Validity and Reliability of Survey Items in Employer Perspective Construct on the Quality of ECCE: Rasch Measurement Model Analysis

Abstract

This paper examines and verifies the reliability of a survey instrument on the long term impact of early childhood and childcare education (ECCE) toward human capital development. The survey consists of separate questionnaires based on four perspectives; Individual Success, Peers, Parents and Employer perspective. This study highlights the reliability of item constructs from the employer perspective questionnaire distributed in the pilot study. This instrument was developed based on 56 items, and was further categorised into three sub-constructs; 1) individual character, 2) soft skills; and 3) good citizenship. Rasch Measurement Model analysis supported by Winsteps software version 3.73 was utilised to examine reliability of item and person, polarity of item and suitability of item. Findings from analysis of reliability of item indicated that Individual Character subconstruct showed a good level of reliability, whilst Soft Skills and Good Citizenship subconstructs showed reliability below par. Further analysis on polarity of item indicated all items scored positive values to measure the construct. While analysis on item fit revealed that a total of 6 items from the three subconstructs were discarded as they did not meet the criteria specified in the Rasch Model.