VERBAL COMPREHENSION INDEX (VCI) SUBTESTS OF WECHSLER INTELLIGENCE SCALE FOR CHILDREN (WISC-IV UK): ADAPTATION AND NORMS DEVELOPMENT IN PAKISTAN

Abstract

Comprehension Index (VCI) subtests of Wechsler Intelligence Scale for Children, UK Fourth Edition (WISC-IV UK; Wechsler, 2004) in Pakistan. The research was completed through three studies. Study-I was concerned with the adaptation of VCI subtests and comprised of two phases. In phase-I a try out was conducted, and an advisory panel is formed to identify the problems in the original VCI subtests regarding the difficulty level, understanding and cultural relevance of the items. Then with the help of two expert panels, identified items were replaced and VCI subtests were adapted. Only five items in vocabulary subtest and four items in information subtests were replaced for adaptation. Order of administration of two items in vocabulary subtest was also changed. Phase-II consisted of a try out and a pilot study conducted to assess the item functioning and psychometric strength of the adapted VCI subtests. This try out explored the response frequency of adapted items in comparison with the response frequency of the original items. The pilot study was conducted on 64 students, boys (n = 32) and girls (n = 32). Sample was further divided into four equal age groups of 10-11 years (n = 16), 12-13 years (n = 16), 14-15 years (n = 16), and 16-17 years (n = 16); and in two parental education level groups of Parental Education Level-I (with average parental education level of less than 14 years, n = 27) and Parental Education Level-II (with average parental education level is equal to more than 14 years, n = 37). The psychometric properties of the adapted Verbal Comprehension Index (VCI-P) subtests were found to be quite satisfactory with an alpha coefficient of .94 for VCI composite. Inter-subtest correlations and subtests-total correlations were also found to be highly significant. For item analysis, item-subtest total correlations and item total correlations yielded significant coefficients for most of the subtest items. Study-II was concerned with the establishment of reliability and validity evidence of the VCI-P and is comprised of three phases. In its phase I, test-retest reliability was established on a sample of 32 students, stability coefficients ranged from .82 (word reasoning) to .92 (vocabulary) for the adapted subtests. Phase II was conducted on 801 students including boys (n = 385, 48.1%) and girls (n = 416, 51.9%). Sample was further divided into four equal age groups of 10-11 years (n = 200, 25%), 12-13 years (n = 200, 25%), 14-15 years (n = 200, 25%), and 16-17 years (n = 201, 25.1%); and in two parental education level groups of P.E.Level-I (n = 384, 47.8%) and P.E.Level-II (n = 418, 52.2%). Alpha reliability coefficients for adapted subtests ranged from .72 (comprehension) to .88 (vocabulary) and standard error of measurement was found to be 3.65 for VCI-P. For structural validity, inter-subtests correlations (ranging from .63 to .78) and subtest-VCI correlations were computed that were .91, .94, and .84 for similarities, vocabulary, and comprehension, respectively. Factorial validity resulted in existence of one single factor as factor loading of all subtests and items of all subtests were quite high on that single factor. In the phase III, relationship of VCI scores was established with academic achievement scores of the students and tetrachoric correlation coefficient was found to be .62. Study-III was concerned with the development of norms of VCI-P in Pakistan. For that purpose age based standard scores (scaled and composite scores), percentile ranks and test-age equivalent norms were developed for the adapted subtests. All norms were quite comparable with the norms of the original version, and confidence interval for VCI equivalents was same (VCI equivalent ± 7) for our sample and for original population as reported in the manual. Mean differences in VCI Equivalent scores showed an increase of 15- 20 points when computed and analyzed with Pakistani norms as compared with the UK norms. Mean differences in VCI subtest scores were also explored on age, gender, and parental education level groups. Mean differences were found to be significant for all subtests in four age groups, and two parental education level groups, where as mean differences were not found to be much significant for all subtest in the two gender groups.