This systematic review aimed to identify the current practices of cross-cultural adaptation of body image assessment scales for young university students. The results showed a variety of practices, which will be discussed below along with recommendations for future investigations on the topic.
Analyzing the studies included in this review, we found that the process of cross-cultural adaptation of body image scales for young university students was first recorded in 2010, followed by an increase in studies over the next few years, with peaks in publications in the years 2013 and 2014, 2016 and 2017, bringing together the largest amount of research.
Another point to be highlighted in this review concerns the locations where the studies were carried out. Brazil appears to be the main place for carrying out this type of research, with 8 out of the 14 studies selected in this review being developed in this country. This finding allows the inference of two counterpoints. On the one hand, Brazil appears to be a powerful example of quantitative reference in the researchers’ instrumentalization through the process of cross-cultural adaptation for the assessment of body image among young university students. This can enable the expansion of cross-cultural studies carried out in the country.
On the other hand, it is possible to state that Brazilian researchers may be too much focused on “importing” instruments created in other countries and perhaps not as well involved in the creation of new assessment scales specific to the Brazilian population and cultural context. Although it is not possible with this study to make a comparison of cross-culturally adapted instruments created in Brazil, there is a great involvement of national researchers in the cross-cultural adaptation of instruments, when compared to researchers from other countries. Morgado et al. (2014) consider that the creation of new measures is recommended when there are no instruments already created and validated correctly in other countries. Perhaps, this justifies the fact that Brazil opts to carry out cross-cultural adaptation of instruments, rather than the creation of them. Possibly, on account of the fact that the international literature has shown an abundance of instruments aimed to evaluate different facets of university students’ body image. Additionally, cross-cultural adaptations enable multicultural studies, in the sense that data comparison among different groups becomes possible (Carvalho, Amaral, & Ferreira, 2014), which might be another justifying reason why Brazil leads the ranking of cross-cultural adaptations.
In this regard, Campana and Tavares (2009) add that Brazil has a lack of instruments about body image, and making a cross-cultural adaptation is important to expand the possibilities of research in our country. These notes and Brazil’s growing interest in researching body image justify Brazil’s leadership in translation, cross-cultural adaptation and validation of already existing instruments in the literature.
Additionally, we observed that Beaton’s guide (Beaton et al., 2000; 2002; 2007) was the most used methodological guide for cross-cultural adaptations of instruments, reported in previous studies as useful and usual in the scale adaptation procedure (Argyrides, Kkeli, & Kendeou, 2014; Swami & Barron, 2019). It describes a cultural adaptation model in medical, sociological, and psychological literature, which is a complete adaptation process that includes translation alongside semantic, idiomatic, and experiential adaptations, as well as the conceptual equivalence between the original instruments and their adaptations (Hendricson et al., 1989; Swami & Barron, 2019; Swami et al., 2019).
Another used guide was that of Guillemin et al. (1993), which was based on previous studies in psychology and sociology. It differs from Beaton’s guide as it also recommends, if relevant, the re-examination of the questionnaire scores beyond the proposed steps. Moreover, two other methodological guides appeared in three studies selected in this review. Both follow the same proposal as the other five guides already highlighted above, and therefore, they will not be detailed. Among the studies gathered in this review, one did not clarify which guides it used in the methodological process. However, it is widely recommended that a methodological guide be used to perform the cross-cultural adaptation of instruments in the area of body image (Swami & Barron, 2019), in order to enable the maintenance of cultural, semantic, idiomatic, conceptual, and experimental equivalences in the adapted instrument (Guillemin, 1995).
Nevertheless, Swami and Barron (2019) argue that the best practices indicate that a combination of translation procedures should be adopted as there is no consensus in the literature about a single guide to be used. According to these authors, while Beaton et al. (2000) provides a structure that several body image scholars will find useful due to its clarity in the steps of the methodological process (Argyrides et al., 2014; Carbonneau et al., 2016; Swami et al., 2019), this structure also requires time and effort, which can lead some researchers to view combined translation techniques as less demanding alternatives in the process of cross-cultural adaptation (Swami & Barron, 2019).
Among the 14 studies included in this review, different psychometric instruments were chosen to be translated and adapted through the cross-cultural adaptation process. These instruments have been commonly used to assess the following components of body image, or associated aspects: sociocultural attitudes related to appearance (Amaral et al., 2011; Barra et al., 2019; Conti et al., 2010); thoughts and desires related to body image (Chakroun-Baggioni et al., 2017; Shoji et al., 2018); body dissatisfaction or feelings directed towards muscularity (Campana et al., 2013; de Carvalho et al., 2013; Rousseau et al., 2014); eating attitudes and behaviors (Bas et al., 2017; Carbonneau et al., 2016; Ulian et al., 2017); dissatisfaction with weight and body shape (Silva et al., 2016); sexual activity (Pakpour et al., 2014); and body change (Conti et al., 2012).
At this point, an interest in the multiple dimensions of the body image held by university students is observed. This can be justified by the fact that body image is a multifaceted and complex construct, requiring different instruments to fully assess. It is also noteworthy that body image is an essential phenomenon in several aspects of human life, ranging from biological aspects related to health and diseases, to psychosocial aspects such as quality of life (Campana & Tavares, 2009; Cash & Pruzinsky, 2002; Thompson, 2004).
Among the two instruments cited as the target of more than one study, the results showed good evidence of psychometric qualities. The Intuitive Eating Scale-2 (Bas et al., 2017; Carbonneau et al., 2016), adapted for Turkey and Canada, assesses the tendency of individuals to follow their tracks of hunger and satiety related to when and how much to eat. The instrument showed good validity and reliability, confirming the psychometric properties of the original study. The Male Body Dissatisfaction Scale (de Carvalho et al., 2013; Rousseau et al., 2014) was adapted for France and Brazil to assess body dissatisfaction towards muscularity. Its internal consistency has been proven, demonstrating the instrument’s good reliability.
The study that evaluated the Tripartite Influence Scale describes different stages of the cross-cultural adaptation of the scale. Conti et al. (2010) performed the translation, back-translation, and assessment of the verbal comprehension of the instrument. Similarly, the study by Amaral et al. (2011) described the stages of adaptation of the Sociocultural Attitudes Towards Appearance Questionnaire-3 to the Portuguese language. Both studies, as well as others identified (Conti et al., 2012; de Carvalho et al., 2013; Ulian et al., 2017), evaluated only the content validity of the instruments. This can be a limitation since the safe use of measurement instruments for a given population is directly related to the assessment of its psychometric qualities. It is worth noting that some of the adapted instruments in the aforementioned studies continued in future studies that assessed their psychometric qualities; this was the case with the Sociocultural Attitudes Towards Appearance Questionnaire-3 (Amaral et al., 2011), from the Tripartite Influence Scale (Amaral et al., 2011), and the Male Body Dissatisfaction Scale (Carvalho et al., 2015).
Furthermore, most of the studies selected in this systematic review recruited a sample of both sexes (N = 10), while two studies were conducted only with women, and the other two studies were conducted strictly with men. This may indicate greater sample heterogeneity, which is considered adequate in the processes of cross-cultural adaptation as it often presents different results and possible new findings, thus representing more reliable results (Barra et al., 2019; Dahl, Wickman, & Wengström, 2014). However, some authors argue that sample heterogeneity can represent a very large diversification of results, which does not represent the totality that is expected to be measured. According to these authors, this can hinder the consensus of the results (Bas et al., 2017; Ferreira, Corazza, Francisco, & Neves, 2018; Swami et al., 2019). Therefore, future studies should consider the pros and cons of recruiting homogeneous or heterogeneous samples, with regard to population sex. These findings are in line with what has been pointed out in previous literature, which has found that, especially in the last two decades, both men and women are dissatisfied with their bodies (de Carvalho & Ferreira, 2014; Hobza & Rochlen, 2009; Kelley, Neufeld, & Musher-Eizenman, 2010), which justifies the availability of instruments for both sexes. Therefore, it is necessary to choose appropriate body image assessment instruments for each of these groups, investigating the specificities of the construct for men and women. It also justifies the use of the psychometric processes of scales for both men and women (Beaton et al., 2000; Gardner & Brown, 2010; Thompson, 2004).
Regarding the assessment of validity, this psychometric quality determines whether, in fact, the test measures what it is supposed to measure (Pasquali, 2009). Barra et al. (2019), Bas et al. (2017), and Pakpour et al. (2014) pointed out in their studies the strength of the instrument to be considered valid and reliable, based on the aforementioned validities, assessed in the study. Among the main validities developed in these studies are the content validity, CFA, construct, and discriminant validities. It is worth noting that CFA (50%) was used more than EFA (28.5%). This can be justified by the fact that the factorial structure of the instruments that have undergone the process of cross-cultural adaptation has already been tested in other countries. Then, the CFA would serve to confirm or refute the previous factorial structure. A similar previous review study concerning the development of scales found different results, with EFA being more recurrent than CFA (Morgado et al., 2014). In comparison with this previous study, we can infer that the process of creating a new instrument requires exploring the new factorial structure, which justifies the EFA. Studies with cross-cultural adaptation demand to confirm or refute previous factorial models. This fact justifies greater use of CFA in the context here investigated.
Reliability is considered one of the main quality assurance criteria of the instrument, as it refers to the ability of the test to measure without errors (Pasquali, 2009; Primi, 2012). This criterion focuses on obtaining consistent and reproducible results when measuring a given attribute (Fayers & Machin, 2007). The most used reliability tests in the studies were internal consistency and test-retest. These tests are related to the coherence, consistency of the results, and the confidence that the test inspires in measuring the phenomena without large fluctuations between repeated measures, which would reflect the presence of measurement errors. In other words, these procedures estimate the level of accuracy of a test and establish an expectation of how erroneous the measurement can be (Kimberlin & Winterstein, 2008; Martins, 2006; Pasquali, 2009; Primi, 2012). Therefore, imprecise or incomplete psychometric procedures, performed in this methodological process, will probably bias the results (Morgado et al., 2014). At this point, it is highlighted that specifically, one study did not specify, or did not clearly show, which reliability method it used (Conti et al., 2012).
Regarding the limitations, three studies did not make this information clear in the text. Therefore, it is possible to highlight a diversity of specific limitations highlighted in each included study. The only limitation that is repeated in the studies (n = 4, 28.7%) is related to the characteristics of the sample, highlighting the homogeneity of the sample because the results cannot be reproduced and generalized to other groups (Bas et al., 2017; Carbonneau et al., 2016; Pakpour et al., 2014; Ulian et al., 2017).
Moreover, the following limitations also stand out: (a) lack of validity and reliability. However, several authors emphasize that validity and reliability are psychometric attributes that must be present in a good instrument and must be considered in studies of psychometric adaptation, since their absence can generate errors in the statistical conclusions of the research and can be considered a limitation of the developed study (Cunha, de Almeida Neto, & Stackfleth, 2016; Pilatti, Pedroso, & Gutierrez, 2010; Swami & Barron, 2019), and (b) the cross-sectional model used in the studies, which does not allow the establishment of cause and effect relationship, representing a certain vulnerability to the simultaneity bias. However, although this does not allow a definition of the temporal sequence between the events studied, we can point out positive points of this methodology, highlighting that it is a method widely used in studies which examine the relationship between events, being simple, low cost, and objective in data collection (Chakroun-Baggioni et al., 2017; Gonçalves & Silvany, 2013).
When we observe the main strengths described in the included studies, the following stand out: (a) the instruments are valid and reliable (Barra et al., 2019; Bas et al., 2017; Conti et al., 2010); (b) they have excellent verbal comprehension (Amaral et al., 2011; Conti et al., 2010, 2012; Ulian et al., 2017); and (c) they promote advances in research (Campana et al., 2013; de Carvalho et al., 2013; Ulian et al., 2017). We can infer, from the results, that researchers recognize the importance of psychometric qualities and also a meticulous semantic adaptation process as outstanding aspects in their studies, together with the recommended procedures for carrying out the process of cross-cultural adaptation (Beaton et al., 2000; Swami & Barron, 2019).
Although several authors recognize the importance of the psychometric qualities of scales, we were able to observe flaws in this process when analyzing the statistical results of the factorial structure of instruments. For example, some studies (35.7%) did not do any type of factorial analysis of the instruments. Others used only EFA or CFA. Among those that performed EFA, most studies presented adjusted reference values. However, one study made no mention of such values at all. Among those who did CFA, some studies presented inadequate reference values, according to what is recommended in the literature, and this can be considered a problem because it weakens the instrument's validation process. Additionally, only two studies combined EFA and CFA. For more consistent results on the psychometric indices of a scale, Morgado et al. (2014) indicate the combined use of EFA and CFA. Most of the studies included in this review are in the opposite direction of the abovementioned recommendation.
Regarding internal consistency, which is the most widely used measure of reliability (Morgado et al., 2014), only one study failed to present this value. Among the others, they all presented values that represent what is recommended in the literature. This measure was the reference in the studies that described it.
In relation to the quality of the articles included, it can be noted that the studies with the lowest scores were the oldest (date of publication between 2010 and 2013). The concentration of studies with low quality in this period may be linked to the increase of studies in the area after 2016, which may have had better support for research through previous studies, greater methodological rigor, as well as advances in research related to the constructs in question. It is important to note, therefore, the importance of publishing the limitations and difficulties found in the studies, in order to further improvements in the development of future research and prevent the same errors from being made by new researchers.
It is also noteworthy that the articles with high-quality scores presented an explicit methodological rigor, as well as a detailed presentation of the information about their research processes. This might offer a better understanding of the study for potential readers, as well as better reliability of the information about the research.
This review has some limitations that must be considered. Initially, when selecting articles, studies may not have been included due to the search terms used, which may have impacted the results. Furthermore, another limitation that needs to be highlighted is the fact that the initial inclusion of the studies was made based on the title. This may have contributed to the loss of studies on cross-cultural adaptation since the title does not always represent the entire content of an article. Finally, our study is only current until February 2020. Studies published after this period could not be included in this review.
Despite these limitations, this study presents important contributions. First, this systematic review analyzed consistent information about the literature in question, updated until February 2020, making the article up to date. In addition, this review presents both the evolution of studies on body image among university students, as well as a broad discussion on this theme, permeating the different constructs that encompass body image. Thus, it is expected to assist future researchers in the knowledge and choice of body image assessment instruments for the public in question.