Paper
Friday, 21 July 2006
This presentation is part of : Research Theory and Methods
Sampling Strategically: Bridging Disconnected Worlds
Elsabeth Jensen, RN, PhD, School of Nursing, York University/Lawson Health Research Institute, Toronto, ON, Canada, Cheryl Forchuk, RN, PhD, School of Nursing, University of Western Ontario/Lawson Health Research Institute, London, ON, Canada, Rick Csiernik, MSW, PhD, RSW, School of Social Work, Kings College, London, ON, Canada, and Heather Atyeo, Health Systems and Health Services Program, Lawson Health Research Institute, London, ON, Canada.
Learning Objective #1: 1. understand the limitations of traditional sampling methods in conducting research in the community
Learning Objective #2: 2. understand the challenges faced by researchers studying health care issues in non-clinical, community populations

Sampling is one of the necessary steps in the conduct of research. Entire books have been dedicated to it (Barnett, 1991; Cochran, 1977; Conway, 1967; Henry, 1990; Lwanga, 1991; Sudman, 1976), and every student of research is familiar with the topic.  Researchers take care to define their population, and then to draw samples that will be most representative of the population. Such care is required in order to secure reliable data. Unfortunately, these sampling requirements can systematically exclude invisible groups such as the homeless and people with a past history of mental illness from study, introducing a different set of biases. Finding ways to include these populations in research is a challenge. This was the experience of a team of researchers studying mental health and housing in a community-based population. We required a sample of 300 stratified by housing type: homeless, group living, supported housing, independent living. For the different subgroups different strategies were used. For example, traditional methods of randomization using lists of residents were used for group and supported homes, but for shelters, random days were chosen.  As health care research moves from clinical to community settings, the approaches to the process of sampling demand increasing creativity and flexibility in order to balance ideal with possible. This paper describes the plan, the problems and the solutions, developed by one group of researchers, in the hope that others can benefit from the lessons learned through this experience.

See more of Research Theory and Methods
See more of The 17th International Nursing Research Congress Focusing on Evidence-Based Practice (19-22 July 2006)