ABSTRACT
Statistical theory is often directly relevant to the comparative approach in psychology. The comparison of treatment groups with control groups is central to classical experimental design. In general, the usefulness of statistical methods in comparing groups depends on the magnitude of the difference between the groups. If groups differ enormously, one has little need for statistical methods to document the fact. A major consideration in employing statistical tests is robustness. Robustness refers to how well a statistical procedure can survive the translation from theoretical cases to real data. Statistical theory has long been able to deal with the effects of multidimensionality of both independent and dependent variables. Practical application of theory had to await the development of modern computers. McCall has recently pointed out that the developmental psychologist is faced with the perplexing statistical tasks of describing both the lawful pattern of change in mean performance over time and the degree of stability or instability in individual differences over time.
