ABSTRACT

Introduction............................................................................................................ 149 10.1 Effect size .................................................................................................... 150 10.2 Probability of a False Positive .................................................................... 152 10.3 Probability of Failing to Detect Something Real ....................................... 154 10.4 Estimating Effect Size from Published Data.............................................. 155 10.5 Methods for Estimating Sample Size ......................................................... 156 10.6 Power and Sample Size for Two Independent Groups .............................. 158 10.7 Comparison of Several Inbred Strains........................................................ 159 10.8 Comparing Specific Groups in a One-Way Design ................................... 161 10.9 Strain × Treatment Experiments ................................................................. 162 10.10 Bridging the Gap: the 2 × 2 Design............................................................ 165 10.11 Sample Size for More Specialized Experiments ........................................ 167 Acknowledgments.................................................................................................. 167 References.............................................................................................................. 167

Planning any experiment involves choice of an experimental design and sample size for each of the groups in the study. The design of a study is dictated by the kind of question one seeks to answer, and the choice of appropriate control groups is determined by logic and established genetic principles. Approaches to design are discussed in numerous articles and chapters, including many in the present volume. Once the design is chosen, the researcher must decide on the sample size. This will commonly be done when preparing a grant or thesis proposal. Increasingly, animal ethics committees also request that the investigator justify the number of animals to be used. Ethical and financial considerations, alike, demand that the minimum effective number be employed. At the same time, if too few animals are studied, the experiment may be unable to detect the very effects it was designed to investigate,

thereby wasting the animals as well as the researchers’ time and the funding agency’s money.