![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
The Dutch Bermond–Vorst Alexithymia Questionnaire (BVAQ) is translated into various languages. The aim of this research was to establish the factor structure of subscales on seven cultural groups. The BVAQ consists of five subscales of eight items each: Emotionalising, Fantasising, Analysing, Identifying, and Verbalising emotions. The BVAQ was administered to a group of Dutch students (n=375), a group of English students (n=175), a group of Australian students, university employees and visitors (n=216), a group of French speaking Belgian students (n=175), a group of Italian people (n=791; a mix of various clinical groups), a group of Polish people (n=427; also a mix of various clinical groups) and a group of Russian people (n=141; general population). The hypothesised two-factor structure of an affective alexithymia dimension (Emotionalising, Fantasising) and a cognitive alexithymia dimension (Insight and Verbalising), with “Analysing emotions” loading on both factors, was clearly supported by confirmatory factor analyses (CFA). Both orthogonal and oblique principal components analyses (PCA), without restriction concerning the number of factors, provided the same two-factor solution in all groups explaining between 55% and 64% of the variance. Oblique rotation further demonstrated that the correlations between these two factors were low in all populations. The combined CFA and PCA results, therefore, indicated that a model with two independent factors has to be preferred over the model assuming two correlated factors.
Acknowledgements
The authors are indebted to Pierluigi Garotti, Paola Gremigni, Elzbieta Scigala, and Emmanuelle Zech for their support in gathering the data.
Notes
1Interestingly, modification indices show that most of the misfit in Model II is not related to the factor correlations, but to the covariance between residuals of Emotionalising and Verbalising in the English language groups. Freeing these parameters greatly improves model fit (lowering the Chi-square by 31, df=2). Further, Wald tests indicated that the correlations between the factors (Model I) are not significantly different from zero in the majority of samples.