2008 Joint Annual Meeting (5-9 Oct. 2008): Contrast Statemets Vs the Slice Option - The Where and When.

722-5 Contrast Statemets Vs the Slice Option - The Where and When.



Wednesday, 8 October 2008: 10:45 AM
George R. Brown Convention Center, 381ABC
Don Bullock, W201E Turner Hall, 1102 S. Goodwin Ave., University of Illinois-Urbana-Champaign, University of Illinois at Urbana-Champaign, Department of Crop Sciences, Urbana, IL 61801-4798
Often in an experiment the scientist is interested in comparing certain groups of means or combinations of groups of means. These are called planned comparisons and they are things that you want to do prior to seeing the data. They are implicit. A common mistake seen in the published literature is the use of multiple comparison techniques such as an LSD for these experiments. A more relevant and appropriate technique is the use of contrasts to make the preplanned comparisons. This talk will cover the basic theory of contrast and estimate statements and methods and code for using them to analyze complex interactions in SAS®. We will also cover the SLICE option as an alternative to contrast statements.  Issues of error rates will be considered for all methods.