Are Instagram Gym Advertisements Working Out? An Experimental Study of Model Body-Size and Slogan-Type

ABSTRACT

Previous health communication research has demonstrated the negative psychological and health effects of depicting thin-sized models in mass media advertisements including on social media sites such as Instagram. However, gym advertisements are one common source for the presentation of lean and thin-sized models on Instagram. Therefore, the current study guided by social comparison theory and signaling theory aimed to experimentally examine the effect of thin-sized models relative to plus-sized models as well as slogan-type (health and wellness versus physique-based) on women’s appearance comparison, body satisfaction, perceived gym fit, and intentions to join the gym. A sample of 217 undergraduate students who identified as women were randomly assigned to one of four Instagram gym advertisement conditions varying in model body-size and slogan-type. Appearance comparisons, perceived gym fit, and intentions to join the gym were measured post advertisement exposure and body satisfaction was measured pre-and-post advertisement exposure. As expected, exposure to Instagram gym advertisements featuring thin-sized models resulted in greater appearance comparisons and lower body satisfaction than exposure to Instagram gym advertisements featuring plus-sized models. Moreover, the combination of plus-sized models with health and wellness slogans in Instagram gym advertisements resulted in greater gym fit perceptions although there was no effect of model body-size and slogan-type on intentions to join the gym. This study supports social comparison theory, signaling theory, and practically the findings indicate that Instagram gym advertisements that depict plus-sized models (versus thin) with health-and-wellness slogans (versus physique) generate fewer body image concerns and lead to greater gym fit perceptions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10410236.2024.2342489.

Notes

1. We verified that the assumptions of the ANCOVA were met. First, there were no significant outliers in each message condition, and there was no reliable difference in physical attractiveness ratings of the models between the four message conditions, F(1, 220) = .008, p = .93, which indicates that the covariate was independent of the experimental effect. Next, condition did not interact with physical attractiveness ratings of the models on appearance comparisons, F(3, 220) = 2.01, p = .11, which indicates that the assumption of homogeneity of within condition regression slopes was not violated. Moreover, the quantile-quantile plots revealed that ratings were normally distributed. Last, we assessed the homogeneity of variance, which was not violated using the Levene's test, F(3, 217) = .57, p = .63.

2. We verified that the assumptions of the ANCOVA were met. First, there were no significant outliers in each message condition, and there was no reliable difference in pretest body satisfaction ratings between the four message conditions, F(1, 217) = 1.81, p = .14, and there was no reliable difference in physical attractiveness ratings of the models between the four message conditions, F(1, 217) = .56, p = .45, both findings indicating that each covariate was independent of the experimental effect. Next, condition did not interact with pretest body satisfaction ratings on posttest body satisfaction ratings, F(3, 217) = 1.52, p = .21, nor did condition interact with physical attractiveness ratings of the models, F(3, 217) = .50, p = .68, both findings indicating that the assumption of homogeneity of within condition regression slopes was not violated. Moreover, the quantile-quantile plots revealed that ratings were normally distributed. Last, we assessed the homogeneity of variance, which was not violated using the Levene's test, F(3, 214) = .32, p = .81.

3. We verified that the assumptions of the ANCOVA were met. First, there were no significant outliers in each message condition, and there was no reliable difference in physical attractiveness ratings of the models between the four message conditions, F(1, 220) = .03, p = .85, which indicates that the covariate was independent of the experimental effect. Next, condition did not interact with physical attractiveness ratings of the models on gym fit perceptions, F(3, 220) = .25, p = .86, which indicates that the assumption of homogeneity of within condition regression slopes was not violated. Moreover, the quantile-quantile plots revealed that ratings were normally distributed. Last, we assessed the homogeneity of variance, which was not violated using the Levene's test, F(3, 217) = 1.79, p = .14.

4. We first verified that the assumptions of the ANCOVA were met. First, there were no significant outliers in each message condition, and there was no reliable difference in physical attractiveness ratings of the models between the four message conditions, F(1, 220) = .40, p = .24, which indicates that the covariate was independent of the experimental effect. Next, condition did not interact with physical attractiveness ratings of the models on intentions to join the gym, F(3, 220) = .22, p = .88, which indicates that the assumption of homogeneity of within condition regression slopes was not violated. Moreover, the quantile-quantile plots revealed that ratings were normally distributed. Last, we assessed the homogeneity of variance, which was not violated using the Levene's test, F(3, 217) = .27, p = .85