Yoga practice among ethnically/racially diverse emerging adults: Associations with body image, mindful and disordered eating, and muscle-enhancing behaviors

Abstract

Background:

This study explores cross-sectional associations between yoga and body image, mindful eating, disordered eating, and muscle-enhancing behaviors among a population-based sample of ethnically/racially diverse emerging adults.

Methods:

An ethnically/racially diverse population-based sample of 1,568 emerging adults (18-26 years) completed surveys as part of EAT 2010-2018 (Eating and Activity over Time). Models adjusted for sociodemographic characteristics and body mass index (BMI).

Results:

Practicing yoga at least 30 minutes/week was reported by 12.7% (n=210) of the sample. Yoga practitioners had higher levels of mindful eating than those not practicing yoga. Although effect sizes were small, yoga practitioners were more likely than non-yoga practitioners to use steroids (3.8% vs. 0.7%, p<.001, h=0.22) or protein powder/shakes (35.1% vs. 25.3%, p<.010, h=0.21) to increase their muscle size/tone. Body satisfaction, unhealthy weight control behaviors, and binge eating tended to be similar among yoga practitioners and non-yoga practitioners. There was a significant interaction between BMI and yoga in predicting body satisfaction with a trend toward a positive impact among yoga practitioners at higher BMI values. Interactions between yoga practice and all body image attitudes and behaviors across gender and ethnicity/race were not statistically significant.

Conclusions:

Young people from diverse ethnic/racial backgrounds who practice yoga are more likely to engage in mindful eating but have equal or elevated levels of unhealthy body image attitudes and behaviors as compared to non-yoga practitioners. Further research should explore how yoga is best taught and practiced to ensure that it is beneficial for body image and related behaviors.

BACKGROUND

Body image concerns, disordered eating behaviors, and behaviors aimed at changing body shape and size are highly prevalent among adolescents and young adults (1-4). Behaviors such as engaging in movement and mindful eating are to be encouraged. For example, mindful eating (e.g., making conscious food choices in response to one’s hunger and satiety cues), has been found to be associated with decreased risk for binge eating, emotional eating, and eating in response to external cues (5). In contrast, body dissatisfaction, unhealthy weight control practices, and use of muscle enhancers, such as steroids, are problematic in that they may increase risk for adverse outcomes such as more severe disordered eating behaviors, clinical eating disorders, excessive weight gain, and cardiometabolic disturbances (6-11). Research suggests that the prevalence of disordered eating and unhealthy muscle-enhancing behaviors may be equally high, or even higher, among young people from ethnically/racially diverse or lower socio-economic backgrounds than among those from primarily white and middle or upper socio-economic backgrounds (2, 12-16).

Yoga is a mind-body practice that involves physical postures, breath work, mindfulness, meditation, self-inquiry, lifestyle practices, and for some, enhanced spirituality (17-19). The practice of yoga has several underlying tenets that may assist in promoting a positive body image and a reduction in harmful disordered eating and muscle-enhancing behaviors (18, 20-24). For example, traditional yoga encourages greater awareness and response to one’s physical, energetic, mental, emotional, and spiritual needs. Learning to connect to one’s body and developing a greater sense of awareness of the body from the inside-out, rather than outside-in, are key components to a positive sense of embodiment (22).

The practice of yoga has been shown to be associated with a number of positive physical and emotional health outcomes (25-29). Research specifically examining the relationships between yoga and body image, disordered eating behaviors, and eating disorders, suggests that yoga may be beneficial, although associations tend to be modest and are not always consistent across studies (21, 30-38). In a qualitative study, when young adults practicing yoga were asked about the impact of yoga on their body image, the vast majority discussed its positive impact of yoga (19). However, some of the responding yoga practitioners, particularly those with higher body mass index, also indicated that the yoga could have a negative impact on their body image, primarily via comparative critique (e.g., comparing one’s body with other students or the teacher). Given that the majority of extant research suggests that yoga may be helpful for body dissatisfaction and related behaviors, with little evidence of harmful effects, it is worthwhile to explore whether young people from ethnically/racially diverse and lower socio-economic backgrounds, who may be underserved with regard to prevention and treatment options, are engaging in yoga.

While the body of research on yoga is rapidly growing, there is a dearth of research on the practice of yoga and associations with body image and related behaviors that has been conducted with young people from diverse ethnic/racial and lower socio-economic backgrounds (39). Learning more about yoga practice among ethnically/racially diverse populations and associations with body image, mindful eating, disordered eating, and the use of supplements to change one’s muscular shape or size, can provide insight into who is practicing yoga and how best to address their needs. Therefore, the aim of the current study is to explore associations between yoga practice and variables of relevance to the promotion of a healthy body image and the prevention of disordered eating/eating disorders and unhealthy muscle-enhancing behaviors in a diverse population-based sample of emerging adults.

METHODS

Study Design and Population

Data were collected from 1,568 emerging adults (mean age=22.0±2.0, range: 18-26 years) as part of the EAT 2010-2018 (Eating and Activity over Time) study. EAT 2010-2018 is a population-based, longitudinal study of factors related to eating, activity, and weight-related health in young people from diverse ethnic/racial and socio-economic backgrounds. Middle school and senior high school students at 20 public schools in Minneapolis-St. Paul, Minnesota completed classroom surveys and anthropometric measures during the 2009-2010 academic year (EAT 2010) (2, 40, 41). Participants completed follow-up surveys in 2017-2018 (EAT 2018). The University of Minnesota’s Institutional Review Board Human Subjects Committee approved all protocols.

Survey Development and Variables

Key items from the EAT 2010 survey were retained on the follow-up EAT 2018 survey, while additions were made to assess new areas of interest (40, 42-44). Focus groups with 29 emerging adults were conducted to pretest the EAT 2018 survey. Test-retest reliability of measures was examined on a subgroup of 112 participants. All variables utilized in the current analysis were assessed on the EAT 2018 survey unless indicated otherwise.

To assess yoga practice, participants indicated if they had ever done yoga over the past year (yes/no) (test-retest agreement=89%). Those who had ever done yoga were additionally asked, “On average, how frequently did you do yoga over the past year?” Seven response options ranged from “less than ½ hour/week” to “10+ hours/week.” Respondents who engaged in yoga at least 30 minutes/week were identified as practicing yoga (test-retest agreement=86%). This cut-off was selected based on prior research using a similar cut-off (19, 38, 45, 46), the distribution of yoga practice in the population, and to ensure that those included in the yoga group were practicing, on average over the past year, for at least 30 minutes/week.

Body Satisfaction was assessed with a modified version of the Body Satisfaction Scale (47). Satisfaction with 13 parts of the body (height, weight, body shape, waist, hips, thighs, stomach, face, body build, shoulders, muscles, chest, overall body fat) was rated using five response categories ranging from “very dissatisfied” to “very satisfied” (Cronbach’s alpha=0.94; range: 13-65; test-retest r=0.80).

Mindful eating was assessed with four items from the Mindful Eating Questionnaire: “I eat so quickly that I don’t taste what I’m eating; I snack without noticing that I am eating; Before I eat I take a moment to appreciate the colors and smells of my food; I taste every bit of food that I eat.” (48). Each item had four response options. The first two items were reverse scored and items were summed to form a score (range: 4-16; test-retest r=0.67).

Body Mass Index (BMI) was calculated (weight in kg/height in m²) based on self-reported weight and height, which have been found to be highly correlated with measured BMI in young adults (males: r=0.95; females r=0.98) (49).

Unhealthy weight control behaviors were assessed by asking: “Have you done any of the following things in order to lose weight or keep from gaining weight during the past year?” (yes/no for each method). Practices categorized as extremely unhealthy weight control behaviors included 1) took diet pills, 2) made myself vomit, 3) used laxatives, and 4) used diuretics. Practices categorized as less extreme unhealthy weight control behaviors included 1) fasted, 2) ate very little food, 3) used a food substitute (powder or a special drink), 4) skipped meals, and 5) smoked more cigarettes. Those who responded “yes” for one or more methods were coded as users of extreme (test-retest agreement=93%) and less extreme unhealthy weight control behaviors (test-retest agreement=76%).

Binge eating was assessed with the questions: “In the past year, have you ever eaten so much food in a short period of time that you would be embarrassed if others saw you (binge eating)”? (yes/no). “During the times when you ate this way, did you feel you couldn’t stop eating or control what or how much you were eating?” (yes/no). Those who responded “yes” to both questions were categorized as engaging in binge eating (test-retest agreement=89%) (50).

To assess muscle-enhancing behaviors, participants were asked: “Have you done any of the following things in order to increase your muscle size or tone during the past year?” and specifically were asked to report (yes/no) on use of protein powder or shakes (test-retest agreement=88%) and steroid use (test-retest agreement=not available) (51-53).

Two items individually assessed moderate-to-vigorous physical activity (MVPA): “In a usual week, how many hours do you spend doing the following activities?” (Six response options ranged from none to six or more hours). Strenuous activity was described as activity during which the heart beats rapidly and moderate activity was described as not exhausting. Examples of activities were given after each question (test-retest r=0.72) (54).

Young people from socially marginalized ethnic/racial and low socio-economic backgrounds may have more adverse health consequences due to factors related to racism and social inequities, such as fewer resources aimed at health promotion, therefore it is important to assess both ethnicity/race and socio-economic status (SES). Ethnicity/race was assessed by self-report at baseline with the question: “Do you think of yourself as…? (You may choose more than one) (1) White, (2) Black or African American, (3) Hispanic or Latino, (4) Asian American, (5) Native Hawaiian or Pacific Islander, (6) American Indian or Native American, or (7) Other” (test-retest agreement=98–100%). The ethnic/racial identity of participants who chose “Hispanic or Latino” and “White” was categorized as Hispanic for analysis. Additionally, participants who reported more than one race (other than White), or reported Hispanic and any race other than White, were coded as “mixed” and combined with participants reporting “other” for analysis. Since very few participants reported “Hawaiian or Pacific Islander,” or did not respond, these participants were also included in the category “mixed or other.” SES classification was determined at baseline and was primarily based on the highest education level of either parent (1=did not finish high school, 2=finished high school or general equivalency diploma, 3=some college, 4=finished college, 5=master's or doctoral degree) with adjustments made for student eligibility for free/reduced-price school meals, family public assistance receipt, and parent employment status (55, 56). Adjustments were made using an algorithm that reduced SES classification by two levels if an adolescent’s family was receiving public assistance, and by one level if an adolescent was eligible for free or reduced-price school lunch or had two unemployed parents (test-retest r=0.90). For the current analysis, the five-level SES variable was collapsed to three groups (SES rank 1=low, SES ranks 2-3=middle, SES ranks 4-5=high). Additionally, on the EAT 2018 survey, participants were given three options for gender: Male, Female, or Different identity (please specify). Age was calculated based on reported birth date and date of survey completion.

Statistical Analysis

Analyses were conducted in SAS software (version 9.4, 2013; SAS, Inc., Cary, NC). Descriptive proportions of young adults practicing yoga by sociodemographic characteristics were summarized. Differences across sociodemographic characteristics were tested by logistic regression of yoga practice (yes/no) as the outcome on each characteristic as the predictors. Models were run unadjusted (e.g., for each sociodemographic characteristic separately), then simultaneously controlling for all sociodemographic characteristics and body mass index (BMI). Body satisfaction, mindful eating, disordered eating, and unhealthy muscle-enhancing behaviors were examined in relation to yoga practice using separate regressions of each measure (as the outcome) on yoga practice as the predictor controlling for sociodemographic characteristics and BMI. Continuous measures used linear regression, while dichotomous measures used logistic regression. Regression-adjusted means and proportions of each measure are presented by yoga practice as well as effect size differences (standardized mean difference Cohen’s d for mean differences, Cohen’s h for proportion differences). Effect sizes are used for comparing magnitudes across various outcomes on different scales. Commonly used labels to describe Cohen’s d and h effect sizes are 0.20 (small), 0.50 (medium), 0.80 (large) effects (57). Additional regression models for steroid use and protein powder/shakes were fit controlling for the potential confounder of physical activity. The association between yoga and each outcome was also assessed by ethnicity/race categories and by gender (only male and female due to small numbers of other identities), including respective interactions into the regressions. Inverse probability weighting (IPW) was used to account for missing data (58, 59). IPW minimizes potential response bias due to missing data and allows for extrapolation back to the original EAT 2010 sample. All regression analyses and percentages were weighted with the non-response weights while raw sample size values are presented. Among the analytic sample (n=1,555 with non-missing yoga practice responses) the weighted demographics (Table 1) were: 28.7% African American or Black, 19.9% Asian American, 18.9% White, 16.9% Hispanic, 3.7% Native American, and 11.9% mixed or other. Most of the Asian American participants reported Southeast Asian heritage; approximately 79.1% of this group was Hmong.

Table 1.

Yoga practice (at least 30 minutes/week) by sociodemographic characteristics

			Unadjusted Results		Adjusted Results2
	Total N (%)1	Number practicing yoga	% practicing yoga	p-value	% practicing yoga	p-value
Age categories
18-21	708 (42.9)	85	11.4		9.2
22-25	794 (53.0)	116	13.5	0.379	12.1	0.109
26+	53 (4.1)	9	15.5		16.2
Gender
Female	901 (53.2)	158	17.2	<.001	16.4	<.001
Male	643 (46.2)	48	7.2		6.5
Other identity	11 (0.6)	4	36.4		27.8
Ethnicity/Race3
White	365 (18.9)	56	14.6	0.043	11.4	0.237
Black	340 (28.7)	42	11.1		9.9
Hispanic	271(16.9)	36	12.4		12.2
Asian American	353 (19.9)	37	10.6		9.5
Native American	62 (3.7)	4	7.0		7.3
Other/Mixed	164 (11.9)	35	19.4		16.1
Socio-economic status
Low	558 (39.3)	64	10.7	0.024	8.6	0.023
Middle	586 (40.0)	78	13.0		11.9
High	374 (20.7)	65	16.9		14.7

¹All percentages are weighted for non-response to be representative of the original EAT-2010 sample.

²Adjusted for age, gender, ethnicity/race, and socio-economic status (SES) and Body Mass Index (BMI) (n=1483; 6.4% dropped due to missing SES or BMI).

³Structurally racialized categories labelled by ethnicity/race.

RESULTS

Yoga practice by sociodemographic characteristics

Practicing yoga (at least 30 minutes/week) was reported by 12.7% (n=210) of the participants. Among these 210 yoga practitioners, the average frequency of practice: 30 minutes to less than one hour/week (n=68; 32.1%); one hour to less than two hours/week (n=84; 40.3%); two to three hours/week (n=39; 18.5%); and ≥ four hours/week (n=19; 8.9%).

Yoga practice did not differ across age in unadjusted analyses nor in analyses adjusted for sociodemographic characteristics and BMI (Table 1). Female young adults were much more likely to practice than male young adults (16.4% vs. 6.5% in adjusted models). There was variation across ethnicity/race in unadjusted analyses, although differences were not statistically significant after adjustment for sociodemographic characteristics and BMI. In both unadjusted and adjusted analyses, yoga practice differed across SES; in adjusted models, yoga practice was reported by 14.7% of young adults in the highest level and 8.6% in the lowest level.

Body satisfaction, mindful eating, BMI, disordered eating, and muscle-enhancing behaviors by yoga practice

As shown in Table 2, differences in body satisfaction across yoga practice were not statistically significant in unadjusted models (yoga practitioners: M=43.9, non-yoga practitioners: M=42.2; d=0.14; p=.070) or in models adjusted for sociodemographic characteristics and BMI (yoga practitioners: M=42.9, non-yoga practitioners: M=42.2; d= 0.06; p=.485).

Table 2.

Body satisfaction, mindful eating, BMI, disordered eating, and muscle-enhancing behaviors by yoga practice (n=1555)

	Unadjusted results				Adjusted results1
	Yoga Practice		ES2	p-value	Yoga practice		ES2	p-value
	No (n=1345) Unadjusted Mean (SE) or % (n)	Yes (n=210) Unadjusted Mean (SE) or % (n)			No (n=1285) Adjusted Mean (SE) or % (n)	Yes (n=198) Adjusted Mean (SE) or % (n)
Body satisfaction (range: 13-65)	42.2 (0.36)	43.9 (0.89)	0.14	0.070	42.2 (.32)	42.9 (.96)	0.06	0.485
Mindful eating (range: 4-16)	11.6 (0.06)	12.2 (0.13)	0.29	<0.001	11.7 (.06)	12.1 (.14)	0.19	0.004
BMI (kg/m2)	27.4 (0.20)	25.5 (0.44)	−0.27	<0.001	27.4 (.20)	25.6 (.45)	−0.26	<.001
Extreme unhealthy weight control behaviors	12.8% (177)	12.4% (25)	−0.01	0.902	11.0% (172)	10.4% (24)	−0.02	0.804
Less extreme unhealthy weight control behaviors	50.5% (671)	56.8% (118)	0.13	0.100	51.8% (650)	62.3% (112)	0.21	0.011
Binge eating	11.5% (159)	12.6% (28)	0.03	0.629	9.9% (151)	10.7% (27)	0.03	0.704
Steroid use	1.3% (14)	5.2% (10)	0.23	<.001	0.7% (11)	3.8% (9)	0.22	<.001
Protein powder/shakes	26.1% (338)	33.1% (66)	0.15	0.044	25.3% (331)	35.1% (63)	0.21	0.010

¹Adjusted for age, gender, ethnicity/race, socio-economic status, and BMI, except when examining associations with BMI; then estimates are only adjusted for sociodemographic characteristics.

²ES (effect size) provides a comparable measure of magnitude across outcomes. It is calculated as the standardized mean difference Cohen’s d (for continuous outcomes) and a scaled difference in proportions Cohen’s h (for dichotomous outcomes).

Young adults practicing yoga had higher levels of mindful eating in unadjusted analyses d=.29, p<.001), and this difference remained statistically significant after adjustment for sociodemographic characteristics and BMI (d=.19, p=.004).

Yoga practitioners had lower BMI values than non-yoga practitioners even after adjustment for sociodemographic characteristics (25.6 and 27.4 kg/m², respectively; d=−0.26, p<.001). Because of the strong association between BMI and body satisfaction, and the non-significant finding for the association between yoga and body satisfaction, we further tested for an interaction between BMI and yoga on body satisfaction. The interaction between BMI and yoga was significant for predicting body satisfaction (p=0.044) indicating those with a higher BMI (30.0 kg/m²) trended toward having higher body satisfaction if they practiced yoga compared to not practicing yoga (41.9 vs. 40.1, d=0.14, p=0.076), while those with a lower BMI (BMI=18.5 kg/m2) showed no difference across yoga practice (e.g. d=−0.09, p=0.456).

Regardless of yoga practice, high percentages of young adults engaged in extreme weight control behaviors, less extreme weight control behaviors, and binge eating (Table 2). In unadjusted analysis, the use of these disordered eating behaviors was similar across young adults practicing and not practicing yoga. However, after adjustment, yoga practitioners were more likely to engage in less extreme unhealthy weight control behaviors than those not practicing yoga (adjusted proportions: 62.3% vs. 51.8%, h=0.21, p=.011).

The use of muscle-enhancing behaviors was higher among yoga practitioners than non-yoga practitioners (Table 2) in both unadjusted models and in models adjusted for sociodemographic characteristics and BMI, although effect sizes were small. In adjusted models, 3.8% of yoga practitioners used steroids as compared to 0.7% of non-yoga practitioners (h=0.22, p<.001) while 35.1% of yoga practitioners used protein powder or shakes as compared to 25.3% of non-yoga practitioners (h=0.21, p=.010). Further analyses indicated that yoga practitioners were still more likely to use muscle enhancers even after further adjustment for physical activity levels (steroids: h=0.24, p<0.001, protein powder: h=0.17, p=.046).

Dose-response relationships with frequency of yoga practice

Restricting to those practicing yoga, associations were further examined to assess if there was a dose-response relationship between frequency of practicing yoga and the various outcomes. No statistically significant dose-response relationships were found in adjusted analyses, except for steroid use (p=.049) indicating a higher prevalence of steroid use among those practicing yoga more often. Of note, BMI decreased across varying levels of yoga frequency from 26.4 units (yoga frequency: 30 minutes to <1 hour/week) down to 24.5 units (yoga frequency: 4+ hours/week), but this trend was not statistically significant (p=0.206).

Interactions with ethnicity/race and gender

Associations between yoga practice and the different outcomes were similar across ethnicity/race with no statistically significant interactions found (Table 3). Small numbers (i.e., cell sizes < 5) precluded testing interactions with ethnicity/race for less prevalent outcomes: binge eating, extreme unhealthy weight control, and steroid use. Associations also did not differ by gender (Table 4). For example, the higher prevalence of steroid use was seen in both female and male yoga practitioners as compared to non-practitioners.

Table 3.

Body image related variables by yoga practice by ethnicity/race (interaction by ethnicity/race)

	Unadjusted results				Adjusted results1
	Yoga Practice
	No (n=1345) Mean (SE) or % (N)	Yes (n=210) Mean (SE) or % (N)	ES	p-value	ES	p-value
Body satisfaction
White	42.9 (0.68)	45.1 (1.74)	0.17	0.257	0.20	0.144
Black	45.8 (0.75)	47.2 (2.13)	0.12	0.520	−0.12	0.560
Hispanic	40.1 (0.79)	40.8 (1.88)	0.06	0.718	0.05	0.772
Asian	37.2 (0.65)	41.7 (2.01)	0.37	0.031	0.28	0.089
Native American	42.0 (1.80)	39.3 (8.62)	−0.22	0.752	−0.52	0.423
Other	43.8 (1.11)	43.4 (1.82)	−0.03	0.863	0.00	0.999
Interaction p-value				0.65		0.51
Mindful eating
White	11.3 (0.12)	12.1 (.28)	0.38	0.008	0.33	0.026
Black	11.5 (0.13)	12.1 (.29)	0.31	0.042	0.25	0.108
Hispanic	12.0 (0.14)	11.9 (.27)	−0.03	0.827	−0.10	0.498
Asian	11.7(0.12)	11.8 (.32)	0.07	0.687	0.02	0.889
Native American	11.8 (0.25)	12.5 (.25)	0.36	0.039	0.23	0.218
Other	12.0 (0.21)	12.9 (.32)	0.47	0.010	0.51	0.007
Interaction p-value				0.17		0.12
BMI
White	25.8 (0.36)	25.4 (1.16)	−0.07	0.699	−0.07	0.661
Black	27.4 (0.43)	24.3 (0.71)	−0.44	0.000	−0.41	0.001
Hispanic	28.6 (0.47)	26.7 (0.84)	−0.27	0.047	−0.24	0.096
Asian	27.6 (0.37)	24.8 (0.97)	−0.41	0.007	−0.40	0.009
Native American	27.4 (0.92)	24.5 (1.45)	−0.41	0.093	−0.50	0.066
Other	27.8 (0.71)	27.1 (1.23)	−0.10	0.640	−0.11	0.605
Interaction p-value				0.42		0.46
Less extreme unhealthy weight control behaviors
White	39.4% (121)	53.4% (30)	0.28	0.056	0.33	0.045
Black	51.2% (149)	44.4% (18)	−0.14	0.427	−0.04	0.801
Hispanic	47.9% (111)	61.1% (22)	0.26	0.151	0.39	0.048
Asian	63.8% (197)	70.9% (25)	0.15	0.398	0.19	0.291
Native American	56.9% (34)	70.4% (3)	0.28	0.620	0.37	0.501
Other	43.9% (59)	59.6% (20)	0.31	0.111	0.27	0.218
Interaction p-value				0.47		0.60
Protein powder
White	28.2% (85)	32.7% (18)	0.10	0.502	0.24	0.101
Black	25.4% (71)	33.8% (13)	0.19	0.273	0.17	0.351
Hispanic	31.8% (73)	26.6% (9)	−0.11	0.544	0.02	0.934
Asian	22.7% (67)	27.9% (10)	0.12	0.503	0.16	0.427
Native American	19.4% (11)	29.6% (1)	0.24	0.647	−0.01	0.979
Other	24.3% (31)	43.4% (15)	0.41	0.031	0.54	0.013
Interaction p-value				0.55		0.61

¹Adjusted for age, gender, ethnicity/race, socio-economic status, and BMI, except when examining associations with BMI; then estimates are only adjusted for sociodemographic characteristics.

Table 4.

Body image related variables by yoga practice by gender (interaction by gender)

	Unadjusted results				Adjusted results1
	Yoga Practice
	No (n=1338) Mean (SE) or % (N)	Yes (n=206) Mean (SE) or % (N)	ES	p-value	ES	p-value
Body satisfaction
Male2	44.3 (0.51)	43.7 (2.03)	−0.04	0.803	−0.16	0.414
Female	40.1 (0.49)	43.9 (0.98)	0.31	<.001	0.14	0.085
Interaction p-value				0.064		0.156
Mindful eating
Male	11.5 (0.10)	12.0 (0.27)	0.24	0.084	0.22	0.133
Female	11.8 (0.08)	12.3 (0.16)	0.24	0.004	0.20	0.016
Interaction p-value				0.989		0.905
BMI
Male	27.0 (0.26)	24.8 (0.64)	−0.32	0.001	−0.24	0.009
Female	27.7 (0.30)	25.8 (0.56)	−0.27	0.003	−0.28	0.007
Interaction p-value				0.747		0.779
Less extreme unhealthy weight control behaviors
Male	45.4% (263)	55.9% (27)	0.21	0.178	0.23	0.159
Female	55.2% (403)	57.4% (89)	0.04	0.620	0.20	0.034
Interaction p-value				0.362		0.891
Extreme unhealthy weight control behaviors
Male	6.5% (37)	10.4% (5)	0.14	0.331	0.14	0.401
Female	19.0% (140)	12.9% (19)	−0.17	0.088	−0.09	0.386
Interaction p-value				0.101		0.264
Binge Eating
Male	5.90% (36)	9.68% (5)	0.14	0.293	.12	0.403
Female	16.92% (122)	12.97% (21)	−0.11	0.231	−.02	0.825
Interaction p-value				0.138		0.397
Protein powder
Male	35.3% (210)	48.1% (23)	0.26	0.090	0.16	0.312
Female	17.1% (127)	28.3% (43)	0.27	0.002	0.24	0.008
Interaction p-value				0.750		0.529
Steroids
Male	1.7% (9)	8.1% (4)	0.31	0.010	0.27	0.015
Female	0.9% (5)	4.2% (6)	0.23	0.008	0.21	0.005
Interaction p-value				0.975		0.947

¹Adjusted for age, gender, ethnicity/race, socio-economic status, and BMI, except when examining associations with BMI; then estimates are only adjusted for sociodemographic characteristics.

²The gender identity, “other”, was not included in the analysis

DISCUSSION

This study examined associations between yoga practice and variables of relevance to the promotion of a healthy body image and the prevention of disordered eating/eating disorders and unhealthy muscle-enhancing behaviors among ethnically/racially diverse, primarily low-income, emerging young adults. Findings suggest that emerging young adults from diverse backgrounds who practice yoga have higher levels of mindful eating, but have equal or elevated levels of problematic body image-related attitudes and behaviors, as compared to non-yoga practitioners. In considering the findings, the small effect sizes should be noted, and any differences should be viewed cautiously. Further analyses examining interactions found that these patterns did not differ significantly across ethnicity/race or gender. Given that prior research has found that yoga may offer many health benefits broadly (25-29) and in terms of body image, disordered eating, and weight-related health (37, 60, 61), it is encouraging to see that yoga is being practiced by those in need of potentially helpful strategies. However, it is also crucial to ensure that the yoga being practiced is beneficial with regard to the promotion of a positive body image and the avoidance of potentially harmful disordered eating and muscle-enhancing behaviors.

To the best of our knowledge, this is the first population-based study that has examined associations between yoga and muscle enhancers. Approximately one-third of young adults practicing yoga used protein powders or shakes, specifically to enhance the size or tone of their muscles. While there is a lack of clarity about what they were consuming, findings suggest a preoccupation with muscle enhancement. The higher use of steroids to enhance muscles among yoga practitioners is of particular concern, given their adverse consequences (62). That said, it should be noted that the percentage of yoga practitioners using steroids was low, as were effect sizes for differences between those practicing and not practicing yoga. In many settings, the practice of yoga has shifted somewhat from the traditional yogic principles and philosophies of Eastern yoga to being more commercialized (63) and focused on fitness and physical strength building, which may feed into young adults thinking of yoga as a way to enhance muscle development. Further research examining what types of muscle enhancers are being used and motivation for their use is needed for greater understanding. It would also be of interest to examine whether higher use of muscle enhancers among yoga practitioners is only found among those engaging in more physically strenuous practices. Further research in different populations is needed to confirm these findings.

Body satisfaction was found to be similar among emerging adults practicing yoga and those not practicing yoga in analyses adjusted for BMI and sociodemographic characteristics. There was a significant interaction between BMI and yoga in predicting body satisfaction with a trend toward a positive impact among yoga practitioners at higher BMI values. Of concern, equal or higher percentages of young adults practicing yoga were found to engage in disordered eating behaviors as compared to those not practicing yoga. It is promising to see that young adults who have body image concerns and engage in disordered eating behaviors are choosing to practice yoga, given the underlying tenets of yoga, in addition to research findings suggesting that yoga may be helpful in terms of promoting a stronger sense of self and positive embodiment (22, 39, 64). Furthermore, some longitudinal and intervention studies have suggested that yoga may be effective in improving body image and disordered eating behaviors (37, 38, 60, 61, 65, 66). Although findings do not consistently show the benefits of yoga (30, 31), we are not aware of research findings showing that yoga can be harmful in terms of worsening these outcomes.

In the current study, yoga was associated with higher levels of mindful eating. In prior qualitative research, young adults practicing yoga discussed how yoga helped with increased awareness of their bodies’ needs, the selection of more nourishing foods, attentiveness to internal signs of hunger or fullness, and greater presence while eating (46). A randomized, controlled trial examining the impact of yoga among adults engaging in binge eating (65) found that yoga was associated with decreased binge eating. Likewise, qualitative interviews with study participants indicated perceived improvements in eating behaviors, including greater presence and mindfulness while eating (67). Further work is needed to determine if yoga can lead to more mindful eating and, in turn can help reduce binge eating behaviors in the general population.

Analyses were adjusted for BMI, in addition to sociodemographic characteristics, given that BMI values were lower among yoga practitioners than non-practitioners. The cross-sectional nature of the study design does not allow for a determination of the directionality of this association. There is some evidence that yoga can help in reducing BMI (68), with two population-based studies suggesting that yoga can help with weight gain prevention (45, 69). However, it is also highly likely that the mean difference in BMI between the yoga practitioners and the non-practitioners is due to young adults with lower BMI values being more likely to choose to engage in yoga. Persons living in larger bodies may not feel comfortable at some settings in which yoga is taught. For example, in one of these population-based studies, among young adults practicing yoga, those with higher BMI values were less likely to practice at yoga studios than yoga practitioners with lower BMI values (45). Yoga settings need to be proactive in helping persons from diverse backgrounds, and of diverse body shapes, sizes, and abilities, to feel more welcome and comfortable (39, 70-73).

Study strengths and limitations should be considered in interpreting these findings. An important study strength is the large size and diverse nature of the sample. There has been a dearth of research on yoga among low-income and ethnically/racially diverse populations; thus, this study fills an important gap in the literature. The assessment of a broad array of attitudes and behaviors related to body image is also a study strength. We are unaware of any studies that have examined yoga in relation to muscle enhancing behaviors in any population-based studies. Young people from socially marginalized ethnic/racial and low socio-economic backgrounds have been found to be at high risk for problematic eating and weight-related outcomes (2, 12-16), but may be underserved with regard to prevention and treatment interventions (74, 75). It is crucial to learn more about how to help in promoting a healthy body image and prevent and reduce unhealthy muscle-enhancing and disordered eating behaviors among diverse and potentially underserved populations.

While the diverse nature of the study population is a strength, it is important to note that in order to properly examine patterns of relatively rare behaviors (e.g., steroid use) by yoga practice, within separate ethnic/racial groups, larger numbers will be needed. Likewise, while utilizing a community-based sample allows for a determination of who is practicing yoga, the mean frequency of practicing yoga among such a broad sample tends to be low, with relatively few respondents reporting practicing an average of more than two hours a week. While we conducted a dose-response analysis, and did not see trends suggesting that those practicing more often were better off, the numbers of young adults practicing frequently was low, leading us to interpret the findings cautiously. Additionally, the cross-sectional nature of the study needs to be considered in interpreting findings; inferences about directionality of associations or causation cannot be inferred. This study allowed for the examination of who is practicing yoga in terms of body image attitudes and behaviors, rather than the impact of yoga on these variables. Further longitudinal work is needed to detect changes over time following a consistent period of yoga practice. Finally, many aspects of yoga practice were not assessed in the current study. Thus, it will be important to replicate these findings in other study populations and to examine differences in associations by characteristics of the yoga practice (e.g., intensity, type), focus of the teaching (e.g., increased emphasis on physicality, emotion regulation, awareness and acceptance of self, or spirituality), and locations of practice (e.g., home vs. studio).

CONCLUSION

In summary, among emerging adults from diverse ethnic/racial backgrounds, disordered eating and muscle-enhancing behaviors were equally or more common among yoga practitioners than non-practitioners. In order to promote yoga among diverse populations, and enhance its safety and benefits, it is crucial for settings where yoga is offered, and yoga teachers, to have awareness of the high prevalence of these concerning behaviors among students and to address them accordingly. Certainly, these settings should not be promoting products aimed at muscle enhancement such as protein powders and shakes. For those practicing yoga on their own, it may also be helpful to learn how to focus more on the internal sensations that occur during one’s yoga practice to enhance the potential for yoga’s positive impact on overall well-being. Increased awareness to self-critique and replacing potentially harmful self-dialogues with increased positive self-talk and greater self-compassion may also enhance the positive impact of one’s yoga practice. Further research exploring yoga practice among diverse populations is needed, including studies that include more details on muscle-enhancing products being used and reasons for using these products; longitudinal studies to examine changes in body image attitudes and behaviors following a consistent yoga practice; qualitative research to better understand yoga practitioners’ perceptions of how the practice of yoga could be enhanced to improve body image and related behaviors; and intervention studies to examine the impact of different teaching and practice styles.