Construct Validity and Measurement Invariance of the Research Skill Inventory

Abstract

This paper aimed to validate the developmental process of instrument to create an instructional design model. Therefore this study was designed to test the construct validity of research skill inventory and examine the measurement invariance of research skills across gender and grade. The effect sampling size was 1969 samples. However the actual samples participated was 1712, given a response rate as 86.94 percent. Data was analyzed to test construct validity and multi-group confirmatory factor analysis across gender and grades. Confirmatory factor analysis supports the six models consists of male, female, Grade 7, Grade 8, Grade 9, and total CFA models. MGCFA was comprised of measurement invariance and structural invariance. Findings indicated that measurement invariance had been proved to support the four models as non-invariance namely gender, Grade 7-Grade 8, Grade 8-Grade 9, and Grade 7-Grade 9 models. Structural invariance testing had been proved to support the three invariance models except Grade 8-Grade 9 model.

DOI: 10.5901/mjss.2016.v7n3p366