Analysis of covariance is a technique used to fix effects of covariates on dependent variable to test treatment effect in experimental studies. In analysis of covariance it is assumed that covariates are measured perfectly reliable. The assumption of perfectly reliable covariates is almost impossible to meet. In reality, it is almost impossible to measure a covariate without error. In this study, a structural model suggested by Bentler and Woodward (1979) was modified both observed and latent and observed models for a reel data set. These modified models do not require perfectly reliable covariates. In these models the amount of errors in covariates are accounted for testing fit of models and significance of treatment effect.
Journal Section | Articles |
---|---|
Authors | |
Publication Date | December 21, 2016 |
Published in Issue | Year 2016 Volume: 36 Issue: 2 |