One of the most important aspects of research is sample size. A common belief when it comes to avoiding statistical issues interfering with a successful research project, is to try and reach the largest representative sample as possible. Though, there are instances when a small sample size is acceptable and will not result in inaccurate data. Prior to choosing a small sample size, it is vital to understand a few key terms that can help achieve your research goals.
To determine a sample size, it is important to first understand the difference between sample and population. In order to create a sample, you need a population.
Population is defined as a number of people living in a particular town, area, or county. The units of a population can be made up by a number of different things. For instance, there can be populations of households, people, events, institutions, and so forth. Anything that is able to be counted can be considered a population unit.
For quantitative or qualitative studies, various targeting (age, geographical, and otherwise) can be placed on a population. Some examples that can be covered by surveys include;
- All participants must live in the United States
- All participants must be 25-40 years old
- All participants must be smartphone owners
If the universe that is surveyed is a high fraction of the desired population, the acquired results should also be true for the larger population. To illuminate, if 95% of homes have a television, the 5% without a television would have to be drastically different, for the survey data not to hold true for the entire population. Populations be simple uncover, but a fundamental point to remember is if you are unable to collect information from a population, then it is not a unit of population that is suitable for market research.
A sample is a portion of the population from which it was drawn. Survey techniques are based on sampling, which involves obtaining information from only a slice of the population. Samples can be drawn in several different ways, such as probability samples, non-probability samples, and quota samples.
Sampling is critical in research to be able to generate accurate data. It is unrealistic and disadvantageous to study an entire population. Yet, if a sample size is too small or excessively large, it can lead to inaccurate results. So, the question remains, when is it appropriate to use a small sample size?
Surveying Small Sample Sizes
There are several scenarios in which small sample sizes can provide meaningful data.
Surveying current clients to gauge customer satisfaction is an essential part of research. Obtaining information about how your clients presently view your products, services, or customer service is extremely valuable. Granted larger sample sizes are useful, but if your clients are unhappy, the data is still important. Depending on the scope of your business, it may be pertinent to learn about any frustrations expressed by your clients so that necessary changes can be implemented.
Similarly, when you lose a client, their responses are significant no matter how large the sample. Every client or customer that does not return has important information to share. Regardless if only a single client was lost over the span of a year, why you lost the client is noteworthy. Absorbing knowledge from these types of clients may provide you with data that assists with progressing your business further.
It is not unusual that the idea of surveying small sample sizes makes people nervous. Combined with the fear that the results will be misleading, it is difficult to believe that small samples are representative. However, we exist in a new age of data consumption and it is possible to value small sample sizes in research. Depending on your research goals, a small sample size may be the exact solution you need for a directional study or client feedback.