Longitudinal models of skill development: conceptual and statistical considerations

Area of Science:

• Developmental Psychology
• Cognitive Science
• Educational Psychology

Background:

• Longitudinal models of skill development require careful conceptual and statistical decisions.
• Conceptual decisions, concerning hypothesized inter-process influences, are often overlooked compared to statistical ones.
• Understanding the interplay between executive function and math skills is crucial for early development.

Purpose of the Study:

• To highlight the importance of conceptual decisions in longitudinal skill development research.
• To examine the cross-domain associations between executive function and math skills in young children.
• To illustrate how conceptual decisions should inform statistical analyses for stronger causal inference.

Main Methods:

• Longitudinal study design with 684 children from low-income households.
• Data collected over four waves, from preschool to kindergarten.
• Statistical analysis of cross-domain associations between executive function and math skills, considering between- and within-person variation.

Main Results:

• Standardized estimates showed executive function predicting math skills (M=.06, SD=.08) and math skills predicting executive function (M=.13, SD=.13).
• Estimated cross-domain associations varied significantly, with EF predicting math ranging from -.24 to .48 and math predicting EF from -.08 to .55.
• Bivariate and model-imposed correlations ranged widely from .00 to .96 (M=.28, SD=.24).

Conclusions:

• Conceptual decisions are critical for accurately modeling skill development and cross-domain associations.
• Findings underscore the need for intentional testing of competing hypotheses to explain developmental pathways.
• Improved conceptualization can lead to more robust theoretical explanations for the executive function-math skill relationship.