Please enable JavaScript.
Coggle requires JavaScript to display documents.
Sampling (Steps of
Sampling Design (Steps in sampling design (in…
Sampling
Steps of
Sampling Design
-
-
-
-
-
5) What is our SAMPLE SIZE? Do we need them to represent the entire population? (ak.a. generalize statistics)
-
definitions
-
-
Census- A study of everything or everyone in the population. E.g. All the employees or students of Murdoch University. Feasible if the population is small and necessary when the elements are quite different from each other.
Sample- A subset of elements from the population- to represent all the units of interest in the population. comes from sampling frame
Unit of analysis- describes the level at which the research is performed and which objects are researched. E.g. People, individuals, employees, organisations, divisions, departments or groups.
Probability Sampling- random selection; it eliminates researcher judgement from the choice of actual participants, provides estimates of precision, requires a complete, accurate and up to date list of cases in population (the sampling-frame).
Simple Random Sampling (SRS)- each population element has an equal chance of being selected into the sample. Easy to implement if population is in a database- requires listing of population and can take time to implement. Uses larger sample size Can produce larger errors and sometimes expensive to sample this way.
Stratified sampling- get an equal number of people from each groupusing SRS . Researcher control sample size, increased stats efficiency, strata can be different representations of research. Can be expensive if strata have to be created by researcher but they are often naturally occurring groups.
Systematic sampling- Begin with a random start then use a sampling interval. Easy to design and easy to use and easy to determine mean and proportion of population. The population and the way we choose may skew our results. To reduce bias: randomize the population before sampling or change the random start several times in the sampling process.
Cluster sampling- When creating a list of population can be difficult, costly or impossible. sample in an economical way. Assumes elements are same around the entire population. Select 10 stores from Perth, Melbourne and Adelaide and 10 customers from each store. Unbiased estimate but can lower statistically efficiency. Save times creating a sample but results in 2 sampling errors
Non-Probability Sampling-non-random and subjective choice based on researchers judgement about population characteristics Meets sampling objective due to lower costs, limited time or unable to get a total population list.
Quota- various sub groups will be represented on important sample characteristics. The research may require 85 males and 15 females- interviewer is responsible for getting the number of people required to meet the quota.
Purposive Sampling- sample of participants that were thoughtfully, purposefully recruited in order to fully answer the research question - most frequency used non-probability technique. Normally used to choose small number of participants. To illustrate a typical case of customer, critical, heterogeneous and theoretical cases.
Snowball sampling- participants recruited by researchers asking well-informed people to identify other people who are relevant and would be interested in the study.
Convenience sampling- obtaining people that are convenient to the situation- but not so representative and not generalisable. Least scientific and lacks intellectual credibility.
benefits of sampling are lower costs- budget, greater speed of data collection, availability of sample elements, greater accuracy of results.
A good sample is accurate- has no systematic variance- that is the variation of measures due to some known or unknown influences that causes the scores to lean in one direction more than another.
A good sample is also precise- the larger the sample the less chance of sampling error. A precise sample has a small standard of error estimate.