Physical inactivity is an important health-threatening issue among young people. This may result in individuals developing obesity (the most common metabolic disorder) [1]. Moreover, over 13% of adults are diagnosed as obese [2], with a growing prevalence globally, including in East Asian countries [3]. For example, a recent population health survey reported nearly 33% of 15–84-year-old individuals in Hong Kong were obese [4]. The prevalence of obesity in Taiwan increased from 11.5% in 1996 to 23% in 2016 [5]. Although obesity (body mass index [BMI] at 30 kg/m² or above for Westerners; 25 kg/m² or above for Asians) and being overweight (BMI between 25 and 29.99 kg/m² for Westerners; 23 and 24.99 kg/m² for Asians) [6] are defined differently, their impacts on health are similar and hereafter, the present study uses the two terms interchangeably. Moreover, it should also be noted that the use of terms like ‘overweight’ and ‘obese’ (as defined by BMI measures) may also contribute to weight stigma (WS). Therefore, neutral word such as ‘higher weight’ may be preferable daily. However, the present study uses the terms ‘overweight’ and ‘obese’ because they are commonly used in the literature.

Individuals who have concerns about weight (especially those who are overweight) are at a higher risk for psychological distress than the general population [7]. Those who are overweight report elevated levels of psychological distress, body dissatisfaction [8], impaired self-esteem, and exposure to bullying [9,10]. Moreover, several longitudinal studies have emphasized that weight stigma rather than weight itself may cause psychological distress [11–13].

WS is defined as a kind of social devaluation among those who have weight concerns, even for those who are not overweight [14]. WS can be classified into different forms. In the present study, two specific types of WS were examined: perceived WS (how individuals perceive other people’s negative thoughts, feelings, and behaviors toward him/her) and weight bias internalization (WBI; how individuals accept the perceived weight stigma toward themselves) [15]. In most cultures and communities, body shapes beyond regular norms are exposed to judgment and frequent teasing in different situations, including education, job, and healthcare settings [16].

In general, WS may be considered as a risk factor for chronic stress, increased psychological distress, and lower engagement in healthy lifestyle, such as physical activity (PA) engagement and having a healthy diet [13,17]. Thoughts of victimization may also lead to suicidal behaviors/thoughts among this population due to being overweight [15]. Moreover, increasing evidence indicates that psychological distress may contribute to the development and continuity of being overweight [18]. For example, individuals having WS may use unhealthy coping strategies such as overeating and tend to engage in sedentary behaviors, which exacerbate weight gain [19].

The severity of perceived WS can include behaviors such as job discrimination [14]. Moreover, individuals who are overweight can experience isolation, rejection, feeling unattractive, feeling less competent, and feeling less intelligent due to stigmatization by strangers, friends, and even members of their own family [11,13]. There is also a significant association between increased risk of stress and stigmatization [19]. Moreover, psychological stress may concurrently be considered as one of the causes and outcomes of WS [8,20].

Frequently experiencing WS may be coupled with the perception of individual negative stereotypes and beliefs of being socially devalued, which may lead to internalizing weight bias [21]. Individuals may engage in a self-devaluation process and experience chronic stress due to being stigmatized for their weight irrespective of excess or not. This may lead to the development of WBI. Unfortunately, there is less information regarding WBI than WS is caused by social discrimination [22]. Nevertheless, studies have shown that WBI exists among individuals with all types of weight status and is associated with psychological distress and eating disorders [23].

However, some scholars have argued that WBI does not exist among individuals without weight problems but as other forms of disturbance (e.g., body image concerns) [24]. Nevertheless, the contribution of WBI (or body image disturbance; hereafter, WBI is used to indicate body image disturbance/concerns) to lowering engagement in health-promoting activities by negatively affecting self-efficacy has also been demonstrated [25]. Indeed, feelings such as not wanting to exercise and laziness are negative weight-related stereotypes may decrease self-efficacy and cause unwillingness to engage in a healthy lifestyle [23,26]. Therefore, a likely factor for this population’s the high prevalence of physical inactivity may be related to WBI.

Adolescents who sought treatment for obesity with less WS have shown a stronger desire to engage in PA than those with frequent WS exposure [27]. Similarly, youth who report more frequent WBI have lower levels of PA [28]. However, other studies have not reported significant associations between WS and PA [29,30]. They proposed other pathways that may mediate the associations between WS and PA. For example, PA and engagement in sports to get fit may provide a protective effect against social discrimination and WS, suggesting an interaction in this association [31]. Similarly, the positive effects of PA to decrease depressive symptoms and stress manifestations have been reported [32]. Therefore, further research is needed to assess the extent to which WS may influence PA and other factors that may affect this relationship is needed.

The present study investigated how perceived WS affects the tendency to engage in PA and whether WBI and psychological health mediate this relationship. Accordingly, the following two hypotheses (H_s) were proposed: (i) there would be a significant association between perceived WS and a tendency to avoid PA (H₁), and (ii) WBI and psychological distress may play mediating roles between direct associations of perceived WS with a tendency to avoid PA (H₂). Since most studies suggest a direction from stigmatization towards the tendency not to engage in PA, the likely causal direction between them was also investigated. Moreover, examining the associations over time provides a better picture of the relationships between these variables (perceived WS, WBI, psychological distress, and tendency to avoid PA) than cross-sectional designs. Therefore, the present study comprised longitudinal data collection over two waves (six months apart) to examine changes in the study’s variables over time.

Among the 592 participants, 388 completed the Time 2 (T2) survey between February and April 2023 (34.4% attrition rate). There were no significant differences between the participants who completed two waves of surveys (n = 388) and those who only completed the T1 survey (n = 204) in terms of their gender (only completed T1: 45.1% male, 54.4% female, and 0.5% other; completed T1 and T2: 45.1% male, 54.6% female, and 0.3% other; p = 1.00) and body mass index (only completed T1: 23.89 [SD ± 4.75]; completed T1 and T2: 23.93 [SD ± 4.22]; p = 0.92). However, those who only completed the T1 survey were significantly younger than those who completed T1 and T2 surveys (28.12 years [SD ± 6.37] vs. 29.65 years [SD ± 6.27]; p = 0.01). Because the attrition rate was lower than 40% with nonsignificant to trivial differences between the retained and lost samples in the present study, the findings were not likely to have any severe biases.

The analyzed data included 388 participants who completed both T1 and T2 surveys. Over half of the 388 participants were female (n = 212; 54.6%), and one participant identified gender as other (0.3%), with the remaining participants being male (n = 175; 45.1%). Detailed information on the participants’ characteristics is presented in Table 1. In particular, over half of the participants did not have perceived weight stigma because 57.5% of the participants scored 0 for the PWSS at T1 and 59.3% at T2. Moreover, all the continuous variable scores were normally distributed (i.e., both skewness and kurtosis values <3).

Table 2 presents the Pearson correlation coefficients between PWSS, WBIS, DASS-21, and TAPAS scores across T1 and T2. All the correlations were significant with the magnitude of r-values ranging between 0.24 and 0.69 (all p-values < 0.001). The mediation model results are presented in Table 3 and Fig. 1. More specifically, (i) PWSS score had significant direct effects on both WBIS (standardized coefficient [β] = 0.25; p < 0.001) and DASS-21 (β = 0.15; p = 0.003); and (ii) WBIS (β = 0.47; p < 0.001) but not DASS-21 (β = 0.08; p = 0.077) had significantly direct effects on TAPAS. In addition, the standardized indirect effect of (i) WBIS in the association of PWSS with TAPAS was 0.12 (bootstrapping SE = 0.04; 95% bootstrapping CI = 0.04, 0.20); and (ii) DASS-21 in the association of PWSS with TAPAS was 0.01 (bootstrapping SE = 0.01; 95% bootstrapping CI = −0.001, 0.03). The standardized total effect of PWSS on TAPAS (i.e., direct effect and indirect effects via WBIS and DASS-21) was 0.13 (bootstrapping SE = 0.04; 95% bootstrapping CI = 0.05, 0.21).

The present study was designed to investigate the associations between weight bias (WS) (including perceived WS and weight bias internalization [WBI]) and the tendency to avoid physical activity (PA) among a convenience sample of the Taiwanese population. Moreover, the indirect associations between these two variables were assessed by including WBI and psychological distress as the mediators. The study findings indicated that while there was a significant association between the main variables of the study (i.e., perceived WS and tendency to avoid PA), there were significantly notable associations between the PWSS and both psychological distress and internalization of weight bias. The mediating role of WBI in the relationship between the main variables was also significant. However, the indirect association between perceived WS and the tendency to avoid PA via psychological distress was not confirmed.

Several studies have assessed the relationships between WS and PA particularly among youth. In a cross-sectional survey, Ajibewa et al. [17] examined the relationship between prior WS experiences and PA intention among college students. The average body mass index (BMI) score was similar to that of the present sample (23.9 ± 4 kg/m²). They found that prior experiences with WS were significantly associated with PA avoidance, while the association between WS and intention to do PA was non-significant [17]. Bevan et al. [48] in a similar study among college students that the development of the Tendency to Avoid Physical Activity and Sport Scale (TAPAS) (the scale used in the present study), recommended that due to the strong association between the WS and PA, any attempts to increase PA participation must firstly address WS concerns particularly among individuals who are overweight or have obesity [48]. A systematic review conducted in 2021 found 38 studies that investigated the associations between WS and PA. It was concluded that both perceived WS and WBI had significant associations with PA behaviors [16]. Moreover, direct and indirect effects of WS on PA were recognized which concur with the present study’s findings. However, one specific matter that differentiates the present study’s findings from previous results is the utilization of a new scale to assess the tendency to avoid PA. Apart from the study by Bevan et al., all the previous studies’ assessments were carried out by assessing PA behaviors through observation or self-report [16], whereas the present study did not assess the actual PA behaviors but rather the tendency to avoid such behaviors. It is believed that because the tendency to engage in or not to engage in a specific behavior may be a prerequisite of that behavior happening, it may provide more accurate information on the likelihood of its occurrence in the future.

To the best of the authors’ knowledge, the present study is the first to investigate the indirect pathways between perceived self-stigma and the tendency to avoid PA. Previous studies have confirmed the association between WS, and the mediators used in the present study (WBI and psychological distress). For example, Romano et al. suggested that due to the strong association between perceived WS and WBI and since WBI may be an underlying factor for general WS, interventions focusing on WBI are recommended for individuals who are overweight or have obesity [21]. Shank et al., in assessing the relationships between WS, WBI, and physical health among military personnel who were mostly at-risk of obesity, found that while WS and WBI were positively associated, only WS predicted a medical condition among these individuals and the WBI was not a significant predictor [52]. This finding confirms that although these variables are very similar, each one assesses different things, and they may be considered as underlying contributors for explaining their associations with other biopsychological variables.

In a study relatively similar to the present one, O’Brien et al. [20] assessed the relationship between WS and eating disorders by examining WBI and psychological distress as the mediators. They recruited 632 college students to complete an online survey and reported a direct association between WS and eating disorders. In addition, indirect associations between these two variables via WBI and psychological distress confirmed their mediating roles [20]. Although the measure used by O’Brien et al. to investigate WS differed from that of the present study, the same measures were used to assess WBI and psychological distress in both studies, suggesting consistency of measurement except for the dependent variable (eating disorders vs. TAPAS). Because diet and PA have determinative roles in weight-related problems, especially among those who are overweight or have obesity (and both are inter-correlated), similar results for both studies would be expected.

However, in the present study, the indirect association between WS and the dependent variable (i.e., PA) via psychological distress was non-significant (unlike the study by O’Brien et al. [20]). In addition to the similarities between these studies, there are also considerable differences between them that may explain relatively different outcomes. First and most important, there were different dependent variables. While there are associations between eating disorders such as polyphagia and the tendency to avoid PA [53], these are related to otherbehavioral patterns and are not necessarily inter-correlated. Second, the present study had a longitudinal design in an attempt to establish causation between variables since it covers the changes over time. In contrast, O’Brien et al.’s study [20] had a cross-sectional design. Third, the cultural differences between the two samples may also be a possible source of difference in outcomes.

Despite the strengths of the present study, such as the longitudinal design with adequate sample size and robust statistical methods to assess the likely associations (mediation analysis), several limitations should be noted when interpreting the results. First, a convenience sample was used with no specific inclusion or exclusion criteria, and all interested individuals could participate. Therefore, the generalizability of the findings specifically for defined categories of the individuals in terms of BMI or other relevant criteria was limited. Second, although the individuals who participated in the study in the second wave of data collection were not so different from the first wave in terms of demographic characteristics, there was an attrition rate of approximately 35% that may have negatively impacted the external validity of the findings. Third, all measures used were subjective and susceptible to reporting biases such as social desirability or recall bias. More objective measures, such as assessing the behaviors through observation or an individual’s actual height and weight, would likely provide more trustworthy data. Moreover, the variables used in the present study were primarily psychological and no objective measures were used to assess PA.

Along with the direct association between perceived WS and the tendency to avoid PA, these variables may also be correlated through bilateral mechanisms, including the WBI. Perceived WS should be considered as a likely source both for WBI as well as avoiding PA. Therefore, any program planning to address physical inactivity across weight status from healthy weight to obesity should include WS reduction strategies to be successful. For example, program planners may suggest environments that are supportive for PA among people with overweight/obesity without any triggers that may arouse weight stigma. Also, when exercising, grouping people with the same anthropometric measures may reduce the chance of weight stigmatization. Moreover, due to the underlying effect of WBI in the association between the PWSS and TAPAS, addressing interventions to reduce WBI would be helpful to improve PA intention, particularly among those who are at higher risk for weight gain and obesity. Therefore, further assessment of other likely effective psychological mediators such as self-esteem and self-concept in such associations are warranted.