Inference on means using the bootstrap (Q794088)

For a linear combination of means of k (\(\geq 2)\) populations (based on independent samples), it is shown that the bootstrap approximation to the distribution of the Studentized estimator is better than that of the estimator itself. In this context, it is assumed that the underlying distribution has finite moments up to the 6th order, and there seems to be scope for weakening this stringent moment condition.