← Capital One Interview Insights

Capital One·Data Scientist·Technical Phone Screen·Senior

SeniorPrefer not to say
Jul 2026

Summary

Capital One data scientist interview that was basically a deep stats exam disguised as an A/B testing question. One long multi-part problem covering everything from t-test mechanics to effect sizes to robustness arguments. Dense.

Questions Asked (5)

Q1

Given two independent samples from an A/B test on per-user daily watch time, compute both the pooled-variance t-statistic and the Welch t-statistic, then report the corresponding p-values for each.

A/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

I knew the formulas but fumbled the pooled variance denominator at first, mixed up whether you use nA+nB-2 or something else.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

First, clarify the data structure and assumptions: two independent samples of per-user daily watch time, likely with unequal variances and sample sizes. Then compute the pooled-variance t-statistic (assuming equal variances) and the Welch t-statistic (not assuming equal variances), and derive p-values from the respective t-distributions with appropriate degrees of freedom. Finally, compare the results and discuss which is more appropriate given the context.

Pro tip: Always check the equal-variance assumption (e.g., with Levene's test or by comparing variances) before choosing the pooled t-test; in practice, Welch's t-test is more robust and often preferred unless variances are clearly equal.

1. Understand the data and assumptions

Identify the two independent samples (control and treatment) of per-user daily watch time. Note sample sizes (n1, n2), means (x̄1, x̄2), and standard deviations (s1, s2). Check assumptions: independence, normality (or large samples), and homogeneity of variances.

2. Compute pooled-variance t-statistic

Calculate the pooled variance: s_p^2 = [(n1-1)s1^2 + (n2-1)s2^2] / (n1+n2-2). Then compute the t-statistic: t = (x̄1 - x̄2) / sqrt(s_p^2 * (1/n1 + 1/n2)). Degrees of freedom = n1+n2-2.

3. Compute Welch t-statistic

Calculate the t-statistic: t = (x̄1 - x̄2) / sqrt(s1^2/n1 + s2^2/n2). Degrees of freedom are approximated using the Welch-Satterthwaite equation: df = (s1^2/n1 + s2^2/n2)^2 / [(s1^2/n1)^2/(n1-1) + (s2^2/n2)^2/(n2-1)].

4. Compute p-values

For each t-statistic, compute the two-sided p-value using the corresponding t-distribution with the appropriate degrees of freedom. Report both p-values and compare them.

5. Interpret and discuss

Compare the results: if variances are similar, the p-values will be close; if not, Welch's p-value is more reliable. Discuss the implications for the A/B test conclusion, considering practical significance and business impact.

Key Points to Mention

  • Assumption of equal variances: pooled t-test assumes equal variances, while Welch does not.
  • Degrees of freedom: pooled uses n1+n2-2; Welch uses Satterthwaite approximation.
  • Robustness: Welch's t-test is more robust to unequal variances and unequal sample sizes.
  • When to use each: pooled when variances are equal (e.g., Levene's test p > 0.05), Welch otherwise.
  • Effect of sample size: with large samples, both tests yield similar results due to CLT.
  • Practical significance: statistical significance does not imply business impact; consider effect size and confidence intervals.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q2

The product goal is to increase watch time. Should you use a one-tailed or two-tailed test, and why? Justify your choice rigorously.

A/B Testing & ExperimentationTechnical Trade-offs
Author's notes

This is where I got a little tripped up trying to be too clever.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying that the choice between one-tailed and two-tailed tests depends on whether the business goal implies a directional hypothesis and whether you care about detecting effects in both directions. For increasing watch time, a one-tailed test is appropriate if you only care about detecting an increase, but you must rigorously justify that a decrease is either impossible or irrelevant. Acknowledge the trade-offs in statistical power, Type I error, and practical decision-making.

Pro tip: Mention that in industry, two-tailed tests are often the default because they guard against unexpected negative effects, but if the business truly only cares about positive lift and the cost of missing a negative is low, a one-tailed test can be justified. Also, note that the choice should be pre-registered to avoid p-hacking concerns.

1. Clarify the hypothesis

State the null and alternative hypotheses. For a one-tailed test, the alternative is that the treatment increases watch time (μ_treatment > μ_control). For a two-tailed test, the alternative is that the means are different (μ_treatment ≠ μ_control).

2. Assess business objective and risk

Determine whether the business only cares about detecting an increase (e.g., if a decrease would not change the decision to launch) or if detecting any difference is important. Consider the cost of a false positive (Type I error) versus a false negative (Type II error).

3. Evaluate statistical implications

Explain that a one-tailed test has greater power to detect an effect in the specified direction, but it cannot detect an effect in the opposite direction. A two-tailed test is more conservative and allows detection of effects in both directions.

4. Consider practical decision-making

Discuss how the test choice aligns with the decision rule. If the decision is to launch only if watch time increases, a one-tailed test may be sufficient. However, if a decrease would lead to abandoning the feature or further investigation, a two-tailed test is safer.

5. Make a recommendation and justify

Provide a clear recommendation (e.g., one-tailed if the goal is strictly to increase watch time and negative effects are not a concern) and justify it with the above considerations, emphasizing rigor and pre-registration.

Key Points to Mention

  • Directional vs. non-directional hypotheses
  • Type I and Type II error trade-offs
  • Statistical power and sample size implications
  • Business context: cost of false positive vs. false negative
  • Pre-registration to avoid p-hacking
  • Potential for negative effects and stakeholder expectations

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q3

Construct a 95% confidence interval for the mean difference in watch time using Welch's method.

A/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

Straightforward once you have the Welch t-stat and degrees of freedom.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

First, clarify that Welch's method is used when the two groups have unequal variances and sample sizes. Then, compute the difference in means, the standard error using the Welch-Satterthwaite formula, and the degrees of freedom. Finally, construct the 95% confidence interval using the t-distribution with the calculated degrees of freedom.

Pro tip: Always check the assumptions: Welch's method is robust to unequal variances but still assumes approximately normal sampling distributions. If sample sizes are small and data are skewed, consider bootstrapping or a non-parametric approach.

1. State the formula and assumptions

Write the formula for the confidence interval: (x̄1 - x̄2) ± t* * sqrt(s1²/n1 + s2²/n2). Mention that Welch's method does not assume equal variances.

2. Calculate the standard error

Compute the standard error as sqrt(s1²/n1 + s2²/n2), where s1² and s2² are the sample variances, and n1 and n2 are the sample sizes.

3. Determine degrees of freedom

Use the Welch-Satterthwaite equation to calculate the degrees of freedom: df = (s1²/n1 + s2²/n2)² / [(s1²/n1)²/(n1-1) + (s2²/n2)²/(n2-1)].

4. Find the critical t-value

For a 95% confidence interval, find the t-value corresponding to the calculated degrees of freedom and a two-tailed alpha of 0.05 (i.e., t_{0.975, df}).

5. Construct and interpret the interval

Plug in the values to get the lower and upper bounds. Interpret the interval in the context of the watch time difference, e.g., 'We are 95% confident that the true mean difference in watch time lies between X and Y minutes.'

Key Points to Mention

  • Welch's method is preferred when variances are unequal and sample sizes differ.
  • The standard error formula accounts for independent samples.
  • The Welch-Satterthwaite degrees of freedom can be non-integer.
  • The t-distribution is used instead of the normal distribution when population variances are unknown.
  • The confidence interval provides a range of plausible values for the true mean difference.
  • If the interval includes zero, the difference may not be statistically significant at the 5% level.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q4

Compute Cohen's d and Hedges' g for the effect size between the two groups.

A/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

Knew Cohen's d, forgot Hedges' g corrects for small sample bias with a scaling factor.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clearly defining the two groups and the metric of interest, then compute the pooled standard deviation and the mean difference. Calculate Cohen's d as the standardized mean difference, and apply the small-sample correction factor to obtain Hedges' g. Finally, interpret the effect size in the context of the business problem.

Pro tip: Always report a confidence interval for the effect size, not just the point estimate, to convey uncertainty and statistical rigor. Also, mention that Hedges' g is preferred for small samples (n < 20 per group) to reduce bias.

1. Define groups and metric

Identify the two groups (e.g., control and treatment) and the continuous outcome variable. Ensure the data is clean and the groups are independent.

2. Compute means and standard deviations

Calculate the mean and standard deviation for each group. These are the building blocks for the effect size calculation.

3. Calculate pooled standard deviation

Compute the pooled standard deviation using the formula: sqrt(((n1-1)*s1^2 + (n2-1)*s2^2) / (n1+n2-2)). This assumes equal variances; if variances are unequal, consider using a different denominator.

4. Compute Cohen's d

Cohen's d = (mean1 - mean2) / pooled_sd. This standardizes the mean difference.

5. Compute Hedges' g

Apply the correction factor J = 1 - 3/(4*(n1+n2)-9) to Cohen's d: g = d * J. This reduces bias in small samples.

Key Points to Mention

  • Pooled standard deviation formula and its assumption of equal variances
  • Cohen's d formula: (mean1 - mean2) / pooled_sd
  • Hedges' g correction factor for small sample bias
  • Interpretation of effect size (small, medium, large) using Cohen's benchmarks
  • Confidence intervals for effect size to quantify uncertainty
  • When to use Hedges' g vs Cohen's d (sample size considerations)

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q5

If the normality assumption is questionable but sample sizes are large, why is the t-test still valid? And if heavy tails are suspected, what robust alternative would you propose and what are the trade-offs?

A/B Testing & ExperimentationTechnical Trade-offs
Author's notes

Central limit theorem saves you with large n, that part was easy.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by explaining the Central Limit Theorem (CLT) and how it ensures the sampling distribution of the mean is approximately normal for large samples, making the t-test robust to non-normality. Then discuss robust alternatives like the Mann-Whitney U test or bootstrap methods, and compare their trade-offs in terms of power, assumptions, and interpretability. Finally, tie it back to practical decision-making in A/B testing at Capital One.

Pro tip: Emphasize that with large samples, the t-test is often robust even with heavy tails, but always check for extreme outliers or skewness that could affect the mean; in such cases, consider trimmed means or bootstrapping to maintain power while addressing robustness.

1. Explain the role of the Central Limit Theorem

Describe how the CLT ensures that the sampling distribution of the sample mean approaches normality as sample size increases, regardless of the population distribution, provided the population has finite variance.

2. Discuss the robustness of the t-test

Highlight that the t-test is relatively robust to violations of normality for large samples because the test statistic's distribution is approximately normal, and the t-distribution converges to the normal distribution.

3. Identify robust alternatives for heavy tails

Propose alternatives such as the Mann-Whitney U test (non-parametric), bootstrap hypothesis testing, or trimmed mean tests (e.g., Yuen's test) that are less sensitive to outliers and heavy tails.

4. Compare trade-offs

Discuss trade-offs: non-parametric tests may have lower power if the distribution is not heavy-tailed, bootstrap methods can be computationally intensive but flexible, and trimmed means reduce the impact of outliers but may lose efficiency if tails are not heavy.

5. Relate to business context

Connect the choice of test to practical considerations in A/B testing, such as sample size, effect size, and the cost of false positives/negatives, to recommend the most appropriate method.

Key Points to Mention

  • Central Limit Theorem and its conditions (finite variance, independence)
  • Robustness of t-test with large samples and its limitations with extreme skewness or outliers
  • Mann-Whitney U test as a non-parametric alternative and its assumptions (e.g., similar shape)
  • Bootstrap methods for hypothesis testing and confidence intervals
  • Trimmed means and Winsorized means as robust estimators
  • Trade-offs between power, Type I error control, and interpretability

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.