文档介绍：388 Resampling: The New Statistics
CHAPTER
How Large a
24 Sample?
Issues in Determining Sample Size
Some Practical Examples
Step-Wise Sample-Size Determination
Summary
Issues in determining sample size
Sometime in the course of almost every study—preferably
early in the planning stage—the researcher must decide how
large a sample to take. Deciding the size of sample to take is
likely to puzzle and distress you at the beginning of your re-
search career. You have to decide somehow, but there are no
simple, obvious guides for the decision.
For example, one of the first studies I worked on was a study
of library economics (Fussler and Simon, 1961), which required
taking a sample of the books from the library’s collections.
Sampling was expensive, and we wanted to take a correctly
sized sample. But how large should the sample be? The longer
we searched the literature, and the more people we asked, the
more frustrated we got because there just did not seem to be a
clear-cut answer. Eventually we found out that, even though
there are some fairly rational ways of fixing the sample size,
most sample sizes in most studies are fixed simply (and irra-
tionally) by the amount of money that is available or by the
sample size that similar pieces of research have used in the
past.
The rational way to choose a sample size is by weighing the
benefits you can expect in information against the cost of in-
creasing the sample size. In principle you should continue to
increase the sample size until the benefit and cost of an addi-
tional sampled unit are equal. 1
The benefit of additional information is not easy to estimate
even in applied research, and it is extraordinarily difficult to
estimate in basic research. Therefore, it has been the practice
Chapter 24—How Large a Sample? 389
of researchers to set up target goals of the degree of accuracy
they wish to achieve, or to consider various degrees of accu-
racy