Sampling

Samples are a defined subset of the population chosen to represent the population under study. It is often impractical to conduct a survey, interview, or focus group with an entire population due to time and cost constraints. Sampling a subgroup of the population can overcome these limitations and make it possible to draw conclusions that are generalizable to the overall population.  

Sample selection

The ability to generalize from a sample to the population depends on the representativeness of the sample.   A sample is representative to the extent that it exhibits the same distribution of characteristics as the population from which it was selected.    For example, is your population all students enrolled in your department's courses or only those students who are majors? You may also wish to use sampling techniques to ensure that certain subgroups of your population are included.

Types of sampling

With proper sampling techniques, a researcher can use the information obtained from the population to infer something about the characteristics of the population.

Probability sampling

Each member of the population has a specifiable probability of being chosen. Probability sampling is critical when you want to make precise statements about a specific population on the basis of the results of your survey.

• Random sampling: Every member of the population has an equal probability of being selected for the sample.

  Example

  If the population has 100 members, each has one chance out of one hundred of being selected.

• Stratified random sampling: The population is divided into subgroups (strata), based on information about the population, and a sample is drawn from each subgroup. Stratified random sampling has the advantage of a built-in assurance that the sample will accurately reflect the numerical composition of the various sub-groups.

  Example

  Students in one university can be separated into 4 strata: freshmen, sophomores, juniors, and seniors. From these, a sample of freshmen will be chosen, a sample of sophomores, and so on.

• Cluster sampling: Rather than randomly sampling from a list of individuals, researchers can sample from clusters they have identified. After the clusters are chosen, all individuals in each cluster are included in the sample.

  Example

  You might identify all English classes as the clusters, then randomly sample from this list of classes and have all members of the chosen classes complete the survey.

Nonprobability sampling

With nonprobability sampling, there is no way to estimate the probability that each member of the population will be included in the sample and there is no guarantee that each member of the population will have the same chance of being included.

• Convenience sampling: These samples are chosen based on criteria such as accessibility or specialized knowledge, which is likely to introduce biases into the sample so that it may not be representative of the population.

  Example

  Students in a class who volunteer to fill out a survey or professors in a department who volunteer to fill out a survey.

• Quota sampling: A sample is chosen based on a quota of the individuals that are to comprise the sample. This sampling technique can involve setting quotas based on race, age, and school classification. Once the quota has been set, the sample may be chosen using random sampling techniques; however, external factors, such as a students’ willingness to participate, could affect who ultimately comprises the final sample.

  Example

  A professor wants to sample 25 students from a course using quota sampling based on the population of each type of student in the class: For example, 8 freshmen (44%), 8 sophomores (32%), 3 juniors (12%), and 3 seniors (12%).

Sample size

However, even if you carefully apply sampling strategies, some error may result from chance variation. This error can be reduced by increasing the sample size. When performing quantitative analysis, you must to choose a sample size that is appropriate for the statistical test you plan to use. Larger samples are more likely to yield results that are more reliable, precise, and representative of the population. Sample size may also depend on the resources available (time, money, staff) and how difficult it is to locate respondents. The table below provides a general guide for determining sample size.  Because your response rate is likely to be less than 100%, be sure to administer your survey to a group larger than your desired sample size to. [more]

Recommended sample size

Population size	Recommended sample size
50	44
75	63
100	80
150	108
200	132
300	168
400	196
500	217
1000	277
3000	340
5000	356
10000	369

Additional information

Kalton, G. (1983). Introduction to Survey Sampling. Newbury Park,CA: Sage Publications, Inc.