Four participants reported that they were colour blind, so their
data were excluded from all analyses. The median scores of the
arrangement tasks for the remaining 97 participants are shown
in Fig. 2. The results of the Kolmogorov–Smirnov tests revealed
that the majority of the data (66 out of 84 measures) were not normally
distributed, D(97) > .09, p < .05. Therefore, we analysed the
effects of Colour and Shape independently, and used
non-parametric statistics by collapsing one factor2 and analysing