Tukey’s Multiple Comparison Adjustment for Partially Balanced Incomplete Block Designs
MetadataShow full item record
Tukey’s multiple comparison adjustment is used as a multiplicity adjustment for all pairwise comparisons of v treatment means. While designed to keep the experiment-wise error rate (EER) at a chosen significance level, this method is suspected to be conservative (i.e. the actual EER is lower than the chosen significance level) for the data that arise from partially balanced incomplete block designs (PBIBDs). In this paper, we use computer simulations to study the Type I error rates for all pairwise comparisons between v treatment means for PBIBD data after Tukey’s adjustment for multiplicity. The simulations cover a wide range of PBIBDs in terms of number of treatments, number of blocks, and block sizes. The simulations performed in the study did not detect statistically significant deviation of the overall experiment-wise error rate from the chosen significance level _ = 0:05.
Franklin and Marshall College Archives, Undergraduate Honors Thesis 2018
- F&M Theses Collection