I am trying to reverse engineer hypothesis test results from an online calculator using Microsoft excel or Googlesheets. The inputs/outputs for the online calculator are shown in the screenshot


control_exposures = 917
control_conversions = 126

variant_exposures = 1002
variant_conversions = 142

control_cvr_rate = control_conversions/control_exposures = 13.74%
variant_cvr_rate = variant_conversions/variant_exposures = 14.17%

control_std_error = sqrt((control_cvr_rate*(1-control_cvr_rate)/control_exposures)) = 1.14%
variant_std_error = sqrt((variant_cvr_rate*(1-variant_cvr_rate)/variant_exposures)) = 1.10%

z_score = (control_cvr_rate- variant_cvr_rate)/sqrt(power(control_std_error,2)+power(variant_std_error,2)) = -0.2724

p_value = normdist(z_score,0,1,TRUE) = 0.3927


Based on this, how do I derive that statistical power value of 5.9% in the screenshot?








stat_sig.PNG