Development and Validation of the Iranian Body Image Scale (IBIS)

Background: Body dissatisfaction has been known as an important public health concern which can inuence physical and psycho-social health of adolescents. Hence, health professionals need a valid and reliable tool to assess this problem and its related factors in youth. This study aims to develop and assess psychometric properties of an indigenous scale for evaluation of body image and to investigate its association with body weight status among Iranian adolescents. Methods: This study conducted on 857 adolescents who completed the Iranian body image scale (IBIS) and self-reported their body weight and height. Face, content and construct validity methods were used to assess validity of the IBIS. Exploratory factor analysis (EFA) was used to assess construct validity. Moreover, internal consistency of the scale was assessed by calculating Cronbach’s alpha coecient (α), and test-retest reliability was examined using intra-class correlation coecient (ICC). To compare the IBIS scores across sex and body weight status groups, the Mann–Whitney and Kruskal-Wallis tests were used, respectively. Results: Mean age and body mass index (BMI) of participants were 16.5±1.1 years and 22.4±4.5 kg/m 2 , respectively. The EFA showed a 10-factor construct which explained 57% of the variance. The Cronbach’s alpha coecient for overall items of the scale was 0.896 and ranged from 0.42 to 0.92 for subscales. The test retest reliability result was acceptable for the IBIS (ICC= 0.805). The ICC values ranged from 0.753 to 0.990 for IBIS subscales. The IBIS total score was signicantly higher in girls compared to boys (median (Q1-Q3): 32.2 (25.5-43.3) vs. 30.3(24.0-38.0); p<0.001). In girls, the IBIS total score was signicantly lower in normal weight girls compared to their overweight (p=0.001), and obese (p=0.001) counterparts. This difference was not observed in boys. Conclusion: These ndings support reliability and validity of the IBIS for assessment of body image in Iranian adolescents;


Introduction
Body image is an important issue during adolescence; pubertal and identity development during this stage make it a critical period for formation of negative body image [1]. A negative body image may adversely in uence physical, psychological and social aspects of adolescents' health; as ndings of previous studies have frequently demonstrated its association with risk of obesity, eating disorders, suicide ideation, low self-esteem, and depressed mood in youth [2][3][4][5][6]. Moreover, negative body image found to adversely in uence adolescents' social interactions with their parents and peers [7][8][9]. Therefore, it is important to identify those adolescents who are at risk of developing body image disturbances.
To assess different aspects of body image, various scales including the Body Shape Questionnaire (BSQ), the Drive for The Muscularity Scale (DMS), the Adolescent Body Image Satisfaction Scale (ABISS), the Body Image Coping Strategies Inventory (BICSI), the Body image scale (BIS), the Multidimensional Body-Self Relations Questionnaire-Appearance Scales (MBSRQ-AS), and the Eating Attitudes Test (EAT-26) have been developed [10][11][12][13][14][15][16]. Most of these instruments have emphasized on evaluation of a unique dimension of this concept including concerns, attitude and behaviors, satisfaction, or coping strategies related to an individual's body image. Past studies conducted on Iranian adolescents with a focus on body image topic, have also been con ned to measurement of certain components of body image such as body dissatisfaction, body image coping strategies, and eating disorders [17][18][19]. Since, body image is a multidimensional constructs and there is consensus about the multi-dimensional properties of this concept by body image scholars [20][21][22]; developing a comprehensive tool covering various dimensions of body image can help experts to better understand and assess this concept in related research.
Previous studies endorsed a signi cant association between individuals' perceptions regarding their bodies and the cultural context of society [23][24][25]. Since, most of the existing scales are developed in other countries, mainly Western ones [10-13, 15, 16], they cannot be favorably be applied in our indigenous population. Therefore, considering lack of a comprehensive and indigenous tool for assessment of body image in Iranian adolescents; developing such scale seems essential. Hence, the current study aimed to rst, develop and assess psychometric properties of an indigenous scale for evaluation of body image in adolescents. Then, its second aim was to investigate the association between body image and body weight status among a sample of Iranian adolescents.

Methods
Participants A total of 903 adolescents (aged 15-18 years) residing in Tehran participated in the current study. A total of 46 adolescents were excluded from the analysis due to having chronic mental or physical diseases (such as depression, social anxiety, anxiety and stress, convulsion, cancer, diabetes, heart diseases, kidney diseases and thyroid disorders); hence, data of 857 adolescents were analyzed.
Prior to data collection, ethic approval was obtained from the Ethics Committee of the Research Institute for Endocrine Sciences (RIES) a liated to Shahid Beheshti University of Medical Sciences, Tehran, Iran. Furthermore, the approvals were obtained from the Ministry of Education in Tehran and selected high schools. All participants provide a written informed consent.

Measurements
Participants were asked to answer a set of questions regarding socio-demographic variables such as their age, sex, and history of diseases as well as their parents' marital status, level of education, and job status and self-report their weight, and height. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m 2 ). The BMI-for-age national percentiles were used to determine the body weight status of adolescents. Underweight, normal weight, overweight and obesity were de ned as BMI values of <5 th percentile, ≥5 th to <85 th percentile, ≥85 th percentile to <95 th percentile, and ≥ 95 th percentile, respectively [26]. In addition, participants were also asked to ll out the Iranian body image scale (IBIS).

Development and scoring of the IBIS
The initial item pool consisted of 95 items that were developed by the authors based on the ndings of a qualitative study and review of existing literature on body image topic. For all subscales except for "emotions and behaviors", a ve-point Likert scale from one to ve was used for scoring answers for each item, where the choice of an answer for "completely agree" was given a ve point and "completely disagree" was given one points for items belonged to "social models", "appearance importance in social interactions", and "perceived cultural values" subscales. Remaining items belonged to other subscales are scored reversely. For "emotions and behaviors" subscale, the choices ranged from "always" to "never" with values of ve and one for always and never, respectively. Then for better interpretation, the 1-5 point scale items are transformed to 0-100 as follows: 1=0, 2=25, 3=50, 4=75, and 5=100. To calculate subscale and total scores of the IBIS, the mean is computed as the sum of the items over the number of items. Hence, the scores ranged from 0 to 100. A higher total score indicates a more negative body image.

Validity assessment
Validity of the scale was assessed using face, content and construct validity methods as described below.

1) Face validity
Face validity was assessed using qualitative and quantitative methods. In qualitative method, a total of ten adolescents were asked to assess di culty, relevancy and ambiguity the preliminary scale. In quantitative face validity method, the same participants were asked to rate each item of the scale by the 5-point Likert scale from completely important (score 5) to not at all important (score 1). Then, the item impact score was calculated using the below formula: Item impact score=frequency (%)×importance Frequency (%) indicates the number of participant who gave the item a score of 4 or 5. Items with an impact score of more than 1.5 were considered appropriate and maintained for next stage.

2) Content validity
Content validity of the scale was assessed by a panel of experts in different disciplines including health education, community nutrition, sociology, psychology and medical sciences. For qualitative content analysis, experts were asked to comment on style, wording and scoring of the items. In addition, for quantitative content analysis, content validity ratio (CVR) and content validity index (CVI) were calculated. To calculate CVR, rstly, each item was scored using a 3-point Likert scale (essential, useful but not essential, not essential) by each expert. Then, using the below formula the CVR was calculated: In this formula, N is the total number of expert panel and ne is the number of individuals who considered the relevant item "essential". Using the Lawshe's table, the CVR higher than 0.62 for 10 individuals (based on number of experts in the current study) indicates the necessity of the item [27].
For calculating CVI, experts were asked to comment independently on degree of the relevance, clarity, and simplicity of each item using a 4-point Likert scale ("not at all" to "completely"). Then, using the below formula CVI was calculated.
Scale's content validity index (S-CVI) was calculated by taking the sum of all item CVIs divided by the total number of items.

3) Construct validity
Construct validity of the scale was evaluated using exploratory factor analysis (EFA). For this purpose, the Principal component analysis with Varimax rotation was used. The Kaiser-Meyer-Olkin (KMO) and Bartlett Sphericity Test were used to show the sampling adequacy. The KMO value of ≥0.8 and the p value for the Bartlett Sphericity test of <0.05 indicate the sampling adequacy for EFA [28]. Scree plot, eigenvalues greater than 1, and number of factors that explain >50% of variance were used to predict the number of factors retained. In this analysis, items with factor loadings over cutoff values of 0.4 were considered important and remained in the model.

Reliability assessment
Reliability of the scale was assessed using Cronbach's alpha coe cient (α) and intra-class correlation coe cients (ICCs).

1) Internal consistency
The Cronbach's alpha coe cient (α) was used to assess the internal consistency of the scale. The Cronbach's alpha coe cients were calculated for total and subscale scores of the BIS and α values greater than 0.7 were considered acceptable [29].
2) Test-retest reliability Test-retest reliability was determined using the intra-class correlation coe cients (ICCs). For this purpose, the scales were completed by 30 adolescents within a 10-14 days interval and intra-class correlation coe cients of total and subscale scores of the two tests were calculated and values greater than 0.5 were considered acceptable [30].

Data analysis
To analyze data, the SPSS software (version 21.0) was used. Descriptive statistics including mean±sd for continuous variables, and frequency (%) for categorical variables were reported. The Chi square test was used to investigate the association between qualitative variables. To compare total and subscale scores of the IBIS between boys and girls, the Mann-Whitney test was used. In addition, the Kruskal-Wallis test was applied to compare total and subscale scores of the IBIS among body weight status groups. For post hoc test, each two groups were compared using the Mann-Whitney test considering corrected p values (the Bonferroni correction method). P values <0.05 were considered statistically signi cant.

Descriptive statistics
Mean age and BMI of participants were 16.5±1.1 years and 22.4±4.5 kg/m 2 , respectively. Descriptive statistics for socio-demographic variables and body weight status of study participants are presented in table 1. There were no signi cant differences between boys and girls in terms of parental marital status, level of education and working status. About half of both mothers and fathers had academic degrees. Most mothers were housewife and about one third of fathers were employee and about half of them were self-employed. In terms of distribution of body weight status, there was a signi cant difference between boys and girls.

Validity assessment
In the quantitative face validity assessment, the importance of each item was assessed and items with impact score of <1.5 were eliminated. In this stage, one item was eliminated. In qualitative content validity assessment, a total of 29 items were deleted due to having overlaps with other items or assessing lots of unnecessary details. Another ve items were deleted because they did not obtain acceptable CVI and CVR levels in quantitative assessment of content analysis.
Then, a total of 60 items entered the construct validity assessment. In EFA, another 8 items were removed due to having low values of communality or factor loading or both. For 52 items (Supplementary table 1) remained in EFA, the KMO value of 0.901, and the Bartlet's sphericity test (p<0.001) con rmed sampling adequacy for EFA. The extracted ten factors were based on scree plot and eigenvalues >1.00. A 10-factor structure explained about 57% of total variance. Factor loadings based on rotated factor matrix and explained variance of each factor were reported in table 2. The factor loading of all items ranged from 0.464 to 0.807 on their corresponding factor.

Reliability assessment
The number of items, intra-class correlation coe cients (ICCs), and Cronbach's alpha (α) for each subscale were reported in  Table 4 shows the median and interquartile range (Q1-Q3) for subscale and total scores of the Iranian body image scale (IBIS). Girls had higher IBIS total score compared to boys. Moreover, except for social models, perceived social support, empowerment, and body evaluation subscales, other subscale scores of the IBIS were signi cantly different in boys and girls; except for appearance importance in social interactions subscale, girls had signi cantly higher scores, compared to boys.

Body image in body weight status groups
In boys, there were no signi cant differences in IBIS total score among different body weight status groups. However, there were signi cant differences in IBIS total score among underweight, normal weight, overweight and obese girls. Based on ndings of post hoc tests, the IBIS total score was signi cantly lower in normal weight compared to overweight (p=0.001), and obese (p=0.001) girls. In terms of subscales, there were signi cant differences in empowerment and body evaluation subscale scores among different body weight stats groups in boys. Further post hoc tests indicated that underweight boys had better scores in empowerment subscale compared to their normal weight (p=0.005) and obese (p=0.006) counterparts; and normal weight had signi cantly better scores in body evaluation subscale compared to obese boys (p=0.002). In girls, there were signi cant differences in body evaluation and emotion and behaviors subscale scores among different body weight stats groups. Further post hoc tests indicated that normal weight had signi cantly better scores in body evaluation subscale compared to obese girls (p=0.004); furthermore, normal weight girls had better scores in emotions and behaviors subscale compared to both overweight (p<0.001) and obese (p<0.001) girls.

Discussion
The current study reported development and psychometric properties of the IBIS, a 52-item scale that can be applied for assessment of body image and its related factors in adolescent boys and girls. The current ndings support validity and relativity of the IBIS. Moreover, current ndings indicated sex and body weight status were signi cantly associated with body image in adolescents. Girls had a more negative body image compared to boys, and overweight and obese girls had signi cantly more negative body image compared to their normal weight counterparts.
Although several tools exist for evaluation of body image; most of them were designed to assess certain dimension of body image. For example, Body Shape Questionnaire (BSQ) is a 34-item scale that measures body shape concerns which can associate with development, maintenance and treatment of eating disorders [10]. The Drive for The Muscularity Scale (DMS) is another scale which focused on attitudes and behaviors related to muscularity. This scale expresses the extent an individual is preoccupied with increasing muscularity and a higher score show more drive for muscularity [11]. The 16-item Adolescent Body Image Satisfaction Scale (ABISS) has been developed to assess body image satisfaction in male adolescents and encompasses three subscales of body competence, body inadequacy and internal con ict [12]. The Body Image Coping Strategies Inventory (BICSI) is a 29-item instrument that developed for assessment of three main coping strategies including avoidance, appearance xing, and positive rational acceptance which are used for management of threats or challenges related to an individual's body image experiences [13]. The Multidimensional Body-Self Relations Questionnaire-Appearance Scales (MBSRQ-AS) is a 34-item instrument which covers several dimensions of body image, and includes ve subscales: appearance evaluation, appearance orientation, overweight preoccupation, self-classi ed weight, and body areas satisfaction scale. It is one the most comprehensive instrument that is widely used across different countries [16]. The Body image scale (BIS) is a 35-item instrument that has been developed for assessment of body dissatisfaction in young adult female in Pakistan. This scale encompasses three main subscales including physical component, psychological component, and strategies used to maintain one's body image [14]. However, the IBIS has been developed using the ndings of a qualitative study focused on exploring different aspects of body image from the perspective of Iranian adolescents which makes it a valuable tool for assessment of adolescents' body image with a more comprehensive view. The IBIS provides a deeper and greater insight to perception of body image and its underlying factors in adolescents which can be used in designing intervention programs targeting adolescents to modify their perceptions, emotions and behaviors towards their bodies.
The IBIS can be applied as a valid and reliable instrument for assessment of different aspects related to body image in adolescents. This 52-item 10 factor solution which has been emerged in EFA can explain an acceptable amount of variance of body image (more than 50%) in participants. The factor entitled as "Emotions and behaviors" with 15 items explained the highest variance followed by "Body evaluation" and "Personal characteristics and strategies" as the second and third factors. Results of assessing internal consistency and test-retest reliability for the whole scale (IBIS), indicate high reliability of the scale for assessment of body image in youth. For IBIS subscales, the ICC values exceeded the acceptable value of 0.5 [30]. In terms of internal consistency of subscales, Cronbach's alpha coe cients for all subscales exceeded the acceptable value of 0.7 [29], except for "priority of health and spiritually", "perceived cultural values", and "empowerment" subscales. The Cronbach's alpha coe cients of 0.68 and 0.67 for the rst two mentioned subscales, were approximately near the acceptable threshold of 0.7. However, the value of 0.42 for "empowerment" subscale showed poor internal consistency for this subscale. In such cases, it is more common to delete the relevant item in the subscale in order to increase the Cronbach's alpha. As there are only two items in this subscale, this approach cannot be applied. We could either remove the subscale or consider another grouping for items. We preferred to retain the related items; hence, considering the existing subscales, we arranged another grouping and merged these two items with existing items in "perceived social support" subscale and then assess the validity and reliability of the new subscale. The performed con rmatory factor analysis (CFA) for the 6-item proposed latent construct and internal consistency assessment showed satisfactory results as follows. The CFA results indicated acceptable t for proposed construct (χ 2 /df= 1.25, GFI= 0.996, CFI= 0.998, RMSEA= 0.017); in addition, the Cronbach's alpha for the new construct increased to 0.67.
Based on ndings of the current study, girls had signi cantly a more negative body image compared to boys. In line with the current ndings, previous studies reported a more negative body image and a lower levels of body satisfaction in females compared to males [4,31,32]. In this regards, ndings of a qualitative study conducted among 12-20 years Iranian adolescents reported that most of the adolescent girls did not have positive feelings about their bodily changes during puberty which led them to feel nervous or ashamed [31]. Hence, in the current study, having a more negative body image in adolescent girls compared to boys, could be a result of body image dissatisfaction due to pubertal changes and its negative in uence on their psychological well-being.
In the current study, overweight and obesity were associated with a more negative body image in girls, but not in boys. These ndings may be due to perceptions of different ideals with respect to body shape and size by adolescent girls and boys [32][33][34]. In this regards, previous ndings demonstrated that while girls tend to be dissatis ed with their bodies when their BMI is average or above average; boys tend to be dissatis ed with their body when they have a BMI either below or above average [33,34]. In another study, girls with excessive weight were more likely to be dissatis ed with their bodies, compared to overweight boys [32]. In addition, previous studies in western societies have shown that, whereas most girls prefer a slim body shape and a smaller body size, most boys prefer a muscular body shape and a larger body size.
In agreement with the current ndings, a previous qualitative study conducted among adolescents residing in Tehran, a number of overweight/obese adolescent boys had a positive self-image rather than a negative one. This perception rooted in beliefs such as having higher resistance to illness and physical blows, and having similar ability to compete in sports and physical activities compared to their normal weight counterparts [35]. This may explain why body image did not differ among overweight and obese boys compared to their normal weight counterparts in the current study.
To the best of our knowledge, this is the rst study that develop a comprehensive scale for assessment of body image in Iranian adolescents covering several aspects related to body image perceptions including social factors, personal characteristics and strategies, attitudes, evaluation, as well as emotions and behaviors. The limitations of this study should also be taken into consideration. First, we did not measure adolescents' body weight and height, and the BMI was calculated using the self-reported data. Hence, some adolescents may misreport their weight and height which may result in misclassi cation of their BMI.
Second, due to the cross-sectional nature of the current study, causal inferences about body image in relation to body weight status is not possible. Finally, participants of the current study were recruited from Tehran (an urban community); therefore, it is recommended to conduct similar study on adolescents residing in suburban and rural areas to replicate the current ndings.

Conclusion
In conclusion, the current ndings indicated that the IBIS can be used as a valid and reliable tool for assessment of body image and its related factors among Iranian adolescents by researchers and healthcare professionals. This scale can help professionals to identify those adolescents who are susceptible to develop a negative body image and its related health consequences like eating disorders. Moreover, in the current study, girls had more negative body image compared to boys. Additionally, overweight and obese girls had more negative body images compared to their normal weight counterparts. Therefore, it is recommended that body image to be incorporated as an important component of future health promotion programs with a focus on adolescent girls as target population, speci cally in those interventions targeting overweight and obesity during adolescence.
Abbreviations BMI: Body mass index; CFA: Con rmatory factor analysis; CVI: Content validity index; CVR: Content validity ratio; EFA: Exploratory factor analysis; ICC: Intra-class correlation coe cient; IBIS: Iranian body image scale Declarations This study was approved by the research ethics committee of the Research Institute for Endocrine Sciences (RIES), Shahid Beheshti University of Medical Sciences. Written informed consent was obtained from all participants. For participants less than 18 years of age/illiterates, informed consent was obtained from their parents/legally authorized representative.

Consent for publication
Not applicable.

Availability of data and materials
Data used in the current study would be available from corresponding author on reasonable request.