← Amazon Interview Insights

Amazon·Data Scientist·Technical Phone Screen·Intermediate

IntermediatePrefer not to say
May 2026

Summary

Amazon stats question, just the one. Felt more like a quick screen than a full interview round.

Questions Asked (1)

Q1

You flip a coin 10 times and get only one head. What null hypothesis would you set up to test whether the coin is fair, and what p-value do you get?

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

The null is straightforward, coin is fair so p=0.5.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

State the null hypothesis that the coin is fair (p=0.5) and compute the p-value as the probability of observing 1 or fewer heads in 10 flips under the null. Use the binomial distribution to calculate the two-sided p-value, and briefly mention that this is a small sample so exact methods are appropriate.

Pro tip: Always clarify whether the test is one-sided or two-sided; in fairness testing, a two-sided test is standard, but Amazon interviewers may appreciate you noting that the alternative could be one-sided if you suspect bias in a specific direction.

1. Define the null hypothesis

Clearly state H0: the coin is fair, i.e., the probability of heads p = 0.5. Also mention the alternative hypothesis H1: p ≠ 0.5 (two-sided) or p < 0.5 (one-sided) depending on the context.

2. Choose the test statistic and distribution

The number of heads in 10 flips follows a Binomial(10, 0.5) distribution under H0. The test statistic is the observed count of heads, which is 1.

3. Compute the p-value

For a two-sided test, calculate the probability of observing a result as extreme or more extreme than 1 head. This includes 0 heads and 1 head, and symmetrically 9 and 10 heads. Sum the binomial probabilities: P(X ≤ 1) + P(X ≥ 9) = 2 * (P(X=0) + P(X=1)) = 2 * (0.0009766 + 0.0097656) ≈ 0.0215.

4. Interpret the p-value

The p-value is approximately 0.0215. At a significance level of 0.05, we reject the null hypothesis and conclude that there is evidence the coin is not fair. However, note that with only 10 flips, the test has low power.

Key Points to Mention

  • Null hypothesis: p = 0.5 (fair coin)
  • Alternative hypothesis: p ≠ 0.5 (two-sided) or p < 0.5 (one-sided)
  • Binomial distribution with n=10, p=0.5 under H0
  • p-value calculation: sum of probabilities for 0, 1, 9, 10 heads (two-sided)
  • p-value ≈ 0.0215, which is less than 0.05, so reject H0
  • Caveat: small sample size limits power and generalizability

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