Statistical and Methodological Consultant
Dale N. Glaser, PhD
Pacific Science & Engineering Group
From my first course in psychology as an undergraduate at the University of California, Irvine, I became absolutely enamored with this discipline. Even though I was relatively apprehensive about taking the required one-year statistics course, the professor (Robert Newcomb, who is still very active with the Southern California Chapter of the American Statistical Association) made the subject matter come alive. At that time, I learned the importance of interweaving concept with algorithms: it was not enough to just memorize, by rote, the various equations without a conceptual and applied thread. Subsequently, I went on to a masters program in counseling, with my primary vocational aim to be a practitioner. I gained experience as a vocational rehabilitation counselor, then proceeded to management. As evident by my career trajectory, activities in research and analysis (except instances of test evaluation and interpretation) were generally subordinated to counseling/personnel responsibilities.
With the aim of obtaining a PhD in psychology always a consideration, I matriculated to a doctoral program in industrial/organizational psychology. Again, my primary aim was to be a practitioner in this domain, though I found, through an internship at the Navy Personnel Research and Development Center (NPRDC), that I not only had an aptitude for statistical analysis (by way of SPSS-X on the mainframe), but that I also enjoyed it. Though I gained employment as an organizational development consultant, research and analysis became my true passion. As circumstances would have it, I was asked to teach a multivariate course in 1994 at United States International University, and from then on I knew exactly what I wanted to do: teach and conduct applied research. I worked in consumer research and clinical research departments at a large health care institution, entailing analysis of patient satisfaction surveys as well as guiding the development of experimental and non-experimental studies. Within this large health care network, my expertise was being actively solicited, to the extent that I was (not by design) serving as an internal consultant on many of the studies being conducted across departments. Concurrently, given the demand for adjunct faculty, I began instructing both undergraduate and graduate statistics and research design/psychometrics at San Diego State University, University of San Diego, and Alliant International University/California School of Professional Psychology.
With my ties to both the applied sector and academic arena, I was making contacts with a wide array of sources, resulting in various consulting opportunities. Though I wasn’t seeking them out, I was referred to consult on a diverse array of projects. What I found most appealing about statistical/methodological consulting was that I wasn’t confined to one topical area. I was consulting for health care institutions, marketing research operations, educational programs, nonprofit organizations, etc. Consulting afforded me the latitude to engage in a variety of projects across varying disciplines. However, with this came the reality that certain disciplines have an orientation in analysis and methodology that may be unique to their domain. It became incumbent that when consulting for a lung/heart/kidney transplant center, I become sufficiently adept at survival analysis. When engaged in a project testing the fit of a longitudinal model with many data points across time, I needed to have more than a passing familiarity with time series analysis. Though a central interest of mine since my internship at NPRDC, I needed to garner more sophisticated skills in structural equation modeling (SEM). Even though my graduate education served as a template for my statistical training, it would not be an overestimate to claim that about 80% of my current knowledge was obtained via self-instruction.
Based on my skills in multivariate and modeling techniques, I was offered a senior statistician position at Pacific Science & Engineering Group, a firm that specializes in applied, experimental, and engineering psychology. Through this company, I have been able to serve as a statistician/methodologist on both internal and external projects. The projects have ranged from fitting various complex quality of life models via SEM to a study examining the group decision-making process in a distributed environment to consulting on multiple health care/nursing research studies for both academic and health care institutions. Since then, I have also been asked to serve as an ad hoc reviewer for various journals (e.g., Psychological Methods) and on editorial boards (e.g., Structural Equation Modeling). Despite my orientation as an applied researcher, my consulting has also culminated in publications in such diverse journals as Human Performance, Heart and Lung, and American Journal of Critical Care.
Even though many of us are not trained as mathematical statisticians, it is my contention that our grounding in methodology, design, and interpersonal communication confers upon those with a quantitative orientation in the psychological sciences to serve aptly as methodologists/statisticians. Admittedly, many of the articles in the statistical literature (e.g., Journal of American Statistical Association, Biometrics, Statistics in Medicine, etc.) may appear arcane; however, the bulk of my consulting has not involved the complexities of mathematical modeling that might be beyond my scope. Rather, much of my work has involved domains that, as a psychological scientist, I am well equipped to consult on (e.g., psychometrics). However, there are occasions when a project dictates an analytical strategy that may be entirely novel. An example of such an undertaking was when I needed to learn the intricacies of spectral analysis for an electromyogram (EMG)/workload study. Though those in health psychology may be versed in this technique, it was a new one for me!
Some global recommendations for those interested in statistics: (1) Become involved in the American Statistical Association (ASA), including your local chapter. (2) Remember, you may be competing against those specifically trained in the discipline you are interested in exploring (e.g., biostatisticians, econometricians, etc.), so it will serve you well to absorb the language unique to that particular environment. (3) A random perusal of the applied statistician positions in AMSTAT News (monthly publication of the ASA) finds that virtually all the large manufacturing/pharmaceutical concerns require SAS programming experience. Thus, the more software you can learn (besides SPSS), the more marketable you will become. I have been absolutely thrilled with my career trajectory. For me, the merging of applied research with academia has fulfilled a life-long ambition!
(Originally published in the November/December 2001 issue of Psychological Science Agenda, the newsletter of the APA Science Directorate.)