American Statistical Association
Hypothesis-testing studies are the “gold standard” of clinical research, but the quality of such studies depends crucially on the “building blocks” of hypothesis-generating studies and pilot studies. However, the distinction among the three types of studies is often obscured by both researchers and reviewers of research proposals or submitted papers. Such confusion wastes time, effort and funding, raises ethical concerns related to using human participants in research that is unlikely to provide new and useful information, and corrupts the research literature with false and confusing findings. I try to clarify the role of hypothesis-generating and pilot studies as essential precursors to hypothesis-testing studies, and to describe each type of early-phase research and discuss common problems and misconceptions. For each approach, I propose a set of criteria that reviewers might use in evaluating proposals. Finally, I consider issues related to publication of the results of such studies.
Dr. Kraemer received her bachelor’s degree in Mathematics from Smith College (Summa cum Laude, 1958), did her first year of graduate study in Statistics as a Fulbright scholar at the University of Manchester, England, and then completed her doctoral studies in the Department of Statistics, Stanford University (1963). She joined the Department of Psychiatry and Behavioral Sciences at Stanford in 1964. Her primary interests concern the applications of biostatistics in the behavioral areas of medicine. In 1964, that seemed largely concentrated in Psychiatry, but in the years since, she has worked in Cardiology, Pediatrics, Radiology, Oncology etc., as behavioral issues have become more prominent in all areas of medicine. She is a Fellow of the American Statistical Association, and of the American College of Neuropsychopharmacology. She was elected a member of the Institute of Medicine, Academy of Sciences, in 2003. She was also the recipient of the Harvard Prize in Psychiatric Biostatistics and Epidemiology in 2001. She has published more than 300 papers in peer-reviewed journals, numerous chapters in books, and 3 books. At various times, she has served as associate editor or on the editorial boards of, for example, Statistics in Medicine, Psychological Methods, Archives of General Psychiatry, and Medical Decision Making, and is a frequent reviewer for journals in Psychiatry and other fields of medicine.
Over the years before retirement, she mentored many young investigators both at Stanford, Pittsburgh, and other universities, providing training in research methods, as well as consultation on their proposals.
Her major current research interests concern the use of statistical methods in risk research, specifically the focus on moderators and mediators, the use of effect sizes to indicate clinical or practical significance to replace the overuse and abuse of statistical significance, and, in general, identifying and trying to rectify common problems in the application of statistical methods in medicine. She became Emerita in 2007, but continued to be active, serving on the NIMH council until 2008, and on the DSM 5 Task Force until 2012. Currently she is serving as Professor on a part time basis in the Department of Psychiatry, University of Pittsburgh, as well as consulting with researchers at various universities, but now focusing only on projects dealing with issues directly relevant to her research interests.
On the personal side, she and her husband of almost 50 years have two daughters (one a statistician who coauthored a book with her), and of two granddaughters.
|Date:||Tuesday, September 27, 2011|
|Time:||3:30 - 4:30 P.M.|
New York State Psychiatric Institute
1051 Riverside Drive
6th Floor Multipurpose Room (6602)
New York, New York