Development of a non-verbal intelligence test for adolescents

Abstract

The study was conducted to develop a non-verbal intelligence test for adolescents of grade 10 in the age range of 15 to 17 years. The test material was selected in such a manner as to measure various aspects of ,general intelligence "~gOO like abstract reasoning, spatial relation, conceptual ability, accuracy of discr imination and educt ion of relation. Items were generally made uf fLgunttivc material and they were costructecl in the framework of urdu writing i.e .• from right to left. All the items of the test were loaded on general ability factor Mg" and flu id ability "gf' . The effectiveness of the items W(lS judged through item analysis by adrnini string the test to a sample of 100 boys und 100 girls of grades 10 lind I I who were 15 to 17 years old. In evalu at ing the items, three major aspects were considered; internal consistency; the disc riminat ion index and level of difficulty of each item. Only internally consistent items at .001 signifi cance level were retaincd. Items having discrimination value less than 03 and level of di fficulty below 0.3 and above 0.8 were discarded. The final version of the test comprized of four sub-tests viz Series, Analogies, Classification and Matrices consisting of 18. 17,9 and 18 items respectively. In the main study. lest was administcred to a sample of 100 boys and 100 gi rls of grades 10 and II. Their ages ranged from [5 to 17 years. The internal consistency reli ability was computed with KR-20. which was .82, t( 198) = 20.08 p< .0001. The spl it-half reliab il ity was .77, P < .001. Both the indices suggest a high internal consistency. homogenei ty and measurement stability of the test. The construct va lidity of the test was computed by co rrelating it with a Group Verballntel!igence in Urdu "GVITU ", which was .76, p< .001. The criterion va lidity was computed by correlating it with school marks, which was .82, p< .001 in case of grade 10 and .71, p< .001 in case of grade II students. Both the indices provided the evidence of high valid ity of the test. Significant differences were found between boys/gi rls and English/Urdu medium samples regarding th eir intell ec tual performance. Percentile norms were developed for grade 10 and for boys and girls separately.