Exercise

Calculating sample size

Let's finish up our dive into statistical tests by performing power analysis to generate needed sample size. Power analysis involves four moving parts:

  • Sample size
  • Effect size
  • Minimum effect
  • Power

In this exercise, you're working with a website and want to test for a difference in conversion rate. Before you begin the experiment, you must decide how many samples you'll need per variant using 5% significance and 95% power.

Instructions 1/3

undefined XP
  • 1

    Standardize the effect of a conversion rate increase from 20% to 25% success using the proportion_effectsize() function.

  • 2

    Calculate and print the needed sample size using the zt_ind_solve_power() function.

  • 3

    Adapt your code to solve for needed sample size using 80% power instead; observe what happens.