It can be somewhat difficult to interpret the partial regression coefficients of a multiple regression model used to perform an analysis of covariance whenever this model includes interaction variables. This difficulty arises because the partial regression coefficient for an interaction variable represents the difference between the partial regression coefficients on a covariate for two groups. It may not be too difficult to interpret such a partial regression coefficient in a simple analysis of covariance model with only one dummy independent variable and one continuous independent variable, but the situation can become much more confusing in analysis of covariance models that contain more than one dummy independent variable or more than one continuous independent variable. Fortunately, there is an alternative procedure for assessing the effect of interaction in analysis of covariance models that yields partial regression coefficients that are much easier to interpret.
CITATION STYLE
Interpreting interaction in analysis of covariance. (2007). In Understanding Regression Analysis (pp. 152–155). Springer US. https://doi.org/10.1007/978-0-585-25657-3_32
Mendeley helps you to discover research relevant for your work.