Code Quiz

Confidence Interval From Sample Mean

Spot the bug in computing a 95% confidence interval for a sample mean.

Codepython

import numpy as np

def mean_confidence_interval(sample, z=1.96):
    n = len(sample)
    mean = np.mean(sample)
    # sample standard deviation
    std = np.std(sample, ddof=1)
    # standard error of the mean
    se = std / n
    margin = z * se
    return mean - margin, mean + margin

data = np.array([4.1, 3.9, 4.4, 4.0, 3.7, 4.2, 4.3, 3.8])
low, high = mean_confidence_interval(data)
print(f"95% CI: [{low:.3f}, {high:.3f}]")

What is the bug in this confidence interval calculation?

← PreviousPeeking at A/B Test Results Next →Interpreting a 95% Confidence Interval