Sampling
I want to study a subset of the population in order to understand the population a whole...
What is it?
A sample is a subset of the population or community that you choose to study in order to understand the population or community as a whole.
Sampling is necessary because gathering data is expensive and time consuming. It is very difficult, perhaps impossible, to speak to every person involved in your project. Thus, appropriate sampling methods help teams collect the right amount of data, from the right respondents to meet their information needs.
There are generally two kinds of sampling:
- Random sampling
- Purposeful sampling
Random sampling is a way of selecting respondents from a list of the entire population of interest (usually your project stakeholders, including participants) so that each respondent has an equal chance of being selected.Random samples give you a broad understanding of the situation. They are designed when you need to state that what is true for your sample is likely true for the entire population, using a process called generalization.
Generalization involves assuming that data gathered from a sample accurately represent the general population from which the sample was drawn.
Purposeful sampling is used when there are particular perspectives you want to capture in your data. It is the intentional selection of respondents based on their experiences or characteristics.
How do I use it?
Random sampling is appropriate when you are collecting quantitative data that you want to analyze statistically. Random samples are created using mathematical calculations to identify how many people will participate in your data gathering efforts. These calculations are developed based on how strong you need your analysis results to be and how varied the population is.
Purposeful sampling is appropriate when you are collecting qualitative data in order to gain a deeper understanding of the experiences of particular groups of people.
An important consideration in creating both random and purposeful samples is the sample size.
The size of a random sample is determined by MEAL experts using precise calculations. These calculations consider the strength of the analysis required to make conclusions about project effects and the diversity of the population. Stronger analysis needs that include multiple sub-groups require larger samples. Similarly, more diverse populations will require a larger samples to ensure that all perspectives are captured.
The size of a purposeful sample is calculated differently. In purposeful samples, you are choosing to speak with certain groups of people because you find their perspectives particularly relevant to your information needs. Purposeful samples are generally smaller than random samples because of the nature of the qualitative method used to collect the data. Focus groups, for example, are best conducted with no more than 8-10 people. Generally, qualitative data collection continues until no more new themes or ideas are collected.
As with random samples, the diversity of the population and the number of different sub-groups with whom you need to speak with will determine the number of focus groups or interviews you need to hold.
Tips:
Tip 1: Work within your resources
Remember that collecting data is expensive. Stakeholders, project team members, and MEAL experts should collaborate to help you design sampling methods that balance your information needs and your resource constraints. As a general rule of thumb, the larger your sample, the more resources you will need to enable you to collect data.
Tip 2: Avoid sampling bias
Sampling bias occurs when some members of the population are more or less likely to be selected for participation in your data gathering efforts than others.
When your sample is biased, you are not taking into consideration all the available perspectives, ideas and opinions. This means that your data will not be as accurate and cannot be easily generalized to the population you want to address. Work with experts to design your sampling methods and sample size so that you avoid bias as much as possible.
Supported by:
Shared by:
Users are free to copy/redistribute and adapt/transform
for non-commercial purposes.
© 2022 All rights reserved.