Some Final Thoughts: Methodological Eclecticism

SOME FINAL THOUGHTS:  METHODOLOGICAL ECLECTICISM

Questions about human behavior and mental processes are among the most interest-ing questions we can ask, and people have been trying to answer them for thousands of years. One of psychology’s great contributions has been the application of the scientific method to these questions—with great benefit, because we have learned an enormous amount.

In this, we have tried to sketch what the scientific method involves in psychology. We’ve emphasized the points that apply to all research, such as the impor- tance of systematic data collection and the need to remove ambiguity or to minimize a procedure’s demand characteristics. But, along with these methodological similarities, we’ve also mentioned the diversity of research. We have emphasized random samples as well as the value of case studies. We have also talked about the contrast between observational studies and experiments.

With an eye on this diversity of research types, it’s important to emphasize that each type has its advantages, and none is better than the others. A random sample allows us to study a small group and then draw broad conclusions from it. Even so, case studies are sometimes appropriate or even necessary. Suppose, for example, an investigator is studying an individual (perhaps someone with brain damage) who truly is unique; in a situation like this, a larger-scale study of multiple participants is just not possible. Often, the case study provides insights or suggests effects that can then be pursued with a larger group—but sometimes the case study is by itself deeply and richly instructive. This feature is, by the way, not unique to psychology: Geologists routinely report “case studies” examining a single volcano; oceanogra-phers study single tsunamis. In both of these disciplines, the investigators under-stand that they’re “merely” describing a single case, but they proceed because they know that a single case can offer powerful insights into more general issues and phenomena.

Similarly, observational studies are immensely valuable because they allow us to learn about events or behaviors that we couldn’t possibly manipulate. Besides that, they’re often done in natural settings that allow us to observe events ranging from how children behave in an actual school setting to how wolves behave in their normal habi-tat. But, as we’ve discussed, observational studies are also limited. Because they usually cannot inform us about cause-and-effect relations, observational studies are weaker in this specific regard than experiments are.

Experiments, in turn, are immensely valuable because they allow us to untangle cause-and-effect relationships. But experiments aren’t always possible. In many situations, manipulation of a variable—or random assignment—is either logistically impossible or forbidden by ethical constraints. When random assignment is possible, it provides a powerful benefit: It virtually guarantees that the two groups being compared were matched to each other at the outset, so that changes between them can be attributed to the experimental manipulation. But this benefit has a cost attached: Randomly assign-ing participants or holding all the variables in a situation constant except the experimen-tal manipulation requires an experimenter to be in control of the research situation. Gaining that control typically means introducing some artificiality into the setting. The artificiality, in turn, raises questions about external validity—that is, questions about whether the experiment accurately mirrors the real-world phenomenon that the investi-gator hopes to understand.

How do researchers manage these trade-offs, or decide which method to use? They make the decision case by case; but for most circumstances, their preferred approach is to use multiple methods in the hope that the different methods will converge on the same answer. In this way, each of the methods complements the other, and the particular strengths of each method can help address concerns that might have been raised by the weaknesses of other methods. This strategy gives researchers a powerful means of arguing that the results are not some peculiar by-product of using this or that research tool; instead, they ’re telling us about the world as it truly is.