Psychometric properties of the weight self-stigma questionnaire and weight bias internalization scale in postpartum women

Abstract

Background

Weight stigma internalization among postpartum women may differ from other populations owing to childbirth, breastfeeding, and postnatal physical recovery. Two instruments to assess the concept of weight stigma internalization—the Weight Self-Stigma Questionnaire (WSSQ) and the Weight Bias Internalization Scale (WBIS)—have not been validated in postpartum women. This study aims to evaluate and compare the reliability, factor structure, and concurrent validity between the WBIS and WSSQ in Chinese postpartum women.

Methods

A total of 476 postpartum women (mean age 30.96 ± 4.67 years) completed the WSSQ and WBIS, as well as other relevant measures regarding perceived weight stigma, psychological outcomes, and eating behaviors. Height and weight were used to calculate the body mass index (BMI). Statistical analyses were performed using internal consistency analysis, confirmatory factor analysis, and correlation analysis.

Results

Both the WSSQ and the WBIS were found to have excellent internal consistency. In addition, both were associated with postpartum BMI, perceived weight stigma, adverse psychological outcomes, and restrained and emotional eating behaviors. Differently, the two-factor structure of the WSSQ was supported among postpartum women, but the one-factor WBIS was not.

Conclusion

The WSSQ is a reliable and valid measurement tool for assessing weight stigma internalization in postpartum women. The structure of WBIS requires further validation and modification in future research.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09528-w.

Introduction

Along with the dramatic increase in obesity rates, weight-based prejudice, discriminatory behavior, or stereotyping—namely weight stigma—have become common social phenomena [1]. For women, pregnancy is often accompanied by weight gain and even postpartum weight retention [2]. The postpartum period is a particularly vulnerable time for concerns about body weight due to natural weight fluctuations, the popularity of the “thin ideal” socio-culture pressure, and the pressure to quickly “bounce back” after birth to pre-pregnancy body weight [3]. About two-thirds of pregnant or postpartum women face weight stigma from at least one source in their daily lives [4], including family, friends, strangers, healthcare providers, and the media [3, 4]. Under the interplay of multiple stressors, the self-evaluation of postpartum women can be distorted [5]. Thus, postpartum women may be more likely to internalize weight stigma as a self-evaluation called weight stigma internalization. Weight stigma internalization has been described as individuals agreeing to stereotypes about weight and applying these stereotypes to themselves, and in turn, self-devaluation [6]. Notably, weight stigma internalization is prevalent among individuals regardless of their actual weight status [7].

Weight stigma is intrinsically linked to a spectrum of adverse physiological and psychological outcomes [8]. Weight stigma internalization may be a key mechanism for understanding the link between weight stigma and negative health outcomes. Tylka et al. [9] developed a theoretical framework positing that weight stigma internalization mediates the relationship between weight stigma experiences and negative health outcomes. Specifically, weight stigma is positively associated with weight stigma internalization, which in turn is associated with negative biopsychosocial outcomes (e.g., disordered eating, depression, and anxiety). Bidstrup et al. [10] reported that internalized weight stigma mediates the relationship between perceived weight stigma and disordered eating, depression, and anxiety symptoms. Weight stigma is harmful, especially when it is internalized [11]. For example, Zancu et al. [12] demonstrated that while the direct relationship between weight stigma and disordered eating behaviors was not significant, there was an indirect relationship between weight stigma and disordered eating behaviors via weight stigma internalization. This finding highlights the deleterious effects of weight stigma internalization on psychological and eating behaviors.

Accurately evaluating weight stigma internalization in postpartum women is crucial for comprehending and enhancing their mental well-being. Two self-report scales have been developed to assess weight stigma internalization, including the Weight Self-Stigma Questionnaire (WSSQ) [13] and the Weight Bias Internalization Scale (WBIS) [14]. The original WSSQ was used to assess self-evaluation and fear of enacted stigma in overweight and obese people. The WSSQ has been validated in overweight and obese adults in China [15], Italy [16], Germany [17], and Spain [18], and has also been extended to children and adults in a wider weight range [19, 20, 21]. The WBIS was designed to assess weight-based self-devaluation, and it has been studied in children, adolescents, and adults of different weight categories [19, 22, 23, 24].

Although the psychometric properties of the WSSQ and WBIS have been evaluated in adults, postpartum women may face weight biases and stresses that are different from those of the general population due to childbirth, breastfeeding, and physical recovery [25]. The positive association between weight stigma and its internalization is well-established across populations [26, 27]. However, the specific sources of weight stigma vary significantly across reproductive stages [28]. During pregnancy, weight stigma often originates from media and healthcare settings [29], whereas during the postpartum period, family-related weight stigma becomes particularly salient [30]. Moreover, the postpartum period imposes a unique sociocultural pressure—the expectation to rapidly “bounce back” to pre-pregnancy weight—that is absent in other life stages [30, 31]. These distinct sources of stigma suggest that the pattern of weight stigma internalization in the postpartum period may differ from that observed in the general population. Consequently, findings from studies of other adult populations may not apply directly to postpartum women. In addition to this population-specific concern, the WBIS and WSSQ may exhibit distinct psychometric properties within the same population. For instance, Hübner et al. [17] observed superior reliability, convergent validity, and predictive power of WBIS over WSSQ in a German bariatric surgery sample; Pakpour et al. [19] compared the WSSQ and the WBIS in Hong Kong children and found that each had strengths in different aspects. These findings point to the necessity of directly comparing the two scales in postpartum women.

To date, however, psychometric validation of these scales in postpartum women remains extremely limited. To our knowledge, only two studies from the United States have reported the internal consistency of the WBIS among postpartum women [31, 32]. Beyond internal consistency, evidence for other psychometric properties of the WBIS in this population—such as construct validity and concurrent validity—is lacking. Additionally, to the best of our knowledge, the psychometric properties of the WSSQ have not yet been examined in postpartum women. Therefore, the current study aimed to evaluate and compare the reliability, factor structure, and concurrent validity of the WBIS and WSSQ in Chinese postpartum women. We hypothesized that both scales would be strongly correlated with each other and significantly associated with perceived weight stigma and negative health outcomes (depression, anxiety, stress, and disordered eating behaviors). We did not make any assumptions about the association between the two measures and body mass index (BMI) due to inconsistent findings [7]. By comparing and analyzing these two measurement tools, we expect to find suitable assessment methods for weight stigma internalization in postpartum women. This will help provide more accurate psychological support and interventions for this group and offer new perspectives for future research.

Methods

Participants and procedure

We conducted a cross-sectional survey between November 2020 and March 2021. With the approval of the study site supervisor, researchers recruited participants from the waiting rooms of four community health centers and a maternal and child health hospital in China using a convenience sampling procedure. To be eligible for this study, participants had to be: (1) 18 years and above; (2) within 12 months postpartum; (3) able to speak and read Chinese fluently; and [4] willing to provide informed consent. Participants with a history of significant mental or physical health disorders were excluded (e.g., cognitive disorders, developmental delay).

The sample size was determined based on established guidelines for factor analysis. A commonly applied rule recommends a minimum of 10 cases per observed variable [33]. Given that the WSSQ and WBIS comprise 12 and 11 items, respectively, a baseline sample of at least 230 participants was required. Moreover, as confirmatory factor analysis typically requires a sample size of no less than 200 to achieve stable model estimation [34], and to account for potential invalid responses, we aimed to recruit a larger sample to ensure adequate statistical power. In total, 530 participants were surveyed. Fifty-four samples with missing values (incomplete questionnaires or missing important information) were excluded to ensure sample uniformity throughout the analysis procedure. Finally, analyses in this study were based on the data of 476 participants.

Ethical approval for this study was obtained from the Ethics Committee of the School of Nursing, Central South University (Approval No. E2021108). All participants provided verbal and written informed consent before study participation. Participants completed the questionnaire face-to-face under the guidance of a trained researcher in a quiet demonstration room. Each questionnaire took 10–15 min.

Measures

Anthropometric information

We measured participants’ height and weight using a portable stadiometer and scale, after removing their shoes. The participants’ BMI was calculated using their weight and height, and their weight status was subsequently categorized according to the Chinese classification criteria [35]: underweight (BMI < 18.5 kg/m²), normal weight (18.5 ≤ BMI < 24.0 kg/m²), overweight (24.0 ≤ BMI < 28.0 kg/m²), or obese (BMI ≥ 28 kg/m²).

Weight Self-Stigma Questionnaire (WSSQ)

The WSSQ is used to measure self-stigma associated with weight. It consists of two domains: self-devaluation and fear of enacted stigma [13]. Each domain consists of six items rated on a five-point Likert scale ranging from 1 (completely disagree) to 5 (completely agree). Higher scores in the WSSQ indicate higher levels of weight-related self-stigma. The WSSQ used in this study is the Chinese version translated by Lin et al. [15]. The Chinese version of the WSSQ has been shown to have adequate reliability and validity in overweight and obese adults.

Weight Bias Internalization Scale (WBIS)

The WBIS was originally designed to assess the degree of weight stigma internalization in overweight and obese individuals [14]. The scale had only one domain, with higher scores indicate higher levels of weight-related self-stigma. The present study used the Chinese versions of the WBIS [19], translated by Pakpour et al. It consists of 11 items rated on a Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). At the recommendation of the WBIS developer, the Chinese version of the WBIS replaced “overweight” with “weight” in the items to enhance its relevance across various weight statuses [19]. The WBIS showed good psychometric properties in Chinese children with different weight statuses.

Perceived Weight Stigma Questionnaire (PWSQ)

The PWSQ is used to assess the severity of perceived weight stigma in the past week. The scale consists of 10 items that are scored on a “yes” or “no” basis, with a score of 1 representing “yes” and 0 representing “no”. Higher PWSQ scores indicate higher levels of perceived weight stigma. The Chinese version of the PWSQ was validated among university students and showed good internal consistency (Cronbach’s α = 0.84) [36]. The Cronbach’s α for PWSQ in this study is 0.804.

Depression Anxiety Stress Scale-21 (DASS-21)

The DASS-21 consists of 21 items that assess an individual’s psychological distress. The scale measures three different forms of psychological distress: depression (7 items), anxiety (7 items), and stress (7 items). Each item is assessed using a four-point Likert scale ranging from 0 (not at all) to 3 (very severe). Higher scores indicate higher levels of psychological distress. The psychometric properties of the DASS-21 have been widely evaluated and supported [37]. The Chinese version of the DASS-21 also had good internal consistency with Cronbach’s α of 0.83 for depression, 0.80 for anxiety, 0.82 for stress, and 0.92 for the overall scale [38]. The Cronbach’s α for the total DASS-21 score in this study was 0.957, with the domains ranging from 0.856 to 0.909.

The Dutch Eating Behavior Questionnaire (DEBQ)

The DEBQ has 33 items, including 3 domains: restrained eating (10 items), emotional eating (13 items), and external eating (10 items) [39]. Restrained eating refers to the deliberate restriction of food intake for the purpose of weight control or maintenance. Emotional eating refers to eating in response to negative emotional states. External eating is triggered by visual or olfactory cues from food. The DEBQ utilizes a 5-point Likert scale ranging from 1 (never) to 5 (always) to assess levels of restrained, emotional, and external eating. Higher scores on each domain indicate greater expression of these behaviors. The DEBQ has good psychometric properties in Chinese adults. The Cronbach’s α for each domain of the Chinese version of the DEBQ ranged from 0.81 to 0.94 [40]. This study reported a Cronbach’s α of 0.943 for the total DEBQ score, with the domains ranging from 0.891 to 0.960.

Statistical analysis

Continuous variables were described as mean ± standard deviation (SD), and categorical variables were presented as frequencies and percentages. The critical ratio method and the item-total correlations were used for item analysis. Total scores for the WSSQ and WBIS are ranked in descending order, respectively. The top 27% and bottom 27% were categorized into high and low subgroups. Item scores of the two groups were compared by independent t-tests. Items with an absolute value of critical ratio ≥ 3 and a statistically significant difference were considered to be discriminatory [41]. Spearman’s correlation was used to calculate the correlation between the items and the total score. Items that were significantly correlated with the total score and had a correlation coefficient ≥ 0.40 were considered representative [42]. In addition, Cronbach’s α was used to assess the internal consistency of the WSSQ and WBIS.

Confirmatory factor analysis (CFA) was used to test the factor structures of the WSSQ (One-factor and two-factor) and the WBIS (One-factor). The maximum likelihood estimation with robust standard errors and mean‑adjusted chi‑square statistic (MLM) was applied. We used several cutoffs to determine the acceptable model fit [43, 44]: comparative fit index (CFI) > 0.9, Tucker-Lewis index (TLI) > 0.9, root mean square error of approximation (RMSEA) < 0.10, and standardized root mean square residual (SRMR) < 0.08. Items with factor loading < 0.50 [45] were assessed to optimize adjustments for enhancing model fit.

Spearman’s correlation analysis was used to explore the relationship between WBIS, WSSQ, PWSQ, BMI, DASS-21, and DEBQ. The correlation coefficients ranged from 0.1 to 0.3, 0.3 to 0.5, and above 0.50, indicating small, medium, and large effects, respectively [46]. Correlation coefficients were examined by Z-tests after z-scores transformation to account for differences in correlation coefficients. Furthermore, partial correlation analyses were performed to adjust for potential confounding variables, including pre-pregnancy BMI and infant feeding method.

The software used in this study was Mplus 8.0 for CFA, and SPSS 29.0 (IBM SPSS Inc., Chicago, IL, USA) for the rest of the analyses. All p-values are < 0.05 suggesting statistical differences for all analyses.

Results

Descriptive analysis

This study included 476 postpartum women. The mean age of participants was 30.96 years (SD = 4.67), with the majority (72.7%) aged 25–34 years. Socio-demographic and obstetric characteristics of the sample are presented in (Table 1). Regarding weight status, the mean pre-pregnancy BMI was 21.09 kg/m² (SD = 2.98), with 85.5% of participants classified as normal or below (< 24.0 kg/m²) and 14.5% as overweight or obese (≥ 24.0 kg/m²). At the time of assessment, the mean postpartum BMI was 22.87 ± 2.97, with 67.0% of participants having a normal or below BMI and 33.0% being overweight or obese. The WBIS, WSSQ, PWSQ, DASS-21, and DEBQ scores are shown in (Table 2).

Table 1.

Socio-demographic and obstetric characteristics of participants (n = 476)

Variable	Categories	n	%
Age	< 25	36	7.6
	25–34	346	72.7
	≥ 35	94	19.7
Ethnicity	Han	440	92.4
	Others	36	7.6
Residence	Rural	70	14.7
	Urban	406	85.3
Education level	Junior high school or below	19	4.0
	Senior high school/technical secondary school	50	10.5
	College degree or above	407	85.5
Occupation	White-collar/Professional	305	64.1
	Self-employed/Freelance	115	24.2
	Blue-collar/Manual	15	3.1
	Other/Unemployed	41	8.6
Marital status	Unmarried	3	6.0
	Married	473	99.4
Feeding method	Exclusive breast feeding	238	50.0
	Formula feeding	48	10.1
	Mixed feeding	190	39.9
Parity	Primiparous	287	60.3
	Multiparous	189	39.7
Mode of delivery	Vaginal delivery	284	59.7
	Cesarean section	192	40.3
Pre-pregnancy BMI	Normal or below	407	85.5
	Overweight or obese	69	14.5
Postpartum BMI	Normal or below	319	67.0
	Overweight or obese	157	33.0

BMI Body mass index

Table 2.

Descriptive analysis of the WSSQ, WBIS, PWSQ, DASS-21, and DEBQ (n = 476)

Variable	Minimum	Maximum	Mean ± SD
WSSQ	12	60	24.93 ± 9.82
Self-devaluation	6	30	13.43 ± 5.10
Fear of enacted stigma	6	30	11.50 ± 5.33
WBIS	11	52	25.33 ± 9.24
PWSQ	0	10	0.47 ± 1.26
DASS-21	0	63	11.04 ± 12.31
Anxiety	0	21	3.68 ± 4.08
Depression	0	21	2.81 ± 4.16
Stress	0	21	4.55 ± 4.71
DEBQ	33	137	80.44 ± 20.19
Emotional eating	13	65	26.59 ± 10.98
Restrained eating	10	49	24.01 ± 8.67
External eating	10	50	29.84 ± 7.19

WSSQ Weight self-stigma questionnaire, WBIS Weight bias internalization scale, PWSQ Perceived weight stigma questionnaire, DASS-21 Depression anxiety stress scales-21, DEBQ The Dutch eating behavior questionnaire

Item analysis

The absolute values of the critical ratio for the WSSQ and WBIS ranged from 13.341 to 27.190 and 9.007–29.512, respectively. The correlations between all items of both scales and their total scores were above 0.4. Therefore, we considered the items to be discriminatory and representative (see Table 3).

Table 3.

Item analysis of the WSSQ and WBIS

Item	Critical ratio			Item-total correlation			Cronbach’s α after item deletion
	WSSQ	WBIS		WSSQ	WBIS		WSSQ	WBIS
Item 1	-13.341***	-9.007***		0.562***	0.409***		0.935	0.917
Item 2	-19.379***	-22.295***		0.705***	0.736***		0.926	0.896
Item 3	-20.485***	-24.406***		0.791***	0.786***		0.920	0.891
Item 4	-17.331***	-21.020***		0.744***	0.717***		0.922	0.897
Item 5	-18.358***	-28.852***		0.724***	0.862***		0.925	0.885
Item 6	-21.365***	-29.512***		0.733***	0.854***		0.926	0.886
Item 7	-19.911***	-15.179***		0.717***	0.676***		0.924	0.897
Item 8	-18.640***	-13.776***		0.761***	0.660***		0.921	0.898
Item 9	-21.217***	-11.815***		0.786***	0.609***		0.920	0.903
Item 10	-27.190***	-25.329***		0.837***	0.818***		0.918	0.889
Item 11	-20.616***	-18.848***		0.802***	0.738***		0.919	0.893
Item 12	-22.417***	-		0.803***	-		0.920	-

^***, p < 0.001. WSSQ Weight self-stigma questionnaire, WBIS Weight bias internalization scale

Internal consistency analysis

In this study, the Cronbach’s α for the WSSQ and its domains were 0.929, 0.838, and 0.927, respectively. The Cronbach’s α for the WBIS was 0.905. Both scales showed excellent internal consistency. Table 3 shows that excluding item 1 improves the internal consistency of both WSSQ and WBIS.

Confirmatory factor analysis

The results of the CFA supported the original two-factor structure of the WSSQ with satisfactory fit indices (CFI = 0.934, TLI = 0.918, RMSEA = 0.084, SRMR = 0.043), whereas the one-factor structure showed suboptimal fit (CFI = 0.905, TLI = 0.884, RMSEA = 0.099, SRMR = 0.053). For the WBIS, the one-factor structure did not exhibit a good fit (CFI = 0.843, TLI = 0.803, RMSEA = 0.140, SRMR = 0.067). All items in the WSSQ (0.654 to 0.890) and WBIS (0.530 to 0.882) showed acceptable factor loadings, except for item 1 of the WSSQ (0.478) and item 1 of the WBIS (0.259). See Table 4 for more details.

Table 4.

Confirmatory factor analysis of the WSSQ and WBIS

Measure	Loading	Fit indices
		CFI	TLI	RMSEA	SRMR
WSSQ (Two-factor)		0.934	0.918	0.084	0.043
WSSQ (One-factor)		0.905	0.884	0.099	0.053
Item 1a	0.478
Item 2a	0.667
Item 3a	0.825
Item 4a	0.811
Item 5a	0.699
Item 6a	0.654
Item 7b	0.682
Item 8b	0.851
Item 9b	0.860
Item 10b	0.855
Item 11b	0.890
Item 12b	0.844
WBIS		0.843	0.803	0.140	0.067
Item 1	0.259
Item 2	0.682
Item 3	0.758
Item 4	0.689
Item 5	0.882
Item 6	0.871
Item 7	0.672
Item 8	0.658
Item 9	0.530
Item 10	0.806
Item 11	0.738
WBIS-10		0.846	0.802	0.152	0.070

^a, embedded in the self-devaluation domain. ^b, embedded in the fear of enacted domain

CFI Comparative fit index, TLI Tucker-Lewis’s index, RMSEA Root mean square error of approximation, SRMR Standardized root means square residual, WSSQ Weight self-stigma questionnaire, WBIS Weight bias internalization scale, WBIS-10 Weight bias internalization scale after deletion of item 1

Given the satisfactory fit of the two-factor WSSQ model, we made no further modifications. Combined with the results of internal consistency, we conducted another CFA on the modified WBIS (WBIS-10) after excluding item 1. The results showed that the factor loadings ranged from 0.527 to 0.882, but the model fit did not get better (CFI = 0.846, TLI = 0.802, RMSEA = 0.152, SRME = 0.070).

Correlation analysis

The Spearman’s correlation analysis showed a strong correlation between WSSQ and WBIS (r = 0.760). WSSQ showed moderate correlations with the DASS-21 (r = 0.429 with anxiety; r = 0.450 with depression; r = 0.433 with stress), the emotional (r = 0.341) and restrained eating behavior (r = 0.380) domain in DEBQ, weak correlations with the PWSQ (r = 0.249). However, it did not correlate with the external eating behavior domain in DEBQ (r = 0.064). WBIS showed similar results to WSSQ (r = 0.375 with anxiety; r = 0.388 with depression; r = 0.361 with stress; r = 0.275 with emotional eating behavior; r = 0.390 with restrained eating behavior; r = 0.043 with external eating behavior; r = 0.205 with PWSQ). Additionally, as shown in Table 5, the correlation between the WBIS and postpartum BMI (r = 0.491) was higher than the correlation between the WSSQ and postpartum BMI (r = 0.375).

Table 5.

Correlations of the WBIS and WSSQ and differences in correlation coefficients

Variable	r		z (p)	r’
	WBIS	WSSQ		WBIS	WSSQ
WSSQ	0.760***	-	-	0.771***	-
PWSQ	0.205***	0.249***	0.714 (0.476)	0.361***	0.399***
BMI	0.491***	0.375***	2.201 (0.028)	0.378***	0.294***
DASS-21	0.389***	0.455***	1.236 (0.217)	0.480***	0.530***
Anxiety	0.375***	0.429***	0.991 (0.322)	0.474***	0.509***
Depression	0.388***	0.450***	1.157 (0.247)	0.460***	0.508***
Stress	0.361***	0.433***	1.316 (0.188)	0.439***	0.497***
DEBQ	0.328***	0.370***	0.736 (0.462)	0.309***	0.326***
Emotional eating	0.275***	0.341***	1.122 (0.262)	0.256***	0.309***
Restrained eating	0.390***	0.380***	0.181 (0.857)	0.359***	0.333***
External eating	0.043	0.064	-	0.038	0.036

^***, p < 0.001. WSSQ Weight self-stigma questionnaire, WBIS Weight bias internalization scale, PWSQ Perceived weight stigma questionnaire, BMI Body mass index, DASS-21 Depression anxiety stress scales-21, DEBQ The Dutch eating behavior questionnaire

To account for potential confounding effects of pre-pregnancy BMI and infant feeding method, partial correlation analyses were conducted. As shown in Table 5, after controlling for these variables, the correlations between the two scales (WSSQ and WBIS) and most outcome variables remained largely consistent in direction and significance compared with the zero-order correlations. Specifically, the correlations between both scales and postpartum BMI were attenuated (WBIS: from r = 0.491 to r’ = 0.378; WSSQ: from r = 0.375 to r’ = 0.294), whereas their associations with PWSQ were increased notably (WBIS: from r = 0.205 to r’ = 0.361; WSSQ: from r = 0.249 to r’ = 0.399). Correlations with DASS-21 and DEBQ remained stable and significant, while external eating behavior remained non-significant. These findings indicate that the relationships between weight stigma internalization and key health outcomes are robust after adjusting for pre-pregnancy BMI and infant feeding method.

Discussion

To our knowledge, this is the first study to compare the psychometric characteristics of the WSSQ and the WBIS among postpartum women. Our study indicated that both the WSSQ and the WBIS had item discrimination, item representativeness, and good internal consistency. In addition, both WSSQ and WBIS were correlated with other measurements, including PWSQ, DASS-21, postpartum BMI, and emotional and restrained eating behavior domains in DEBQ. Differently, the WSSQ has an acceptable factor structure, but the WBIS did not satisfy a good fit for a single factor.

In our study, the WBIS showed acceptable item characteristics and excellent internal consistency, but item 1 (“No matter how much I weigh, I can do just as much as everyone else”) was significantly less discriminatory and representative than the other items. The exclusion of item 1 was supported by both the internal consistency analysis and factor loadings of WBIS. Item 1 has shown similar results in previous studies, leading to its exclusion in research involving children [22], adolescents [23, 47], and adults [48]. Additionally, two studies conducted in the United States assessed the reliability of the WBIS in postpartum women. In the first study, the 11-item WBIS showed high internal consistency (Cronbach’s α = 0.814) [32], whereas the second study showed better internal consistency (Cronbach’s α = 0.95) after removing item 1 [31]. The challenge with item 1 may be that its reverse scoring makes it more difficult for respondents to understand, thus affecting the reliability and validity of the WBIS. Follow-up research could further explore the controversial aspects of the item through cognitive interviews.

Neither the initial 11-item WBIS nor the WBIS-10 with item 1 removed achieved the expected model fit has also been reported in a Norwegian adult sample [49]. As noted by Angela Meadows, the one-factor assumption oversimplifies the multidimensional nature of weight stigma internalization. The original 19-item WBIS was developed around self-devaluation [14], but subsequent studies suggest it may not fully capture self-deprecation and substantially overlaps with body image and self-esteem [50, 51, 52]. After re-evaluating the original items, Angela Meadows proposed a 13-item, two-factor WBIS [53]. Thus, even in the general population, the uni-dimensional WBIS model faces structural validity concerns. Given that weight stigma internalization is a complex concept, we suggest that the original model of the WBIS may have been too simplified to capture the complex relationships in the data. Another possible reason concerns sample specificity. The WBIS was primarily designed to measure weight stigma internalization in the general population and does not adequately account for the unique sociocultural pressures faced by postpartum women, including the expectation to return to pre-pregnancy body shape, physiological weight fluctuations, and the transition to postpartum roles [3]. These contextual characteristics reshape how weight stigma internalization is expressed, making the generic one-factor WBIS poorly suited to the psychological characteristics of postpartum women. Notably, removing the problematic reverse-scored Item 1 did not get a better model fitting. This finding further demonstrates that the structural inadequacy of the WBIS in this population is systematic rather than caused by a single item. In summary, this study supports the conclusion that the one-factor WBIS is not applicable to postpartum women. Future empirical research is needed to validate and optimize this scale specifically for the postpartum women.

Likewise, item 1 of WSSQ had negative impacts on its internal consistency. However, the bifactorial structure of the WSSQ was supported by the CFA, demonstrating a multidimensional explanation for the weight stigma internalization in postpartum women. Other studies on children, adolescents, and adults have shown that weight stigma internalization can be measured using the latent variables of self-devaluation and fear of weight stigma [19, 20, 54]. We found that item 1 (“I will always go back to being overweight”) of the WSSQ had a low level of contribution to the self-devaluation domain in our sample, which is consistent with the results of the Italian adult study [16]. One explanation is that despite being in a specific postnatal phase, most women in the study were not overweight, and item 1 did not specifically relate to postpartum women of normal weight. Furthermore, a systematic review found that when using the English and translated versions of the WSSQ compared to the original WBIS, the high degree of overlap between weight bias internalization and negative body image was smaller [55]. This avoids the jangle fallacy to some extent (i.e., constructs with different names measure the same thing). Therefore, our findings suggest that the WSSQ has satisfactory construct validity.

Our findings align with previous studies [20, 56], indicating both the WSSQ and the WBIS were significantly associated with perceived weight stigma measures. It has been suggested that the fear of enacted stigma items on WSSQ may lead to stronger associations between weight stigma internalization and perceived weight stigma [55], but this was not found in our study. Notably, after controlling for pre-pregnancy BMI and infant feeding method, the partial correlations between the two scales (WBIS and WSSQ) and perceived weight stigma increased substantially. This suggests that failing to account for these two covariates may underestimate the true relationship between perceived and internalized weight stigma. Therefore, pre-pregnancy BMI and infant feeding method should be considered as statistical adjustments in future postpartum weight stigma research. Furthermore, a systematic review by Pearl and Puhl [7] found that higher-weight individuals tended not to show a consistent linear relationship between weight and weight stigma internalization, but that this correlation was consistent across a wider range of weights. Our findings support the latter: there was a correlation between higher levels of weight stigma internalization and higher BMI in postpartum women with different weight statuses. Notably, a slightly stronger correlation was shown between WBIS and postpartum BMI compared to WSSQ. Thus, our results suggest that the WBIS is slightly more sensitive to postpartum BMI than the WSSQ.

Regarding psychological outcomes, our findings further indicated that weight stigma internalization is associated with increased psychosocial problems (depression, anxiety, and stress), which were consistent with previous studies [20, 57, 58]. In terms of eating behaviors, we replicated evidence from studies in adults that both the WSSQ and WBIS correlated significantly with emotional eating [57, 59] and restrained eating behaviors [60, 61]. However, contrary to our hypothesis, neither measure correlated with external eating behavior. This may be because external eating behavior is primarily shaped by the external environment (e.g., taste of food, aroma of food, advertisements, etc.), rather than the internal psychological state of the individual. Overall, the WSSQ and WBIS have similar sensitivity to psychological outcomes and eating behaviors in our study.

Based on the psychometric findings from the WSSQ and WBIS, we suggest that healthcare professionals and researchers prioritize the WSSQ for assessing weight stigma internalization in postpartum women due to its good psychometric properties. The WBIS also performed well in some aspects (particularly sensitive to postpartum BMI). As the WBIS did not show good construct validity in this study, we believe that further adaptation is required to make it suitable for postpartum women. Moreover, we found that studies have modified the WBIS to assess weight stigma internalization during pregnancy (e.g., replacing “weight” with “weight gain during pregnancy”) [62, 63], but there is insufficient evidence on its psychometric properties. Considering the specificity and high concern of weight stigma internalization in perinatal women, we recommend that future research develops a weight stigma internalization measurement tool specifically for perinatal women.

There are some limitations in this study. First, the study sample from a single city in China may not be representative of the entire country. Future multi-center studies are needed. Second, the cross-sectional design used does not allow for determining the direction of the relationship between weight stigma internalization and adverse health outcomes. Prospective studies are necessary to examine the causal effects of weight stigma internalization on the psychological outcomes and eating behaviors of postpartum women. Third, several variables that may influence weight stigma internalization in postpartum women—such as gestational weight gain and weeks since childbirth—were not collected in the original survey. Future research should consider incorporating these variables to provide a more comprehensive understanding.

Conclusion

In conclusion, the WSSQ is a valid measurement tool for assessing weight stigma internalization in postpartum women. The structure of the WBIS needs to be validated in further studies. Our findings provide the first evidence for the psychometric properties of the WSSQ and WBIS for weight stigma internalization in postpartum women. Future research needs to consistently evaluate measurement tools for weight stigma internalization in new samples of postnatal women.

Abbreviations

WSSQ

Weight Self-Stigma Questionnaire

WBIS

Weight Bias Internalization Scale

BMI

Body Mass Index

PWSQ

Perceived Weight Stigma Questionnaire

DASS-21

Depression Anxiety Stress Scale-21

DEBQ

Dutch Eating Behavior Questionnaire

CFI

Comparative Fit Index

TLI

Tucker–Lewis Index

RMSEA

Root Mean Square Error of Approximation

SRMR

Standardized Root Mean Square Residual

CFA

Confirmatory Factor Analysis

Authors' contributions

PT: Conceptualization, Investigation, Formal analysis, Data Curation, Writing-Original Draft, Writing-Review and Editing, Visualization. PJY: Investigation, Formal analysis, Data Curation, Writing-Original Draft, Writing-Review and Editing, Visualization. TXM: Validation, Writing-Review and Editing. SM: Conceptualization, Investigation, Writing-Review and Editing, Visualization, Supervision, Funding acquisition.

Funding

This study was supported by the Hunan Province Science and Technology Innovation Program (2026RC9046), the China Medical Board Open Competition Program (22–464).

Data availability

The datasets analyzed during the current study are not publicly available due to ethical reasons but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the School of Nursing, Central South University (Approval No. E2021108). All participants provided verbal and written informed consent before study participation.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.