prediction
Definition
Prediction refers to the use of psychological variables or network-derived metrics to forecast or estimate outcomes in other psychological domains, requiring researchers to distill complex construct representations into single or composite variables that can model associations between constructs. In psychometric network models, prediction involves identifying which summary variables—whether network-structure properties (such as density or global strength), network-informed metrics (such as an individual's average value of the most frequently central node), or simple sum scores—best capture a construct's relationship to external psychological variables. Across four intensive longitudinal datasets, network-derived metrics showed varying predictive utility, with an individual's average value of the most frequently central node emerging as the most consistent and effective predictor of psychological outcomes, whereas traditional network-structure variables such as density, global strength, and modularity were not significant predictors for most outcomes.
Sources: Johal & Rhemtulla (2024)
Related Terms
Applications
Prediction and Network Structure
Network structure variables such as density, global strength, and modularity coefficient can be evaluated for their predictive utility in relation to external psychological outcomes. Across multiple longitudinal datasets, however, network-structure variables were not significant predictors of psychological outcomes for most cases, suggesting that the structural properties of networks alone may not effectively capture the relationship between network-instantiated constructs and external variables.
Sources: Johal & Rhemtulla (2024)
Prediction and Centrality
An individual's average value of the most frequently central node consistently demonstrated stronger predictive associations with external psychological variables compared to other network-derived metrics across multiple datasets.
Sources: Johal & Rhemtulla (2024)
