Adaptation of the Eating Expectancy Inventory and Thinness and Restricting Expectancy Inventory to the Turkish, and Examination of Psychometric PropertiesSeda Sapmaz Yurtsever, Günay Serap Tekinsav Sütcü
Understanding the cognitive, social, and emotional motivations underlying eating behaviors is important. Equally important, we must recognize eating-related problems sooner and more definitively to develop appropriate intervention programs. In this study, it was aimed to adapt two important inventories into Turkish. The first of these was the Eating Expectancy Inventory (EEI), which evaluates the different expectations underlying eating behavior. Also, adaptation of the Thinness and Restricting Expectancy Inventory (TREI), which evaluates the motivations underlying dietary behavior and thinness, into Turkish was aimed. Following the adaptation, their psychometric properties were examined. The sample consisted of 329 female undergraduate students. In addition to the previously mentioned inventories, Eating Disorder Examination Questionnaire, Three-Factor Eating Questionnaire, Bulimic Investigatory Test Edinburgh, and Palatable Eating Motives Scale were used to test the construct validity of the adapted scales. Confirmatory factor analyses conducted separately revealed that both adapted scales have consistent factor structures with the original formats. For the Turkish form of EEI (T-EEI), five-factor [χ2 (94, N = 329) = 198,71 p <.000, χ2 /df = 2.11, GFI =.93, NFI =.94, CFI = .97 RMSEA=.058]; for the Turkish form of TREI (T-TREI), one factor [χ2 (2, N = 329) = 5,318 p <.000, χ2 /df = 2.66, GFI =.99, NFI =.99, CFI = .99 RMSEA=.071]. Reliability analyses showed that the internal consistency coefficients for the sub-dimensions of T-EEI ranged between .75 and .95, and the test–retest reliability scores ranged between .63 and .83. The study found that the internal consistency coefficient of T-TREI was .98, and its test–retest reliability coefficient was .70. Also, significant positive correlations were found between the scales used to test the construct validity and the measurement tools. Results showed that Turkish version of both inventories had good reliability and valid psychometric properties. Both adapted measures can be used as an effective assessment tool in diagnosing and treating of eating disorders.
Yemeden Beklentiler Ölçeği ile Zayıflıktan ve Yemeyi Kısıtlamaktan Beklentiler Ölçeğinin Türkçeye Uyarlanması ve Psi̇kometri̇k Özelli̇kleri̇ni̇n İncelenmesiSeda Sapmaz Yurtsever, Günay Serap Tekinsav Sütcü
Yeme bozuklukları, en temelde aşırı yeme ve yiyeceği aşırı kısıtlama olmak üzere iki farklı davranış örüntüsü ile kendini göstermektedir. Yeme bozukluklarına uygun müdahale yaklaşımlarının geliştirilmesinde yeme davranışının altında yatan bilişsel, sosyal ve duygusal motivasyonların anlaşılması ve yeme ile ilgili belirli problemlerin daha iyi tanınması önemlidir. Bu bakımdan bu çalışmada yeme davranışı ile ne tür kazanımların hedeflendiğini değerlendiren Yemeden Beklentiler Ölçeği (EEI) ile diyet davranışı ve zayıflık isteğinin altındaki beklentileri değerlendiren Zayıflıktan ve Yemeyi Kısıtlamaktan Beklentiler Ölçeği’nin (TREI) Türkçeye uyarlanması ve ölçeklerin psikometrik özelliklerinin incelenmesi amaçlanmıştır. Çalışmanın örneklemini 329 kadın üniversite öğrencisi oluşturmaktadır. Söz konusu ölçeklere ek olarak, ölçeklerin yapı geçerliklerini sınamak amacıyla Yeme Bozuklukları Değerlendirme Ölçeği, Üç Faktörlü Yeme Ölçeği, Edinburgh Bulimiya Araştırma Testi ve Lezzetli Şeyler Yeme Nedenleri Ölçekleri kullanılmıştır. Ölçekler için ayrı ayrı yürütülen doğrulayıcı faktör analizleri, EEI’nın Türkçe Formu (T-EEI) için beş faktör [χ2 (94, N = 329) = 198,71 p <.000, χ2 /df = 2.11, GFI =.93, NFI =.94, CFI = .97 RMSEA=.058]; TREI’nın Türkçe formu (T-TREI) için ise tek faktör [χ2 (2, N = 329) = 5,318 p <.000, χ2 /df = 2.66, GFI =.99, NFI =.99, CFI = .99 RMSEA=.071] olmak üzere ölçeklerin orijinal faktörleri ile tutarlı yapıları ortaya koymuştur. Güvenirlik analizleri sonucunda, T-EEI’nın alt boyutları için Cronbach alfa iç tutarlılık katsayılarının .75 ile .95 arasında, test-tekrar test güvenirlik değerlerinin ise .63 ile .83 arasında değiştiği saptanmıştır. T-TREI’nin Cronbach alfa iç tutarlık katsayısının .98, test tekrar test güvenirlik katsayısının ise .70 olduğu bulunmuştur. Ölçüm araçlarının yapı geçerliklerini sınamak amacıyla kullanılan ölçeklerle pozitif yönde anlamlı korelasyonlar elde edilmiştir. Sonuçlar, her iki ölçeğin de Türkçe formlarının geçerli ve güvenilir olduğunu ortaya koymuştur. Alanyazınla tutarlı sonuçlardan hareketle, uyarlanan ölçüm araçlarının yeme bozuklukların teşhisi ve tedavisinde etkili birer değerlendirme aracı olarak kullanılabileceği önerilmektedir.
According to Tolman’s (1932) expectancy learning theory, what determines our behavior is our expectation for the outcome of our action. These expectations are behavior-outcome association learnings that one acquires through directly experience or observation in the early years of development (Miller et al., 1990). Many believe that certain expectations about eating and dieting may have an effective role in developing and maintaining eating disorders (Annus et al., 2007; Pearson et al., 2012). According to this view, a person has learned to use eating or diet behavior with the aim of meeting many expectations such as relaxation, relieving negative emotion, feeling safer, and being more attractive. Indeed, research in the literature revealed that unrealistic expectations regarding eating are strongly correlated with anorexia nervosa, bulimia nervosa, and binge eating disorder (Hohlstein et al., 1998; Williams-Kerver et al., 2019). Further, they mediate the effect of other factors such as negative impulsivity on eating disorders (Fischer et al., 2013; Pearson et al., 2012; Pearson et al., 2018; Smith et al., 2018). Studies have also shown that eating expectations also predict the course of treatment of the disorder (Fitzsimmons-Craft et al., 2013) and that symptoms of eating disorder are reduced by changing expectations relating to thinness (Annus et al., 2008). In this respect, understanding the expectations placed on eating, dieting, and thinness appear to be extremely important in prevention and treatment of eating disorders.
Accordingly, Hohlstein and colleagues (1998) developed two different inventories to understand the underlying motivations of individual eating and dieting behaviors. To make use of this work, the lack of a measurement in Turkish for such assessments was the starting point of this study. Therefore, this study aimed to adapt these two inventories, the Eating Expectancy Inventory (EEI) and the Thinness and Restricting Expectancy Inventory (TREI), into Turkish. In summary, because these inventories share the same theoretical background and method, the scale (EEI) measuring the expectations from eating, and another scale (TREI) measuring the expectations from dieting and thinness were selected and adapted into Turkish and examined with respect to their psychometric properties.
The study sample consisted of 329 female students between the ages of 19 and 24 (M = 21.41; SD = 1.46), studying at different education levels and different departments of the Ege University. Sixty-three female students (M = 22.44; SD = .98), who could be reached again three weeks after the first application, represent the sample of the test–retest reliability analysis. Participants filled out all the Turkish forms including the following measurements: EEI, TREI, Eating Disorder Examination Questionnaire, Three-Factor Eating Questionnaire (TFEQ-R21), Bulimia Investigatory Test Edinburgh, and Palatable Eating Motives Scale.
The researchers performed descriptive analyses, criterion-related validity, and correlation analyses for test–retest reliability using SPSS 17, and conducted confirmatory factor analysis using AMOS 25 software. Confirmatory factor analyses were conducted separately to assess the construct validity of the scales and demonstrated that both measuring instruments revealed consistent structures with the original factors of the scales (five factors for T-EEI; one factor for T-TREI). Accordingly, the data revealed that the model fit indexes for the eating expectancy scale fit well in five-factor solutions of the data [χ2 (94, N = 329) = 198,71 p < .000, χ2/df = 2.11, GFI =.93, NFI =.94, CFI = .97 RMSEA=.058]. For the thinness and restricting expectancy scale, the model showed a good fit in the single-factor solution of data [χ2 (2, N = 329) = 5,318 p <.000, χ2/df = 2.66, GFI =.99, NFI =.99, CFI =.99 RMSEA=.071]. The internal consistency coefficients for the sub-dimensions of T-EEI ranged between .75 and .95, and test–retest reliability values ranged between .63 and .83. The internal consistency coefficient of T-TREI was .98 and the test–retest reliability coefficient was .70.
This study was conducted with female university students, as in the original study of the scales. To perform the factor analysis, the researchers followed the universally accepted person–item ratio rule (Worthington & Whittaker, 2006), and applied with the current sample size. As a result, consistent with the original measurement tool, five dimensions for T-EEI and the only dimension for T-TREI were confirmed. In addition, the item-test correlation analyses showed that the values for T-EEI and T-TREI had significant relationships, meaning the items were related to the same structure. CFA and item-total correlation values confirmed that both T-EEI and T-TREI measured the structure they intended to measure.
Consequently, Turkish forms of the scales were found to be valid and reliable tools in terms of psychometric properties . For future studies, it would be also appropriate to test the validity and reliability of the scales for other age groups, genders, and clinical samples. The authors believed that these inventories would meet an important need in the field of eating disorders, and that they will make wide-range studies possible. In addition, it would be beneficial to evaluate whether there are any other significant expectations from food and to add these variables to the treatment goals of eating disorders.