18.104.22.168.3. Data analysis
A sampled randomization test is used because the distribution of the variables to be compared is unknown. Such test is preferred over a conventional parametric statistics since they often lead to an overestimate of the significance of differences.
For each arena, the number of times the mite is observed in the treated and control wells, respectively, over the 30 min period are used as scores for the statistical analysis. This is done regardless of whether the mite had changed wells between observations or just stayed in the same well. Then, a matrix is constructed with as many rows as the number of mites used in the bioassay, and two columns containing the scores for treated and control wells for each of the tested mites. The treated and control scores in a given set of data are compared by a sampled randomization test (Manly, 1997; Sokal and Rohlf, 1995). The randomization distribution should be resampled a sufficient number of times (e.g. 106 times).