← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Senior

Senior
Jun 2026

Summary

A technical screen for a DS role at Meta that went deep on experiment design. The whole session was basically one long power analysis problem broken into pieces, and it got harder as it went.

Questions Asked (5)

Q1

Given a baseline friend-accept rate of 9% and a target of detecting a 0.5 percentage point absolute lift, calculate the per-arm sample size and how many calendar days the experiment would need to run, assuming a two-sided test, 80% power, equal allocation, and 10 million eligible users per day.

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

I knew the formula but fumbled the variance calculation at first.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clearly stating the assumptions and the formula for sample size per arm for a two-proportion z-test. Then plug in the baseline rate (9% = 0.09) and the target rate (9.5% = 0.095), compute the required sample size per arm, and finally divide by daily eligible users (10 million) to get the number of days, accounting for equal allocation.

Pro tip: Always sanity-check the sample size with a quick approximation: for small lifts, the required sample per arm is roughly 16 * p(1-p) / (delta^2). This helps catch calculation errors and shows you understand the underlying statistics.

1. Define parameters and hypotheses

State the baseline conversion rate (p1 = 0.09), target rate (p2 = 0.095), absolute lift (delta = 0.005), significance level (alpha = 0.05, two-sided), power (1 - beta = 0.80), and equal allocation ratio.

2. Choose the sample size formula

Use the standard formula for two-proportion z-test: n per arm = ( (z_{1-alpha/2} * sqrt(2 * p_bar * (1-p_bar)) + z_{1-beta} * sqrt(p1*(1-p1) + p2*(1-p2)) )^2 ) / (delta^2), where p_bar = (p1+p2)/2.

3. Compute the sample size per arm

Plug in z_{0.975} = 1.96, z_{0.80} = 0.84, p1 = 0.09, p2 = 0.095, delta = 0.005. Calculate p_bar = 0.0925. Then compute the required n per arm. (Result: approximately 1,048,000 per arm, but show the calculation.)

4. Calculate total sample and experiment duration

Total sample needed = 2 * n per arm. With 10 million eligible users per day and equal allocation, daily users per arm = 5 million. Days = n per arm / 5,000,000. Round up to the nearest whole day.

5. Interpret and discuss practical considerations

Mention that the calculated duration assumes all eligible users are exposed and that the metric is stable. Discuss potential adjustments for novelty effects, seasonality, or multiple testing, and the importance of checking for sample ratio mismatch (SRM).

Key Points to Mention

  • Two-sided test at alpha = 0.05 and power = 80% (beta = 0.20).
  • Use of pooled variance under the null for the first term and unpooled variance for the second term in the sample size formula.
  • Baseline rate 9% and target rate 9.5%, so absolute lift = 0.5 percentage points.
  • Equal allocation means 50% of users in control and 50% in treatment.
  • Daily eligible users = 10 million, so 5 million per arm per day.
  • Resulting sample size per arm and total days, with rounding up to ensure sufficient power.

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

Q2

How does your sample size estimate change if the true standard deviation turns out to be 20% higher than what you assumed?

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

Sample size scales with variance, so a 20% higher SD means variance goes up by 44% (1.2 squared).

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by recalling the sample size formula for comparing means, which is proportional to the variance (σ²). Then explain that a 20% increase in standard deviation leads to a 44% increase in required sample size (since 1.2² = 1.44). Finally, discuss practical implications such as the need to adjust the experiment duration or accept reduced power.

Pro tip: Mention that in practice, you can mitigate the impact by using a more sensitive metric (e.g., relative instead of absolute) or by employing sequential testing to allow early stopping, but always pre-register the analysis plan to avoid p-hacking.

1. Recall the sample size formula

State the standard formula for sample size per group in a two-sample t-test: n = 2*(Zα/2 + Zβ)² * σ² / Δ², where σ is the standard deviation and Δ is the minimum detectable effect.

2. Identify the relationship between σ and n

Note that sample size is directly proportional to the variance (σ²). Therefore, a 20% increase in σ results in a 44% increase in required sample size (1.2² = 1.44).

3. Quantify the new sample size

If the original sample size was n, the new required sample size becomes 1.44n. For example, if n was 10,000 per group, it would increase to 14,400 per group.

4. Discuss practical implications and trade-offs

Explain that this increase means longer experiment duration, higher cost, or reduced power if the sample size is fixed. Suggest options like increasing the sample size, accepting lower power, or using a more sensitive metric.

5. Recommend proactive measures

Advise using pilot data or historical data to better estimate σ, and consider variance reduction techniques (e.g., CUPED) to mitigate the impact of higher-than-expected variance.

Key Points to Mention

  • Sample size formula: n ∝ σ²/Δ²
  • 20% increase in σ → 44% increase in n (since 1.2² = 1.44)
  • Impact on power: if n is fixed, power decreases
  • Trade-offs: longer duration, higher cost, or lower power
  • Variance reduction techniques (e.g., CUPED, stratification)
  • Use pilot data to estimate σ more accurately

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

Q3

For a three-arm experiment with a shared control group, where you want to detect a 0.3 percentage point lift for either treatment versus control at a familywise error rate of 5%, which multiple comparison correction would you use and why? Then recalculate per-arm sample size and total duration with 20% traffic to control and 40% each to the two treatment arms.

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

This is where I got a bit tangled.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by explaining that with two treatment arms compared to a shared control, you have two primary comparisons, so you need to control the familywise error rate (FWER) at 5%. Recommend a multiple comparison correction such as Bonferroni or Holm-Bonferroni, and justify why it's appropriate for this small number of comparisons. Then, recalculate the per-arm sample size using the corrected alpha, and adjust for the unequal traffic allocation (20% control, 40% each treatment) to determine the total sample size and duration.

Pro tip: Mention that while Bonferroni is simple and controls FWER, it can be conservative; if the comparisons are independent, you might consider a slightly less conservative approach like Holm-Bonferroni, but for two comparisons the difference is minimal. Also, emphasize that the shared control means the control data is used for both comparisons, so the effective sample size for control is larger, which can be leveraged in power calculations.

1. Identify the number of comparisons and error rate control

Recognize that there are two primary comparisons (Treatment A vs Control, Treatment B vs Control). To maintain a familywise error rate of 5%, you need to adjust the per-comparison alpha.

2. Choose a multiple comparison correction method

Select Bonferroni or Holm-Bonferroni. For two comparisons, Bonferroni splits alpha into 0.025 per comparison. Explain the trade-off between FWER control and power.

3. Calculate per-arm sample size with corrected alpha

Use the standard sample size formula for two proportions with alpha = 0.025 (two-sided) and power = 80% (or 90%). Compute the required sample size per arm assuming equal allocation first.

4. Adjust for unequal traffic allocation

Given 20% to control and 40% to each treatment, the allocation ratio is 1:2:2. Recalculate the total sample size needed to achieve the same power, accounting for the unequal split. The control group will need to be larger relative to each treatment to maintain power.

5. Compute total duration

Using the total sample size and the daily traffic rate (if provided, otherwise assume a rate), calculate the number of days required to complete the experiment.

Key Points to Mention

  • Familywise error rate (FWER) vs per-comparison error rate
  • Bonferroni correction: alpha divided by number of comparisons (0.05/2 = 0.025)
  • Holm-Bonferroni as a step-down procedure that is uniformly more powerful than Bonferroni
  • Sample size formula for two proportions: n = (Z_{alpha/2} + Z_{beta})^2 * (p1(1-p1) + p2(1-p2)) / (p1-p2)^2
  • Unequal allocation ratio: control:treatment1:treatment2 = 1:2:2, so the control group needs to be larger in absolute terms to achieve the same power
  • Duration calculation: total sample size divided by daily traffic, considering the allocation proportions

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

Q4

What are the consequences of checking your experiment's p-value every day against a naive p < 0.05 threshold, and how would a sequential testing approach like O'Brien-Fleming change your stopping boundaries?

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

Classic question and I knew it well.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by explaining the multiple comparisons problem and how daily peeking inflates the Type I error rate. Then describe how sequential testing methods like O'Brien-Fleming adjust boundaries to control the overall alpha while allowing early stopping. Finally, discuss the trade-offs and practical implications for experiment design and decision-making.

Pro tip: Mention that O'Brien-Fleming is conservative early on, which is ideal when you want to avoid stopping too early on noise, but it requires pre-planning the number and timing of looks. Also, note that many tech companies use sequential testing to enable continuous monitoring without inflating false positives.

1. Explain the problem with naive daily checks

Describe how repeatedly testing at p < 0.05 each day increases the chance of a false positive (Type I error) due to multiple comparisons. Quantify the inflation: e.g., with 20 peeks, the false positive rate can exceed 30%.

2. Introduce sequential testing and O'Brien-Fleming

Define sequential testing as a method that adjusts significance thresholds at each interim analysis to control the overall alpha. Explain that O'Brien-Fleming uses conservative boundaries early (very small alpha) and more lenient boundaries later, minimizing the risk of early stopping on noise.

3. Describe how O'Brien-Fleming changes stopping boundaries

Detail that O'Brien-Fleming boundaries are based on the cumulative alpha spent, often using a function like the Lan-DeMets approach. Early looks require extremely small p-values (e.g., p < 0.0001), while the final look uses a threshold close to the nominal alpha (e.g., p < 0.05).

4. Discuss implications for experiment design and decision-making

Explain that sequential testing requires pre-specifying the number and timing of interim analyses. It allows for early stopping if effects are strong, but may delay decisions if effects are moderate. Also, note that it controls false positives but can increase sample size if the effect is small.

5. Conclude with practical recommendations

Summarize that naive daily checks are invalid, and sequential testing like O'Brien-Fleming provides a rigorous framework for monitoring experiments. Recommend using it when continuous monitoring is desired, and always pre-register the analysis plan.

Key Points to Mention

  • Multiple comparisons problem and inflation of Type I error rate with repeated testing
  • Definition of sequential testing and its goal to control overall alpha
  • O'Brien-Fleming boundaries: conservative early, lenient later
  • Lan-DeMets alpha spending function for flexible monitoring
  • Trade-offs: early stopping benefits vs. increased sample size and complexity
  • Importance of pre-specifying the number and timing of interim analyses

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

Q5

When would you switch from a standard variance reduction approach to cluster-robust standard errors, for example clustering at the feed level rather than the user level, and how does that affect your required sample size?

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

I talked about intraclass correlation and how if users within a cluster behave similarly, your effective sample size is smaller than the raw count.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Explain that the choice between standard variance reduction and cluster-robust standard errors depends on the correlation structure of the data and the unit of randomization. Discuss how clustering at a higher level (e.g., feed) accounts for within-cluster correlation, and describe the impact on sample size requirements, including the design effect and effective sample size.

Pro tip: Always consider the trade-off between bias and variance: cluster-robust standard errors are consistent but less efficient, so you need a larger sample to achieve the same power. Also, remember that the clustering level should match the level at which treatment is assigned or where interference occurs.

1. Identify the correlation structure

Determine if there is correlation among observations within clusters (e.g., users within a feed, or multiple observations per user). This often arises from group randomization or interference.

2. Assess the impact on standard errors

Explain that ignoring clustering leads to underestimated standard errors and inflated false positive rates. Cluster-robust standard errors adjust for this by allowing arbitrary correlation within clusters.

3. Decide when to switch

Switch when the intra-cluster correlation (ICC) is non-negligible, or when the treatment is assigned at a cluster level. Also consider if there is interference between units within a cluster.

4. Quantify the effect on sample size

Introduce the design effect (DEFF) = 1 + (m-1)*ICC, where m is the average cluster size. The required sample size increases by this factor to maintain power.

5. Consider practical implications

Discuss how larger sample sizes may be infeasible, and potential alternatives like using more clusters, reducing cluster size, or using more efficient estimators (e.g., fixed effects).

Key Points to Mention

  • Intra-cluster correlation (ICC) and its role in determining the need for cluster-robust standard errors.
  • Design effect formula and how it inflates required sample size.
  • The difference between clustering at user level vs. feed level and when each is appropriate.
  • The trade-off between bias and efficiency: cluster-robust SEs are consistent but less efficient.
  • The importance of matching the clustering level to the randomization unit or interference structure.
  • Potential solutions when sample size is limited, such as increasing the number of clusters or using fixed effects.

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