Predicting online participation through Bayesian network analysis.

Area of Science:

• Political Science
• Computational Social Science

Background:

• Previous research on political participation has not fully distinguished direct causal factors from correlated variables.
• Understanding the network of preconditions for political participation is crucial for accurate prediction.

Purpose of the Study:

• To investigate the causal relationships between variables influencing online activism.
• To develop and validate a predictive model for political participation using a novel three-step approach.

Main Methods:

• Bayesian network analysis to model causal structures.
• Structural equation modeling to stabilize causal relationships.
• Parameter fitting for predictive accuracy.

Main Results:

• Age, political interest, and internal political efficacy were identified as the sole direct predictors of online activism participation.
• The direct effect of political interest is mediated by internal political efficacy and age.
• The predictive performance of the model significantly improved when incorporating key direct effects.

Conclusions:

• Direct causal factors for online activism are limited to age, political interest, and internal political efficacy.
• While the model shows improved prediction, uncertainty remains in forecasting online participation.
• Further research is needed to identify factors driving novel forms of political engagement.