The AB/BA Design with Normal Data
Stephen John Senn · 2002
This chapter contains sections titled: An Example A Simple Analysis Ignoring the Effect of Period Student's Approach Assumptions in the Matched-pairs t Approach Adjusting for a Period Effect: Two-sample t Approach Adjusting for a Period Effect: The Hills–Armitage Approach Examining Period Effects Testing for Carry-over and/or Treatment by Period Interaction A Model for the AB/BA Cross-over Carry-over or Treatment by Period Interaction? Confidence Intervals for Carry-over Are Unbiased Estimators of the Treatment Effect Available? Can We Adjust for Carry-over? The Two-stage Analysis Correcting the Two-stage Procedure Use of Baseline Measurements A Bayesian Approach Computer Analysis Further Reading Recommendations Appendix 3.1 Analysis With GenStat® Appendix 3.2 Analysis With S-Plus® Author's note: “Some passages of text are indicated with an asterix (*). This is either because the material is rather more difficult than the general level of the book or because, whatever the theoretical interest, its utility to the analyst is low.”