Early identification of dropouts during the special forces selection program.

Area of Science:

• Psychology
• Military Science
• Performance Psychology

Background:

• Special forces selection involves extreme psychological and physical stress.
• High dropout rates (up to 80%) necessitate predictive models for attrition.

Purpose of the Study:

• To develop an interpretable model predicting recruit dropout during special forces selection.
• To identify key psychological predictors of attrition.

Main Methods:

• Weekly assessment of 249 recruits on self-efficacy, motivation, stress, and recovery.
• Application of linear regression and machine learning techniques for prediction.
• Cross-validation to evaluate model performance and identify significant predictors.

Main Results:

• Linear regression achieved good predictive performance (AUC 0.69) with interpretable results.
• Low self-efficacy and motivation were identified as significant predictors of dropout.
• Dropout could be predicted several weeks in advance of occurrence.

Conclusions:

• Predictive models can effectively identify recruits at risk of dropping out.
• Early identification allows for timely intervention and support.
• Findings enhance understanding of psychological factors in high-stress selection programs.