ABSTRACT
When the same form of model is used to describe data from several populations, treatments, or treatment combinations, many questions about the models can be answered by testing hypotheses and constructing confidence intervals about functions of parameters or a single parameter from each model. For example, if a simple linear regression model is fit to data from each treatment, comparisons between the slopes can answer questions about parallelism or can be used to group treatments together that have common slopes. The beta-hat model can easily be used to extract the necessary information. Beta-hat models are extremely useful in investigating the necessary form of the analysis of covariance model when there are a large number of treatments and computing resources are limited.
