See discussions, stats, and author profiles for this publication at: [Link]

net/publication/281107209

Multistage sampling

Article in The BMJ · July 2015

CITATIONS READS

42 75,697

1 author:

Philip M. Sedgwick
St George's, University of London
437 PUBLICATIONS 9,778 CITATIONS

SEE PROFILE

All content following this page was uploaded by Philip M. Sedgwick on 20 August 2015.

The user has requested enhancement of the downloaded file.

Endgames

ENDGAMES

STATISTICAL QUESTION

Multistage sampling
Philip Sedgwick reader in medical statistics and medical education
Institute for Medical and Biomedical Education, St George’s, University of London, London, UK

Researchers investigated the views of the British public on the The aim of the study was to seek the views of the British public
National Cancer Registry’s use of personal medical data for on the use of personal medical data by the National Cancer
public health research and surveillance without individual Registry for the purposes of public health research and
consent. A cross sectional study with a face to face survey was surveillance without individual consent. A cross sectional study
performed by the Office for National Statistics. Participants design, described in a previous question,2 was used. Cross
were selected using multistage sampling of adults in the United sectional studies are observational in design; the investigators
Kingdom during March and April 2015. In each month a sample do not intervene in any way but simply record the health,
of postal districts was selected at random, with the probability behaviour, attitudes, or lifestyle choices of the study participants.
of selection proportionate to size. Within each district, a sample As the name suggests, the purpose of using a cross sectional
of private households was chosen at random. During March study design is to obtain a representative sample by taking a
2762 households were selected with a further 1819 households cross section of the population. The sample in the above study
in April. At the start of the interview, the interviewer determined was obtained using multistage sampling of adults in the UK.
the household composition and selected the respondent from The study was run in two successive months, and a different
among all those aged 16 or more. In households with more than sample was obtained in each month.
one adult, one person was selected at random. If the person Multistage sampling entails two or more stages of random
selected was unavailable or declined to be interviewed it was sampling based on the hierarchical structure of natural clusters
recorded as a non-response. Face to face interviews were carried within the population. Clusters are natural groupings of
out with 1703 (62%) adults in March and 1252 (69%) adults in people—for example, electoral wards, general practices, schools,
April. The data were combined for analysis.1 or households. A different type of cluster is randomly sampled
Of the 2955 respondents, 72% (95% confidence interval 70% at each stage, with the clusters nested within each other at
to 74%) did not consider any of the following to be an invasion successive stages. The final stage of sampling involves choosing
of their privacy by the National Cancer Registry: inclusion of a random sample of people in the clusters selected at the
postcode, inclusion of name and address, and the receipt of a penultimate stage. Multistage sampling methods can be used to
letter inviting them to a research study on the basis of inclusion recruit participants in experimental or observational studies.
in the registry. It was concluded that most of the British public In the above study, the sample was obtained using multistage
considers the confidential use of personal, identifiable patient sampling with three stages (a is true). The first stage involved
information by the National Cancer Registry for the purposes a random sample of postal districts in the UK, with the
of public health research and surveillance not to be an invasion probability of selection proportionate to size. Therefore, postal
of privacy. districts with a greater number of households had a larger
Which of the following statements, if any, are true? probability of selection. In the second stage, a random sample
a) The sampling technique constituted a multistage sampling of households nested within the selected postal districts was
method with three stages obtained. At the final stage, an adult was selected at random
within each household and invited to take part in the study. The
b) Multistage sampling meant that resources could be
postal districts are referred to as the first stage units, the
concentrated in a limited number of areas of the country
households as the second stage units, and the adults living within
c) By definition, multistage sampling constitutes probability the households as the third stage units. It would have been
sampling necessary to construct a sampling frame of postal districts—a
Answers list of all postal districts in the UK—and then sampling frames
listing all private households in each selected postal district.
Statements a, b, and c are all true. The researchers did not indicate how many postal districts were

Correspondence to: P Sedgwick [Link]@[Link]

For personal use only: See rights and reprints [Link] Subscribe: [Link]
 BMJ 2015;351:h4155 doi: 10.1136/bmj.h4155 (Published 31 July 2015) Page 2 of 2

ENDGAMES

selected at the first stage, although the number would have been region in the population. Cluster sampling is sometimes
fixed in advance. combined with stratified sampling, whereby a fixed number of
Multistage sampling was used in preference to simple random clusters are selected at random from a stratum within the
sampling because the population was geographically diverse. population. The strata are often based on, for example,
Simple random sampling would have involved obtaining a geographical regions or age groups. An example is given in a
random sample of a fixed size from a sampling frame—a list previous question.4
of all adults in the UK population. Each adult in the population Two types of sampling methods can be used to recruit
would have had the same probability of selection. However, it participants to a study—random sampling (sometimes called
would have been impractical and too expensive to survey such probability sampling) and non-random sampling (sometimes
a sample because the UK population is geographically diverse. called non-probability sampling). Random sampling involves
Multistage sampling meant that resources and efforts could be some form of random selection of the population members.
concentrated in a limited number of areas of the country, because Simple random sampling (sometimes referred to simply as
the sampling of households was undertaken only in those postal random sampling), described above, is the most straightforward
districts selected at the first stage (b is true). type of probability sampling. By definition, multistage sampling
Multistage sampling and cluster sampling are often confused. constitutes probability sampling (c is true). As described above,
As described above, multistage sampling is based on the multistage sampling entails two or more stages of random
hierarchical structure of natural clusters within the population. sampling based on the hierarchical structure of natural clusters
A different type of cluster is randomly sampled at each stage, within the population. The final stage of sampling involves
with the clusters nested within each other at successive stages. choosing a random sample of people in the clusters selected at
The final stage of sampling is to choose a random sample of the penultimate stage.
people in the clusters selected at the penultimate stage. Cluster
sampling, described in a previous question,3 involves a random Competing interests: None declared.
sample of clusters from the population. All those people in the
selected clusters are then invited to be in the sample. Cluster 1 Barrett G, Cassell JA, Peacock JL, Coleman MP. National survey of British public’s views
on use of identifiable medical data by the National Cancer Registry. BMJ 2006;332:1068.
sampling does not involve successive stages of sampling based 2 Sedgwick P. Cross sectional studies: advantages and disadvantages. BMJ
on the hierarchical structure of natural clusters in the population. 3
2014;348:g2276.
Sedgwick P. Cluster sampling. BMJ 2014;348:g1215.
Cluster sampling can be time consuming, expensive, and 4 Sedgwick P. Stratified cluster sampling. BMJ 2013;347:f7016.
impractical, not least because the clusters will be geographically
diverse. Therefore, cluster sampling sometimes uses a random Cite this as: BMJ 2015;351:h4155
sample of clusters from a conveniently selected geographical © BMJ Publishing Group Ltd 2015

For personal use only: See rights and reprints [Link] Subscribe: [Link]
View publication stats