Construct validity and composite reliability of students' motivation instrument toward science in UNNES

Abstract

This instrument manuscript was arranged with unstandardized stimulus and would come to a very subjective interpretation. This study aimed to test (1) construct validity and (2) factor reliability of students' motivation instruments toward Sciences in Mathematics and Natural Sciences Faculty (FMIPA), UNNES. The response approach method was used in the instrument development. This approach would put response categories in points along the defined psychological continuum. The variable of this research is student's motivation toward science. Research respondents were some students of FMIPA, UNNES who were studying in the even semester of academic year 2016/2017. Respondents was taken by proportional random sampling technique. This research was carried out in FMIPA UNNES. Loading factor on each variable in this instrument is significant or valid. That is observed from the p-value of the t-test <0.05. The minimum loading factor in each factor is 0.3, so qualitatively all the loading factors within the instrument are valid. Thus, the instrument of students' motivation toward science is valid. As factorial, the coefficient of instrument reliability in science is very high. In composite, the coefficient of instrument reliability of students' motivation toward sciences is also very high. Thus, the instrument of students' motivation of FMIPA UNNES toward sciences is reliable.