Sum influence test

The sum influence test compares sum of a measure in a category with the average sum value for the measure across all categories of explanatory field.

Considering the significance value and effect size, the test identifies influential categories.

The sum influence z-statistic uses the error of the sum per category that is obtained as a square root of the squared error of the sum per category that is described in the sum comparison test. Corresponding z-statistic is specified as the difference between category sum and the average sum that is divided by the error of the sum per category.

The z-statistic value is compared to a theoretical standard normal distribution with mean zero and variance one to determine the probability of obtaining the z-statistic value by chance.

• This probability is the significance value.
• If the significance value after a Bonferroni adjustment is less than the significance level, the category is judged to be influential. The Bonferroni adjustment is necessary because multiple z tests are conducted, one for each category.

The effect size is influential category strength. It is computed as absolute value of z-statistic divided by the square root of the category count. Meaningful differences highlight the categories with the highest effect size.