The most comprehensive bank Interview Answer Review tooling available online.

Sample size is based on a few factors - type 1 error rate threshold, type 2 error rate threshold and effect size. Based on these factors, I will compute the sample size that is needed to detect the effect from the experiment. I will check if all the experiment parameters are satisfied, the sample ratio is along expected lines and all the guardrail metrics are as expected. Then, I'll conduct a statistical test to see if the change is statistically significant. I will also evaluate if the change is practically significant for the costs incurred to implement this. Also, I will check if the effect size is seen across user groups or is restricted to specific set of users.