SOCECOL 10 Chapter Notes - Chapter 10: Ronald Fisher, Ovulation, Analysis Of Variance
Document Summary
Methods toward a science of behavior and experience. Chapter 10: extending the logic of experimentation: within-subjects and. We have stressed two themes: the logical necessity of beginning our experiments with groups that can be assumed to be equal. The random assignment of participants is one of the best techniques for ensuring that no systematic differences are present at the onset of our experiment. However, there are times when random assignment alone may not be the most appropriate approach: the statistical importance of reducing variation, especially the within-groups variation (error variance) component of the f ratio. In this chapter, we discuss two approaches to experimental design that do not use random assignment alone: within-subjects designs. The same participant is exposed to different experimental conditions: matched-subjects designs. Matching can be used both as a control procedure and as an experimental procedure for analyzing difference between groups.
