Psychometric Properties of a Reading Comprehension Test for Primary School Students

Abstract

This article presents the psychometric properties of the test of Reading Comprehension for Intervention in Primary School (CLIP), which was developed based on theoretical models of cognitive processes from current research on text comprehension (Kintsch, 1998). The purpose of the test is to assess reading comprehension in elementary school students and provide useful information to plan the teaching that should follow the evaluation. To this end, several texts are used (narrative and expository, short and long), along with 3 types of questions (local, global, and inferential), based on the classification by Magliano et al. (2007). To answer, readers should need to use basic comprehension processes to construct the basic text and situational model. Through a convenience sampling process, 1958 3rd to 6th grade students were recruited from 10 subsidized private and public schools located in the municipalities of Villarrica and Loncoche (Araucanía region of Chile). The internal consistency analysis (Kuder-Richardson 20) of CLIP scores yielded values equal to or greater than 0.79. The confirmatory factor analysis performed finds evidence of the presence of 3 factors in the comprehension of short texts: Local Processing, Global Processing, and Inferential Processing. In long texts, model fit was unsatisfactory. Differential item functioning analysis revealed no sex or socioeconomic bias in any of the items.