Introduction

People with obesity are highly stigmatized and face discrimination and prejudice because of their weight [1]. The weight stigma about obesity is thought to be rooted in the misconception that body weight is easily controlled by changing the diet and physical activity levels [2]. Weight stigma internalization refers to “internalization of negative weight stereotypes and subsequent self-disparagement” [3]. Weight stigma internalization is associated with poor psychological [3,4,5], physical [5,6,7,8], and social [8,9,10] outcomes. Therefore, it is important for healthcare professionals [8, 11], social media [12, 13], and public health policy makers [14] not to exacerbate weight bias.

There are several self-reported questionnaires about internalized weight bias, among which weight bias internalization scale (WBIS) [15] is the most used. A previous study compared the WBIS and the Weight Self-Stigma Questionnaire and showed a higher reliability of the former [16]. The original version of the WBIS is in English, and it has been translated into several languages [17,18,19], including Asian languages [20]. However, a Japanese version of the WBIS has not yet been developed.

In Japan, the government began a health checkup for abdominal obesity in 2008, but the prevalence of obesity has not decreased between 2009 and 2019, with approximately 30% in males and 20% in females [21]. Although the health checkup has promoted awareness of obesity, previous studies have raised concerns that awareness-based approaches to obesity prevention could lead to unintended consequences from weight stigma [22, 23]. Japanese young females perceived high pressure from social media to be thin [24] and, dissatisfied with and concerned about their bodies [25]. In Japanese adolescents, 16% of males and 32% of females experienced body-related teasing in school and/or at home, and the experience associated with the perception to be overweight [26]. Despite this society-wide obsession against obesity, little is known about the potential harm of weight bias internalization in Japan. Weight bias toward individuals has been shown to exist in Japan as in other countries [27]. A previous study has shown that weight bias internalizes individuals when they perceived negative responses from others regarding their weight or body shape [3]. We hypothesized that weight bias internalization can play an important role in Japan. The aim of this study is to develop a Japanese version of the WBIS and to verify its reliability and validity.

Methods

Participants

We conducted a baseline and two-week follow-up online survey in April 2021. An Internet survey company, Macromill, Inc. [28], recruited participants based on the following inclusion criteria: (a) those who live in Japan and whose native language is Japanese, (b) those aged ≥ 20 years, and (c) those who perceive themselves to be obese. The perception of obesity was assessed by the same question as in the original WBIS (Appendix for details). Macromill could access to over 2,000,000 monitors representing all prefecture in Japan. Participants of this study were drawn from their monitor registered as respondents of the company. Of the available respondents, 285 participants completed a web-based questionnaire in the baseline survey. After two weeks, the company invited the responders of the baseline survey to participate in the follow-up survey in order of arrival and 100 participants completed the follow-up survey. Responses were anonymous.

Measures

Baseline survey included the Japanese version of the WBIS and scales for obesity-related quality of life, self-esteem, self-efficacy, and depression. Follow-up survey included the Japanese version of the WBIS.

The original version of the WBIS is an 11-item questionnaire that measures weight-related self-stigma and has a one-factor structure [15]. These items are scored on a 7-point Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree). Higher scores indicate more internalization of weight bias. The Japanese version of the WBIS was developed through three steps. First, two authors who were native Japanese speakers independently forward translated the original WBIS into Japanese and combined the two Japanese translations into one. Second, back-translation was conducted by a native English translator who was blinded to the original scale. Third, the original scale developer reviewed the English translation produced in the second step. Based on the suggestions of the developer, several items were modified by repeating the forward and backward translation procedures to reflect the original meaning after translation. Finally, we obtained the permission of the scale developer and used the Japanese version of the WBIS. The back-translation of the Japanese WBIS is provided in Appendix.

Obesity-related quality of life was evaluated using the Japanese version of the Obesity and Weight Loss Quality of Life Questionnaire (OWLQOL) [29, 30]. As the same as the original, it is composed of 17 items rated on a 5-point Likert scale. As a negative correlation was previously demonstrated between weight bias internalization and weight-related quality of life [31], a negative correlation would be expected between the WBIS and the OWLQOL.

Self-esteem was evaluated using the Japanese version [32] of the Rosenberg Self-Esteem Scale (RSES) [33]. As with the original, the Japanese version [32] is composed of 10 items rated on a 4-point Likert scale. A strong negative correlation was reported between the WBIS and the RSES [15].

Self-efficacy was evaluated using the General Self-Efficacy Scale (GSES) [34], which has 16 items rated on dichotomous (yes/no) scale. Weight bias internalization was reported to be associated with lower self-efficacy [35]; therefore, the present study predicted a negative correlation between the WBIS and the GSES.

Depression was evaluated using the Japanese version [36] of the Center for Epidemiologic Studies Depression Scale(CES-D) [37]. Both the original and its Japanese version [36] are composed of 20 items rated on a 4-point Likert scale. A strong positive correlation was reported between the WBIS and the CES-D [19].

Statistical analyses

We calculated the Cronbach’s alpha for internal consistency and the intra-class correlation coefficient (ICC) for test-retest reliability. Confirmatory factor analysis (CFA) was also conducted for structural validity. In CFA, we assumed a one-factor model as observed in the previous study [38]. The indicators of model fit were chi-square, comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR).　A good　fit was indicated by a value of 0.95 or more for the CFI, 0.06 or less for the RMSEA, and 0.08 or less for SRMR [39]. To evaluate convergent validity, we calculated Pearson’s correlation coefficients between the WBIS and OWLQOL, RSES, GSES, or CES-D in the baseline survey. The minimum effect size for detection in this study was 0.20. Based on a sample size calculation using G-power version 3.1.9.7, the necessary sample size was estimated to be more than 255 in the case of an alpha error probability of 0.05 and p power (1-β) of 0.90. In the test-retest reliability analysis, the sample size could be considered excellent when more than 100 participants were recruited [40]. Therefore, the numbers of participants in this study, 285 at baseline survey and 100 at follow-up survey, were adequate.

P-value below 0.05 was regarded as statistically significant. Data were analyzed using SPSS ver. 23, and CFA was performed using AMOS version 23.0.

Results

Demographic characteristics of the participants at baseline and follow-up survey are in Table 1. In the baseline survey (N = 285, 151 females and 134 males, mean age = 47.7 ± 13.5), the mean BMI was 25.0 (SDs = 3.5). In BMI categories, about half of the participants were normal weight. Characteristics of the participants at the follow-up survey (N = 100, 55 females and 45 males, mean age = 46.5 ± 13.5) did not differ from the baseline survey.

Table 1 Demographic characteristics of the participants.
Full size table

Cronbach’s alpha of the Japanese WBIS, as a measure of internal consistency, was 0.91 at baseline and 0.92 at follow-up. The ICC, as a measure of test-retest reliability, was 0.87 (p < 0.01) calculated using longitudinal data of 100 participants.

The results of CFA are shown in Fig. 1. The one-factor hypothesized model demonstrated acceptable fit (χ2 (44) = 158.6, CFI = 0.93, RMSEA = 0.096, SRMR = 0.051). Table 2 shows the correlation coefficients between the Japanese WBIS and OWLQOL, RSES, GSES, or CES-D. The total scores of the Japanese WBIS had a moderate positive correlation with CES-D, a moderate negative correlation with GSES, and strong negative correlations with OWLQOL and RSES.

Fig. 1
figure 1

CFA path diagram of the Japanese version of the WBIS (N = 285). Factor loadings were standardized. χ2 (44) = 158.6, CFI = 0.93, RMSEA = 0.096, SRMR = 0.051

Full size image
Table 2 Convergent validity of the Japanese version of the WBIS (N = 285).
Full size table

Discussion

The purpose of this study was to develop a Japanese version of the WBIS and to verify its reliability and validity. We finished the translation process and conducted the analysis. The results supported most of our hypotheses indicating that the Japanese WBIS showed good internal consistency, test-retest reliability, and convergent validity. The CFA showed acceptable results although RMSEA was higher than the cut off value of 0.06. This may be due to the sample size of the present study as RMSEA tends to be high with small sample size such as 250 [39].

The mean (SD) of the Japanese version WBIS was 43.1 (12.1), close to that of the original version [15]. The existence of obesity stigma was widely confirmed in Japan [27], although Japan has a low obesity rate. Social discrimination against people with obesity can lead to further obesity [40] and can prevent weight loss [41]. Flint et al., recommended that healthcare professionals working with people with obesity need to be educated about reducing weight bias internalization [42]. Cognitive-behavioral intervention study targeting weight stigma was conducted and produced short-term reductions in WBIS [43]. Measuring WBI is therefore needed both in society and in individuals to help reduce the number of people with obesity and overweight in Japan. The use of the Japanese version of the WBIS would be valuable to assess the levels of weight bias internalization in individuals in daily clinical practice and intervention studies.

In conclusion, we developed the Japanese version of the WBIS, and demonstrated that it has adequate psychometric properties, internal consistency, test-retest reliability, structural validity, and convergent validity for Japanese adults.

Limitations

First, this study was conducted thorough an Internet-based survey, which might lead to a selection bias as those with high internalized weight stigma may be reluctant to participate in the Internet survey. In addition, the generalizability of this study could be questioned. Second, the response rate and dropout rate could not be calculated since the Internet survey company recruited participants until the target number of participants.