Comparision of the effects of classical (frequentist) and bayesian approaches on sample size

Sample size calculation is one of the essential step in design of clinical studies. Major purpose of sample size calculation is to detect the differences in clinical parameters, treatmet effects and associtations between variables. It is not uncommon to see underpowered studies which could not detect differences because of inadequete sample size. In contrast, a large sample size may cause an amplification in detection of differences resulting in statistical differences that are not clinically relevant. In the present study we aimed to evaluate effect of sample size on clinical results of study by using frequentist and Bayesian approaches. For the evaluation of differences between small and large sample sizes, we constituted two samples. All the samples were consisted of male patients who were 40-50 years old age. Small sample was consisted of 32 patients with ischemic heart disease (IHD) and 37 control subjects. Large sample was consisted of 355 IHD patients and 545 controls. Biochemical parameters of the patients were compared by frequentist and Bayesian approaches. Type I error rate of 5% was accepted for statistical analysis. All the analyses were conducted by using IBM SPSS 26. Our study results suggested that there were no differences between frequentist and Bayesian approach results when the sample size was large and the power of the study was high.