Sampling in epidemiological research: issues, hazards and pitfalls
Authors:
Open Access
- 1 April 2016
- journal article
- editorial
- Published by Royal College of Psychiatrists in Bjpsych Bulletin
- Vol. 40 (2) , 57-60
Abstract
Summary: Surveys of people's opinions are fraught with difficulties. It is easier to obtain information from those who respond to text messages or to emails than to attempt to obtain a representative sample. Samples of the population that are selected non-randomly in this way are termed convenience samples as they are easy to recruit. This introduces a sampling bias. Such non-probability samples have merit in many situations, but an epidemiological enquiry is of little value unless a random sample is obtained. If a sufficient number of those selected actually complete a survey, the results are likely to be representative of the population. This editorial describes probability and non-probability sampling methods and illustrates the difficulties and suggested solutions in performing accurate epidemiological research.Keywords
This publication has 5 references indexed in Scilit:
- The case of the lady who risked exploding: A study of multiple consequences and contested valuesHealth, Risk & Society, 2012
- Severe mental illness in 33 588 prisoners worldwide: systematic review and meta-regression analysisThe British Journal of Psychiatry, 2012
- Large scale questionnaire survey on respiratory health in Sweden: Effects of late- and non-responseRespiratory Medicine, 2009
- Using Survey Participants to Estimate the Impact of NonparticipationPublic Opinion Quarterly, 1995
- The Psychiatric Status ScheduleArchives of General Psychiatry, 1970