Identification of Variables Affecting Employee Performance Using Ordinal Logistic Regression

Abstract

This study aims to identify the variables that most influence employee performance at PT. XYZ using ordinal logistic regression. Data was collected through an employee performance evaluation questionnaire covering variables such as professionalism, openness, integrity, kinship, and occupational health and safety (K3L). The research sample consisted of 36 employees selected using the random sampling method based on Slovin's formula. Data analysis was conducted through simultaneous tests, partial tests, and model fit tests to ensure the suitability of the ordinal logistic regression model used. The results showed that out of the five tested variables, only the kinship variable had a significant effect on employee performance (p-value < 0.05). The resulting model met the goodness-of-fit test (p-value = 0.801), making it applicable for human resource management decision-making. These findings indicate that enhancing the culture of kinship in the workplace can be a key strategy for improving employee performance at PT. XYZ.