al. trial, and d=.07 in the current study; neither study found differences for implicit attitudes). Unique to the design of the present study, we found significantly greater eating (d=0.72) in response to watching normal weight vs. overweight/obese individuals exercise. Indeed, consumption following viewing the no-exercise video fell approximately between these two exercise video conditions, suggesting suppression of appetite following viewing of overweight/obese exercisers, and a relative enhancement of appetite following viewing of fit exercisers. This latter effect is consistent with a compensatory eating response to exercise cues (Albarracin, et al., 2009).
Research has found that pairing food images with obese body images reduced liking of foods (Lascelles, Field, & Davey, 2003). Trends toward lower eating in overweight/obese exercise viewing condition may reflect these effects. There was not a hypothesized interaction between BMI and video condition; however, overweight/obese participants consumed more calories across video conditions (d=0.44). Yet, a limitation of our study is that 74% of the sample fell into normal range of BMI (<25), which is somewhat healthier than the US young adult population at large (60% normal weight range; Ogden, Carroll, Kit, & Flegal, 2014). In a larger sample of overweight/obese participants, the BMI main effect may have been stronger, with power to detect interaction effects between BMI and condition.
A second limitation is that our university-based sample lacked diversity with respect to age and educational status. Given that viewership of the television shows used in our study targets the 18–49 age group, this sample is representative of the low end of the range. Additionally, we cannot rule out that the lack of effects observed for implicit or explicit attitudes towards exercise may have been due to the progression of outcome tasks; the taste test occurred before these measures, making the exercise assessments more distal from television stimuli. Future research may wish to counterbalance outcomes to identify more proximal effects. Also, this study did not control for hunger. Finally, though the overall sample size was relatively similar to that of Berry et al. (2013), the addition of a third video significantly reduced sample sizes within the groups. Despite this difference in power, effect sizes did not support the effects reported by Berry et al. (2013).
Overall, this study draws further attention to the negative health behaviors that may be influenced by the content of television shows, particularly consumption of high calorie snack foods. Replication under more naturalistic conditions (e.g., home environment with participants’ preferred food options) is needed. If confirmed under more naturalistic conditions, information on effects like compensatory eating after viewing these television programs may be an important part of educational programs to aid weight-loss and to promote healthy eating and exercise behaviors.
References
- Albarracin D, Wang W, & Leeper J (2009). Immediate increase in food intake following exercise messages. Obesity (Silver Spring), 17(7), pp. 1451–1452. doi: 10.1038/oby.2009.16 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Berry TR, McLeod NC, Pankratow M, & Walker J (2013). Effects of biggest loser exercise depictions on exercise-related attitudes. American Journal of Health Behavior, 37(1), pp. 96–103. [DOI] [PubMed] [Google Scholar]
- Conroy DE, Hyde AL, Doerksen SE, & Ribeiro NF (2010). Implicit attitudes and explicit motivation prospectively predict physical activity. Annals of Behavioral Medicine, 39(2), pp. 112–118. [DOI] [PubMed] [Google Scholar]
- Domoff SE, Hinman NG, Koball AM, Storfer-Isser A, Carhart VL, Baik KD, & Carels RA (2012). The effects of reality television on weight bias: an examination of The Biggest Loser. Obesity (Silver Spring), 20(5), pp. 993–998. doi: 10.1038/oby.2011.378 [DOI] [PubMed] [Google Scholar]
- Hyde AL, Doerksen SE, Ribeiro NF, & Conroy DE (2010). The independence of implicit and explicit attitudes toward physical activity: Introspective access and attitudinal concordance. Psychology of Sport and Exercise, 11(5), pp. 387–393. [Google Scholar]
- King NA, Caudwell P, Hopkins M, Byrne NM, Colley R, Hills AP, … Blundell JE (2007). Metabolic and behavioral compensatory responses to exercise interventions: barriers to weight loss. Obesity (Silver Spring), 15(6), pp. 1373–1383. doi: 10.1038/oby.2007.164 [DOI] [PubMed] [Google Scholar]
- Krantz-Kent R (2018). Television, capturing America’s attention at prime time and beyond. Beyond the Numbers: Special Studies & Research, 7(14) [Google Scholar]
- Lascelles KR, Field AP, & Davey GC (2003). Using foods as CSs and body shapes as UCSs: A putative role for associative learning in the development of eating disorders. Behavior Therapy, 34(2), pp. 213–235. [Google Scholar]
- Maibach E (2007). The influence of the media environment on physical activity: looking for the big picture. Am J Health Promot, 21(4 Suppl), pp. 353–362, iii. [DOI] [PubMed] [Google Scholar]
- Melanson EL, Keadle SK, Donnelly JE, Braun B, & King NA (2013). Resistance to exercise-induced weight loss: compensatory behavioral adaptations. Med Sci Sports Exerc, 45(8), pp. 1600–1609. doi: 10.1249/MSS.0b013e31828ba942 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moshier SJ, Landau AJ, Hearon BA, Stein AT, Greathouse L, Smits JA, & Otto MW (2016). The Development of a Novel Measure to Assess Motives for Compensatory Eating in Response to Exercise: The CEMQ. Behav Med, 42(2), pp. 93–104. doi: 10.1080/08964289.2014.955077 [DOI] [PubMed] [Google Scholar]
- Ogden CL, Carroll MD, Kit BK, & Flegal KM (2014). Prevalence of childhood and adult obesity in the United States, 2011–2012. Jama, 311(8), pp. 806–814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- World Health Organization. (2000). Obesity: preventing and managing the global epidemic: World Health Organization. [PubMed] [Google Scholar]