C4D: Sample

What is it?

In some cases, it might be possible to gather data on an entire population (for example, some data might be available from every participant, or about every project), but in most cases, it will be necessary to take a sample of projects, sites, events, or people. Deciding on sampling strategies is an important part of an R,M&E design. The decision should be strategic and well-considered, informed by the purpose, the nature of the initiative, the nature and requirements of particular methods, and the resources available.

Three broad types of sampling are: random sampling (which uses random or quasi-random methods to select the sample and then uses statistical inference to draw conclusions about the population); purposeful sampling (which selects information-rich cases to study and then use analytical inference to draw conclusions with wider applicability; and convenience sampling (which selects readily accessible cases and is at greatest risk of bias).

General information 

More information on sampling methods is available in the Rainbow Framework. This page is recommended background reading before considering options to apply to C4D. Sampling should be considered alongside issues of response rate and coverage - results are more accurate from a well-chosen sample with a high response rate than from a population with a poor response rate that usually is biased.

Applying the C4D principles

Situations that influence sampling decisions:


Samples should include multiple perspectives, to understand differences in experiences in different settings. Complex interventions might need sampling strategies that can be adapted to suit emerging issues and understandings, such as using ‘purposeful’ sampling (selecting based on what is useful or most interesting) to follow up emerging patterns and findings.


Thoughtful and thorough sampling helps to make the R,M&E design more rigourous. In quantitative (numbers based) methods sampling the sample size and the sample selection are key to making credible claims about the findings. In qualitative (words, stories,visual) methods, sharing details about the sample and selection process increases credibility and trustworthiness.


More generally, sampling should pay attention to equity dimensions, and ensure that the most vulnerable groups are represented and that the data is able to be disaggregated. Additional effort might be needed to get adequate coverage of more remote, more disadvantaged groups due to known biases such as: roadside bias, seasonal bias, pro-literacy bias, etc.

Recommended methods and adaptations for C4D


'C4D: Sample' is referenced in: