Can bootstrap help reduce variance

An interviewer might ask: “Can bootstrap help reduce variance?”

Respond precisely. Separate these two ideas:

1. Applying the bootstrap to estimate variance or standard error.

2. Using bootstrap-inspired techniques to actually lower an estimator’s variance.

Provide at least one concrete case in which the bootstrap does not reduce variance, and one case in which bootstrap-style resampling can reduce variance (or produces a lower-variance predictor), then explain the reason.

Overview:

This question tests your grasp of bootstrap resampling, the difference between estimating an estimator’s variance and techniques that truly lower estimator variance, and related ensemble thinking for a Data Scientist position in Statistics & Math.

Read the full data science interview experience this question came from.