10.10. Summary

In writing a single chapter on sample-size analysis, one can either strive for breadth or depth. We opted to cover two examples in depth, and thus we failed to even mention any of the vast array of other tools now available to help investigators carefully assess and justify their choices for sample sizes across the statistical landscape. What have we not discussed? The list of topics and references is too long to begin and would soon be outdated anyway.

What readers need to understand is that if they have a sample-size analysis issue, there may be good methodological articles and strategies that address it. If no such help can be found, then Monte Carlo simulation can provide results that are entirely satisfactory. In fact, some excellent statisticians now use simulation for all such problems, even for traditional ones that have sound "mathematical' solutions that are widely used.

We hope the two examples given here provide a rich context to fashion good strategies to address other problems one may encounter. Though the methods may vary widely, the core concepts and issues do not.