Sampling

• Sampling:
• the art of learning about a very large group of people by getting information from a small set of people

• population:
• the entire set of individuals, events, units with specified characteristics.
• examples: citizens of Canada; students; adults; men.


• parameter
• a summary description of a particular aspect of the entire population.
• Example: the mean age of citizens in the country.

• sample
• the subset of the population from which data is collected and used as a basis for making statements about the entire population.
• Example: the people interviewed in a poll

• statistic
• a summary description of a particular aspect of a sample.
• Example: the mean age of the people in the sample.
• Statistics are used to describe samples and to estimate population parameters.

• census
• a sample that includes the entire population
• very expensive, time-consuming
• often impossible to do

• sampling frame
• the list (or quasi-list) of units comprising a population from which a sample is to be selected.
• If the sample is to be representative of the population, the sampling frame should include all members of the population.
• Example: telephone book

• There are two kinds of samples:
• The Good
• Probability samples:
• You know everyone's chance of being in the sample . . .
• so you know which population the sample represents
• The Bad & Ugly
• Non-Probability samples:
• You don't know everyone's chance of being in the sample . . .
• so you don't know which population, if any, the sample represents

• Non-Probability samples
• Accidental/Convenience Samples
• Take whoever is convenient; whoever comes along
• Purposive samples
• "Handpicked" samples
• The researcher decides who will be included as "typical" or "representative"
• Quota or proportionate sample
• A sample that is as similar to the population as possible
• Ex: if 23 percent of the population are single men, 40 percent have brown eyes, and 19 percent drive blue cars

• What's wrong with Non-Probability samples
• What's wrong with them?
• Convenience samples do not accurately represent the population from which they are drawn.
• Purposive samples do not accurately represent the population from which they are drawn.
• Although they are sometimes called "representative" samples, neither quota nor proportionate samples are representative.
• The worst thing is that you don't know which population these samples represent!

• Probability samples
• Simple random samples
• Systematic samples
• Stratified random sample
• Cluster samples

• Simple random samples
• Equal probability
• All members of the sampling frame have an equal chance of being selected.
• Independence
• The fact that one member is selected has no effect on any other member's chance of being selected.

• Systematic samples
• If you are able to line up all members of your population and move down the line one member at a time, you can do a systematic sample.
• Start at one end of the line and take every k^th member (for example, every tenth member).
• If you choose the first member randomly, your sample becomes a systematic sample with a random start.
• Systematic samples are almost equivalent to simple random samples.

• Stratified random sample
• Divide your population into a number of strata (subsets) . . .
• where each stratum is internally homogeneous with respect to the characteristic being studied . . .
• and take a random sample from each stratum.
• A stratified random sample can be more efficient than a simple random sample.

• Cluster samples
• A sample in which clusters of people are sampled.
• You choose the clusters randomly, and then take all of the members of the selected clusters.
• Example:
• sample university students with a cluster sample of courses
• you would choose courses randomly; then take everyone in each chosen course
• Cluster samples are less efficient than simple random samples or stratified random samples;
• there is more variability from sample to sample with cluster samples