Home / Data Collection / Sample Source & Quality / Sample Types

Probability And Non-Probability Sampling Types, Four Sampling Methods Each

Probability Market Research Sampling Methods:

1) Simple Random Sampling - a random selection procedure to ensure that each respondent has the same chance of being selected into the sample (e.g. random digit dialing, plus-one dialing, random number selection of sample frame database, etc.).

2) Systematic Sampling - a random starting point in the sample frame is chosen, and then a constant skip interval is used to select each respondent for the sample. This is more efficient than simple random sampling and the formula often used to compute the skip interval is the population size/sample size.

3) Cluster Sampling – population is divided into clusters, or subgroups, which are very similar to each other and then one of two sampling methods are used: perform a census on one or a few of the clusters of the entire population, or randomly select more clusters and take a sample of them. Area sampling (clusters are a form of geographic location) is an often-used form of cluster sampling. There is a danger of error in cluster sampling if the clusters are not actually homogenous.

4) Stratified Sampling – identify strata, or subpopulations (e.g. income, gender, etc.), and perform simple random sampling on each strata, and then weights are applied to estimate the population’s findings. Stratified sampling is best used if a population is not a normal population or has a skewed distribution.

Non-Probability Market Research Sampling Methods:

1) Convenience Sampling – use a high-traffic area to recruit participants (e.g. mall). The error that occurs is those of the population that do not visit the area chosen have no chance of being included in the sample.

2) Judgment Sampling – use of judgment or educated guess by someone who is knowledgeable about the population of who should be in the sample. This is highly subjective, so there is likely to be error. This method is often used for recruiting focus groups.

3) Referral Sampling – ask respondents to identify people like themselves to participate in the survey; most often used when there is very small sample frame; sometimes called “snowball sampling.” The error that occurs is those of the population who aren’t well known, disliked, or have different opinions as the person being asked are not included in the sample.

4) Quota Sampling – identify quota characteristics, such as demographics, to use as the sample, and the quotas are defined by the research objectives

Contact us to learn more about sampling types, Polaris' sample sources, and sample quality.