Toronto Alexithymia Scale–20: Examining 18 Competing Factor Structure Solutions in a U.S. Sample and a Philippines Sample

Abstract

The Toronto Alexithymia Scale–20 is arguably the most utilized measure of alexithymia. Although a three-factor solution has been found by numerous studies, these findings are not universal. This article examined and compared 18 competing factor structures for the Toronto Alexithymia Scale–20, which included between one and four correlated latent factor structures, common methods models that accounts for negatively worded items, and bifactor models. Although the two-factor bifactor model with a common methods factor had the better model fit compared with the other 17 models examined, it still did not achieve the requisites of a good model fit across all model fit indices. Issues stemmed primarily from the externally oriented thinking factor and the negatively worded items. Post hoc analyses indicated that a two-factor bifactor model with the negatively worded items dropped achieved the requisites of a good model fit and can be treated as a unidimensional measure despite the presence of multidimensionality. Multiple-group analysis indicated that the factor loadings were invariant across U.S. and Philippines samples. After controlling for noninvariance at the item intercept level, the Philippines sample had a higher alexithymia general score compared with the U.S. sample.

Assessing alexithymia in forensic settings: Psychometric properties of the 20‐item Toronto Alexithymia Scale among incarcerated adult offenders

“…Our findings are therefore similar to most older work in non‐forensic samples (e.g., Bagby et al, 1994), and Parker et al’s (2005) offender study, which also supported the presence of correlated DIF, DDF and EOT factors within the TAS‐20. More recent studies with nonoffenders also highlight the presence of a prominent reverse‐scored item method factor in the TAS‐20 (e.g., Meganck et al, 2008; Preece et al, 2018a, 2020b; Tuliao et al, 2019), and thus our findings are highly consistent with these previous patterns. Indeed, we found this factor structure to be invariant across our offender and community samples.…”

“…In the general psychometric literature (see Van Sonderen, Sanderman, & Coyne, 2013), reverse‐scored items are often found to have problematic effects, in terms of reducing levels of reliability and creating a ‘method factor’ in the factor structure (i.e., where some variance in these items is accounted for by their unique response format, rather than the substantive latent construct they were designed to measure). This may raise potential content problems with some EOT items (e.g., Gignac et al, 2007; Kooiman, Spinhoven, & Trijsburg, 2002; Meganck, Vanheule, & Desmet, 2008; Preece et al, 2018a; Tuliao, Klanecky, Landoy, & McChargue, 2019). Consistent with these views, more recent CFA studies of the TAS‐20 have tested factor models with a reverse‐scored item method factor added (where all the reverse‐scored items were specified to load on this method factor as well as their intended alexithymia factor).…”

“…This allowed the researchers to partial out the variance attributable to the reverse‐scored method effect when looking at whether the items are good indicators of their intended alexithymia factor. Adding the method factor has always improved model fit, usually to adequate levels according to most fit index values (e.g., Preece et al, 2018a; Tuliao et al, 2019; Watters, Taylor, Ayearst, & Bagby, 2016). Some reverse‐scored EOT items have loaded more highly on the method factor than the intended EOT factor, indicating that the reverse‐scored format may unduly affect the validity of these items as accurate markers of EOT.…”

“…Studies with non‐forensic and forensic samples have found that the internal consistency of the TAS‐20 total scale and DIF and DDF subscale scores is usually acceptable to good, but that the EOT subscale is not (<0.70; for a review, see Kooiman et al, 2002; if the reliability coefficient falls below 0.70, more than 50% of the variance in that score is attributable to error, so the score is judged not suitable for research or clinical/forensic use; e.g., Groth‐Marnat, 2009; Preece et al, 2019). In non‐forensic samples, splitting the EOT items into IM and PR subscales, or removing all the reverse‐scored items, has not improved their reliability (e.g., Preece et al, 2018a; Tuliao et al, 2019), but these alterations have not been tested in forensic samples.…”

Interoceptive Abilities in Inflammatory Bowel Diseases and Irritable Bowel Syndrome

“…Although the EOT dimension appears to assess difficulties in emotion regulation (44), we decided not to use this dimension for two reasons. Firstly, its internal reliability is usually poor (47)(48)(49)(50). Secondly, the dimension might better reflect the social norms that guide emotional behaviors rather than a cognitive style of thinking (51).…”