Investigating the Dimensionality of Early Numeracy Using the Bifactor Exploratory Structural Equation Modeling...

Area of Science:

• Educational Psychology
• Developmental Psychology
• Psychometrics

Background:

• Previous research on the factorial structure of early number skills yielded inconsistent results, with varying conclusions on its dimensionality.
• Confirmatory factor analysis (CFA) and exploratory structural equation modeling (ESEM) have limitations in accounting for complex structures like early numeracy.
• The bifactor exploratory structural equation modeling (bifactor-ESEM) framework offers a robust method for analyzing hierarchical and multidimensional constructs.

Purpose of the Study:

• To investigate the dimensionality of early number skills measurement using the bifactor-ESEM framework for the first time.
• To compare the fit and structure of various measurement models, including CFA, ESEM, and their bifactor counterparts.
• To identify the optimal model for understanding the general and specific factors underlying early numerical ability.

Main Methods:

• Applied the bifactor exploratory structural equation modeling (bifactor-ESEM) framework to data from 644 prekindergarten and kindergarten children (ages 4-6).
• Contrasted several competing models: one-factor CFA, independent cluster model (ICM-CFA), ESEM, bifactor-CFA, and bifactor-ESEM.
• Assessed model fit using established indexes and evaluated factor loading structures and measurement invariance across gender.

Main Results:

• The bifactor-ESEM model with one general factor and three specific factors (Counting, Relations, Arithmetic) demonstrated the best model fit.
• This model also exhibited the most coherent factor loading structure and achieved full measurement invariance across gender.
• While a general factor underpins early numerical ability, specific latent dimensions with significant value were also identified.

Conclusions:

• The bifactor-ESEM framework is highly relevant for disentangling general and specific factors in early numerical ability.
• Early numerical ability is primarily influenced by a general factor, but specific facets (Counting, Relations, Arithmetic) are also crucial.
• Identifying these specific facets is vital for enhancing the quality of early numerical ability measurement, predictive validity, and practical applications.