Journal of Statistics: Advances in Theory and ApplicationsAuthor's: Dieter Rasch and Klaus D. Kubinger
Pages: [43] - [59]
Received Date: February 27, 2010
Submitted by:
After explaining seven steps in empirical research, we demonstrate the procedure by using an example from psychological testing. This practical problem can be solved traditionally by first determining the size of an empirical study, performing the study, and then analysing it. But, we show an alternative, sequential approach, where sampling and analysing data take turns. On average, the sequential approach results in smaller sample sizes than the traditional one. In our example for the traditional approach, 59 observations are needed in each of two groups (i.e., 118 in total), whereas the sequential approach required 62 observations in total. The sequential procedure should be used especially, when observations are expensive but recruiting times are short.
planning experiments and research studies, calculating sample size, type-I-and type-II-risk, requirements of precision, sequential testing.
