Gender differential item functioning on the Kentucky Inventory of Mindfulness Skills instrument using logistic regression

Abstract

The item differential function (DIF) describes a situation in which testees of similar ability but from different demographic groups have varying chances of achieving the same result. This study aims to identify the function of uniform and non-uniform differential items on the Kentucky Inventory of Mindfulness Skills Instrument using logistic regression techniques and determine the impact of DHF on construct validity. This study uses a survey method with a quantitative approach. The study involved 602 people, divided into two groups based on gender: 301 women and 301 men. The Kentucky Inventory of Mindfulness Skills (KIMS) is a 39-item online questionnaire that measures mindfulness. KIMS has been proven to meet content, construct, and factor validity and has good test-retest reliability and internal consistency estimators. This study uses Regression Logistics to detect DIF, analyzed with R Studio 4.1.3 software. Research results found 17 DIF items detected using logistic regression, 13 uniform DIF items, and four non-uniform DIF. Through CFA, we have succeeded in proving that DIF-free items are proven to have better construct validity. The implications of this study are expected to inspire counseling psychologists to be more careful in using rating scales or instruments. The validity and reliability of the measures are not strong enough to justify that all measuring instruments are correct. However, it is also necessary to check for item bias or functional differential items to ensure that each item on the scale or instrument is understandable to all demographic groups and does not benefit only certain demographic groups.

Keywords

• item biases
• differential item functioning
• logistic regression