The Use of Rasch Measurement Model for The Validity and Reliability Teacher Leadership Instrument

Abstract

A pilot study was conducted to validate and examine the reliability of the instrument for teacher leader evaluation. The instrument consisted of 109 items and was distributed to 50 teachers. This instrument was developed to measure three (3) constructs; i) teacher commitment in the concept of teacher leadership ii) competence of teachers in the concept of teacher leadership and iii) emotional intelligence of teachers in the concept of teacher leadership. Through this approach, respondent and item reliability been measured with more accurate than using Alpha Cronbachs' value. Rasch Measurement Model (using Winstep version 3.69.1.11) measured the reliability and separation item- respondent, Point Measure Correlation (PTMEA CORR), fitness of item to measure the construct (outfit Mean-Square or MNSQ) and Standardized Residual Correlation. This approach allows expelled of item that not follow the requirement based on the diagnosis. Final result of analysis, out of 109 item, 17 item need to be expelled and the rest of the item, which is 92 item are suitable to measure three (3) constructs in the concept of teacher leadership. A total of 50 teachers were selected as respondents. All respondents involved in this pilot study were not selected in the actual sample.