ABSTRACT
A rough glance is encouraging, the two distributions being close. The chi-square test, with three end values grouped at the left and four at the right to avoid frequencies less than five, accepted normality with 97 percent confidence (df=20) until the poorest scores were included, only 72 percent (df=21) thereafter. A D’Agostino software package was also used. It includes a plot, linear for normal data. The plot appears as Fig. 1, the upward bend meaning skewness to the right. It also includes a test combining measures of skewness and kurtosis which firmly rejected normality. Tests were also run for the scratch level of play with very similar results. In general, normality may be an acceptable assumption except for the worst end of the scoring spectrum, the end which, fortunately, may be the least important for most forms of play.
