To speed up the work, we can perform individual tests in sets. The quantitative data will be able to be further described by means, medians, etc., and the qualitative data by counts and percentages.

The settings window with the Test summaries can be opened in Stistics→Summaries→Test summaries and then the selected group of analyses.

At our disposal we have:

1. Comparison of two dependent groups:
   • The t-test for dependent groups
   • The Wilcoxon test (matched-pairs)
   • The Bowker-McNemar test
   • Normality of distribution test Kołmogorov-Smirnov (or another one suggested by the user)
   • and others …
2. Comparison of two independent groups:
   • The t-test for independent groups
   • The Mann-Whitney U test
   • The Chi-square tests, Fisher exact, OR/RR
   • Normality of distribution test Kołmogorov-Smirnov (or another one suggested by the user)
   • and others …
3. Korelację:
   • Pearson linear correlation
   • Spearman's monotonic correlation
   • The Chi-square tests, Fisher exact, correlation co.
   • Normality of distribution test Kołmogorov-Smirnov (or another one suggested by the user)
   • and others …

In the program, for each of the analyzed variables, depending on whether they are quantitative or qualitative, we can return results:

• selected tests - automatically according to the rule described below returned in the table report;
• all tests and accompanying coefficients and measures regardless of whether the minimum conditions for their use are met.

Notes on the program's automatic test selection

Note 1!

If the user does not describe each variable with the appropriate scales before analysis, the quantitative data will be treated as an interval scale, and the qualitative data as a nominal scale.

Note 2!

Testing the normality of the distribution is based on the results of the normality test selected by the user when setting the descriptive statistics.

Note 3!

If the user chooses not to indicate tests that assess the normality of the distribution in the window of these statistics, then it will be checked based on the Kolmogorov-Smirnov test. The analyses we propose are robust to small deviations from the normal distribution, and the Kolmogorov-Smirnov test is the most conservative among the available tests, by which we show the non-normality of the distribution only when the tested distribution differs greatly from the normal distribution. In this situation, we test the normality of the distribution (1) for the comparison of two independent groups based on the data in each group, (2) for the comparison of two dependent groups based on the difference in measurements, (3) for correlation based on the model residuals.