Dataford
Interview QuestionsInterview GuidesExperiencesMock InterviewsPricing
Get started

P-Value and Power for Button Redesign

EasyStatistics & Probability00:00
Practice interviewer
In session
5 left
00:00

Your question is P-Value and Power for Button Redesign. Take a moment with it on the right.

Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).

You need to log in / sign up to chat or submit.

Problem

Business Context

PixelCraft, a design collaboration app, tested a new onboarding button style proposed by the product design team. The PM wants a simple explanation of whether the observed lift is "real," and whether the test was large enough to detect a meaningful improvement.

Problem Statement

You need to explain p-value and statistical power in plain language, but also quantify them using the experiment results below. Treat this as a standard A/B test on signup conversion.

Given Data

GroupSample SizeSignupsConversion Rate
Control (old button)8,00096012.0%
Treatment (new button)8,2001,06613.0%

Additional test design inputs:

ParameterValue
Significance level0.05
Test typeTwo-sided
Minimum meaningful lift for planning1.0 percentage point
Baseline conversion rate for planning12.0%

Requirements

  1. State the null and alternative hypotheses.
  2. Compute the observed difference in conversion rates.
  3. Calculate the two-proportion z-statistic and p-value.
  4. Explain the p-value in non-technical language appropriate for a product designer.
  5. Estimate the test's statistical power to detect a 1.0 percentage point absolute lift (from 12.0% to 13.0%) at α=0.05\alpha = 0.05α=0.05.
  6. Explain statistical power in non-technical language and say whether this experiment was adequately powered.
  7. Give a recommendation on whether PixelCraft should roll out the new button.

Assumptions

  • Users were randomly assigned and counted once.
  • The normal approximation for proportions is valid.
  • Ignore segmentation and multiple-testing issues for the core calculation.