Analysis of Switchback Data Using Statistical Analysis System, Inc.® Software

Open Access
Publisher Website
Google Scholar
Abstract
Models are presented to analyze data from switchback design trials with and without blocking using the General Linear Models procedure of the Statistical Analysis System, Inc. software. Appropriate analyses for these designs may be obtained using a computational technique that simultaneously deploys the same variable (period) in both continuous and classification forms. With blocked data, the SAS model contains variables for block, cow within block, period (linear) by cow within block, period (class variable) within block, and treatment. With unblocked data, the model contains variables for cow, period (linear) by cow, period, and treatment. Immediate advantages of models presented include increased ease and accuracy of data analyses, particularly with unequal subclass data.