Evaluating Teacher Attitudes Toward Language in Science Education: A Factor Analytic Approach

• In Event: Research in Progress Roundtable Session 2
  In Invited Roundtable: Teacher Attitudes, Identity, and Inclusive Learning in Science Education (Table 11)

Abstract

In this study, Item Factor Analysis (IFA) is used to evaluate 15 categorical items measuring teacher attitudes toward language in science education. Data from 148 teachers at one time point were analyzed using a weighted least squares mean and variance adjusted (WLSMV) estimator. A single factor model showed poor fit at first, prompting revisions with theoretically grounded error covariances, which improved fit but fell short of adequacy. Marginal maximum likelihood (MML) estimation compared 1PL Rasch and 2PL graded response models, with the 2PL model showing superior fit via likelihood ratio testing. Findings underscore the need for advanced modeling techniques to refine latent trait measurement in categorical data.

Author

• Zeynep Mentesoglu, University of Iowa