Correcting for Type I Errors in Crosstabs


A Type I error is the likelihood of observing statistically significant results when no significant relationship actually exists. The chance of Type I error increases with multiple simultaneous Contrast or Pairwise tests. If the option to ”Correct for Type I errors in all comparisons” is enabled, MarketSight conservatively adjusts the test results by multiplying the p-value for each test by the degrees of freedom for the cross-tab. The degrees of freedom for a crosstab are given by the following formula: df = (R - 1)(C - 1), where R = the number of rows in the crosstab and C = the number of columns in the cross-tab.

When crosstabs have more than two columns or rows, the corrected p-values for all tests will always be larger and thus less significant. As a result, the p-values are appropriately adjusted for the increased chance of Type I error stemming from the fact that multiple comparisons were run.

No correction is applied for Multiple Response Variables.

When Do I Use This Option?

Select this option to correct for the higher chance of Type I error when running multiple statistical tests.

By selecting this option, MarketSight automatically adjusts statistical tests to control the overall Type I error rate for a set of related tests. Correcting for Type I error is not necessarily ”better,” however, because it increases Type II error and therefore reduces the power of tests.

Hence, running statistical tests with and without this option selected and then observing those cells that are only significant without using the correction can help to identify statistical relationships that might be marginal.

Did this answer your question? Thanks for the feedback There was a problem submitting your feedback. Please try again later.

Still need help? Contact Us Contact Us