Which statement is NOT a standard characteristic of bootstrap resampling?

• A. It relies on resampling from the observed data to approximate the sampling distribution.
• B. It does not assume a specific theoretical distribution for the population.
• C. It generates synthetic data from a fitted model as the primary resampling method.
• D. It is used to estimate confidence intervals.

Answer: (C)

Bootstrap resampling centers on using the observed data itself to approximate the sampling distribution of a statistic by repeatedly sampling with replacement. This approach lets us estimate how a statistic would vary if we could repeatedly draw samples from the population, without needing to specify a particular theoretical distribution for the population.

That’s why it’s correct to say the method relies on resampling from the observed data and does not assume a specific theoretical population distribution. It’s also standard to use bootstrap methods to construct confidence intervals, by looking at the distribution of the statistic across many resampled samples. What isn’t a standard characteristic is generating synthetic data from a fitted model as the primary resampling method. Generating data from a fitted model describes a parametric or model-based bootstrap variant, rather than the typical nonparametric bootstrap that resamples the actual data.