Diagnostic model for preschool workers' unwillingness to continue working: Developed using machine-learning...

Area of Science:

• Early Childhood Education
• Occupational Psychology
• Human Resources Management

Background:

• Preschool teacher turnover has significantly increased globally over the last decade.
• High attrition rates negatively impact childcare quality and parental workforce participation.
• Understanding the drivers of preschool worker attrition is crucial for retention strategies.

Purpose of the Study:

• To develop a diagnostic model for predicting preschool workers' intention to leave their jobs.
• To identify key predictors associated with unwillingness to continue working in the preschool sector.
• To provide actionable insights for improving preschool workforce stability.

Main Methods:

• A diagnostic model was developed using logistic regression, comparing it with LASSO and XGBoost algorithms.
• Data was collected from 1002 full-time preschool workers via general and mental health questionnaires.
• SHapley Additive exPlanations (SHAP) were employed to visualize predictor-outcome relationships, particularly work experience.

Main Results:

• The logistic regression model identified "not living with children," "human relation problems with boss," "high risk of mental distress," and "work experience" as significant predictors.
• A risk score was developed, with an optimal cutoff value of 14 points indicating a high likelihood of unwillingness to continue working.
• The model effectively predicts preschool worker attrition risk based on identified factors.

Conclusions:

• The developed diagnostic model offers a practical tool for identifying at-risk preschool workers.
• Supervisors can utilize this model for targeted interventions to address employee difficulties and improve retention.
• Addressing mental distress, workplace relationships, and work experience can mitigate preschool teacher turnover.