Abstract
This study examined the association of weight perception and weight satisfaction with body change intentions and weight-related behaviours in 928 overweight adolescents (aged 11–18 years, 44% female). Accurate perception of weight was associated with trying to lose weight and inaccurate perception was associated with trying to gain muscle. Weight dissatisfaction was associated with trying to lose weight and gain muscle. Accurate weight perception and weight dissatisfaction were not associated with healthy weight-related behaviours. Awareness of overweight and body dissatisfaction may be detrimental to the adoption of healthy weight-control behaviours. Interventions with overweight adolescents should encourage body satisfaction, rather than promoting awareness of overweight.
It is estimated that between 21 and 25 per cent of Australians aged 2–18 years are overweight (Olds et al., 2010). Obesity tends to track into adulthood; therefore, overweight adolescents are at increased risk of developing type 2 diabetes, heart disease and certain cancers in adulthood (Deshmukh-Taskar et al., 2005). It is of concern then that many adolescents do not recognize that they are overweight (Edwards et al., 2010; Martin et al., 2010; Wang et al., 2009), as this lack of awareness may have implications for weight-related behaviours such as diet and physical activity. Adolescents who do not recognize that they are overweight may not be motivated to make changes to their lifestyle to prevent further weight gain.
Overweight adolescents who are aware of their weight status (Edwards et al., 2010; Khambalia et al., 2012; Lenhart et al., 2011) or dissatisfied with their body (Wang et al., 2009) are more likely to report that they are trying to lose weight. This suggests that perception of overweight and body dissatisfaction engender a desire to alter weight or shape. Adolescents’ intentions to change their weight and shape are also likely to be influenced by gender-specific body ideals. For girls, the ideal body tends to be thin (Keel et al., 2007); thus, for girls who are aware of their overweight status, their ideal body is much slimmer than their actual body. Accordingly, overweight girls are most likely to be dissatisfied with their body (Wang et al., 2009). In addition, girls who are dissatisfied with their body are typically trying to lose weight (Neumark-Sztainer et al., 2006).
Overweight boys may not perceive a discrepancy between their ideal body and their actual body. The male ideal tends to be lean and muscular (Gray and Ginsberg, 2007); yet there is also evidence to suggest that adolescent boys feel inadequate if they are not physically large (Lubans and Cliff, 2011). Thus, perception of overweight may equate to an unfavourable perception of having excess fat tissue or it may be consistent with a favourable perception of being ‘large’ and muscular. Weight perception research has tended to focus on weight-loss intentions; however, the drive to develop muscularity is a chief component of male body image (Ricciardelli and McCabe, 2004); therefore, it is important to understand how perception of weight relates to muscle intentions. While previous research suggests that boys who are dissatisfied with their body will be trying to gain muscle (McVey et al., 2005), it is unknown how perception of overweight is related to intention to gain muscle in adolescents. It is possible that overweight boys who underestimate their weight status may be of the view that they are not large enough and consequently, will be trying to gain muscle mass. Thus, engagement in certain weight-related behaviours that build muscle may have more to do with desire to gain muscle, than to lose weight.
An understanding of the relationship between perception of weight status and intentions to gain muscle or lose weight may help to understand the dietary behaviour and physical activity that overweight adolescents engage in. Currently, it is unclear whether accurate perception of overweight has an impact on actual weight-related behaviours, such as physical activity and diet. Adolescents of both genders who are aware of their overweight status may be more likely to engage in physical activity to lose weight (Edwards et al., 2010; Khambalia et al., 2012; Lenhart et al., 2011). However, perception of overweight has also been found to negatively predict engagement in exercise in both healthy weight and overweight adolescents (Gillison et al., 2006). The evidence for the impact of accurate weight perception on dietary behaviour in overweight adolescents is also mixed. Accurately perceiving overweight is associated with dieting behaviour and lowered calorie intake (Edwards et al., 2010), but also with poorer dietary choices (Khambalia et al., 2012). Weight-related behaviours may differ by gender; studies have found that among overweight adolescents who accurately perceive their weight status, girls are likely than boys to diet in order to lose weight (Kurdak et al., 2010; Lenhart et al., 2011).
The equivocal evidence for the impact of weight perception on behaviour may be related to other factors, including body satisfaction and gender. As noted above, dissatisfaction with weight is associated with trying to lose weight and some researchers have suggested that dissatisfaction with weight among overweight adolescents may actually motivate individuals to make healthy lifestyle changes (Heinberg et al., 2001). However, other research suggests that body dissatisfaction is a predictor of reduced fruit and vegetable intake and lowered levels of physical activity (Neumark-Sztainer et al., 2006; Skemp-Arlt, 2006). Very few studies have taken into account body dissatisfaction when examining the association of weight perception with weight-related intentions and behaviour. The relationship of perception of overweight with an individual’s intentions and behaviour may also depend on the individual’s gender. There is evidence to suggest that the body change strategies that adolescents employ may differ by gender. Boys may be more likely to engage in physical activity (Yost et al., 2010), whereas girls restrict their food intake (Forman-Hoffman, 2004). These differences may exist because girls simply want to lose weight, whereas boys want to engage in activities that will build muscle.
The aim of this study was to examine the relationship of accurate weight perception, gender and weight satisfaction with body change intentions and weight-related behaviours among overweight and obese adolescents. It was hypothesized that intention to lose weight would be more common among girls and would be associated with accurate perception of overweight and dissatisfaction with weight. It was hypothesized that overweight boys who underestimated their weight status would be trying to gain muscle. Furthermore, it was hypothesized that behaviours aimed at reducing weight, such as reduced intake of unhealthy food and increased intake of fruit and vegetables and increased physical activity, would be more prevalent among overweight adolescents who accurately perceived their weight status and were dissatisfied with their weight.
Method
Participants
Participants were drawn from a quasi-experimental, longitudinal intervention study called ‘It’s Your Move!’, which was part of the larger Pacific Obesity Prevention in Communities (OPIC) study. The original sample consisted of 3040 adolescents aged between 11 and 18 years who were attending secondary schools in the Barwon-South Western region of Victoria, Australia. The area was selected because it had sufficient number of youth, sufficient settings to conduct interventions and reasonable proximity to the research team. For this study, a subsample of overweight and obese (N = 928) adolescents was examined. The analysed sample was ethnically homogenous; 93.4 per cent of participants were European Australian, 1.9 per cent were Indigenous Australian and the remaining group identified with a range of other ethnicities.
Measures
Weight and height were measured by trained researchers using standardized methods (Swinburn et al., 2011). Body mass index (BMI; weight (kg)/height (m2)) was used to define weight status. BMI was converted to standardized BMI (BMI-z) and adolescents’ weight status was defined as overweight/obese using World Health Organization cut-offs (De Onis et al., 2007).
Weight perception, weight satisfaction and demographic information were derived from the Adolescent Behaviour, Attitudes and Knowledge Questionnaire (ABAKQ; Mathews et al., 2009). The ABAKQ consisted of 87 questions and was developed for the OPIC study and was pilot tested for acceptability, appropriateness and feasibility of completion prior to implementation. The question about weight satisfaction was phrased, ‘How happy or unhappy are you with your body weight?’ Responses were measured on a scale from 1 (very unhappy) to 5 (very happy). Intention to change weight and change muscle was assessed by the question ‘Which of these statements most closely applies to you?’ For intention to change weight, the responses ranged from ‘trying to gain weight’, ‘trying to stay at my current weight’, ‘not doing anything about my weight’ and ‘trying to lose weight’. For intention to change muscle, the response options were as follows: ‘trying to gain muscle’, ‘trying to stay at the same muscle size’ and ‘not doing anything about my muscles’. Both of these variables were dummy coded so that responses were either equal to 1 (intention to lose weight/gain muscle) or 0 (no intention to lose weight or gain muscle; which was all other responses for that question combined).
Activity after school was assessed via the question, ‘In the last 5 school days, on how many days after school did you do sports, dance, cultural performances or play games in which you were active?’ Response options ranged from 0 to 5 days. Current activity guidelines recommend that adolescents should engage in at least 60 minutes of moderate to vigorous physical activity each day (Australian Department of Health, 2004). Thus, this variable was dichotomized into 0 (3 days or less) and 1 (4 days or more) to capture those adolescents who were most likely to be meeting the requirements while still retaining enough responses in each category to conduct analysis. For takeaway food, the question was ‘In the last 5 school days, on how many days did you buy snack food from a shop or takeaway after school?’ Available responses ranged from 0 to 5 days. This variable was dichotomized into 0 (bought takeaway food after school on one or more days) versus 1 (not bought takeaway food after school) to reflect dietary guidelines that recommend limiting foods that are high in saturated fat or have added salt or sugar (NHMRC, 2013). To assess intake of fruit, participants were asked, ‘How many serves of fruit do you usually eat each day?’ The question assessing vegetable intake was phrased the same way. The responses were 1 serve or less, 2 to 3 serves and 4 serves or more. These variables were dichotomized to reflect recommended daily intake of fruits and vegetables for adolescents, which is two serves and five serves, respectively (NHMRC, 2013). Responses were split into 1 serve or less (0) versus 2 serves or more (1) for fruit. For vegetables, the existing categories did not match exactly with daily recommended intake and were dichotomized at the point which was closest to the recommendations: 3 serves or less (0) versus 4 serves or more (1). All variable were coded so that unhealthy behaviours were coded as 0 and the more healthy behaviours were coded as 1.
To determine accuracy of weight perception, participants were also asked how they would describe their weight using a 5-point scale where 1 = very underweight, 2 = slightly underweight, 3 = about the right weight, 4 = slightly overweight and 5 = very overweight. Participants who reported they were slightly overweight or very overweight were classified as accurate; all other participants were classified as underestimating.
Procedure
The ‘It’s Your Move!’ study was designed to prevent unhealthy weight gain in adolescents in 12 schools in the Barwon-South Western region of Victoria, Australia. The methods used have been reported previously (Millar et al., 2011; Swinburn et al., 2011). Briefly, five secondary schools participated in the intervention and the comparison sample was drawn from seven schools. The study was conducted between 2005 and 2008; students completed the ABAKQ and had their anthropometry measured at baseline and at follow-up. Baseline data from both the intervention and comparison groups were used in the cross-sectional analysis. Ethical approval for this study was provided by the University’s Human Research Ethics Committee.
Analysis
Of the original subsample of 928 overweight adolescents, 11 were excluded from the analysis because of missing data on weight satisfaction. A chi-square test was used to test for differences between boys and girls for accuracy of perception, and a t-test was used to examine gender differences for weight satisfaction. Two separate bivariate logistic regression models, adjusted for age, were conducted to determine the relationship between weight misperception and body change intentions and behaviour. The data were tested to ensure that the requirements for logistic regression were met. Specifically, the continuous variables, BMI-z, weight satisfaction and age, were found to be linearly related to the log of the outcome variables used in the logistic regression models. In addition, collinearity diagnostics revealed that all tolerance values were above 0.2, indicating that there was no substantial multicollinearity between the variables. Model 1 examined the association of weight perception, BMI-z, weight satisfaction and gender with intention to lose weight, intention to gain muscle, physical activity after school, takeaway eaten in the past week and fruit and vegetable intake. Model 2 included the additional effect for the interaction of weight perception with gender. The outcomes of logistic regression models were reported as odds ratios and 95 per cent confidence intervals. All analyses were performed using SPSS (V21) and statistical significance was accepted as p < 0.05.
Results
Characteristics of the study sample
Characteristics of the adolescents in this study are presented in Table 1. The adolescents’ BMI ranged from 20.01 to 42.82. Just over half of all overweight adolescents correctly identified their weight status. A non-significant trend suggested that boys were more likely to underestimate their weight status than girls, χ2(1, N = 917) = 3.28, p = 0.07. Boys were more satisfied with their weight than girls, t(915) = 8.84, p < 0.01.
Participant characteristics: full overweight subsample and gender subgroups.
SD: standard deviation; BMI: body mass index.
Weight satisfaction was measured on a 5-point scale from 1 (very unhappy) to 5 (very happy).
Correlates of body change intentions
The results of the logistic regression models for body change intentions are presented in Table 2. The factors that were associated with trying to lose weight were consistent across both model 1 and model 2. Adolescents who were aware of their overweight status were more than twice as likely to be trying to lose weight compared to those who underestimated. In addition, adolescents with higher BMI-z scores were significantly more likely to be trying to lose weight. As BMI-z increased, the odds that the adolescents would be trying to lose weight became higher. Lower weight satisfaction was also associated with increased likelihood of trying to lose weight and boys were half as likely to report that they were trying to lose weight compared with girls. The interaction of weight perception with gender was not significant, indicating that adolescents of both genders were likely to be trying to lose weight if they perceived themselves to be overweight.
Association of weight perception, weight satisfaction, weight status and gender with body change intentions for subsample of overweight adolescents (n = 917).
OR: odds ratio; CI: confidence interval; BMI-z: standardized BMI.
Regression models also adjusted for age.
Weight satisfaction was measured on a 5-point scale from 1 ‘very unhappy’ to 5 ‘very happy’.
p < 0.05; **p < 0.01.
In both model 1 and 2, participants who were satisfied with their weight were less likely to be trying to gain muscle, with each unit increase in satisfaction; the odds that adolescents would be trying to gain muscle were lowered. There was no association with BMI-z and trying to gain muscle. In model 1, boys were 8 times more likely to report that they were trying to gain muscle. Trying to gain muscle was significantly associated with inaccurate weight perception in model 1; adolescents who misperceived their weight were more likely to report trying to gain muscle. However, in model 2, when the interaction of gender and weight perception was added to the model, the association between underestimation and trying to gain muscle was no longer statistically significant. This indicates that the association of trying to gain muscle with underestimation was relatively weak. In addition, the interaction of weight perception with gender was not significant, indicating that there was a weak association between trying to gain muscle and perception of overweight in both genders.
Correlates of weight-related behaviour
The results of the logistic regression models for weight-related behaviours are presented in Table 3. In model 1, only boys were significantly more likely to be active after school compared to girls. In both model 1 and 2, underestimation of weight was also associated with higher levels of physical activity; adolescents who underestimated their weight were more than 1.5 times more likely to be active after school at least 4 days of the week. There was no association between BMI-z and level of activity in either model.
Association of weight perception, weight satisfaction, weight status and gender with weight-related behaviours for subsample of overweight adolescents (n = 917).
OR: odds ratio; CI: confidence interval; BMI-z: standardized BMI.
Regression models also adjusted for age.
Weight satisfaction was measured on a 5-point scale from 1 (very unhappy) to 5 (very happy).
p < 0.05.
In both models 1 and 2, adolescents who were satisfied with their weight were less likely to have eaten takeaway in the past week. Weight perception, BMI-z and gender were not associated with having eaten takeaway in the past week. In both models, adolescents who were satisfied with their weight were more likely to report eating the recommended number of daily serves of fruit per day. In model 1, only girls were 1.4 times more likely to meet the recommended daily intake of fruit per day compared to boys. In model 2, there was no association with gender, weight perception or BMI-z. Eating the recommended daily serves of vegetables was not associated with weight perception or any other factors in the models.
Discussion
This study examined the association of weight perception, gender and weight satisfaction with body change intentions and weight-related behaviour. The main findings supported the prediction that accurate perception of being overweight and dissatisfaction with weight would be associated with trying to lose weight. The findings, however, did not support the prediction that accurate perception of overweight would be associated with healthy dietary intake and physical activity. In addition, satisfaction with weight was associated with healthy dietary intake.
Intention to lose weight was associated with accurate perception of overweight and dissatisfaction with weight. This suggests that both awareness of overweight and dissatisfaction with weight status prompted a desire to lose weight. Adolescents with a higher BMI-z were the most likely to report trying to lose weight, perhaps because their excess weight was more obvious. In addition, girls were more likely than boys to be trying to lose weight and were more dissatisfied with their weight. This suggests that girls were influenced by the thin ideal body, and the discrepancy between an idealized thin body and their actual body may have provoked dissatisfaction and a desire to lose weight. In addition, the prevalence of trying to lose weight was 10 per cent higher than the prevalence of accurate perception among girls. This suggests that some girls who underestimated their weight also wanted to lose weight. This is consistent with research that indicates girls want to lose weight despite having a perception of healthy weight (Duncan et al., 2011) and reflects a desire to obtain an ideal body that is underweight rather than healthy.
Boys were more likely than girls to report they were trying to gain muscle, and there was an association of underestimation of weight with muscle gain intentions. Previous research suggests that boys with smaller bodies have less favourable self-perceptions (Lubans and Cliff, 2011), so having a perception of a smaller body might engender a desire to increase their size through gaining muscle. However, the relationship between misperception and intention to gain muscle was weak and did not occur in boys only. Overall, weight perception did not account for the gender differences in terms of intentions and behaviours. Furthermore, both boys and girls who perceived themselves to be overweight were more likely to want to lose weight. This suggests some similarities between the body shapes that boys and girls are aiming to achieve. Overall, the data are consistent with boys trying to achieve a body ideal that is lean and muscular rather than ‘large’.
In this study, awareness of overweight did not inspire healthy dietary choices nor did it inspire poor dietary choices. However, weight dissatisfaction did have a negative impact on dietary choices. Dissatisfaction with weight may encourage overweight adolescents to restrict their intake of healthy foods and increase their intake of unhealthy foods. This pattern of eating is consistent with disordered eating patterns seen in bulimia or binge eating disorder, where body dissatisfaction leads an individual to restrict their intake of food throughout the day. In turn, the hunger caused by restriction leads them to eat energy dense foods later in the day (Murphy et al., 2010). In this study, adolescents who were dissatisfied with their weight may have bought takeaway for themselves at the end of the day in response to food restriction throughout the rest of the day. Body dissatisfaction has previously been associated with reduced consumption of fruit and vegetables (Neumark-Sztainer et al., 2006) and perception of overweight has been associated with unhealthy weight-control behaviour such as skipping meals (Felts et al., 1996), eating fewer vegetables (Nystrom et al., 2005) or restricting food and fasting (Al Sabbah et al., 2010). While in this study, weight perception had no direct impact on dietary intake; adolescents who were dissatisfied with their weight were more likely to make poor dietary choices.
Adolescents who underestimated their weight were more active than adolescents who were aware of their overweight status. This indicates that accurate perception of overweight may not be a motivating factor to maintain engagement in exercise. In a study by Gillison et al. (2006), perception of overweight was associated with perceived pressure to lose weight, exercise being an extrinsic goal and reduced engagement in exercise. In this study, participants who recognized that they were overweight may have answered in the affirmative when asked if they wanted to lose weight because they knew that was expected of them, but not because weight loss was an intrinsic goal for them. Therefore, their perception may have influenced their intentions but not their behaviour. Underestimating overweight seemed to be a protective factor in terms of encouraging engagement in exercise in this sample. The participants who underestimated their weight may have felt less external pressure to lose weight and so they engaged in physical activity for reasons other than weight loss.
Another possible explanation for the higher levels of physical activity among underestimating adolescents could be linked to the definition of overweight. Some participants may have had a high level of lean mass relative to their height which could cause them to be misclassified as overweight. These individuals would be more likely to both report their weight as healthy and have a high level of participation in physical activity. However, the pattern of association with BMI-z does not support this explanation. BMI-z was associated with intention to lose weight, but it was not associated with physical activity. If the relationship of underestimation to activity was due to body composition, it would be expected that BMI-z would be negatively correlated with activity, that is, the lower the BMI-z, the more likely they would be to engage in physical activity. However, as this relationship was not present, it suggests that underestimation itself is a factor in higher activity independent of body composition.
Overall, the findings demonstrate that factors that were associated with intention to lose weight often did not correlate or correlated negatively with positive health behaviours. There was evidently some disconnect between what adolescents reported they were trying to do compared to their actual behaviours. This may reflect the difficulty of translating intentions to behaviour or it may signify the pernicious influence of perception of overweight itself. Accurate perception of overweight and dissatisfaction with weight were actually related to unhealthy weight behaviours. Furthermore, when examining the overall potential benefits of accurate perception of overweight versus the potential harms, it is worth considering that adolescents who accurately perceive overweight are more likely to report bullying and suicide attempts (Lenhart et al., 2011) compared to those who underestimate. In addition, perception of overweight predicts the development of depressive symptoms in adolescents (Rawana, 2013). Encouraging adolescents to recognize their overweight status may cause greater harm than good.
Strengths of this study include the use of a large sample of overweight adolescents, consideration of the factors that influence intention to gain muscle and examination of weight dissatisfaction in addition to weight perception. There were several limitations to this study, one being that dietary and exercise behaviours were self-reported. Lowry et al. (2002) suggested that self-report of fruit and vegetable intake is acceptable as long as it is systematically underestimated. However, in this sample, there may be a difference in the reporting between adolescents who were accurate and those who underestimated their weight. It is possible that the individuals who favourably self-reported their weight status may also over-report their engagement in healthy dietary and exercise behaviours. This may account for the correlation between underestimation of weight and higher engagement in exercise. In addition, the main outcome variables for weight-related behaviour were dichotomous and did not always correspond exactly with the government recommendations for weight-related behaviour. This resulted in dichotomous measures that were not very sensitive or robust. In addition, the physical activity variable was based on a single-item measure and so it does not give a comprehensive account of adolescents’ physical activity patterns.
In this study, no distinction was made between overweight and obese adolescents because the measure of weight perception was not a standardized measure. Therefore, a perception of ‘very overweight’ could not be confidently described as being equivalent to obese BMI. As a result, some of the details of the magnitude of an individual’s misperception were lost. Excess fat was described in terms of overweight only, so obese individuals who reported that they were overweight were classified as accurate, when arguably they may have underestimated their weight. Another limitation of examining the behaviour of adolescents is that they are often subject to some degree to school and parental control. For example, not all adolescents are given opportunities to be active after school or have fruit and vegetables available at home. These environmental factors are likely to have a large impact on their engagement in these behaviours, regardless of intentions. This may explain why some of the factors that correlate with intentions did not correlate with behaviour. The impact of weight perception on behaviour may change over time; however, as this was a cross-sectional study, it was not possible to examine this possibility. It is possible that some relationships were missed because they take some time to develop. Future studies should use a longitudinal design to enable inferences to be made on the direction of the relationship between weight perception, intentions and behaviour.
Perception of overweight itself may not be sufficient to affect healthy lifestyle changes in overweight adolescents. Despite its impact on intentions, accurate perception of weight does not seem to have the desired effect on behaviour. This suggests that making adolescents aware of their overweight status is not sufficient to engender behaviour change. In addition, there is potential for harm if adolescents feel dissatisfied with their bodies in terms of adopting maladaptive dieting practices. Overall, it appears that making overweight adolescents aware of their weight status is not a good strategy to promote the adoption of healthy behaviours. Health practitioners should emphasize health and body satisfaction when promoting lifestyle change rather than emphasizing weight or appearance. Public health initiatives that can alter an adolescent’s environment to promote healthy eating and exercise may be more useful in encouraging appropriate behaviour change as opposed to those that promote awareness of overweight.
Footnotes
Funding
The study was funded by the Victorian Department of Health and the National Health and Medical Research Council.