A variety of allele-sharing methods was used to assess linkage to disease for a whole-genome screen in each of ten replicates from the Genetic Analysis Workshop 12 simulated data. Analysis of each replicate produced many false positive results in addition to a few "true" positives. Conditioning on the strongest initial linkage and reanalyzing using a variety of conditional methods did not improve the power for detection or help discriminate between true and false positive signals.