The WSSQ is a 12-item Likert-type measure of weight-related self-stigmatization (Table 1). It has two subscales that individually measure (1) weight-related self-devaluation and (2) fear of enacted stigma. The original version of the questionnaire had good psychometrics and a valid preliminary construct. Cronbach α coefficients were acceptable for the full scale (α=.878) and the two subscales (α=.869 and α=.812) [9]. Analyses of principal components revealed a two-factor structure [9]. WSSQ items are rated on a scale of 1 (completely disagree) to 5 (completely agree). Sum scores are calculated for the full scale and each subscale. Items 1 to 6 constitute the self-devaluation subscale (factor 2), and items 7 to 12 constitute the fear of enacted stigma subscale (factor 1). There are no reverse-scored items.

Standard backward and forward translation was performed. One nutritionist and 2 research professionals independently conducted the forward translation, and 2 professional translators independently conducted the backward translation. A focus group composed of 8 participants was asked to answer and discuss the questionnaire, and their comments and understanding of each item’s meaning were discussed in the first focus group. Another round of focus group discussions was conducted with a total of 2 focus groups, each with 8 participants. Based on these discussions, the language of the questionnaire items was edited and clarified further, and the final version produced received universal approval among the group members. The final translated items can be found in Multimedia Appendix 1.

Based on published literature, the recommended sample size for test-retest reliability is 20 to 40 participants [18,19]. The original study that established test-retest reliability of the WSSQ used 44 participants [9]. The recommended sample size for exploratory factor analysis is between 100 and 250 participants, while some studies recommend more than 300 participants to ensure rigorous psychometric validation [20]. The suggested sample size includes a minimum of 2 and a maximum of 20 participants per item. For this study, the recommended minimum sample size was 240 participants based on 20 participants per each of the 12 items on the WSSQ. Thus, the targeted sample size range was 240 to 300 participants.

Similar to the original WSSQ validation process, internal consistency was assessed using the Cronbach α coefficient, and the test-retest reliability was assessed with the intraclass correlation coefficient. To assess the suitability for conducting an exploratory factor analysis, analyses of the correlation between the scale items were conducted using the Kaiser-Meyer-Olkin sample adequacy measure (nonsignificant results mean the data are suitable for factor analysis) and Bartlett test (significant results mean the data are suitable for factor analysis) [21-24]. To examine the factorial structure of the scales, an exploratory factor analysis using principal factor extraction was performed. The oblique rotation method was used to obtain clear factorial structures and enable comparison with the original study results; an item loading cutoff of greater than 0.15 was adopted, as in the original study, and factors with eigenvalues greater than 1.0 were retained [9]. There were no missing data, as all questions needed to be completed before submitting the questionnaire electronically. All statistical tests were performed using the IBM SPSS statistical package (version 20).

The ethics committee of the Sharik Association for Health Research, Riyadh, Saudi Arabia, approved this research project according to national research ethics regulations. Participants gave consent electronically, which was recorded by the data collection system.

Of the 43 participants in study stage 1 (for test-retest reliability), 48.8% (21/43) were male and the mean age was 34.4 years (range 18-66). Of the 295 participants in study stage 2 (for exploratory factor analysis), 48.5% (143/295) were male and the mean age was 33.6 years (range 18-70).

Internal consistency measures for both the subscales and the overall scale were good. The internal consistency of the full Arabic WSSQ was α =.898, and for the subscales—factor 1 and factor 2—internal consistency was α=.819 and α=.847, respectively. These values are similar to those observed in the original English version (α=.878 overall, and α=.869 and α=.812 for the enacted stigma and self-devaluation subscales, respectively) [9]. In the analysis of test-retest reliability, the intraclass correlation coefficient was α=.982.

Correlation coefficients among the 12 items were all 0.3 or higher, except those for item 12, which were below 0.3 for some other items. The Kaiser-Meyer-Oklin value was 0.891, and values for Bartlett’s test of sphericity reached statistical significance (P<0.001).

An analysis of principal components revealed two components that explained 47.7% and 10.6% of the variance. The scree plot (Figure 1) revealed a break after the second component. Oblique rotation (Table 2) was performed to compare the results with the original scale, and identical variance explanations of 47.7% and 10.6% were found.

Test-retest reliability/stability with the intracorrelation coefficient was high. Exploratory factor analysis revealed that loading of the items for the Arabic WSSQ was consistent with that of the original English WSSQ, except for the loading of item 9 (Table 2), which loaded positively in both factors but more strongly in factor 2 than in factor 1 [9].

This study involved translating the WSSQ into Arabic and testing the reliability and validity of the Arabic-translated version. The Arabic WSSQ demonstrated good psychometric properties, which were congruent with the original English version of the WSSQ. The internal consistency of the Arabic WSSQ and its subscales (ie, fear of enacted stigma and self-devaluation) was good: α=.898, α=.819, and α=.847, respectively. These values were consistent with the original English version of the WSSQ and its subscales: α=.878, α=.869, and α=.812, respectively [9]. Test-retest reliability of the Arabic WSSQ was high at 0.982. Exploratory factor analysis showed that the loading of the majority of the questionnaire items (except item 9) was consistent with the original English version of the WSSQ.

We searched PubMed for any published research articles from Saudi Arabia about weight stigma in general and found no related articles. The lack of research conducted on weight-related stigma in Saudi Arabia may be the result of a lack of validated tools for assessing this topic. The validation of such a tool, as presented in this study, provides an initial means to fill that gap. More tools to measure weight-related stigma for different age groups and dimensions (eg, public weight stigma) are needed to expand this research field and generate more understanding about the impact of weight self-stigma on quality of life and psychological stress. Increased use of the Arabic WSSQ has the potential to close the gap between the well-known high prevalence of obesity and the scarcity of knowledge on the frequency and severity of weight self-stigma among Saudi adults. To enable the expansion of WSSQ utilization to Arabic-speaking countries, we have provided the first Arabic translation of the WSSQ.

This study was not without notable limitations and strengths. The utility of the Arabic-WSSQ is not generalizable to children, as the study included only adults. A larger sample size is needed to assess the cross-cultural validity of this scale using confirmatory factorial analysis. Also, convergent validity was not assessed, and future studies are needed to assess correlations between the Arabic WSSQ and other scales that measure weight self-stigma.

However, the study had a high response rate among participants. The study’s sample size was in line with published sample size recommendations for psychometric validation and was consistent with the sample size of the initial study that established the psychometric properties of the original English version of the WSSQ [9]. To our knowledge, this is the first Arabic translation of the WSSQ and the first demonstration of its psychometric properties among Arabic-speaking Saudi adults. The Arabic WSSQ not only allows researchers and clinicians to assess weight self-stigma among Arabic-speaking adults, but it provides researchers and clinicians with an empirical measure to assess the effectiveness of interventions to reduce weight self-stigma.

The Arabic WSSQ has good internal consistency and reliability. The factorial structure is similar to the original English WSSQ, which endorses its value for use in cross-cultural studies. The Arabic WSSQ appears to be a reliable measure for assessing weight-related self-stigma in Arabic-speaking people.