Personality and Individual Differences 51 (2011) 683–686

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier .com/locate /paid

Short Communication

Are alexithymia and Type D personality distinct or overlapping constructs?A confirmatory factor analysis of the Toronto alexithymia and Type D scales

Lynn Williams ⇑, Cindy Curren, Gillian BruceSchool of Social Sciences, University of the West of Scotland, Paisley, PA1 2BE, UK

a r t i c l e i n f o

Article history:Received 20 February 2011Received in revised form 7 June 2011Accepted 7 June 2011Available online 2 July 2011

Keywords:AlexithymiaConfirmatory factor analysisType D personality

0191-8869/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.paid.2011.06.007

⇑ Corresponding author. Tel./fax: +44 (0) 141 8483E-mail address: lynn.williams@uws.ac.uk (L. Willia

a b s t r a c t

Theoretically and conceptually the constructs of alexithymia and Type D personality share many com-mon characteristics. Despite both measures being utilized widely in psychosomatic research, to-dateno study has examined the constructs simultaneously. The present study was undertaken to determineif alexithymia and Type D personality are distinct or overlapping constructs. A cross-sectional sampleof 1016 healthy participants completed the 20-item version of the Toronto alexithymia scale (TAS-20),and the Type D personality scale (DS14). Confirmatory factor analysis of the 14 items of the DS14, and20 items of the TAS-20 was conducted. Based on the underlying theoretical structures of the DS14and TAS-20 a five-factor model was examined (corresponding to the five subscales of the DS14 andTAS-20). The analysis revealed that the five factor model was an acceptable fit for the data (v2/df = 4.7; CFI = .94, RMSEA = 0.06). These results provide evidence to suggest that alexithymia and TypeD are distinct and separate constructs. However, future research should aim to replicate these findingsin population and clinical samples.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Alexithymia is a personality construct defined by difficultiesidentifying and describing emotions; an impoverished fantasy life,and externally-oriented thinking (Taylor, Bagby, & Parker, 1991). Itis associated with depression and anxiety (Tselebis et al., 2010),substance use disorders (Thorberg, Young, Sullivan, & Lyvers,2009), and has also been implicated in the pathogenesis of physicalillnesses, including coronary heart disease (Todarello, Taylor,Parker, & Fanelli, 1995). Similarly, Type D personality (the conjointeffects of negative affectivity (NA) and social inhibition (SI))(Denollet et al., 1996), has been associated with symptoms ofpsychological distress (including anxiety and depression) and pre-dicts poor clinical outcome in cardiac patients (Pedersen &Denollet, 2006).

In the field of cardiovascular health, there is considerable de-bate regarding the distinctiveness and overlap of psychological riskfactors (Kudielka, von Kanel, Gander, & Fischer, 2004), with theo-rists raising concerns over the conceptual differences between psy-chological constructs. Despite measures of alexithymia and Type Dbeing widely used within the field of psychosomatics, to-date nostudy has examined Type D personality and alexithymia simulta-neously. The definitions of alexithymia and Type D suggest thatthere is a potential conceptual overlap between the constructs,

ll rights reserved.

956.ms).

with the constructs sharing several key characteristics. It alsoappears as though there is a potential theoretical overlap betweenthe constructs. De Fruyt and Denollet (2002) demonstrated thatthe Type D components of NA and SI are closely related to thefive-factor model components of neuroticism and extraversion.Similarly, Luminet, Bagby, Wagner, Taylor, and Parker (1999) dem-onstrated that alexithymia is also positively correlated with neu-roticism, and negatively correlated with extraversion. It istherefore possible that both instruments represent different mea-sures of the same underlying constructs (i.e. high neuroticismand low extraversion). Accordingly, the aim of the current studyis to determine if alexithymia and Type D personality are overlap-ping or distinct constructs.

2. Method

2.1. Participants and procedure

There were 1016 participants in the study (241 males),recruited through convenience sampling from a Scottish Univer-sity. The mean age of the participants was 26.1 years (Median = 22;SD = 9.14) and the ages ranged from 17 to 62 years. The men(M = 27.0, SD = 9.94) and women (M = 25.8, SD = 8.8) did not differsignificantly in terms of age, t(1, 1014) = 1.86, ns. They were givena brief introduction of what the study would require and invited toparticipate. Ethical approval had been obtained from the Univer-sity of the West of Scotland’s ethics committee prior to testing.

Table 1Correlations, means and standard deviations of all variables.

Variable 1 2 3 4 5 6 7

1. DDF –2. DIF .71⁄ –3. EOT .29⁄ .14⁄ –4. Alexithymia .86⁄ .84⁄ .59⁄ –5. NA .41⁄ .57⁄ .04 .47⁄ –6. SI .45⁄ .40⁄ .12⁄ .43⁄ .45⁄ –7. Type D

(NA � SI).48⁄ .56⁄ .09⁄ .51⁄ .80⁄ .81⁄ –

Mean 10.94 13.31 19.35 43.57 10.36 9.95 116.58SD 4.01 5.5 4.32 10.66 5.73 5.2 106.68

Note: ⁄p < .01. DDF = difficulty describing feelings; DIF = difficulty identifying feel-ings; EOT = externally oriented thinking; NA = negative affectivity; SI = socialinhibition.

684 L. Williams et al. / Personality and Individual Differences 51 (2011) 683–686

2.2. Measures

Socio-demographic variables including age and gender were col-lected. In addition, all participants were asked to complete the TypeD personality scale (DS14; Denollet, 2005), and the Toronto alexi-thymia scale (TAS-20; Bagby, Parker, & Taylor, 1994). The DS14 is a14-item scale comprising of two subscales. A 7-item subscale whichmeasures negative affectivity (e.g. ‘I often feel unhappy’) (NA: items2, 4, 5, 7, 9, 12, 13) and a 7-item subscale measuring social inhibition(e.g. ‘I find it hard to start a conversation’) (SI: items 1, 3, 6, 8, 10, 11,14). Respondents rate their personality on a 5-point Likert type scalewhich ranges from 0 = false to 4 = true (items 1 and 3 were reversescored). Type D can be scored as a categorical variable, with partic-ipants who score highly on both NA and SI using a cut-off point ofP10 classified as having a Type D personality (Denollet, 2005).Recently, it has been suggested that Type D may be a dimensionalconstruct, therefore Type D can also be conceptualized and scoredusing the multiplicative NA � SI term (Ferguson et al., 2009;Williams, O’Connor, Grubb, & O’Carroll, 2011). It is the dimensionalapproach that is used in the current study. Cronbach’s a = 0.87 and0.81, respectively for NA and SI indicating good internal consistencyin the current sample.

The TAS-20 is a 20-item scale comprising of three subscales; dif-ficulty describing feelings (DDF) (e.g. It is difficult for me to find theright words for my feeling) (DDF: items 2, 4, 7, 12, 17), difficultyidentifying feelings (DIF) (e.g. I am often confused about whatemotion I am feeling) (DIF: items 1, 3, 6, 9, 11, 13, 14), and exter-nally-oriented thinking (EOT) (e.g. Looking for hidden meaning inmovies or plays distracts from their enjoyment) (EOT: items 5, 8,10, 15, 16, 18, 19, 20). Responses are made on a five-point Likert-typescale which ranges from ‘strongly agree’ to ‘strongly disagree’. A to-tal alexithymia score can be computed by using the sum of responsesto all 20 items. Standardized cut-off scores can be used to categorizepeople as non-alexithymic (scores of <51), alexithymic (scored ofP61), or possible alexithymic (scores of 52–60) (Bagby et al.,1994). Cronbach’s a = 0.82, 0.85 and 0.67, respectively for the DIF,DDF, and EOT subscales indicating acceptable internal consistencyin the current sample.

2.3. Statistical analyses

First, we calculated a series of zero-order correlations in orderto determine the degree of association between the subscales ofthe TAS-20 and DS14, and of alexithymia and Type D (when scoreddimensionally). Next, in order to assess the factor structure of thedata we conducted a confirmatory factor analysis (CFA) (using LIS-REL 8.8). CFA was computed to test if the underlying five factorstructure (i.e. NA, SI, DDF, DIF, EOT) was a good fit for the data.Parameters were estimated using maximum likelihood (ML) esti-mation. Three goodness-of-fit indices were used to evaluate thefit of the model v2, comparative fit index (CFI), and the root meansquare error of approximation (RMSEA). With large sample sizes asin our study, the v2 test statistic is prone to model rejection, wetherefore used the v2/df ratio was used to adjust the v2 statisticby sample size (Kline, 1998). Criteria for acceptable fit were setat v2/df < 5.00 (Schumacker & Lomax, 2004; Wheaton, Muthen,Alwin, & Summers, 1977), CFI > .90 (Ullman, 2007), RMSEA with90% confidence intervals with values ranging from <.06 to .08, sug-gests an acceptable fit (Hu & Bentler, 1999).

3. Results

3.1. Prevalence of alexithymia and Type D personality

From the sample of 1016 participants, 332 participants wereclassified as Type D (32.7%), and 72 were classified as alexithymic

(7.1%), with a further 126 participants classified as possible alexi-thymics (12.4%). Forty-nine participants (4.8%) were found to beboth Type D and alexithymic, and 72 participants were found tobe both Type D and possibly alexithymic (7.1%). Chi-square analy-sis revealed a significant effect of Type D on alexithymia classifica-tion (v2 (2, N = 1016) = 94.1, p < .001).

3.2. Correlations

As shown in Table 1, there were positive and significant corre-lations between the subscales of the TAS-20 and DS14. DDF andDIF were both correlated with NA, SI and Type D (range ofrs = .40–.57), while EOT was correlated with SI (r = .12) and TypeD, but not with NA. Total alexithymia score and total Type D scorewere also positively and significantly correlated (r = .51).

3.3. Confirmatory factor analysis

Factor loadings and fit indices for the CFA are shown in Table 2.The fit indices suggest that the five-factor model is an acceptablefit for the data, v2/df = 4.7; CFI = .94, RMSEA = 0.06 (90%CI = 0.062; 0.067). The five-factor solution, with minor overlap onthe SI and DDF subscales (items 10 and 11 of the SI subscale ofthe DS14 loaded on the DDF subscale) was also supported by theresults of an additionally conducted exploratory factor analysis(which are not reported here).

4. Discussion

The present study has demonstrated that although alexithymiaand Type D personality are conceptually and theoretically overlap-ping constructs, they possess distinct factor structures. CFA analy-sis suggested that the theoretically driven five-factor structure(NA, SI, DDF, DIF, EOT) was an acceptable fit for the data. Therefore,although there is conceptual overlap between the constructs (andthere were substantial correlations between the subscales of theTAS-20 and DS14), the present results suggest that in general alex-ithymia and Type D are distinct constructs.

The factor loadings of the NA, SI, DIF and DDF subscales weresignificant and higher than the recommended cut-off. However,the factor loadings of the EOT subscale were more problematic,three of these items were functioning poorly and explain littlevariance in the latent EOT factor. In addition, the internal consis-tency for the EOT subscale was just below the acceptable level.Other authors (e.g. Thorberg, Young, Sullivan, Lyvers, & Hurst,2010) have also questioned the reliability of the EOT sub-scale.The current findings add to this, and suggest that a refinement ofthe EOT subscale may be needed. It should also be noted that the

Table 2Results of the confirmatory factor analysis of the DS14 and TAS-20.

NA SI DIF DDF EOT

DS14 negative affectI often make a fuss about unimportant things .52I often feel unhappy .83I am often irritated .78I take a gloomy view of things .83I am often in a bad mood .70I often find myself worrying about something .71I am often down in the dumps .88

DS14 social inhibitionI make contact easily when I meet people .62I often talk to strangers .40I often feel inhibited in social interactions .71I find it hard to start a conversation .79I am a closed kind of person .81I would rather keep other people at a distance .68When socializing I don’t find the right things to talk about .68

TAS-20 difficulty identifying feelingsI am often confused about what emotion I am feeling .80I have physical sensations that even doctors don’t understand .45When I am upset, I don’t know if I am sad, frightened, or angry .76I am often puzzled by sensations in my body .86I have feelings that I can’t quite identify .62I don’t know what’s going on inside me .78I often don’t know why I am angry .76

TAS-20 difficulty describing feelingsIt is difficult for me to find the right words for my feelings .83I am able to describe my feelings quite easily .63I find it hard to describe how I feel about people .55People tell me to describe my feelings more .67It is difficult for me to reveal my innermost feelings, even to close friends .75

TAS-20 externally-oriented thinkingI prefer to analyze problems rather than describe them .35I prefer to just let things happen rather than to understand why they turned out

that way.31

Being in touch with emotions is essential .50I prefer talking to people about their daily activities rather than their feelings .45I prefer to watch ‘light’ entertainment shows rather than psychological dramas .28I can feel close to someone, even in moments of silence .42I find examination of my feelings useful in solving personal problems .67Looking for hidden meanings in movies or plays distracts from their enjoyment .21

Fit indicesv2 2442.03df 517v2/df 4.7CFI .94RMSEA 0.06

Note: NA = negative affectivity; SI = social inhibition; DIF = difficulty identifying feelings; DDF = difficulty describing feelings;EOT = externally oriented thinking; df = degrees of freedom; CFI = comparative fit index; RMSEA = root mean square error ofapproximation.

L. Williams et al. / Personality and Individual Differences 51 (2011) 683–686 685

TAS-20 (particularly the EOT subscale) contains longer and morecomplex items than the DS14, this additional cognitive complexitymay increase reading and thinking times, potentially influencingthe factor structure.

The present study should be considered as preliminary withinthis line of research. It is limited as the sample is predominantly fe-male and is comprised of University students. A further limitation isthat the items in the questionnaire were not randomized (i.e. theywere presented as separate scales). It is important that future re-search determines whether the distinctiveness of alexithymia andType D also exists in clinical populations (e.g. cardiovascular pa-tients). In addition, as other measures of alexithymia are available(e.g. Bermond–Vorst Alexithymia questionnaire; Vorst & Bermond,2001) it may be useful to examine the overlap of alexithymia andType D using these measures, in addition to the TAS-20.

The current study is the first to examine the potential overlapbetween alexithymia and Type D. We confirm that the five-factor

model is an acceptable fit to the data. Future research should seekto replicate the five-factor structure in other samples. The findingssuggest that although alexithymia and Type D share some commoncharacteristics they are separate constructs.

References

Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994). The twenty-item TorontoAlexithymia Scale-I. Item selection and cross-validation of the factor structure.Journal of Psychosomatic Research, 38, 23–32.

De Fruyt, F., & Denollet, J. (2002). Type D personality: A Five-Factor Modelperspective. Psychology and Health, 17, 671–683.

Denollet, J. K. L. (2005). DS14: Standard Assessment of negative affectivity, socialinhibition, and Type D personality. Psychosomatic Medicine, 67, 89–97.

Denollet, J., Sys, S. U., Stroobant, N., Rombouts, H., Gillebert, T. C., & Brutsaert, D. L.(1996). Personality as independent predictor of long term mortality in patientswith coronary heart disease. Lancet, 347, 417–421.

Ferguson, E., Williams, L., O’Connor, R. C., Howard, S., Hughes, B. M., Johnston, D. W.,et al. (2009). A taxometric analysis of Type-D personality. PsychosomaticMedicine, 71, 981–986.

686 L. Williams et al. / Personality and Individual Differences 51 (2011) 683–686

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structureanalysis: Conventional criteria versus new alternatives. Structural EquationModeling, 6, 1–55.

Kline, R. B. (1998). Principles and practice of structural equation modeling. New York:Guilford Press.

Kudielka, B. M., von Kanel, R., Gander, M.-L., & Fischer, J. E. (2004). Theinterrelationship of psychosocial risk factors for coronary artery disease in aworking population: Do we measure distinct or overlapping psychologicalconcepts. Behavioral Medicine, 30, 35–43.

Luminet, O., Bagby, R. M., Wagner, H., Taylor, G. J., & Parker, J. D. (1999). Relationbetween alexithymia and the five-factor model of personality: A facet-levelanalysis. Journal of Personality Assessment, 73, 345–358.

Pedersen, S. S., & Denollet, J. (2006). Is type D personality here to stay? Emergingevidence across cardiovascular disease patient groups. Current CardiologyReviews, 6(2), 205–213.

Schumacker, R. E., & Lomax, R. G. (2004). A beginner’s guide to structural equationmodeling. Mahwah, NJ: Lawrence Erlbaum.

Taylor, G. J., Bagby, R. M., & Parker, J. D. A. (1991). The alexithymia construct: Apotential paradigm for psychosomatic medicine. Psychosomatics, 32, 153–164.

Thorberg, F. A., Young, R. M., Sullivan, K. A., & Lyvers, M. (2009). Alexithymia andalcohol use disorders: A critical review. Addictive Behaviors, 34, 237–245.

Thorberg, F. A., Young, R. M., Sullivan, K. A., Lyvers, M., & Hurst, C. (2010). Aconfirmatory factor analysis of the Toronto Alexithymia Scale (TAS-20) in analcohol-dependent sample. Psychiatry Research, 178, 565–567.

Todarello, O., Taylor, G. J., Parker, J. D. A., & Fanelli, M. (1995). Alexithymia inessential hypertension and psychiatric outpatients: A comparative study.Journal of Psychosomatic Research, 39, 987–994.

Tselebis, A., Kosmas, E., Bratis, D., Moussas, G., Karkanias, A., Ilias, I., et al. (2010).Prevalence of alexithymia and its association with anxiety and depression in asample of Greek chronic obstructive pulmonary disease (COPD) outpatients.Annals of General Psychiatry, 9, 16.

Ullman, J. B. (2007). Structural equation modeling. In B. G. Tabachnick & L. S. Fidell(Eds.), Using multivariate statistics (5th ed.. Boston, MA: Allyn and Bacon/Pearson.

Vorst, H. C. M., & Bermond, B. (2001). Validity and reliability of the Bermond–VorstAlexithymia Questionnaire. Personality and Individual Differences, 30, 413–434.

Wheaton, B., Muthen, B., Alwin, D. F., & Summers, G. (1977). Assessing reliabilityand stability in panel models. Sociological Methodology, 8, 84–136.

Williams, L., O’Connor, R. C., Grubb, N., & O’Carroll, R. E. (2011). Type D personalityand illness perceptions in myocardial infarction patients. Journal ofPsychosomatic Research, 70, 141–144.