آیا آلکسیتیمیا، دوسوگرایی در ابراز هیجانی و ساختار های ناامنی اجتماعی با هم تداخل دارند؟

Abstract Objective The aim of the present study was to analyze the relationship and differential validity of three constructs related to reduced emotional expression. Methods One hundred six patients of a psychosomatic clinic completed questionnaires assessing alexithymia (TAS-20, BVAQ), ambivalence over emotional expression (AEQ-G18), and social insecurity (UQ). Results A second-order principal component analysis with the scales of all questionnaires yielded three factors and revealed that the scale Competence Ambivalence assessed by the AEQ-G18 loaded on the same factor as the TAS-20 and BVAQ scales measuring Difficulties Describing and Identifying Feelings. A high correlation between the factor Social Insecurity (composed of all UQ scales) and the factor Difficulty Identifying and Describing Feelings (composed of BVAQ, TAS-20, and AEQ-G18 scales) was found. In contrast to this, the factor Emotionalizing and External Thinking showed only low correlations with the remaining factors. Conclusion The results of the present study did not support the view that the alexithymia facets related to difficulties identifying and describing feelings and Competence Ambivalence are distinct constructs, when measured by self-report. This might be explained by methodological problems with the assessment of alexithymia and ambivalence. Furthermore, the results indicate that social insecurity is strongly related with the “difficulty identifying and describing feelings” facets of alexithymia and with effect ambivalence.

Introduction The term alexithymia, coined by Sifneos [1], describes a set of affective and cognitive characteristics. It was defined as a multi-facet construct composed of the following salient features: (a) difficulty identifying feelings and distinguishing them from the bodily sensations of emotional arousal, (b) difficulty describing feelings, (c) paucity of fantasies, and (d) a stimulus bound, externally orientated cognitive style [2]. In this context, “difficulty identifying feelings” describes the diminished ability or confusion in identifying feelings or bodily sensations or in distinguishing between different feelings. Difficulty describing feelings to others means a diminished capacity to use language to communicate feelings and to symbolize emotions, i.e., it refers to the ability to put feelings into words (into a lexical dimension) and not the inclination or lack thereof to disclose feelings to others. The 20-item Toronto Alexithymia Scale (TAS-20 [3]) is the most frequently used instrument for the assessment of alexithymia and measures the characteristics (a), (b), and (d). Another instrument is the Bermond-Vorst Alexithymia Questionnaire (BVAQ [4]), which was developed to measure alexithymia more comprehensively than the TAS-20. The BVAQ covers the facets (a)–(d) and, as an additional aspect, the degree to which someone is emotionally aroused by emotion-inducing events (emotionalizing). A problem which might be relevant in clinical settings is the question whether self-report measures such as the TAS and the BVAQ assess state-dependent disorder related characteristics rather than stable personality traits in clinical subjects, i.e., the distinction of primary vs. secondary alexithymia [5]. Empirical data regarding this issue yielded mixed results (e.g., Ref. [6] vs. [7]). Although research generally suggests that the TAS-20 is a reliable and valid measure for alexithymia (cf. Ref. [8]), the discriminant validity of the TAS-20 and, to a greater extent, the BVAQ has not yet been satisfactorily investigated (cf. Ref. [9]). With regard to the TAS-20, the validity of the factor Difficulty Describing Feelings was doubted because it was argued to evaluate aspects of social anxiety or shame [10]. Own unpublished data [11] suggest that self-reported social insecurity (U-Questionnaire [12]) correlated highly (up to r=.74) with the TAS-20 and the BVAQ in a sample of depressive psychiatry patients (n=67), raising the question of the distinctness of both constructs. However, it is also possible that alexithymia and social insecurity might be theoretically related and not only because of an instrumental artifact. For example, alexithymic individuals have been described to experience decreased pleasure in social settings [13] and to be socially distant, conforming, and incompetent, avoiding close interpersonal relationships [14]. Ambivalence over emotional expression is another construct which is related to inhibited emotional expression. It describes manifest ambivalent cognitions such as when a person wants to express positive and negative emotions in a social context [15]. A closer look at the definition of the construct, i.e., wanting to express emotions and being unable to do so or expressing emotions and later regretting it [15], reveals that it may partly overlap conceptually with alexithymia. Furthermore, ambivalence over emotional expression, like alexithymia, is related to the report of bodily symptoms, somatizing, depression, and neuroticism [15], [16], [17] and [18]. The ambivalence construct can be assessed by the Ambivalence over Emotional Expression Questionnaire (AEQ [15]). A factor analysis revealed a two-factor structure with the factors “ambivalence over the expression of positive emotions” and of “emotions of entitlement.” In the German 18-item version, the factors Effect Ambivalence (predominantly related to negative emotions) and Competence Ambivalence (predominantly related to positive emotions) emerged [16]. Effect Ambivalence is the perceived ability to express one's emotions, but expected negative consequences impede this expression, while Competence Ambivalence is the desire to express one's emotions and the concomitant perceived inability to express these emotions. It has been shown that both factors correlate differently with other constructs and with measures of health [16]. Empirically, alexithymia and ambivalence are moderately to strongly correlated [16], [19], [20] and [21]. It might be the case that alexithymic individuals who may have a decreased ability to put emotions into words are also ambivalent in trying to communicate their emotions to others. The high correlations between alexithymia and ambivalence lead some researchers to conclude that both constructs together would depict emotional inhibition in a broader sense [16]. However, the relationship of all facets of alexithymia with the two different aspects of the ambivalence construct has not yet been investigated. A comparison of the items of the instruments for the measurement of alexithymia (TAS-20 and BVAQ) and ambivalence (AEQ) reveals that the contents of several items of the AEQ factor Competence Ambivalence is very similar to the items of the TAS-20 and to a lesser degree to the BVAQ Difficulty Describing Feelings factor (AEQ-G18 Items 14, 17, 9, 10, 13: “It's hard to find the right words to indicate to others what I am really feeling,” “I often cannot bring myself to express what I am really feeling,” “Often I find that I am not able to tell others how much they really mean to me,” “I want to tell someone when I love them, but it is difficult to find the right words,” or “I would like to be more spontaneous in my emotional reactions but I just can't seem to do it,”; TAS-20 Items 2, 4, 11, 17: “It is difficult for me to find the right words for my feelings,” “I am able to describe my feelings easily,” “I find it hard to describe how I feel about people,” “It is difficult for me to reveal my innermost feelings, even to close friends”; BVAQ Items 1, 11, 31: “I find it difficult to verbally express my feelings,” “Even with a friend, I find it difficult to talk about my feelings,” “I can express my feelings verbally”). Furthermore, inspection of the items of the scale Competence Ambivalence of the AEQ-G18 shows that a conflict or ambivalence between the desire to express the feelings and the lack of ability or competence to do so (e.g., Item 10 “I want to tell someone when I love them, but it is difficult to find the right words”) is not formulated in several items (e.g., Items 14, 17, 9: “It's hard to find the right words to indicate to others what I am really feeling,” “I often cannot bring myself to express what I am really feeling,” “Often I find that I am not able to tell others how much they really mean to me”). This might blur the difference between the expression facets of the questionnaires for the assessment of alexithymia and competence ambivalence. This inspection shows that the instruments for the assessment of facets of alexithymia and ambivalence might partly assess very similar contents. This leads to the question of the discriminant validity of these instruments. The purpose of the present study was to investigate the discriminant validity of self-rated alexithymia on the one hand and social anxiety and ambivalence over emotional expression on the other hand. To this end, the correlations between the assessment instruments for these constructs were calculated and a common factor analysis with the scales of all measures was performed. Because of the overlap in the contents of the items, we expected high correlations between the alexithymia scales Difficulty Describing Feelings of the TAS-20 and the BVAQ on the one hand and the AEQ-G18 scale Competence Ambivalence on the other hand. We also predicted high correlations between the UQ scales and both alexithymia scales (the BVAQ and the TAS-20) since our own unpublished results already revealed high correlations between these constructs. Similarly, the UQ scales and the AEQ-G18 scale Effect Ambivalence were predicted to be clearly distinguishable from the alexithymia scales and the AEQ-G18 scale Competence Ambivalence in a common factor analysis. Furthermore, because of their conceptual overlap, the AEQ-G18 scale Competence Ambivalence was expected to load on the same factor as the alexithymia scale Difficulty Describing Feelings of the TAS-20 and the BVAQ. All factors within the factor analysis should correlate at least moderately. We also hypothesized that the scales Difficulty Identifying Feelings and Difficulty Describing Feelings of the TAS-20 and the BVAQ on the one hand and Externally Oriented Thinking of the TAS-20 and the BVAQ as well as the BVAQ scales Fantasizing and Emotionalizing on the other hand would load on different factors.

Results Correlations We first investigated whether social insecurity and ambivalence over emotional expression and alexithymia are highly related constructs. Therefore, we analyzed the correlations of the alexithymia measures (TAS-20 and BVAQ) with the UQ and AEQ-G18. They are presented in Table 2. As expected, the results revealed that the highest correlations have a large effect size [28] and occurred between the alexithymia scales Difficulty Describing Feelings and Difficulty Identifying Feelings (of the TAS-20 and BVAQ) and the AEQ-G18 scale Competence Ambivalence and most scales of the UQ. In contrast to this, the BVAQ subscales Fantasizing and Emotionalizing were almost uncorrelated to both the UQ and the AEQ-G18. Table 2. Pearson correlations between the alexithymia measures and the UQ and AEQ-G18 UQ (Social Insecurity) AEQ-G (Ambivalence) Failure Contact Demand Negate Guilt Decency Total Effect Competence TAS-20 total .57 .67 −.50 .51 .48 .41 .67 .47 .75 Identifying .57 .60 −.52 .48 .42 .36 .62 .45 .66 Describing .50 .67 −.44 .49 .37 .38 .70 .47 .78 Thinking .27 .32 −.22 .25 .35 .23 .31 .19 .36 BVAQ total .36 .55 −.26 .35 .31 .40 .42 .23 .53 Cognitive .49 .62 −.43 .41 .39 .43 .57 .39 .63 Describing .44 .61 −.37 .39 .27 .39 .59 .39 .66 Identifying .41 .46 −.41 .31 .39 .31 .42 .30 .46 Analyzing .30 .39 −.24 .25 .28 .33 .32 .23 .35 Affective −.02 .15 .14 .07 .03 .14 −.03 −.14 .08 Fantasizing −.01 .07 .15 .08 −.05 .13 −.09 −.16 .00 Emotionalizing −.03 .19 .07 .03 .12 .09 .06 −.03 .15 AEQ-G18 Total .69 .69 −.64 .62 .42 .55 Effect ambivalence .61 .50 −.65 .60 .38 .53 .90 Competence ambivalence .63 .74 −.49 .51 .38 .45 .90 .61 All correlations above .20 are significant at P<.05, above .26 at P<.01, and above .34 at P<.001. Correlations with large effect sizes are shown in italics. Table options Principal component analysis In order to analyze the correlations more systematically, we conducted a principal component analysis with all subscales of the used questionnaires. The pattern matrix is presented in Table 3. The MAP test yielded three factors to be extracted. The first factor (F1) consisted of all scales of the UQ and the AEQ-G18 scale Effect Ambivalence. This factor could be labeled as “Social Insecurity”. The second factor (F2) contained the scales Difficulty Identifying Feelings and Difficulty Describing Feelings of the TAS and the BVAQ, and, as expected, additionally, the AEQ-G18 scale Competence Ambivalence. Thus, this factor was named “Difficulty Identifying and Describing Feelings”. The third factor (F3) was composed of the BVAQ scale Emotionalizing, the External Thinking scales of TAS and BVAQ, and the BVAQ scale Fantasizing. This factor could be labeled as “Emotionalizing and External Thinking.” The scales were well represented by the factors since all communalities were high (above .60) except those for the UQ subscale “feeling of guilt” and the BVAQ subscale Fantasizing. Indicated by the communalities, the scale variance explained by the factors reached or exceeded the reliability of the scales in most of the cases. This means that there is not much systematic variance left to be explained by other variables. In some cases, the explained variance by the factors exceeded the reliability indicating that the reliability estimate (Cronbach α) is only a lower bound of true reliability. The three extracted factors accounted for a total of 65% of the explained variance. Table 3. Loadings and communalities of the common principal component analysis (Promax rotation) of the TAS-20, BVAQ, AEQ-G18, and UQ subscales Test Scales Component h2 α F1 F2 F3 F1: Social Insecurity UQ Decency .98 .75 .54 UQ Failure to negate .91 .74 .85 UQ Fear of failure .86 .81 .92 UQ Ability to demand −.84 .74 .81 AEQ-G Effect ambivalence .71 .61 .76 UQ Fear of social contact .60 .77 .90 UQ Feeling of guilt .41 .34 .68 F2: Difficulty Identifying and Describing Feelings TAS-20 Identifying feelings .88 .74 .85 BVAQ Identifying feelings .88 .57 .76 TAS-20 Describing feelings .86 .81 .69 BVAQ Describing feelings .74 .72 .84 AEQ-G Competence ambivalence .72 .74 .85 F3: Emotionalizing and External Thinking BVAQ Emotionalizing .83 .62 .67 BVAQ External thinking .67 .61 .55 TAS-20 External thinking .67 .60 .50 BVAQ Fantasizing .53 .24 .79 Eigenvalues 7.24 2.00 1.17 Percentage of explained variance 45.26 12.50 7.33 Only loadings above .30 in the pattern matrix are displayed. α=Internal consistency (Cronbach α). Table options The highest correlation (r=.65, P<.001, one-tailed) was found between F1 (Social Insecurity) and F2 (Difficulty Identifying and Describing Feelings). F2 and F3 (Emotionalizing and External Thinking) correlated moderately (r=.39, P<.001, one-tailed); the correlation between F1 and F3 was low only (r=.24, P<.01, one-tailed). This indicates that both alexithymia factors (F2 and F3) correlated less strongly with each other than F2 with F1. As the correction for the attenuation of the correlations by unreliability of the scales shows, the low internal consistencies of the scales constituting F3 are not mainly responsible for the low correlations with F1 and F2.