Letters to the Editor
To the Editor:
We thank Colliver et al for a number of meritorious statements in summarizing the results of 11 studies examining decline in empathy during medical school and residency.1 Namely, the studies do call for further critical examination of the issue, and as a matter of course, all researchers should consider threats to the validity of their findings. However, there are a number of methodologic limitations in Colliver and colleagues' study that call their conclusions into question. First, these authors seem to believe that response bias accounts for the findings of declines in empathy found in all of the studies reviewed. In our own study demonstrating declines in empathy in dental school, we noted that our overall response rate was 61%.2 This was better for the first year of training than for subsequent years but was still considered a response rate with adequate generalizability by sources cited in our text. The response rate was even better in all of the other reviewed studies where the information could be garnered.
Second, these authors treat response bias as the only consideration for quasi-experimental research validity and ignore external validity. Paramount to external validity is replication of findings across time and different populations. We have replicated our findings of decline in empathy during dental school in Taiwan,3 declines in empathy have been shown at other dental schools using other measures,4 and these declines have been documented in a number of different medical schools described in their own study. Similarly, good instrumentation is integral to internal validity. The factor structure of the Jefferson Scale of Physician Empathy has been replicated in our own and others' research in the United States2,5 and abroad.6,7 Combined, there is an enormous amount of evidence for internal and external validity for this research.
Finally, calculation of Cohen's effect size (d) for our study resulted in a d = 0.80 between first- and second-year classes—by no interpretation a “weak,” “small,” or “trivial” change. The mathematical transformation that Colliver et al chose serves only to dilute findings in lieu of generally accepted effect-size calculations. These authors themselves have cautioned against reviews that obscure biases and confound by averaging. Ironically, they did so in their report.
They chose to end their discussion by paraphrasing Mark Twain. But as we read their report, we were reminded of another phrase popularized by Twain about inappropriate use of statistics simply to prove a point.
Jeffrey J. Sherman, PhD
Clinical associate professor, Department of Rehabilitation Medicine, University of Washington Medical Center, and Department of Oral Medicine, University of Washington School of Dentistry, Seattle, WA; firstname.lastname@example.org.
Adam P. Cramer, DDS
President, Dr. Adam P. Cramer, P.C.