SAMPLING
==Sampling== is the process of selecting the sample or a portion of the population. A sample is a subset of the population elements. An important characteristic of a sample that must be considered is Representativeness. A representative sample is one whose key characteristics closely approximate those of the population.
There are no fixed rules with regard to the number of samples. But there are recommendations and this is because of the possibility of sampling error. Sampling error occurs if the selection of the sample does not take place in the way that it was planned. This will result in the overrepresentation or
underrepresentation of some segments of the population. How big the sampling error is depends on the size of the sample. The smaller the sample size, the bigger the chance of sampling errors. The appropriate sample size also depends on the heterogeneity or homogeneity of the group. A heterogeneous group requires a bigger size and a homogenous group, a smaller one.
| PROBABILITY SAMPLING | NON-PROBABILITY SAMPLING |
|---|---|
| There is a random selection of sample | There is a form of bias in the selection of the sample |
| Each element in the population has the same equal chance of being selected as a sample | There is no assurance that each element in the population has the same equal chance of being selected as a sample. |
| There is greater representation in each unit in the population. | There is no assurance that each unit in the population is properly represented. |
| The findings can be generalized to the findings | The findings are limited to the sample |
Types of Non-Probability Sampling
- Convenience Sampling - this is a selection of samples based on the convenience of the researcher. With this, the researcher involves the most conveniently available people to participate in the study. It is sometimes called Accidental Sampling. For example, stopping people in the street to conduct an interview or to administer a survey questionnaire
- Snowball Sampling - works the same way as the referral system. With this sampling technique, initial sample members are asked to refer other people who meet the criteria required by the researcher. This is based on the assumption that people who share the same traits or experiences know each other. This is also particularly useful for participants who are hard to find (e.g. women who earn at least 3 million per year). Just like a snowball which gets larger as you roll it over the snow, snowball sampling starts with a few participants and continues to get larger until the desired sample size is met.
- Purposive Sampling – in this technique, the selection of the sample is based on the selective judgment of the researcher. That is why it is also called Judgmental Sampling. With this, the researcher sets a set of criteria that is relevant to the topic under study. People possessing the set criteria are invited to participate in the study.
- Quota Sampling - is a sampling technique where the researcher identifies population sections or strata and decides how many participants are required from each section. This technique allows a better representation each of unit in the population. Usually, the stratification is based on variables that are relevant to the study. For example, is stratification based on gender, age, and educational attainment.
Types of Probability Sampling
- Simple Random Sampling - This is the most basic probability sampling technique. Succeeding probability sampling techniques which are more complex usually incorporate features of simple random sampling. In this technique, a sample selection is purely based on chance. Each member of the population has the same equal chance of being selected as a sample. According to Burns in 2012, simple random sampling happens through any of these two methods:
- Have a list of the members of the population, write each name on a card, and choose cards through a pure-chance selection. One may use the fishbowl technique.
- Have a list of all members of the population, known as the sampling frame (Polit and Beck, 2003); give a number to members and then use randomized or unordered numbers in electing names from the list by
- Systematic Random Sampling - this method uses the kth interval formula. The researcher sets the desired sample (n). The size of the population is then known or estimated (N). By dividing N by n, the sampling interval width (k) is established. The sampling interval is the standard distance between elements chosen for the sample.
For example, the researcher sets the sample size to 100 from a population of 20,000 senior high school students listed in the student directory. In computing the sampling interval, the computation would look like this: k=20,000/100 k= 200
- Stratified Random Sampling - the population is divided into subgroups or strata. Just like in quota sampling, stratification is based on variables that are relevant to the study.
For example, stratification based on gender, age, educational attainment, etc. After the stratification, an appropriate number of elements are randomly selected from each stratum. Example: Senior High School students are divided into subgroups based on their academic tracks and samples were randomly obtained from each track.
- Cluster Sampling - is particularly useful when the population is large and widely dispersed. In cluster sampling, the sampling of units is done in several stages. This is the reason why it is also called multi-stage sampling. The resulting design is described in terms of the number of stages (e.g. three-stage cluster sampling.
For example, In drawing a sample of Business Students, the researcher will first draw a random sample of business schools in the National Capital Region and then obtain a sample of students from the selected schools. The final selection from within a cluster may also be performed by simple or stratified random sampling.