This study showed social inequality in overweight and obesity, especially among boys < 13 years, but not in regard to perception of overweight, which was commonly found among girls ≥ 13 years. These age- and gender-specific findings might contribute to a better understanding of the developmental pathways leading to weight-related disorders.
Age and gender-specific social inequality in overweight/obesity
The higher prevalence of overweight and obesity found among younger children of low-SES parents, especially among those with low-SES mothers, could be explained by lifestyle characteristics in addition to strong hereditary predisposition . In families and especially among mothers with lower SES, less emphasis on, as well as limited access to, healthy eating habits and physical activity may be important risk factors for developing overweight and obesity [5, 9, 22, 23]. In addition, since lower SES has been found to be associated with higher prevalence of overweight among adults , parental overweight/obesity may affect perception of their children’s weight and motivation for their children to avoid further weight gain [25, 26]. It has been suggested that overweight mothers in particular have greater difficulty assessing weight in younger children when using themselves as a reference point [27, 28]. The greatest maternal misperceptions of weight status among children were observed among those children categorized as overweight/obese . One study shows that parents believe that being “chubby” at a young age is “cute” and that overweight in childhood is not a problem prior to adolescence . Moreover, research suggests that some physicians may lack the confidence to identify childhood overweight and do not always inform mothers that their children’s weight is of concern .
In addition, younger children spend more time with their parents, especially their mothers. Parents shape social norms and serve as models for behavior, likely reflecting the values of their social class . However, the type and extent of parental influence on the child will vary in accordance with the perceived maturity of the child . Parental influence decreases, while peer influence increases as children grow older, for which reason parents are likely to have a stronger influence on health behaviors in younger children . It is well known that girls mature earlier than boys, suggesting that parents have less influence on girls than on boys of similar age. The latter might explain why we only found social inequality to be an influential factor in younger boys, but not in younger girls.
Among boys ≥ 13 years, we only found a significantly higher BMI and WC when their fathers had a low level of education. Lower SES is associated with an increase in prevalence of overweight and obesity among adults in developed countries , suggesting that fathers with a low level of education in this study are at increased risk of being overweight themselves due to the social norms and behaviors of their social class. In addition to hereditary factors, adolescent boys may be at increased risk for overweight and obesity by adopting their fathers’ norms and lifestyle when they serve as role models . This in turn could presumably lead to increased social inequality as reflected by overweight that continues into adulthood.
No social inequality, but age and gender differences found in perception of overweight
From the perspective of obesity prevention, one positive finding is that perception of overweight among boys and younger girls was associated with actual overweight according to both ISO-BMI and WC, and not with paternal or maternal SES. This finding stands in contrast to those of Park, who reported underestimation of overweight among boys, especially from low-income households . Misconception 2, in which overweight is underestimated, was not found in our study. In the WHO HBSC study , no significant relationship between perception of body weight status and family SES was seen in most participating countries, with the exception of a few countries and regions, especially in Western Europe and North America, where the perception of overweight/obesity was found to be related to low family SES . Our results show a realistic perception of overweight even among children and adolescents from low-SES backgrounds that had a higher prevalence of overweight, thereby suggesting that the strongest predictor for body dissatisfaction is overweight prevalence, at least among boys and younger girls .
Meanwhile, boys going through puberty have a natural potential to become more muscular and develop broader shoulders after their growth spurt, which both contribute to the positive ideal of the male body. Boys ≥ 13 years may therefore develop a more positive self-image with age and pubertal development [35, 36], which could explain the lack of misconception 1 among boys.
At the same time, the increase in body fat that naturally accompanies puberty in girls conflicts with stereotypes of the ideal female body. Girls in puberty have therefore been shown to develop a negative perception of their body , confirmed strongly in this study by the high prevalence of misconception 1. The WHO HBSC study  also found that in many countries 15-year old girls were significantly more likely to report that they were too fat than 11-year olds. This age difference in perception of overweight was not found among boys neither in this study nor in the HBSC study . Compared with their mothers, adolescent girls have a less realistic and more negative perception of their body weight status . However, the female desire to be thin shared by both mothers and girls [39, 40] may have a negative impact on perception of the body among girls.
Meanwhile, a comparison of prevalence of overweight based on self-reports in the HBSC study and on measured ISO-BMI in the present study may indicate a reporting bias among Swedish girls . The prevalence of overweight according to ISO-BMI based on self-reports in the HBSC study was 9 % among Swedish girls, compared with 15 % based on measured ISO-BMI in the present study. This difference is not seen between self-reported or measured ISO-BMI among boys [both 14 %] . The incongruence between perception of the body and actual body size in girls, particularly at older ages, was corroborated in the present study by the findings of a lower OR for perception of overweight with regard to measured overweight according to ISO-BMI and WC. Also, the increased OR for perception of overweight for age ≥13 when adjusted for measured ISO-BMI, along with the high prevalence of misconception 1 among girls ≥ age 13, confirm that factors other than actual body size determine perception of overweight in adolescent girls.
Strengths and limitations
In the present study, perception of body weight status was self-reported, but is nonetheless considered to be a strength since self-reported perception of overweight is a determinant for actual and future weight and health behaviors . In our study, the self-reported perception of overweight was compared not only with actual body size according to anthropometric measurements, but also with the Swedish and international results of the WHO HBSC study performed the same year . The latter compensated for the small study population regarding perception of body weight status, but also confirmed the importance of comparing physical measurements with self-reported weight and height, especially in girls. In boys, the comparison of self-reported weight and height in the HBSC study  and the measurements in this study led to the same prevalence of overweight and obesity, implying that there was no clear bias in the selection of participants regarding being overweight, despite the small study population.
Our small study population limited generalization of the results, and also allowed only classification into two SES groups. Another limitation was the fact that parental SES was self-reported and not obtained through high-quality register data for ethical reasons. In this study, levels of parental education and occupational status were treated as two separate indicators of SES, despite the fact that they correlate with each other to some extent. Consequently the results should be interpreted with caution. The same applies to the separate analysis for paternal and maternal SES, even though paternal and maternal impact on children and adolescents was a partial goal of this study.
The extent to which social inequality increases the risk for overweight/obesity, and affects age and gender specific differences in perception of overweight, requires further investigation through larger quantitative and complementary qualitative studies. Age and gender differences in awareness of social norms and behaviors for development of overweight and perception of overweight should receive special attention.