Type of distribution of the test data of the optokinetic nystagmus.

In statistical analysis, the type of distribution of the population under study dictates the choice of analytical methods. Exploratory data analysis was undertaken to evaluate the type of distribution for several variables in optokinetic nystagmus: the total numbers of nystagmus (NYS), the algebraic sums of the slow-phase amplitude (AMP) and of the slow-phase velocity (VEL), and the means of the fast-phase velocity (FM). The samples were collected from 814 healthy subjects as reference data. The 5-number summaries were first calculated for each variable, and the mid-summaries for the six letter values were examined for fine structure to reveal the presence of some outliers. The upper and lower limits of the 5-number summaries were then defined to reject the outliers. After outlier rejection the distribution was examined by histogram and probability plots, and by evaluation with the chi-square test, the Kolomogorov-Smirnov test and also with skewness and kurtosis. All the variables proved to fit the normal distribution by at least three methods, except by skewness and kurtosis. Of all methods, the probability plot seems to be the best in the present evaluation.