Sampling is an activity that involves the selection of individual people, data or things, from a target population/universe.
A population, or universe, is the entire set people data or things that is the subject of exploration.
A census involves obtaining information, not from a sample, but rather from the entire population or universe.
A sample (as opposed sampling) is a subset of the population/universe.
For Marketing Research purposes, sampling usually involves people, not data or things.
Sampling Plans are strategies and mechanics for selecting members of the sample from the population:
1. Define the population. It is usually limited based on some set of characteristics, e.g., males, aged 21-39, who have consumed alcoholic beverages within the past 3 months for a beer study. 2. Choose data collection methodology. What kind of information do you require from the sample, how will they be identified, where are they available, etc. 3. Set sampling frame. This is as exhaustive a list as operationally and economically possible that represents the population and is also accessible utilizing the selected methodology. 4. Choose sampling method.
• Probability samples are those that allow all members of the sampling frame an equal opportunity of selection. Probability samples include Simple Random, Systematic, Stratified and Cluster sampling • Nonprobability samples do not allow all members of the sampling frame an equal opportunity of selection. Nonprobability samples include Convenience, Judgment, Quota and Snowball sampling. 5. Determine sample size (subject of Chapter 13)
6. Develop operational procedures for extracting sample from the population (logic and controls) 7. Executed the operational plan
Sampling error. Random sampling error always exists.
Administrative errors are generally controllable when properly identified and monitored. Results from two samples, drawn from the same population, will have random sampling errors that can be estimated.
Probability sampling methods:
1. Simple random: Selection to the sample is sample size/population size. Select records from table of random numbers.
2. Systematic: Selection to sample is population size/sample size. Then, every nth record is selected.
3. Stratified: Ensures that explanatory (independent variable) characteristics are properly represented in the selected sample. • Identify classification factor(s)
• Determine proportion in population (need not be balanced, can be disproportionate) • Divide population and select based on target proportion
4. Cluster: Generally geographically based. Save $. Counties, blocks, households/locations (polling).
Nonprobability sampling methods.
1. Convenience: From biased sampling frame, e.g. SSI LITe. Hard to find populations.
2. Judgment: Researcher determines who is in sample
3. Quota: Not selected randomly, must have characteristic of interest
4. Snowball: referral from others in sample (focus groups)
Internet sampling is loaded with biases, but that is not necessarily a bad thing. Depends on research objective
DETERMINING SAMPLE SIZE
Sample size determination is computed using three inputs:
• The estimate of the population standard deviation (often obtained from earlier studies). • The acceptable level of sampling error
• The desired confidence level
Generally, research practitioners utilize the following sequence and inputs in computing sample size:
1. Survey respondents will split 50/50 in response to dichotomous (e.g. yes/no) questions. 2. The desired level of confidence will be 95%, or 1.96 standard deviations from the mean or .05 possible
Py = Proportion responding “yes”
Pn = Proportion responding “no”
Standard error is the acceptable amount of error/confidence interval. In...
Please join StudyMode to read the full document