Main Article Content
Many independent variables related te developrnent are better approached with group than single-subject methods, either because experimental manipulation would be unethical or because they are only indexes of the real causal variables. Both kinds of variables can be studied in correlational group research, but correlational research deals with effect-effect relations, not functional relations, which are the subject-matter of behavior analysis. Understanding is the goal of behavíor analysis; it is testedby rather than equivalent te the demonstration of control, and the dernonstration can come from experimental group research in addition to single-subject research. Deviations of individual performance from group means is not necessarily a problem; variability can be dealt within experimental group research; between-group experiments can dernonstrate functional relations; and statistical inference done properly is objective.