| |
 |
| Mike Saks and Judith Allsop |
Chapter 9 - Health Research Sampling Methods
Peter Davis and Alastair Scott |
| |
|
| |
| Contributor
biography |
| Peter Davis is Professor of Sociology,
and Director of the Social Statistics Research Group,
at the University of Auckland, New Zealand. He has masters
degrees in sociology and in statistics from the London
School of Economics and a PhD in community health from
the University of Auckland. His main interests are in
research methods, social structures, and policy, particularly
health policy and health services. He has collaborated
with colleagues in health research and in social statistics
on a number of major surveys since the 1970s. He is Senior
Editor (Health Policy) on the international journal, Social
Science and Medicine. |
| Alastair Scott is an Emeritus Professor
of Statistics at the University of Auckland, New Zealand.
His main research interests are in sample survey theory
and methods, particularly in the development of methods
for the analysis of survey data in medical statistics,
especially in the design and analysis of retrospective
studies, and in health services research. He has collaborated
with Peter Davis on a number of large health studies,
including the New Zealand Quality of Healthcare Survey
on adverse events in public hospitals and the National
Primary Medical Survey on the work of general practitioners. |
| |
| Chapter overview |
| This chapter outlines the basic features
of sampling in health research. The different forms of
non-probability sampling and the techniques employed in
probability sampling are discussed together with issues
that arise in sampling and ways of reducing error. |
| |
| Chapter links |
Chapter
10 - Quantitative Survey Methods in Health Research
Chapter 11 - Statistical Methods
for Health Data Analysis |
| |
| Suggested
Online Readings |
| Field,
L., Pruchno, R.A., Bewley, J., Lemay, Jr., E.P. and Levinsky,
N.G. (2006) ‘Using Probability vs. Non-probability
Sampling to Identify Hard-to-Access Participants for Health-Related
Research: Costs and Contrasts’, Journal of Aging
and Health, 18 (4): 565-83 |
| This article compares the recruitment
costs and participant characteristics associated with
the use of probability and non-probability sampling strategies
in a longitudinal study of older haemodialysis patients
and their spouses. Probability-based sampling was found
to be more time-efficient and cost-effective than non-probability
sampling. There were no significant differences between
the respondents identified through probability and non-probability
sampling in terms of age, gender, years married, education,
work status, and professional job status. It is argued
that researchers should consider representativeness and
external validity when designing sampling and recruitment
plans for health-related research. |
| |
| Fredman,
L., Tennstedt, S., Smyth, K.A., Kasper, J.D., Miller,
B.A., Fritsch, T., Watson, M. and Harris, E.L. (2004)
‘Pragmatic and Internal Validity Issues in Sampling
in Caregiver Studies: A Comparison of Population-Based,
Registry-Based, and Ancillary Studies’, Journal
of Aging and Health, 16 (2): 175-203. |
| This paper looks at issues related
to sampling in the case of caregivers. It argues that
studies of caregivers illustrate a classic sampling dilemma:
maximizing recruitment without compromising study validity.
Because caregivers are defined in relation to a care recipient,
sampling methods are often determined by pragmatic decisions
such as access, efficiency, and costs. However, overlooking
validity may result in selection bias, misclassification
of caregiver status, and the confounding of results. In
a review of a number of studies, the authors found that
all used task-based inclusion criteria. Caregiver participation
rates ranged from 81% to 96%. The paper argues that standard
task-based inclusion criteria to define caregivers may
enhance validity. |
| |
| O’Connell,
A.A. (2000) ‘Sampling for Evaluation: Issues and
Strategies for Community-Based HIV Prevention Programs’,
Evaluation and the Health Professions, 23 (2):
212-34. |
| The paper argues that sampling methods
are important in the evaluation of community-based HIV
prevention initiatives because responsible sampling procedures
provide a valid model of the population and reliable estimates
of behaviour change can be determined. It provides an
overview of sampling with particular focus on the needs
of community-based organizations. Several probability
and non-probability sampling designs are reviewed and
issues of bias, cost and feasibility factors in design
selection discussed. Six sampling guidelines for programme
evaluations are provided. |
| |
| Further Reading |
| Aday, L.A. and Cornelius, L.J. (2006)
Designing and Conducting Health Surveys: A Comprehensive
Guide. 3rd edition. New York: Jossey-Bass. |
| This is a standard reference book
written with the non-technical user in mind, drawing substantially
on recent methodological research on survey design and
cognitive research on question and questionnaire design
and presenting a total survey error framework. |
| |
| Fowler, F. J. (2002) Survey Research
Methods. 3rd edition. Thousand Oaks: Sage. |
| This book provides a concise overview
of the entire survey research process, using clear and
easy to understand language. |
| |
| Korn, E. L. and Graubard, B. I. (1999)
Analysis of Health Surveys. New York: Wiley. |
| This is a more advanced book dealing
with the technical aspects of the analysis of data from
complex surveys, illustrated with many examples from real
health surveys. |
| |
|
|
| |
| |
| |
© Sage Publications Ltd. |
|
