Genetic factors play important role in the development of Alzheimer’s disease (AD). It is substantial to identify these factors. To reveal them from collected data, appropriate sample size should be used in the statistical analysis, so the results are more powerful. However, these tests are often carried out without power and sample size estimation.

The chi-square test is the most frequently used statistical method to detect difference between proportions. Recently power calculation became more important in statistical analysis to determine the appropriate sample size, or to get a power probability which shows the reliability of the statistical test based on preliminary data.

Our aim was to check how powerful was the statistical comparison of the interaction of two investigated alleles and the risk for AD applying power analysis.

Allele and genotype distributions of the sigma non-opioid intracellular receptor 1 (SIGMAR1) gene were compared between patients diagnosed as AD and healthy controls (HC) to check the calculation of post-hoc power of chi-square tests in various software. We used Cohen’s formula for the effect size calculation and constructed a calculator in MS Excel for a 2×3 table to apply this formula in statistical analysis. Based on total of 569 patients (320 in AD and 249 in HC groups) we calculated w=0.111 effect size using 3 subgroups of SIGMAR1 diplotypes (interaction of 2 alleles). The same Cohen formula is applied in the commonly used GPower software. Our MS Excel based calculator gave the same effect size of chi-square analysis as GPower, however it is difficult to use GPower for effect size estimation. From sample size and effect size we had a power of 0.659 with GPower and R software as well.

In conclusion, as the usage of GPower is complicated for those who are not expert in this field we suggest an easy MS Excel calculator based on contingency table.