Beruflich Dokumente
Kultur Dokumente
Sampling design
What is it?
– Sampling is a statistical practice used to select individual
observations from the target population.
Issues to consider
– Size of study population (Census for small populations)
– Previous studies (literature review)
– Published sample tables (gives recommended SS for different
population sizes
– Calculated sample size
– Costs implications of research
STUDY POPULATION
SAMPLE
TARGET POPULATION
6
Sampling designs
This must be consistent with the research design.
– Systematic sampling
– Stratified sampling
• Proportionate
• Disproportionate
– Cluster sampling
Guiding principles:
– Population is divided into mutually
exclusive categories called strata.
Patients
Male Female
Heterogeneous
Homogeneous Homogeneous
Stratified Random Sampling
Place of residence
Clients
Urban Rural
Strata
Heterogeneous
Homogeneous Homogeneous
Systematic sampling
Guiding principle:
– This is a method which relies on arranging the target population in a
sequential order and selecting the elements at regular intervals.
– The nth element falls on the interval which is obtained by diving the
total number of elements by the sample size.
– Steps:
• Population is divided by number of cases required for the sample,
yielding the sampling interval (k=N/n).
• A number less than or equal to the sampling interval is selected
randomly to identify the first case to be sampled
• Every nth case is selected, where n is the sampling interval
Cluster sampling
This is a sampling design in which the target population is
divided into mutually exclusive clusters or areas.
- Purposive sampling
- Convenience sampling
- Snowball sampling
Purposive sampling technique
This is a sample selected at the discretion of the researcher based
on the interest of the study.
- Saturation: confidence that you are learning little that is new from
subsequent interviews.
Convenience
This is a sample where inclusion is based on the availability of
the participants.
Saves time, money, and effort; but could result in poor quality
data; lowest credibility.
Snowball
This is used when a sample requires participants with
specific characteristics who are difficult to locate or
identify.