MEASURING THE CONSTRUCT VALIDITY OF A SURVEY ON PARENTAL INVOLVEMENT TOENHANCE CHILDREN’S READING SKILLS: A RASCH MODEL ANALYSIS

Abstract

This study aimed to assess the construct validity of a novel four-point Likert survey following Epstein’s (1987) overlapping spheres of influence on parental engagement to enhance children’s reading skills with the Rasch model analysis. Five constructs were identified with 56 items. A total of 34 parents completely addressed the distributed survey. The derived raw data were fed into the Statistical Package for Social Science (SPSS) Version 16 and assessed with WINSTEP software version 3.72 through item fit analysis, item and person separation index (item and person),reliability index (item and person), and unidimensionality. The item fit analysis revealed that several retained items required further optimization. The item and person separation index and item and person reliability index reflected values of 5 and 2 and .72 and .97, respectively. Meanwhile, aCronbach’s alpha value of .97 was highlighted. Lastly, the unexplained variance in the first contrast denoted13.6% while theraw variance explained by measures was 53.8%.Conclusively, the aforementioned analyses implied the survey to havefair to excellent construct validity. This survey could be employed to gather the outcomes on parental engagement in preschoolers’ reading.