March 2003
Volume 16, Number 3
'Quantitative Methods Travel Well' from Educational Psychology to Nursing
 In 1970, Steve Owen earned a PhD in educational research from Purdue University and joined the educational psychology department at the University of Connecticut, where he taught courses in social cognitive theory, human motivation, learning, educational psychology, and multivariate analysis. In 1993, he was inducted as a lifetime honorary member of the nursing honor society, Sigma Theta Tau International. In 1999, Owen began a new career as a biostatistician in the University of Texas Medical Branch's School of Nursing. |
During almost 30 years in educational psychology at the University of Connecticut, I gradually layered skill in quantitative methods on top of knowledge of general psychology content, such as learning theory and motivation). The slow shift, fueled more by interest than by careful career planning, was almost entirely unintentional. But it proved to be good fortune. I retired from the University of Connecticut and took a position as a biostatistician in the School of Nursing at the University of Texas Medical Branch. I imagined it would be a dramatic career shift, but quantitative methods travel well.
Still, there are some interesting differences.
- Discriminant function analysis has fallen from favor in the health professions. In its place, logistic regression dominates, even though odds ratios are harder to interpret than many imagine.
- In the social sciences, there is more attention to the harm of chopping continuous variables (say, with median splits) in order to force data into an ANOVA structure. In the health professions, there is little hesitation about converting age, income, or depression scores into high versus low. A variation occurs in health surveys, where questions are often constructed to force a dichotomous response (e.g., Do you get short of breath when bathing or dressing? Yes or No).
- More than liberal arts schools, health professions universities are financed by grant support. Although applications to federal agencies typically require a power analysis, in which effect sizes play a crucial role, health researchers generally avoid reporting effect sizes when they publish results. So many published studies, focusing solely on statistical significance to draw conclusions, exaggerate (with large N) or overlook (with small N) findings.
- At the same time, health researchers are far more likely to include confidence intervals in reporting research results.
- Theory does not drive most research in the health professions; it tends to be in the back seat. The emphasis is more on practical applications: Does drug X work better than drug Y? Does surgery method A decrease length of hospitalization?
- Large epidemiological databases invite mindless data mining. Researchers often report their discovered associations, sometimes urging causal interpretations, and journalists extend the causal exaggerations.
Liberal arts campuses, of course, have concentrations of psychologists, and quite a few associated faculty who have a nodding acquaintance with psychological science. A health science school will typically have a psychiatry department, but those faculty concentrate on their practice and are usually not that well rounded in basic psychology. But this is not an academic desert where a psychologist feels isolated and lost. In such a place, there is great opportunity to spread some messages about what psychological science is all about. Nurses, allied health faculty, and medical faculty sometimes harbor quaint stereotypes about who academic psychologists are and what they do (therapy, right?). Media carry interesting stories about discoveries in psychology, but many readers seem unaware that there is psychological science behind the articles.
Faculty here are more tolerant than I had expected of anon-medical associate in their midst. They're very patient with my stumbling through the vocabulary of a medical school. But at the same time, they are often fascinated to hear about, say, principles of social cognitive theory. It is not just that they are amused by such theory, but because they come to understand that it can inform and broaden their research efforts. For example, self-efficacy measurement has now caught the attention of many of my colleagues, and they have begun to plug it into their own research and grant applications. And, because self-efficacy measures are tailored to the particular behavior under study, they are discovering that instrument development is not some high-powered talent that only a precious few can accomplish.
My former life has not faded entirely to memories. I still have-and use-shelves of books on cognition, self-regulation, motivation. Even E.G. Boring's History of Experimental Psychology still gets its pages turned from time to time. I can't resist buzzing around Current Directions in Psychological Science, even when grant deadlines are bearing down on me. And thanks largely to the electronic age, I continue to write and collaborate on research with earlier partners.
So this has been a fascinating foray into the health professions. I love my colleagues' impression that I am offering valuable statistical assistance. What I haven't revealed, until now, is the symbiosis: they are eager to absorb ideas from a distant field, and I am learning more from them than I ever expected or hoped.
