The research on the nature of the underlying reading sub-skills is still unsettled. Nor is there any consensus over the divisibility of reading into a set of component skills. Improving CONSTRUCT VALIDITY of tests of L2 reading comprehension requires further investigation on the variables underlying reading ability and the possibility of operationalizing the underlying traits. To this end, a Test of Reading Sub-skills (TRS) was developed which operationalized the components of the Comprehensive Taxonomy of Reading Sub-skills (CTRS) derived from major taxonomies in the literature through a detailed comparative analysis of their proposed components. The TRS employed a wide variety of test methods in two general categories of the Expected Response Format, namely, ‘Selected Response’ and ‘CONSTRUCTed Response’. Data was collected on the performance of 1606 Iranian EFL learners including English majors and non- English majors, each subject randomly answering one of the eight forms of the TRS. Data were analyzed using Exploratory and Confirmatory Factor Analysis. The results showed that L2 reading ability is composed of a number of underlying latent traits or macro-skills. It is maintained that a componential approach to reading assessment and pedagogy should be informed by empirical evidence on the nature of variables underlying reading ability.

Objectives: In view of the existing methodological drawbacks in the previous investigations of NEO-Personality Inventory-Revised, the present study aimed at reinvestigating the CONSTRUCT VALIDITY of this inventory in Iran.Method: 334 students 231 female and 98 male, 5 not specified) with a mean age of 23.1 were selected via convenience sampling from Tehran, Elm-o-Farhang and Shahid Beheshti Universities, and completed NEO-Personality Inventory-Revised. Inclined rotation and parallel analysis was used for exploratory and confirmatory factor analysis.Results: Exploratory factor analysis resulted in the extraction of five factors, which altogether accounted for 59.25 percent of the total variance of the data. As a result of exploratory factor analysis, a noticeable number of cross-loads and wrongly loaded subscales were observed. Confirmatory factor analysis also resulted in weak to moderate fit indexes.Conclusion: Considering the results of the present study and with regard to the medium and weak alpha coefficients in the Garousi Farshi study, the primary applicability of this questionnaire in the Iranian culture is questionable. It seems that factor analysis at the level of characteristics could be helpful.