In an ongoing series in which the APS Student Caucus talks with leading professors, Drew Westen shares his advice for success and challenges facing graduate students. Before becoming professor of psychology at Emory University, Westen was at the Center for Anxiety and Related Disorders at Boston University. He has been chief psychologist at Cambridge Hospital and associate professor at Harvard Medical School. His major areas of research are personality disorders, eating disorders, emotion regulation, implicit (unconscious) processes, psychotherapy effectiveness, and adolescent psychopathology. Westen talks about the connections between clinical and empirical psychology, and their mutual benefits, and offers advice to students who are interested in both.
APSSC: Your work addresses the interface between personality psychology and psychopathology. What can modern clinical practice provide to empirical psychology?
WESTEN: I think clinical practice has two things to offer empirical psychology. The first is a wonderful source of theories and hypotheses. It is no accident that clinicians took unconscious processes as axiomatic a century ago, because the gaps between what their patients were doing (binging and purging, to use a modern example) and what they explicitly understood about what they were doing (“I want to stop, but I just can’t”) rendered the hypothesis of unconscious processes essential.
The philosopher and cognitive scientist Paul Thagard has written about the way our minds equilibrate to solutions or inferences that confer “explanatory coherence” to observations that would otherwise be less coherent or intelligible. I think we do ourselves a disservice when we don’t take seriously the hypotheses and ways of conferring explanatory coherence to clinical data of experienced, highly trained observers.
The data of psychotherapy, though obviously far less useful for establishing causation than the data of replicable research, have three advantages from a scientific point of view: They are by definition externally valid, given that they deal with real problems in people’s lives; they are emotionally charged and personally meaningful in ways that are almost impossible to replicate in the laboratory; and they occur in a longitudinal context as clinicians attempt to understand over time the associative networks that become activated under various conditions and are associated with symptoms or problems in living.
The second thing clinical practice has to offer empirical psychology is an unparalleled natural laboratory for identifying treatment strategies that may be useful to test. Given the wide variation in what clinicians do, and the generally large effect size that we know generic psychotherapies obtain in comparison to controls, it is difficult to imagine that if we identified clinicians who, empirically, obtain the best outcomes with particular kinds of patients, we would have nothing to learn from studying the way they intervene with their patients. The dominant model of treatment research today is unidirectional — researchers take their best guess at what works, test it, usually find that it works better than a control condition selected for its expected lack of efficacy, and then attempt to convince clinicians to use it. It seems to me that alongside these experimenter-initiated interventions, which employ empirical methods for testing treatments, we should think more empirically about how we develop our treatments and intervention strategies, and turn the clinic into a laboratory for developing and identifying promising interventions.
APSSC: Likewise, what is the most important thing, in your mind, that empirical psychology can offer clinical practice?
WESTEN: Just as clinical practice is useful for hypothesis generation, systematic empirical research is essential for hypothesis testing. I think, though, that we’ve understated what clinicians can and should learn from basic science by focusing primarily on randomized clinical trials [RCTs] as the primary data that should inform clinical practice. There is a great deal of basic science on psychopathology — let alone personality, emotion, memory — that is of tremendous relevance to clinical practice. As well, correlational findings in treatment research — such as the consistent finding that the quality of the therapeutic relationship is strongly associated with outcome — can provide substantial guidance for practice.
I would by no means minimize the utility of RCTs, which are essential to evidence-based practice. But I would not assume, for example, that clinicians who keep up with the latest outcome studies by Dave Barlow do better work than those who keep up with his broader corpus of research, which is much wider than just clinical trials.
For example, he and his colleagues have definitively shown that personality variables, such as negative affect, account for much of the comorbidity between seemingly distinct mood and anxiety disorders. A clinician who thinks about how to alter these diatheses in psychotherapy, tests out various strategies with his or her patients, and carefully monitors the outcome is clearly practicing in an empirically-informed, evidence-based way. I think we have thought too narrowly over the last several years about evidence-based practice and need to broaden our understanding of how clinicians and researchers can make use of all available evidence.
APSSC: What advice would you give students interested in both clinical practice and research?
WESTEN: That’s an increasingly tough question to answer, as programs have become increasingly dichotomized into those that aspire to train researchers and those that aspire to train clinicians. I think this polarization is really problematic, because it robs aspiring young researchers of one of the most useful contexts for scientific discovery — immersing themselves in the minds and lives of people seeking help for just the kinds of problems they are interested in studying as researchers — and robs aspiring young clinicians of the opportunity to become participants in the enterprise of science. Would we have research on psychopathy if Cleckley hadn’t immersed himself in the lives of psychopaths?
My suggestion to students who really want to understand psychopathology is to become proficient as clinicians, and my suggestion to students who really want to understand how to treat people effectively is to become proficient in thinking like a scientist in framing and testing clinical hypotheses. And, for those who want to do both, consider getting adequately trained clinically to become licensed, and practice a few hours a week even when the pressures of academia and grant writing make this difficult. It is precisely because I am a stalwart empiricist that I believe we do well to observe from multiple vantage points, and the vantage point of clinical observation allows us to observe different things than the vantage point of the laboratory.
APSSC: More broadly, what advice would you give to graduate students on how to become first-rate researchers?
WESTEN: Most importantly, find something you really want to understand and are excited about. Affect tends to drive cognition. Apprentice yourself to a first-rate mentor or, preferably, work with a primary mentor and do some secondary work with one or two other faculty who are doing interesting work and use methodologies you want to learn. Keep your mind open, and always attend to any difference between the way your data came out and the way you hypothesized (and wanted) them to come out, because that’s where you learn things you didn’t already believe at a subjective probability of p < .01. And always understand the broader context of your research, and ask yourself whether your findings are likely to apply at different ages, in different cultures, and in different historical epochs.