054. Types of Samples

1 Simple Random Sample. Several difference samples of a given size can be selected from a population. A Simple random sample involves a method by which each possible sample of some given size has the same likelihood or probability of being selected.

Example 5.3. A national fast-food chain wishes to randomly select 5 of the 50 states to sample consumer taste. (1) If the five states are chosen in such a way that any one of the 50C5 = 2,118,760 possible samples of five states is as likely to be chosen as any other sample of five states, a simple random sample has been taken. (2) The simplest technique is merely to list the states on 50 identical pieces of paper, put them in a hat, and draw out five of them.

There is an alternative approach of random sampling. It is a Random number table. It is often generated by a computer program and formed through the repeated selection of numerical digits. Each time, all 10 digits (0 – 9) have an equal likelihood of being selected.

2 Systematic Sampling. A Systematic sampling is formed by selecting every ITh item from the population. If I is equal to 10, a systematic sample consists of every tenth observation in the population. The population must be ordered or listed in a random fashion. For example, if somebody is needed to select a sample of size 100 from a population of 1,000, I must be 10.

3 Stratified Sampling. A Stratified sample is taken by forcing the proportions of the sample from each strata (division of all units of interest into subgroups) to conform to the pattern of the population. It is commonly employed when the population is heterogeneous, or dissimilar, yet certain homogeneous subgroups can be isolated. In this manner the researcher can increase accuracy beyond that obtained by a simple random sample of similar size.

4 Cluster Sampling. Cluster sampling offers certain advantages over other methods. It consists of dividing the entire population into clusters, or groups, and then selecting a sample of these clusters. All observations in these selected clusters are included in the sample.

< Предыдущая		Следующая >