Why Statistics Matters interview questions

19 questions interviewers ask on this module’s topics, from first principles to the follow-ups. Each comes with a model answer.

What is the difference between descriptive statistics and statistical inference?●
What are a sample and a target population? Give an example where they differ.●
Why declare filters, thresholds and time windows before looking at the outcome?●●
What is the "unit" of an analysis, and why state it first?●

Why evaluate a forecast on data that was not used to build it?●
Why split by time rather than at random for a forecast like this?●●
What is mean absolute error, and why take the absolute value?●
Why must two forecasts be compared on the same rows?●●
One trip favors forecast A and another favors forecast B. What do you conclude?●●
The zone forecast wins overall. Does it win in every zone?●●
What should a forecast do for a category it has never seen?●●
A forecast's advantage is 1.1336 minutes over four weeks. How would you report it?●●

What goes in the numerator and the denominator of a rate?●
The median trip is 12.875 minutes and the promise is 20. Is a 90% target met?●
Why count against a threshold row by row instead of reading a histogram?●●
What does an empirical cumulative distribution show?●
After seeing a result, someone widens the time window. What is the problem?●●
How do you write a defensible one-sentence claim for a measured rate?●●
What three questions do you ask about any reported model score or rate?●●

Directory listings