Research methods as a way of talking about behavior

I have more than a passing interest in research methods, particularly those that have driven the basic and applied research literature of behavior analysis over the years. A fourth edition of “S&T” is well underway, which means I’ve been writing about the methods for studying behavior for more than 35 years (see Johnston & Pennypacker, 1980, 1993, 2009). Along the way, I’ve observed that how I talk about behavior is greatly informed by my understanding of how we study it. Although our philosophy of science might seem fairly distinct from the experimental and clinical methods of our science, they are inseparable verbal repertoires.

Assumptions and implications

When people talk about behavior in a casual, everyday manner, they implicitly assume many things about it. For example, they accept that the particular piece of behavior at issue has certain characteristics of form and function. They may throw out a common label or familiar description – “I cooked bacon for breakfast” – but such accounts imply certain features and parameters. The behavior of cooking bacon involves specific topographies and functional outcomes, and effective communication requires a listener who shares a history that allows them to respond appropriately.

Furthermore, colloquial references to behavior assume these characteristics are consistent, meaningful, and describable in terms of its dimensional features. If someone says their dog loves to chase squirrels, it is implicit that this behavior occurs with some regularity. A comment about how much a child loves to play video games implies a way of looking at the distribution of the behavior over time in which extended periods of responding are relevant. Any quantitative specifications also imply questions of how the values were determined. Of course, statements about how much or little of a behavior there is raises matters of observation and sampling. Arguing that “He does it all the time.” or “She almost never does it.” implies some basis for the assessment. Asking, “How do you know?” usually provides no more than a vague hint about how the estimate was determined.

Everyday comments about behavior are very likely to imply causal relationships as well, which raise the matter of how they were evaluated and how true they are. Our colloquial dialect offers endless possibilities, but most of the easy ones involve mental influences. Even when environmental or biological factors are acknowledged, mental processing often has the last – if not the most powerful – say in the matter. So someone might propose, “She only does that when people are around,” adding “but that’s because she is so insecure.”

A question of methods

Only behavior analysts are likely to be aware of all of these assumptions and implications. After all, we’re trained to minimize casual ways of talking about behavior in favor of more science-based alternatives, which depend on the methods of science for their credibility. We see methodological issues in everyday mentions of behavior, and though we don’t usually have any way of resolving them at the moment, we take them into account in how we think about the behavior at issue.

For instance, everyday references to particular behaviors almost inevitably raise questions about their features. Even when a referenced behavior is fully physical and evidently free of any mentalistic taint, we wonder about the key features of the response class. What counts as instances of the behavior? Is the speaker defining it topographically or functionally? What are the outcomes that hold it together? What does “cooking bacon” actually involve? What are its limits and consequences? Surely they are not important in the vernacular moment, but behavior analysts might wonder about such things.

Then there are the matters of what dimensions might best describe the behavior. Rate? Duration? Latency? What sampling procedure is implied in references to the behavior? If a dog “loves to chase squirrels,” what measure might best evaluate this assessment? Should “loving to chase squirrels” be defined in terms of how fast it runs upon seeing a squirrel, the latency between seeing a squirrel and starting out after it, or how often chasing occurs across opportunities? How would we sample such behavior to get the evidence?

When colloquial mentions of behavior come with causal attributes, we naturally wonder what is really going on. Our speculations are largely formed by our understanding of learning processes, but that’s only the beginning. What environmental events are playing what role? What evidence might there be for one alternative versus another? How might our observations be misguided? We cannot know the answers to such questions in the moment, but it is our business to be curious, even as we participate in the ongoing conversation.

Finally, all of these questions are relatively straightforward when the behavior at issue really is behavior, but what about the many circumstances in which no physical features are suggested. What do we wonder when the referenced “action” is evidently mentalistic in its “features?” Of course, we know to skip past the mentalistic provenance and look for actual behavior, and this is when our special perspective really shines. When a friend says that their child always wants to play video games, we know better than to think of “wanting” as a behavior. Instead, we reconceptualize the description in terms of the behavior of “playing video games” and consider how what our friend calls “wanting” might be turned into measures of video game playing that reflected its strength.

A methodological sieve

I’ll concede that I may be more methodologically aware than most behavior analysts, but I think about this kind of stuff all the time. When I’m watching the local news on TV, for example, I’m filtering the way I think about some mentioned behavior through a methodological sieve in a way that intertwines a radical behavioristic point of view and scientific method. A report on Viet Nam veterans going back to Viet Nam as a way of ameliorating lingering feelings about their war time service becomes an occasion for wondering about what these “feelings” might really be and how we might characterize them as behavior analysts – a curiosity that can’t be resolved without recourse to methodological questions.

The objective of talking about behavior in a way that integrates our philosophy of science and research methods is insuring that the sources of control over our behavioral interests are appropriately dominated by our scientific training, rather than by our colloquial history. It is not enough to shun mentalism. By itself, radical behaviorism only guides us toward the phenomenon of behavior. The scientific methods of behavior analysis are then required to take full advantage of this focus and make our way of talking about behavior fully operational.

Reference

Johnston, J. M. & Pennypacker, H. S. (1980). Strategies and tactics of human behavioral research. Hillsdale, NJ: Erlbaum Associates.

Johnston, J. M. & Pennypacker, H. S. (1993). Strategies and tactics of behavioral research. Hillsdale, NJ: Erlbaum Associates.

Johnston, J. M. & Pennypacker, H. S. (2009). Strategies and tactics of behavioral research. New York, Routledge.