The Impact of Z-Score Transformation Scaling on the Validity, Reliability, and Measurement Error of Instrument SATS-36

Abstract

The Likert scale is a psychometric scale commonly used for response measurement. This measurement scale includes techniques for designing and administering surveys as well as coding and analyzing data. However, Likert scaling has various limitations that can affect the resulting data. This study aims to reprove the number of dimensions of the SATS-36 instrument, prove the validity, and estimate the reliability of the statistical attitude instrument (SATS-36) on students at religious universities in Indonesia using Z-Score Transformation Scaling. The latent constructs of cognitive competence, value, difficulty, effect, and effort were constructed using a Likert scale according to the pattern of statements on each item. This study uses confirmatory research with a quantitative approach. For students at religious universities in Indonesia, 243 respondents were selected using a stratified one-stage cluster random sampling technique. Proof of validity and estimation of reliability was done using confirmatory factor analysis. The results of this study show that the rescaling method can improve the validity of the factors but cannot increase Cronbach's coefficient of internal consistency and cannot reduce the standard error of measurement for each item. This research implies that it is not enough to rescale or transform the data to improve the validity and reliability of a measuring instrument. However, it is necessary to calibrate the statement sentence or item question so that the item measures its construct. Further research also needs to test the effectiveness of rescaling in addition to the Z-Score in improving the validity and reliability of measuring instruments.