Fax +41 61 306 12 34 E-Mail karger@karger.ch www.karger.com Innovations Psychother Psychosom 2006;75:25–39 DOI: 10.1159/000089224 The Development of the Toronto Structured Interview for Alexithymia: Item Selection, Factor Structure, Reliability and Concurrent Validity R. Michael Bagby a, b Graeme J. Taylor a, c James D.A. Parker d Susan E. Dickens b a Department of Psychiatry, University of Toronto, b Centre for Addiction and Mental Health, c Department of Psychiatry, Mount Sinai Hospital, Toronto, and d Department of Psychology, Trent University, Peterborough, Canada tural configuration resulted in the Toronto Structured Interview for Alexithymia (TSIA) with two domain scales and four facet scales. The TSIA and its six scales demon- strated acceptable levels of interrater, internal, and retest reliability. The TSIA and its scales correlated modestly but significantly with the 20-item Toronto Alexithymia Scale and its three factor scales, providing some support for the concurrent validity of this interview. Conclusion: The TSIA appears to be a promising structured interview for assessing alexithymia. Copyright © 2006 S. Karger AG, Basel Introduction The alexithymia construct, as formulated by Nemiah et al. [1], encompasses difficulties in identifying and de- scribing subjective emotional feelings and an operative thinking style (pensée opératoire), which is characterized by ‘an absence or paucity of fantasies referable to drives and feelings, and … a preoccupation with the details of external events’ (p. 433). Although the construct was de- rived from clinical observations in the early 1970s [2] , for many years research evaluating the construct and its pur- ported association with various medical and psychiatric Key Words Alexithymia  Toronto Structured Interview for Alexithymia  Toronto Alexithymia Scale  Concurrent validity Abstract Background: Assessments of personality constructs in- creasingly use self-report and structured interview in- struments, which allow for a multimethod measurement approach and decrease specific measurement method bias. The aim of this study was to develop a valid and reliable structured interview for assessing the alexi- thymia construct. Methods: Sixty interview questions were written initially, each with a set of scoring criteria and prompts and probes to elicit information assisting in the scoring of the respondents’ answers. Results: After pilot testing, the number of questions was reduced to 43, which were administered to 136 community participants and 97 psychiatric outpatients. A series of item and scale analyses further reduced the item pool to 24 items. Prin- cipal component analysis and confirmatory factor analy- sis of these 24 items revealed preliminary evidence of a hierarchical, four-factor structure, with four lower factors nested within two higher-order latent factors. This struc- R. Michael Bagby, PhD, C. Psych. Centre for Addiction & Mental Health, College Street Site 250 College Street Toronto, Ont. M5T 1R8 (Canada) Tel. +1 416 535 8501, ext. 6939, Fax +1 416 260 4125, E-Mail michael_bagby@camh.net © 2006 S. Karger AG, Basel 0033–3190/06/0751–0025$23.50/0 Accessible online at: www.karger.com/pps