Dataford · Statistics Report

The Statistics & A/B Testing Interview Report 2026

Statistics is the topic candidates fear most — and the data agrees with them. We broke down 725 probability, statistics, and experimentation questions to show why it is the hardest subject in the data interview, who faces it, and where it bites hardest.

Author
Amney, Founder at Dataford
Dataset
725 questions
Period
2026 question bank
725
statistics and experimentation questions analyzed
35%
rated hard — the hardest topic in the data interview
41%
are A/B testing and experimentation
26%
rated easy — the lowest easy share of any topic

Summary · Key findings

01

Statistics is the hardest topic in the data interview. 35% of its questions are rated hard — the steepest difficulty profile of any subject — and only 26% are easy, the lowest easy share anywhere. This is the round that humbles strong candidates.

02

It is really two subjects. Probability and statistics make up the majority, and A/B testing and experimentation the other 41%. The first tests how you reason about uncertainty; the second, how you design and read a controlled test.

03

Data Scientists and Data Analysts own it. These two roles carry almost all of the topic. It barely touches engineering seats — a clear signal that statistical rigor is what separates a data hire from a software one.

04

Experimentation is where product-minded analysts shine. A/B design, common pitfalls, and interpreting ambiguous results are a large, learnable slice — and the part most directly tied to product decisions.

05

The companies that run experiments at scale ask the most. Meta leads, with Intuit, Uber, Google, and Netflix behind — the firms whose products are tuned through constant experimentation, and who interview for that discipline.


Ask a room of candidates which round scares them most and the answer is almost always statistics. It turns out that fear is well-calibrated: by the numbers, this is the hardest part of the data interview.

We separated the topic into its two real halves — the probability and statistics that test how you reason about uncertainty, and the A/B testing that tests how you run a controlled experiment — and looked at how hard each gets, who faces it, and which companies lean on it.

The encouraging part is that the topic is finite and learnable. The hard part is that there is nowhere to hide in it. The full method is at the end.


The bar

The steepest profile in data

Every topic has hard questions. Statistics has the most of them, in proportion — and the fewest easy ones. There is no soft on-ramp.

Share of questions by difficulty
The hardest topic in the interview
25.9%
38.8%
35.3%
Easy 25.9%Medium 38.8%Hard 35.3%

Part of what makes it hard is that it is genuinely two subjects, each with its own way of catching you out.

The two halves of the topic
Reasoning about uncertainty, and testing it
59%
41%
Probability & statistics 59%A/B testing & experimentation 41%

Probability and statistics reward clear thinking under uncertainty; A/B testing rewards knowing the traps — peeking, novelty effects, misread significance. A candidate strong in one and weak in the other still has a visible hole, and this is the topic where holes show.


Who and where

The line between data and software

Statistics is where a data role stops looking like a software role. Two seats carry almost all of it, and a handful of experiment-driven companies ask the most.

Statistics questions by role
Scientists and analysts carry it
Data Scientist
53
Data Analyst
40
Business Analyst
6
Research Scientist
6
Companies that test it most
Built on experimentation
Meta
39
Intuit
22
Uber
17
Google
11
Netflix
8

Meta, Intuit, Uber, Google, and Netflix lead — products that are tuned through constant experimentation, so they interview for the discipline that makes experimentation trustworthy. If one of these is your target, the statistics round is not optional polish; it is central.


Outlook

How to prepare for the statistics round

Treat it as the topic you cannot skim. Because it is the hardest and the least forgiving, it gives the best return on deliberate practice — a candidate who is genuinely solid here stands out precisely because so many are not.

Cover both halves. Get fluent with the probability and inference fundamentals, then drill experimentation until the common pitfalls are reflexes. For product-facing roles especially, the ability to design a clean A/B test and defend its reading is one of the strongest signals you can send.


Drill the round candidates fear most

Real statistics and experimentation questions, from A/B test design to experiment pitfalls.

Practice statistics questions

FAQ

Frequently asked questions

What is the hardest topic in data interviews?+

Statistics and probability. 35% of its questions are rated hard — the steepest profile of any topic — and only 26% are easy, the lowest easy share anywhere. It is consistently the round candidates find most punishing.

What do statistics interview questions cover?+

Two halves: probability and statistics (reasoning about uncertainty, distributions, inference) and A/B testing and experimentation (designing a test, avoiding pitfalls, interpreting results). Experimentation is about 41% of the topic.

How important is A/B testing in data interviews?+

Very, for product-facing roles. It is roughly 41% of the statistics topic and the part most tied to real product decisions — designing experiments, spotting biases, and reading ambiguous results.

Which roles get statistics questions?+

Mainly Data Scientists and Data Analysts. Statistical rigor barely appears in engineering loops, which makes it one of the clearest dividing lines between a data role and a software one.

Which companies ask the most statistics questions?+

The companies that live on experimentation: Meta leads, followed by Intuit, Uber, Google, and Netflix — products tuned through constant A/B testing, interviewing for that discipline.


Methodology

How this report was built

This report draws on 725 published questions across Dataford's Statistics & Probability and A/B Testing & Experimentation categories, each tagged with a difficulty, the roles it applies to, and any associated companies.

Role and company figures use those tags; a question can apply to several roles, and company coverage is uneven, so company counts indicate emphasis rather than a precise ranking.

The bank reflects the statistics and experimentation questions companies ask for these roles as captured and structured by Dataford. Figures are current as of June 2026.