Quiz: Statistics

This round covers statistics, experiments, and causation. These are the questions that separate strong candidates from everyone else.

• Statistics and experiments
• Causation
• Separates strong candidates

Question 1

Question 1. What does a p-value measure? A, The chance the result is true. B, How surprising the data would be if nothing was going on. C, The effect size. D, The sample size. Pause the video and choose. The answer is B. It assumes that there's no effect.

Question 2

Question 2. To cut the error of an average in half, how much more data do you need? A, Twice. B, Four times. C, Ten times. D, The same. Pause the video and choose. The answer is B. Error shrinks with the square root.

Question 3

Question 3. Why does randomisation work so well? A, It cancels hidden differences. B, It adds data. C, It removes noise. D, It is cheap. Pause the video and choose. The answer is A. Any remaining gap comes from the change.

Question 4

Question 4. You stop a test as soon as it looks significant. What's the risk? A, More false wins. B, Less data storage. C, Better accuracy. D, None. Pause the video and choose. The answer is A. Peeking inflates false positives.

Question 5

Question 5. You test twenty ideas. What happens? A, A false win is likely. B, All will be right. C, None can win. D, Nothing changes. Pause the video and choose. The answer is A. Correct for multiple tests.

Question 6

Question 6. Ice cream and drownings rise together. Why? A, Ice cream causes drowning. B, A confounder, hot weather. C, Pure chance. D, Bad data. Pause the video and choose. The answer is B. A hidden cause drives both.

Question 7

Question 7. What is Simpson's paradox? A, A trend flips when you split the data. B, A kind of chart. C, A SQL error. D, A sampling method. Pause the video and choose. The answer is A. A hidden variable can reverse a result.

Question 8

Question 8. What assumption does difference in differences need? A, Parallel trends. B, Normal data. C, No outliers. D, Large samples only. Pause the video and choose. The answer is A. Both groups would have moved together.

Question 9

Question 9. A coefficient of three on visits means what? A, Three more, holding others still. B, Three percent always. C, Causation for sure. D, Nothing. Pause the video and choose. The answer is A. It's an association, unless the design supports cause.

Question 10

Question 10. Which is the honest way to give a forecast? A, With a range. B, As one exact number. C, In a pie chart. D, Without a baseline. Pause the video and choose. The answer is A. Decisions need to know the uncertainty.

Summary