Title: Studying longitudinal change in teacher practices using the multilevel model and latent growth model with an examination of alternative covariance structures
Authors: Jaime Maerten-Rivera; Nicholas D. Myers; Okhee Lee
Addresses: School of Education, University of Miami, 1507 Levante Avenue, Room 223, Coral Gables, FL 33146, USA ' School of Education, University of Miami, 1507 Levante Avenue, Room 223, Coral Gables, FL 33146, USA ' New York University, 239 Greene Street, 6th Floor, New York, New York 10003-7599, USA
Abstract: The purpose of this research was to compare results of the multilevel model (MLM) and latent growth model (LGM) for examining change over time. The study came out of the field of education and used data collected from 191 teachers through a professional development intervention in science education. Teachers' reported use of reform-oriented practices (ROP) was used as the outcome. Change was examined using a piecewise change model and different error covariance structures were examined. Parameter estimates obtained from a model using the error covariance structure commonly assumed in the MLM framework (i.e., random slopes, homogeneous level 1 variance) were nearly identical as were the results of models with various alternative covariance structures commonly associated with the LGM framework. Most of the model fit information was in agreement regarding the best fitting model, with the exception of the standardised root mean square residual (SRMR) fit index.
Keywords: latent growth models; multilevel models; piecewise models; structural equation modelling; SEM; longitudinal change models; hierarchical linear models; error covariance structure; teacher change; teaching practices; professional development; science education.
International Journal of Quantitative Research in Education, 2014 Vol.2 No.2, pp.89 - 112
Received: 20 Apr 2013
Accepted: 10 Feb 2014
Published online: 30 Aug 2014 *