The relative efficiencies of Latin square and randomized complete block designs for insect trapping experiments: An investigation using field data on tsetse flies

Abstract

1. The Latin square design (LSD) is usually preferred over the randomized complete block design (RCBD) in insect trapping experiments because of its ability to cater for variations due to both sites and days in addition to treatments. 2. The efficiency of LSD relative to RCBD is quantified for tsetse trapping experiments using a data set comprising 2200 trap-days. LSD was found to be more efficient than RCBD when blocks ≡ days, but less efficient when blocks ≡ sites, especially when experiments involve only three or four treatments. 3. The importance of sites relative to days as sources of variation in tsetse trapping experiments is quantified. Effects due to site differences were shown to be more important than day to day variability. 4. By relating LSD relative efficiency to the relative importance of sites, it was found that LSD efficiency depended on the relative magnitude of effects of sites and days.