The Dimensionality of Reading Self-Concept: Examining Its Stability Using Local Structural Equation Models.

Area of Science:

• Educational Psychology
• Psychometrics
• Large-scale Survey Research

Background:

• Accurate interpretation of survey items is crucial for valid results in large-scale assessments.
• Ensuring consistent understanding of psychological constructs, like reading self-concept, is as vital as for achievement measures.
• Reading self-concept significantly influences various student outcomes.

Purpose of the Study:

• To examine the dimensionality of reading self-concept across different reading achievement levels.
• To investigate how students with varying reading abilities interpret items related to reading competence and difficulty.
• To highlight the importance of construct validity in large-scale psychological surveys.

Main Methods:

• Utilized data from the Programme for International Student Assessment (PISA) 2018.
• Employed local structural equation models (LSEMs) to analyze the factor structure of reading self-concept.
• Analyzed responses from a large sample (N = 529,966) to assess dimensionality across achievement strata.

Main Results:

• Reading self-concept dimensions demonstrated variability across distinct reading achievement levels.
• Students with low reading achievement exhibited differentiated responses to competence and difficulty items (high competence-high difficulty).
• Students with high reading achievement showed reconciled responses (high competence-low difficulty).

Conclusions:

• Local structural equation models (LSEMs) are valuable for assessing factor structure generalizability in large surveys.
• The dimensional structure of reading self-concept is not uniform across all reading achievement levels.
• Increased cognitive testing during item development is recommended to improve construct interpretation consistency.