Psychometric properties of the Chinese version of the Body Esteem Scale for Adolescents and Adults (BESAA) among adolescents

Abstract

Background

Adolescents’ body image concerns are related to their mental and physical health. Reliable instruments are fundamental to understanding body image concerns, but there are some concerns about the adolescent body image instruments currently used in China. One of the obvious concerns is the inadequate psychometric properties of the extant Chinese translations of the Body Esteem Scale for Adolescents and Adults (BESAA), such as the lack of evaluation of factorial validity and construct validity. BESAA is an instrument that encompasses partly positive body image and partly negative body image, and the broad conceptualization of appearance is considered one of its unique values. The main purpose of this study was to translate the BESAA into Chinese and to preliminarily evaluate its reliability and validity in a sample of Chinese adolescents.

Methods

A total of 1368 adolescents (age: mean = 14.94, standard deviation = 2.08; 70.18% females, 29.82% males) were recruited through a convenience sampling method. Exploratory Factor Analysis (EFA), Confirmatory Factor Analysis (CFA), Exploratory Structural Equation Modelling (ESEM), test–retest reliability, convergent validity, measurement invariance, and McDonald’s Omega were used to evaluate the psychometric properties of the Chinese version of the BESAA. The total sample was used for the item analysis, convergent validity, and reliability. The total sample was randomly divided into two equal-sized subsamples; Sample 1 (n = 684; girls n = 459, boys n = 225) was used for the EFA, while Sample 2 (n = 684; girls n = 501, boys n = 183) was used for the CFA, the ESEM, and the measurement invariance.

Results

A three-factor, 15-item model was obtained through EFA. The model showed satisfactory goodness of fit in terms of both CFA and ESEM (χ² = 172.071, df = 63, χ²/df = 2.73 (p < 0.001), Comparative fit index (CFI) = 0.972, Tucker-Lewis index (TLI) = 0.953, Root mean square error of approximation (RMSEA) = 0.050, and Standardized root mean residual (SRMR) = 0.024). The Chinese version of the BESAA demonstrated good internal consistency (McDonald’s Omega: 0.881), and its convergent validity is supported.

Conclusion

The current findings provide evidence for the validity and reliability of the Chinese version of the BESAA, and support the use of the BESAA among Chinese adolescents.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-025-02955-y.

Background

Research on body image has a rich history and in recent years has flourished in a variety of cultural contexts. Body image is a holistic concept that includes both positive body image (e.g., appreciation, respect, and acceptance of the body as it is) [1] and negative body image (conceptualized as poor body esteem, body dissatisfaction, or body shame) [2]. Positive and negative body image are independent constructs that do not have a negative linear relationship [3, 4]. Prior studies have shown that negative body image is a prevalent issue in adolescent populations globally [5], in both developed and developing countries [6].

China is the world’s second most populous country, with 158 million teenagers, according to the Seventh National Census in 2020 [7]. Recent studies have shown that body image concerns are widespread among Chinese adolescents, and that both male and female adolescents experience body dissatisfaction to varying degrees [8, 9]. A study of children and adolescents aged 8–12 in Guangzhou, China, found that 78.10% of them had varying degrees of body dissatisfaction [10]. Negative body image is harmful to adolescents. Studies among Chinese adolescents have found that body dissatisfaction is associated with low self-esteem [11], social anxiety [12], depressive symptoms [13], and loneliness [13]. Moreover, studies have shown that Chinese adolescents adopt unhealthy behaviors for self body image regulation, including restrained eating [14] and increased willingness to undergo cosmetic surgery [15]. These findings suggest that the body image concerns among Chinese adolescents is serious, and that it is associated with adolescents’ mental health and maladaptive behaviors. Reliable instruments are fundamental to understanding body image concerns and developing relevant interventions. However, there are some concerns about the adolescent body image instruments currently used in China.

A variety of instruments are currently being used to assess body image in Chinese adolescents. The more widely used instruments include the Multidimensional Body-Self Relations Questionnaire (Original version: 69 items; Chinese version: 93 items) [16, 17], the adolescent students’ physical self scale (34 items) [18], the Negative Physical Self Scale (NPSS) (48 items) [19], and the Body Esteem Scale for Adolescents and Adults (BESAA) (Original version: 23 items) [20]. These instruments have made important contributions to the field, but their limitations need to be noted. Some of these instruments are lengthy [17–19]. Instruments with more items take more time to complete, and tend to have more missing data and higher refusal rates [21]. Shorter instruments would reduce the burden on participants, lower the cost of data collection and potentially improve data quality. Some instruments have specificities, such as the NPSS, which is only used to assess negative aspects of body image [19]. Given this, there is a need for an instrument that is relatively short and encompasses broad aspects of body image, and the BESAA fulfills this need.

The BESAA was developed as a self-assessment measure for Canadian adolescents and adults in 2001 [20]. The BESAA measures an individual’s self-evaluation of their body or appearance and indirectly reflects their body satisfaction or dissatisfaction [20]. Body dissatisfaction is defined as the discrepancy between the ideal body and the actual body [22]. It is important to note that body dissatisfaction and body esteem are different and independent constructs. The BESAA comprises 23 items, divided into the three subscales of appearance (evaluation of general feelings and satisfaction with overall appearance, without specifying a particular appearance ideal), weight (evaluation of general feelings and satisfaction with weight), and attribution (evaluations attributed to others about the respondent’s appearance). The original BESAA developed by Mendelson has been translated and validated in different cultural contexts, including Indonesian [23], Portuguese [24], Turkish [25], Spanish [26], Arabic [27], Persian [28], and more. The English version has also been adapted to meet the needs of adolescents in urban India [29]. Validation studies have shown acceptable internal consistency of the BESAA in other cultures, but the factor structure has changed. A large number of versions showed a new structure that differed from the original three-factor structure, which consisted of the Appearance-Positive/Appearance Satisfaction, Appearance-Negative/Appearance Anxiety, and Weight factors [23, 24, 26, 28]. These findings indicate that the BESAA may be a culturally sensitive instrument, implying the need for rigorous psychometric evaluation before using it in different cultural contexts. However, the Chinese versions of the BESAA currently in use have some issues that cannot be ignored.

To the authors’ knowledge, there are three Chinese translations of the BESAA [30–32]. These translations were used directly after forward translation and provided limited information on psychometric properties. In fact, most of them only had information about internal consistency, without any information about the exploration of the factor structure of the BESAA. Furthermore, the samples used for validation were mostly college students, and the applicability of these translations among adolescents is unknown [30–32]. Given the limitations of the current Chinese translations of the BESAA (i.e., may not be applicable to adolescents and their inadequate psychometric properties), the strengths of the BESAA compared to other instruments (i.e., brief but comprehensive), and the unique value and importance of the BESAA (i.e., broad conceptualization of individual appearance and assessment of cognitive and affective body image [20, 24]; one of eight key measures of body image in the field [33]), it would be useful to develop a Chinese version of the BESAA that follows the good practice guidelines [34].

This study aimed to (1) investigate the factor structure of the BESAA using Exploratory Factor Analysis (EFA), Confirmatory Factor Analysis (CFA) and Exploratory Structural Equation Modelling (ESEM); (2) examine its measurement invariance across gender; (3) examine its test–retest reliability and internal consistency; and (4) test its convergent validity by examining correlations of the BESAA with positive affect and life satisfaction. We anticipated good validity and reliability of the Chinese version of the BESAA and hypothesized that body esteem is positively related to positive affect and life satisfaction.

Method

Participants

Chinese adolescents aged 11–19 years were recruited between April and July 2024 from three public schools in Fuzhou City, Fujian Province, Huainan City, Anhui Province, and Huizhou City, Guangdong Province. The inclusion criteria were: (1) age 10–19 years [35]; (2) basic literacy in the Chinese language; and (3) voluntary participation. The exclusion criterion was: (1) a history of mental disorders or currently diagnosed with a mental disorder (including but not limited to schizophrenia, bipolar disorder, etc.).

Differing criteria for sample size adequacy in EFA and CFA can be found in the literature. For EFA, the most commonly used guideline is a 10:1 participant-to-item ratio [36, 37]. Worthington and Whittaker [38] suggested that, if the item communalities are ≥ 0.50 or there are 10:1 items per factor with factor loadings of around 0.40, then a sample size of 150–200 may be adequate; more generally, a sample size of 300 or more is sufficient. For CFA, a sample with a participant-to-item ratio of at least 5:1 and a sample size greater than 100 is considered to have an adequate sample size. Therefore, a minimum of 400 samples was required for this study.

Procedure

The researchers contacted schools in China and recruited adolescents through a convenience sampling method. The researchers communicated with school headmasters and grade directors, providing them with Chinese versions of the BESAA for their review and explaining the main aims of the study. The questionnaires and informed consent forms were distributed to the students, with the consent of the headmaster. Students under 18 years took the questionnaires and informed consent forms home on the same day and completed them with the consent of at least one parent. On the following day, the students brought their questionnaires to school to be collected by the researchers. To prevent adolescents with a history of mental disorders or a current diagnosis of a mental disorder from feeling discriminated against, their questionnaires were collected and labelled as well. The students were informed of their rights and obligations and the anonymity of their data before participating in the study. They were told they had the right to refuse to answer entries that caused discomfort and were free to ask any questions they had about the study. In addition, the students were informed they had the right to withdraw from the study at any time without reason. Two weeks after the first survey, some participants from one of the schools were invited to complete the questionnaire again, for purposes of assessing test–retest reliability.

Translation and adaptation

Following Swami and Barron’s recommendations regarding translation and changes of cultural context, the following process was used to translate the original English version of BESAA into Chinese [34, 39]. First, the scale was independently forward-translated into Chinese by two master’s students who were informed of the BESAA, and two bilingual certified translators who were not, and all four of whom were native Chinese speakers. Second, the differences between the four independent translations and the original version were discussed and reviewed by the researchers and the four forward-translators to synthesize a unified translation. Third, a PhD in medicine and a native English translator, both uninformed of the purpose of the study, translated the forward-translated Chinese version of the BESAA back into English. Fourth, a psychologist, an expert on Chinese language and literature, and the six translators involved in the previous steps compared and reviewed all the translated versions to produce the first draft of the Chinese version of the BESAA. The back-translated version was also provided to the original author of the BESAA, and adjustments were made to the Chinese version based on the original author’s suggestions to ensure semantic equivalence. Finally, the researchers conducted pre-tests and cognitive interviews with 20 adolescents (10 girls, 10 boys) aged 11–19 years using the first draft.

During the review and the first round of cognitive interviews, the experts, headmasters, teachers, and adolescents pointed out some problems with items 10, 17, and 20 of the original scale. Specifically, they considered that item 10 “I really like what I weigh” duplicated item 8 “I am satisfied with my weight”. The meanings expressed by these two items are especially similar in Chinese, and so they were merged. They also noted that item 17 “I feel ashamed of how I look”, might arouse feelings of shame or negative self-evaluation in adolescents. The experts and teachers considered that this item may not be appropriate for use in a non-clinical setting and suggested that this item be deleted. Item 20 “My looks help me to get dates” is problematic because Chinese adolescents do not usually date openly. Most elders (e.g., parents, educators) do not promote dating behaviour until children reach adulthood. In China, dating and and “early love” are considered to be essentially the same. To avoid possible misunderstandings or encouraging inappropriate behaviour, this item was changed to “My looks make me popular”. The above adjustments were made with the consent of the original author of the scale, resulting in a 21-item Chinese version of the BESAA. Following this, a second round of cognitive interviews (four girls, four boys) was conducted using the 21-item Chinese version of the BESAA. Adolescents who participated in the second round reported no further problems, and the headmasters then agreed to officially test the adapted version in the school.

Instruments

Body esteem

The BESAA was developed by Mendelson in 2001 to assess the level of body esteem [20]. The original version of the BESAA has 23 items, and the adapted Chinese version with 21 items (one merged and one deleted) was used in this study. All items are scored on a five-point Likert scale from 0 = never to 4 = always. The original version of the BESAA has good internal consistency (Cronbach’s alpha was 0.81 to 0. 94).

Positive affect

The Chinese version of the Positive Affect (PA) subscale of the Positive Affect and Negative Affect scale for Children-10 (PANAS-C10), developed by Wei in 2017, was used to assess the level of positive affect [40]. The PANAS-C10-PA has five items that assess five aspects of an individual’s positive affect (e.g., joyfulness, activeness). Each item is rated on a five-point Likert scale ranging from 1 = not even a little to 5 = very much. Adolescents were asked to indicate the extent to which they had experienced each emotion in the past 2 weeks. The mean scores of the items were used as an indicator of the participants’ positive affect, with higher scores representing stronger positive affect. The PANAS-C10-PA has been validated for use with Chinese adolescents, demonstrating good internal consistency (Cronbach’s alpha = 0.91) [40].

Life satisfaction

The Satisfaction With Life Scale (SWLS) was developed by Diener in 1985 [41]. It is used to assess an individual’s level of life satisfaction. The SWLS comprises five items (e.g., “In most ways my life is close to my ideal”) which are scored on a seven-point Likert scale ranging from 1 = strongly disagree to 7 = strongly agree. An individual’s overall level of life satisfaction is calculated as a total score, with higher scores indicating higher levels of life satisfaction. The Chinese version of the SWLS was developed by Xiong in 2009 [42]. The Chinese version of the SWLS has good internal consistency (Cronbach’s alpha = 0.78) and has been validated in adolescents (Cronbach’s alpha = 0.87) [42, 43].

Demographics

Participants reported their age, gender, and ethnicity.

Data analysis

EFA was performed using SPSS version 26. CFA and measurement invariance tests were performed using AMOS version 24. Exploratory structural equation modelling (ESEM) was performed using Mplus version 7. The skewness and kurtosis of these items were evaluated based on Kline’s reference values [44]. If the skewness value is less than 3 and the kurtosis value is less than 8, it is considered acceptable. The total sample was used for calculation of the item analysis, convergent validity, and reliability. The total sample was randomly divided into two subsamples of equal size: Sample 1 (n = 684; girls n = 459, boys n = 225) was used for the EFA, while Sample 2 (n = 684; girls n = 501, boys n = 183) was used for the CFA, the ESEM, and measurement invariance.

Content validity

Eight experts (including a psychologist and a Chinese language and literature expert who were involved in the translation process, and six additional experts (from the fields of psychology and public health)), were invited to assess the scale. Each expert evaluated the translation accuracy, semantic clarity, and cultural appropriateness of each item based on their theoretical knowledge and practical experience.

Item analysis

The items were analyzed using the critical ration method and the correlation coefficient method with data from the total sample (n = 1368; girls n = 960, boys n = 408) as follows: (1) Critical ration method: the questionnaire scores of all participants were ranked in descending order. The data in the top 27% were classified as the high group, while the data in the bottom 27% were classified as the low group. An independent samples t-test was conducted, and items with p > 0.05 or t < 3 were deleted. (2) Correlation coefficient method: the correlations between each item of the questionnaire and the total score (i.e., the item-total correlation (ITC)) were calculated, and items with p > 0.05 or r < 0.30 were deleted.

Factorial validity

For EFA, the Kaiser–Meyer–Olkin (KMO) method and Bartlett’s test of sphericity were used to examine the factorability of the data. A KMO value greater than or equal to 0.70 is considered to be the ideal value, the items meeting that criterion are considered suitable for factor analysis [45]. Factor analysis was then conducted using principal axis factoring and oblique (promax) rotation, as the factors of BESAA were expected to be correlated. The Guttman-Kaiser criteria were used to initially determine the number of factors to be retained; the criterion for retention was an eigenvalue greater than 1 [46, 47]. Parallel analysis was used to further identify factors to be retained, as this is known to be one of the most accurate methods of factor retention [48, 49]. The following criteria were used for item deletion: a) items with low loadings on the primary factor (i.e., less than 0.5) [50], and b) items that cross-loaded on a secondary factor (i.e., greater than 0.30 and/or the difference between the primary and secondary loadings was less than 0.20) [51].

CFA was used to test the fit of the factor models derived from EFA. The factor loadings were used as local indices of goodness of fit, as was the ratio of chi-square to degrees of freedom (χ²/df), the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). The model’s fit was considered good if CFI ≥ 0.95, TLI ≥ 0.95, RMSEA ≤ 0.07, and SRMR < 0.08; whereas CFI and TLI ≥ 0.90, and RMSEA and SRMR ≤ 0.10 were taken to indicate a merely adequate fit [52–55]. A χ²/df ratio close to or less than 2 was taken to indicate a good model fit, whereas values less than 5 indicated an acceptable model fit [56]. When the initially hypothesized model was revealed to have an inadequate fit, the modification indices were used to improve the overall model fit.

ESEM, a popular alternative to CFA, was used to test model fit. ESEM has advantages over CFA when cross-loading is present, because CFA limits the cross-loadings to zero, resulting in a high degree of correlation among the factors, whereas ESEM estimates all the cross-loadings [57].

Measurement invariance

Sample 2 was used to determine the gender invariance of the Chinese version of the BESAA. The invariance was tested at three levels: configural (i.e., whether the numbers of factors and loading patterns were the same across the two gender groups), factor loading (i.e., whether the magnitudes of the factor loadings were the same across the two gender groups), and intercept (i.e., whether the regression intercepts were the same across the two gender groups) [58]. We assessed several criteria as well, including the chi-square difference (i.e., ∆χ²), CFI, SRMR, and RMSEA model fit indices. A significant change in the χ² value at p < 0.05 is a sign of non-invariance. For testing loading invariance, ∆CFI < 0.010, ∆RMSEA < 0.015, or ∆SRMR < 0.030 indicate that measurement invariance holds; for testing intercept invariance, ∆CFI < 0.010, ∆RMSEA < 0.015, or ∆SRMR < 0.010 indicate that measurement invariance holds. These criteria were developed by Chen [58] for use with adequate samples (i.e., total sample size > 300). After establishing intercept invariance, the scale scores from the Chinese version of the BESAA were compared between the two gender groups to assess the effect size using an independent samples t-test and Cohen’s d (0.20, 0.50, and 0.80 are the cutoff criteria for small, moderate, and large effects, respectively) [59, 60].

Internal consistency and test–retest reliability

McDonald’s Omega coefficient was used to assess internal consistency and a value greater than 0.8 indicates satisfactory reliability [61, 62]. The intraclass correlation coefficient (ICC) was used to test the reliability of the retest after two weeks. The ICC values were interpreted according to the guidelines provided by Portney and Watkins [63]: an ICC value of 0.75 or higher is considered to be excellent; 0.50 to 0.75 is moderate or good; 0.25 to 0.50 is fair; and < 0.25 is weak.

Convergent validity

The convergent validity of the Chinese version of the BESAA was assessed by calculating the Pearson correlation coefficients between the BESAA and the measures of positive affect and life satisfaction. Absolute correlation coefficients (r) values < 0.3, 0.3 < values < 0.5, and values > 0.5 are considered weak, moderate, and strong correlations, respectively [59].

Ethics and participants’ consent

This study was approved by the Branch of Medical Research and Clinical Technology Application, Medical Ethics Committee, First Affiliated Hospital of Fujian Medical University ((2024QT) No. (111)). All participants provided informed written consent after being informed of the aims of the research, and all study data were kept confidential. For adolescents under 18 years, informed consent was obtained from the adolescent and at least one parent. All procedures were followed in accordance with the Declaration of Helsinki.

Results

Sociodemographic characteristics of the samples

Of the total of 1508 participants who agreed to the face-to-face survey and completed the questionnaire, 15 selected an incorrect response to the attention check item, 8 had an excessive number of missing values, and 117 selected the same responses over a large number of items. After excluding these respondents, data from the remaining 1368 participants who responded to all scale items were retained for analysis. The final sample comprised 408 men and 960 women, with a mean (M) age of 14.94 years (standard deviation (SD) = 2.08). All participants were born in China. A summary of the participants’ demographic characteristics can be seen in Table 1.

Table 1.

Participants’ demographic characteristics (N = 1368)

Characteristic		n(%)
Gender	Male	408(29.82)
	Female	960(70.18)
Age	11	10(0.73)
	12	346(25.29)
	13	69(5.04)
	14	94(6.87)
	15	88(6.43)
	16	355(25.95)
	17	342(25.00)
	18	62(4.53)
	19	2(0.14)
Nation	Han ethnic group	1337(97.73)
	Hui ethnic group	26(1.90)
	Buyei ethnic group	1(0.07)
	Hmong or Miao ethnic group of southwest China	1(0.07)
	Yi ethnic group	1(0.07)
	She ethnic group	1(0.07)
	Dai ethnic group	1(0.07)

The total sample of 1368 participants had no missing data. Table 2 presents the descriptive statistics (i.e., M, SD, skewness, and kurtosis) of all items for the total sample of the Chinese version of BESAA. The absolute values of skewness and kurtosis for all items were within the acceptable range of normal distribution.

Table 2.

The descriptive statistics of all items for the total sample of the Chinese version of BESAA

Item	Total sample (n = 1368) M (SD)	Girls’ sample (n = 960) M (SD)	Boys’ sample (n = 408) M (SD)	Skewness (total sample)	Kurtosis (total sample)	ITCa (total sample)
Item 1	2.01 (1.17)	2.23 (1.09)	1.50 (1.20)	0.05	-0.64	0.42
Item 2	1.74 (1.07)	1.86 (1.00)	1.46 (1.18)	0.34	-0.26	0.595
Item 3	1.41 (1.16)	1.44 (1.11)	1.35 (1.28)	0.59	-0.31	0.688
Item 4	2.42 (1.23)	2.38 (1.20)	2.52 (1.30)	-0.33	-0.84	0.388
Item 5	0.78 (1.00)	0.78 (0.95)	0.77 (1.12)	1.36	1.48	0.408
Item 6	1.95 (1.23)	2.19 (1.14)	1.38 (1.25)	0.06	-0.86	0.499
Item 7	2.30 (1.25)	2.14 (1.22)	2.67 (1.27)	-0.23	-0.90	0.3
Item 8	1.63 (1.25)	1.65 (1.20)	1.58 (1.36)	0.40	-0.77	0.697
Item 9	1.28 (1.26)	1.08 (1.15)	1.74 (1.37)	0.63	-0.65	0.169
Item 10	3.16 (1.15)	3.12 (1.13)	3.27 (1.18)	-1.38	1.05	0.254
Item 11	1.38 (1.10)	1.50 (1.02)	1.12 (1.21)	0.65	-0.05	0.571
Item 12	3.12 (1.03)	3.00 (1.04)	3.41 (0.95)	-1.08	0.59	0.494
Item 13	1.61 (1.18)	1.65 (1.13)	1.53 (1.30)	0.48	-0.52	0.643
Item 14	1.86 (1.20)	1.88 (1.15)	1.81 (1.33)	0.24	-0.70	0.752
Item 15	1.72 (1.30)	1.73 (1.27)	1.70 (1.37)	0.27	-0.95	0.599
Item 16	2.94 (1.21)	2.82 (1.22)	3.22 (1.14)	-0.98	-0.02	0.532
Item 17	3.02 (1.17)	2.93 (1.18)	3.22 (1.13)	-1.09	0.33	0.541
Item 18	1.31 (1.08)	1.33 (1.01)	1.25 (1.24)	0.74	0.16	0.632
Item 19	2.76 (1.13)	2.57 (1.11)	3.20 (1.04)	-0.68	-0.27	0.538
Item 20	1.54 (1.18)	1.56 (1.12)	1.49 (1.30)	0.48	-0.49	0.721
Item 21	1.46 (1.19)	1.44 (1.13)	1.49 (1.33)	0.61	-0.38	0.718

^athe correlations between each item of the questionnaire and the total score (the item-total correlation)

Content validity

The content validity index of the individual items (I-CVI) ranged from 0.875–1.00, and the content validity index of the entire scale (S-CVI) was 0.958.

Item analysis (total sample)

The critical ration results indicate that all entries exhibited significant differences between the high-score and low-score groups. Pearson’s correlation coefficient was used to assess the correlations between each item of the scale and the total score. As shown in Table 2, most items were significant, except for items 9 and 10. Those two items were recommended for deletion because their ITC values were less than 0.30. However, the final decision to delete these two items was postponed until after the EFA, because relying on ITC values alone as a deletion criterion is not recommended.

Exploratory factor analysis (Sample 1)

The KMO (0.903) and Bartlett’s test of sphericity (χ² (210) = 7161.77, p < 0.001) results indicated that the Chinese version of the BESAA was suitable for factor analysis. When EFA was conducted on the first subsample, parallel analysis was used to assess the number of factors to be retained. The results are presented here as a scree plot, which indicated that three factors (Appearance-Positive, Appearance-Negative, and Weight) should be retained. These three factors explained 56.593% of the variance.

The results of the EFA showed low factor loadings or cross loadings for some items. For example, items 8 “I am satisfied with my weight” (0.430(Appearance-Positive)-0.464(Weight)), 15 “I feel I weigh the right amount for my height”(0.384(Appearance-Positive)-0.335(Weight)), and 20 “I think I have a good body”(0.529(Appearance-Positive)-0.404(Weight)) cross-loaded on two factors, and items 4 “I am preoccupied with trying to change my body weight”(0.364) and 10 “I wish I looked like someone else”(0.381) had low factor loadings. Therefore, these items were individually removed over multiple rounds of EFA. The three-factor, 16-item model was found to fit the data poorly: χ² = 835.651, df = 101, χ²/df = 8.274 (p < 0.001), CFI = 0.862, TLI = 0.836, RMSEA = 0.103, and SRMR = 0.1012.

Two problematic items (i.e., items 9 and 10) were identified during the item analysis. The justification for the deletion of item 10 is described above; the deletion of item 9 resulted in an improved model fit: χ² = 671.088, df = 87, χ²/df = 7.714 (p < 0.001), CFI = 0.884, TLI = 0.860, RMSEA = 0.099, and SRMR = 0.0834. A three-factor 15-item model was finally obtained. The factor loadings for that model ranged from 0.523 to 0.933 (Table 3).

Table 3.

Factor loadings for the Chinese version of the BESAA (EFA)

Item	APa	ANb	Wc
1 我喜欢照片里的自己 I like what I look like in pictures	0.702
2 别人认为我长得好看 Other people consider me good looking	0.772
3 我为自己的身体感到骄傲 I’m proud of my body	0.643
5 我认为我的外表有助于我找到一份工作 I think my appearance would help me get a job	0.562
6 我喜欢镜子里的自己 I like what I see when I look in the mirror	0.765
11 同龄人喜欢我的长相 People my own age like my looks	0.811
13 我和大多数人一样好看 I’m as nice looking as most people	0.635
14 我很满意自己的长相 I’m pretty happy about the way I look	0.642
18 我的长相让我很受欢迎 My looks make me popular	0.707
21 我看起来和我想象中的一样好看 I’m looking as nice as I’d like to	0.662
7 如果可以的话, 我会对自己的容貌做出很多改变 There are lots of things I’d change about my looks if I could		0.523
12 我的长相让我很难过 My looks upset me		0.665
19 我担心自己的长相 I worry about my looks		0.803
16 称体重让我很沮丧 Weighing myself depresses me			0.933
17 我的体重让我不开心 My weight makes me unhappy			0.858

^aAppearance-Positive subscale

^bAppearance-Negative subscale

^cWeight subscale

Confirmatory factor analysis (Sample 2)

Judged according to the criteria proposed by Hu and Bentler (1999) [54], the three-factor original model obtained by EFA did not have a good fit to the data from Sample 2, which was used for CFA: χ² = 576.703, df = 87, χ²/df = 6.629 (p < 0.001), CFI = 0.895, TLI = 0.874, RMSEA = 0.091, and SRMR = 0.0790. The modification indices (MIs) were then scrutinized to determine whether freeing the error covariances would improve the model fit. This procedure identified three pairs of items (Items 1 and 5, 2 and 6, 6 and 9) whose residuals were correlated. This may be due to the introduction of method effects when the scale items were adjusted for purposes of cross-cultural adaptation. Specifically, similar content used between items may have led to method effects. It is common and sometimes essential to re-correct of the model using correlated error covariances. After freeing the error covariances between item pairs 1 and 5, 2 and 6, and 6 and 9, the modified model had adequate fit: χ² = 407.381, df = 84, χ²/df = 4.85 (p < 0.001), CFI = 0.931, TLI = 0.914, RMSEA = 0.075, and SRMR = 0.0756. Therefore, the re-specified model was retained, with error covariances between the three pairs of items allowed.

ESEM (Sample 2)

To further validate the factor structure, ESEM analysis was conducted, based on the three-factor 15-item model obtained from EFA by target rotation. The fit indices improved compared to the fit of the two models in CFA: χ² = 172.071, df = 63, χ²/df = 2.73 (p < 0.001), CFI = 0.972, TLI = 0.953, RMSEA = 0.050, and SRMR = 0.024. The factor loadings for each item in the ESEM are shown in Supplementary Material 1.

Measurement invariance (Sample 2)

Table 4 presents the cross-gender measurement invariance analysis conducted on the Chinese version of the BESAA. The RMSEA and SRMR values indicate that the Chinese version of the BESAA showed gender invariance at the configural, metric, and scalar levels, while the CFI value indicated that the Chinese version of the BESAA showed gender invariance only at the configural and metric levels. These results suggest that caution is needed when comparing the total scores of females and males, as the gender invariance of the Chinese version of the BESAA is not fully supported. Chen stated that caution should be used when applying establishment criteria for measurement invariance, and that many factors (e.g., pattern of non-invariance, sample size, ratio of sample size, and model complexity) can affect the magnitude of the change in the fit statistics [58]. With this in mind, one should remember that these criteria may not be absolute.

Table 4.

Tests of measurement invariance of the Chinese version of the BESAA

Model	χ2	df	CFI	RMSEA	95%CI	SRMR	∆χ2	∆df	∆CFI	∆RMSEA	∆SRMR
Females and males
Configural	441.274	168	0.94	0.049	0.043–0.054	.0710
Metric	468.153	180	0.937	0.048	0.043–0.054	.0735	26.878	12	0.003	0.001	0.0025
Scalar	636.359	195	0.903	0.058	0.053–0.063	.0738	168.206	15	0.034	0.010	0.0003

Reliability (Total sample)

The McDonald’s Omega values of the Chinese version of the BESAA are presented in Table 5. The internal consistency of each subscale was found to be good. The females’ test–retest reliability was 0.79 (n = 107), and the males’ test–retest reliability was 0.74 (n = 46).

Table 5.

McDonald’s Omega of the Chinese version of the BESAA

Variables (females/males)	McDonald’s Omega
Appearance-Postive	0.915 (0.912/0.922)
Appearance-Negative	0.839 (0.833/0.820)
Weight	0.960 (0.962/0.952)
BESAA	0.881 (0.888/0.883)

Convergent validity (Total sample)

Table 6 presents the correlation coefficients between the Chinese version of the BESAA and the other research variables. The total BESAA score was moderately positively correlated with positive affect and life satisfaction in both females and males. The Appearance-Positive subscale of the BESAA was moderately positively correlated with positive affect and weakly positively correlated with life satisfaction in both females and males. The Appearance-Negative subscale of the BESAA was moderately negatively correlated with positive affect and life satisfaction in females, but only weakly negatively correlated with positive affect and life satisfaction in males. The Weight subscale of the BESAA was weakly negatively correlated with positive affect and life satisfaction in both females and males. These findings provide support for the convergent validity of the Chinese version of the BESAA.

Table 6.

Pearson correlations of the BESAA subscales with PANAS-C10-PA and SWLS

Variables (females/males)	APa	ANb	Wc	BESAA
PANAS-C10-PA	0.373**(0.432**/0.353**)	-0.297**(-0.300**/-0.208**)	-0.191**(-0.193**/-0.208**)	0.435**(0.472**/0.383**)
SWLS	0.238**(0.287**/0.251**)	-0.398**(-0.413**/-0.260**)	-0.226**(-0.227**/-0.254**)	0.359**(0.396**/0.311**)

^aAppearance-Positive subscale

^bAppearance-Negative subscale

cWeight subscale

^**p < 0.01

Discussion

The present study is the first to culturally adapt and preliminarily validate the BESAA in Chinese adolescents. Our findings provide evidence for the psychometric properties of the Chinese version of the BESAA. Specifically, we evaluated the factor structure, internal consistency, test–retest reliability, convergent validity, and gender invariance of the Chinese version of the BESAA.

A three-factor, 15-item model was obtained by EFA and found to be a good fit to the data in CFA and ESEM. The three factors are Appearance-Positive (10 items), Appearance-Negative (3 items), and Weight (2 items). The factor structure of the Chinese version of the BESAA is similar to that of the Indonesian, Portuguese, Spanish, and Persian versions [23, 24, 26, 28], but different from that of the original version. This finding is interesting. In terms of the construct of the Chinese version of the BESAA, the presence of the Appearance-Positive subscale and the Appearance-Negative subscale strongly supports the proposition that positive and negative body image may be two distinct constructs [24, 26, 64]. Furthermore, the Turkish version replicated the original three-factor structure of the BESAA in children, and the Arabic version did the same in adolescents [25, 27]. It is important to note that these two versions differ from the others in their method of structural exploration. These two versions used principal component analysis (PCA) to explore the structure of the BESAA rather than EFA [25, 27]. PCA tends to overestimate factor loadings relative to EFA and underestimate the strength of correlations between factors [34]. In the case of using the same sample dataset, the results obtained through PCA may be different from those obtained through EFA. Therefore, this may be one of the important reasons for the emergence of different factor structures of the BESAA.

Six items were deleted from the factor analysis, namely items 4, 8, 9, 10, 15, and 20. Items 8 “I am satisfied with my weight”, 15 “I feel I weigh the right amount for my height”, and 20 “I think I have a good body” loaded onto the Appearance-Positive subscale and the Weight subscale. This phenomenon is conceptually plausible, as these items refer to both the positive emotional experiences that individuals have with their bodies and the weight-related psychological concerns. Moreover, the finding on item 20 was also observed in Brazilian adolescents [24]. Item 5 “I think my appearance would help me get a job” had a good factor loading, which is inconsistent with the Latin-American Spanish version [26], but consistent with the Brazilian Portuguese, Indonesian, Turkish, and Arabic versions [23–25, 27]. A potential explanation might be that there are cross-age differences in respondents’ understanding of this item. Only the Spanish version was developed among adults, while the other versions were developed among adolescents or children [26].

In addition to examining the factor structure of the BESAA in samples of Chinese adolescents, this study also aimed to develop a gender invariant version of the BESAA that would allow comparisons of total scores across genders. Although the RMSEA and SRMR values indicate that the three-factor, 15-item Chinese version of the BESAA has measurement invariance across gender, the CFI value exceeds the recommended cut-off range. This may be because the items in the current version of the Chinese version of the BESAA are not sufficient to capture body esteem as perceived by both females and males. Chinese adolescent females and adolescent males are known to differ significantly in their negative perceptions of their appearance [65]. In China, adolescent females are more likely than males to be dissatisfied with their appearance and tend to have a stronger tendency toward appearance perfectionism, whereas adolescent males are more likely to show dissatisfaction with their weight and are less concerned about appearance dissatisfaction than females are [65–67]. Future studies could use qualitative methods in a bottom-up approach to explore how adolescents of different genders conceptualize body esteem. This may provide new items for the Chinese version of the BESAA, which may improve the overall quality of the scale. Furthermore, given the known limitations of testing measurement invariance using multigroup CFA, it is essential that future studies test the gender invariance of the Chinese version of the BESAA using other methods such as multigroup ESEM and multiple-indicators multiple-causes modelling. To the authors’ knowledge, it appears that only the Spanish version establishes gender invariance through multigroup CFA [26].

For the convergent validity, BESAA scores were moderately positively correlated with positive affect and life satisfaction, thus demonstrating good convergent validity. For Chinese adolescents, the strongest correlation was found between positive affect and the Appearance-Positive subscale, which is consistent with the findings in Brazilian adolescents [24]. Life satisfaction had the strongest correlation with the Appearance-Negative subscale of the Chinese version of the BESAA. This relationship is not easy to explain as there does not seem to be a similar finding before. In sum, the findings on convergent validity may provide valuable information for the future development of body image intervention programs for adolescents.

Limitations

This study has some limitations that should be acknowledged. First, our sample was recruited from three provinces in southern China. To improve the generalizability of the data, future studies should recruit participants from more geographically diverse adolescent populations. Second, The evidence for convergent validity provided in this study may not be comprehensive. Only the Chinese PANAS-C10 positive affect subscale was used in this study. This is because the negative emotions subscale was not approved by the school headmasters. Moreover, future use of measures of body appreciation, self-esteem, or eating disorders that are directly related to body esteem may strengthen the evidence for construct validity. Third, the present study was a cross-sectional study, and thus was not sensitive to changes in body esteem over time. Future longitudinal studies could provide a deeper understanding of how body esteem develops over the course of adolescence in Chinese adolescents. Finally, the gender invariance of the Chinese version of the BESAA has tentatively been established but needs to be further tested, as the change values of the CFI did not meet acceptable standards. We suggest that future studies use other methods (e.g., multi-group ESEM or multiple-indicators multiple-causes modelling) to test for gender invariance with gender-balanced samples in the future.

Conclusion

In summary, the current findings provide evidence for the validity and reliability of the Chinese version of the BESAA, and support the use of the BESAA among Chinese adolescents. It is important to note that the evidence on gender invariance is weak, and future studies should recruit participants from more geographically diverse adolescent populations.

Abbreviations

BESAA

Body Esteem Scale for Adolescents and Adults

EFA

Exploratory factor analysis

CFA

Confirmatory factor analysis

ESEM

Exploratory structural equation modelling

ICC

Intraclass coefficient

ITC

Item-total correlation

χ²/df

Chi-square to degrees of freedom

CFI

Comparative fit index

TLI

Tucker-Lewis index

RMSEA

Root mean square error of approximation

SRMR

Standardized root mean residual

Authors’ contributions

KZ, DL and QW conceptualized the study. KZ, DL and AH analysed, interpreted the data and wrote the main manuscript. QW reviewed the related literature and wrote the manuscript. DL, KZ, AH, LY, HH, JF, FL collected and curated the data. AH edited figures. FL and NY contributed to analysis and interpretation. KZ got project-administrated, supervised the study, and reviewed the manuscript critically. The authors have read and approved the final manuscript.

Funding

Not applicable.

Data availability

The datasets used and analyzed in the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Branch of Medical Research and Clinical Technology Application, Medical Ethics Committee, The First Affiliated Hospital of Fujian Medical University ((2024QT) No. (111)). All participants submitted an informed consent form after being informed of the aims of the study, and minors (< 18 years old in China) participated after obtaining informed consent from their parents (guardians). The study data were kept confidential All procedures were followed in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.