Choosing Prediction Over Explanation in Psychology: Lessons From Machine Learning

Area of Science:

• Psychological Science
• Machine Learning Applications

Background:

• Traditional psychology prioritizes explaining causal mechanisms of behavior.
• The "gold standard" of randomized controlled experiments often yields intricate theories with limited predictive accuracy.
• A strong focus on explanation over prediction may hinder advancements in behavioral science.

Purpose of the Study:

• To propose machine learning (ML) as a tool to enhance psychology's predictive capabilities.
• To explore how ML principles and techniques can be applied to psychological research.
• To advocate for a shift towards prediction-focused research questions in psychology.

Main Methods:

• Review of fundamental machine learning concepts and tools.
• Examination of existing psychological research that utilizes ML for predictive questions.
• Conceptual framework for integrating ML into psychological research methodologies.

Main Results:

• Machine learning offers powerful techniques for improving behavioral prediction.
• Examples demonstrate the successful application of ML in psychological research for predictive insights.
• Current psychological research often lacks sufficient predictive power despite strong explanatory models.

Conclusions:

• Adopting a prediction-focused approach, augmented by machine learning, can significantly advance psychological understanding.
• Increased emphasis on prediction can lead to more robust and applicable psychological theories.
• Machine learning provides a valuable methodological avenue for developing a more predictive behavioral science.