Application of a Bayesian Network Learning Model to Predict Longitudinal Trajectories of Executive Function...

Area of Science:

• Developmental Psychology
• Child Psychiatry
• Educational Psychology

Background:

• Executive function (EF) is crucial for children's academic success and school adjustment.
• Difficulties in executive function (EFDs) can significantly impact a child's development.
• Understanding longitudinal EFD trajectories is vital for early intervention.

Purpose of the Study:

• To classify longitudinal trajectories of executive function difficulties (EFDs) in Korean children.
• To identify predictors associated with different EFD patterns.
• To inform strategies for supporting children's development and preventing risks.

Main Methods:

• Latent Class Growth Analysis (LCGA) was employed to identify EFD trajectories.
• Bayesian Network Learning was used to determine predictors of EFDs.
• Longitudinal data from Korean children aged 7-10 from a Panel Study were analyzed.

Main Results:

• Three distinct EFD trajectories were identified: low, intermediate, and high.
• The high EFD group demonstrated excellent model performance (AUC = .91).
• Key predictors included child's gender, emotionality, happiness, anxiety (DSM), maternal depression, and coparenting conflict.

Conclusions:

• LCGA and Bayesian network learning effectively classify longitudinal EFD patterns in elementary students.
• Childhood EFDs are influenced by persistent emotional issues in both children and parents.
• Early identification of at-risk children can facilitate targeted support and risk prevention.