The New Statistics

“Exploration has a second meaning: Running pilot tests to explore ideas, refine procedures and tasks, and guide where precious research effort is best directed is often one of the most rewarding stages of research. No matter how intriguing, however, the results of such pilot work rarely deserve even a brief mention in a report. The aim of such work is to discover how to prespecify in detail a study that is likely to find answers to our research questions, and that must be reported. Any researcher needs to choose the moment to switch from
not-for-reporting pilot testing to prespecified, must-be-reported research.”


Comments are closed.