Sampling

Now that you have chosen the most appropriate techniques and tools to collect your research data, it is important to know how many people you need to approach to participate in your research. This is called the ‘sample size’. In general, when using quantitative research tools, you need to ensure that you recruit enough people to provide an accurate and reliable estimate of what you are studying. When using qualitative research tools, the aim is to reach enough individuals that you can represent the prevalent opinions, experiences and knowledge in the study population. In this section, we review the sampling designs used in both quantitative and qualitative research tools.

Sampling design in quantitative methods

Quantitative studies require a representative sample of the study population to be able to accurately portray the characteristics of the population and to yield maximum precision of such population parameters. The following criteria are critical when designing a sampling strategy: (1) What are the research objectives? (2) What are accurate estimates of sampling variability? (3) Is it feasible to apply the sampling strategy and obtain the calculated sample size? (4) Is it possible to mininize costs (or to achieve research objectives for minimum cost). As these criteria can conflict with each other, research teams must find a balance between them.

Sample size

A representative sample requires an adequate sample size, taking into account statistical power parameters. Power is the probability of rejecting the ‘null’ when an alternative hypothesis is true. In simple terms, this is the probability of actually detecting an effect under study. Different sample size calculations should be used for the various study design types. Sample size calculation formula and calculation procedures can be found in standard biostatistics reference materials.17 Further discussion with a statistician will also help to confirm and calculate the appropriate sample size needed for various types of research methods.

Sampling strategy

As quantitative studies require a representative sample with regard to population characteristics, a ‘probability’ sampling is preferable. This enables every individual in the population to have a certain chance of being included in the sample. Probability sampling also allows estimates of sampling error to be calculated. There are several probability sampling strategies (Table 12).

In some situations, random sampling is not the preferred option due to lack of specific resources (e.g. a list of the entire population), time, costs or ethical constraints. In other situations, the research requires some ‘weighting’ to the information being collected (e.g. a survey among experts). In this scenario, nonprobability sampling is preferable. There are several commonly used non-probability sampling strategies (Table 13).

Sampling in qualitative methods

Sampling strategy

Sampling in qualitative research uses quite different approaches to those used in quantitative studies. The aim in qualitative research is not to have a representative sample, but rather one that reflects the characteristics and richness of the context and/or study population. Whatever sampling method is used, the IR team will need to justify their sampling frame selection. Table 14 reviews the different kinds of sampling techniques used in qualitative research.

TDR Implementation research toolkit(Second edition)

© 2024 TDR. All rights reserved

References