Complete group of entities that share common sets of characteristics
All that meet specific conditions of membership Population elements = those that meet population characteristics Number of criteria increase = cost and time increase
Defining a population?
Right population is selected based upon research questions research
design The target population must be very clearly defined through population parameters
What is a sample?
Researches use samples to infer characteristics of populations without
consulting every member A sample is a subset of a larger population Uses a portion of the population to make a conclusion regarding the whole population (representable) Results of a good sample should have the same characteristcs as the population as a whole (generisable and replicable)
Why use samples?
Practical reasons Accuracy and reliability
Sampling frames
A list from which the sample is drawn
Derived from population elements E.g. customer databases, member lists, phone books, employees No perfect frame exists may need to develop own sampling frame
Sampling procedures
Process of drawing the right sample
Divided into non-probability and probability procedures Necessary for your research proposal
Non-probability sampling procedures
Not ever population member has the opportunity to be sampled
A correct sampling frame does not exist Researchers who make decisions on who they will sample from, with limitaions Includes: - Convenience Being included is a matter of convenience
Right place at right time, or easy to access
Suited to exploratory research Cannot necessarily draw important conclusions Judgement Hand-picked by researchers Possess certain characteristics May be suitable procedure or invalidate results Snowball Respondents to identify and select other respondents Relies on ability to locate initial set of respondents Useful for niche or other hard to reach population May lead to bias due to similarity or familiarity Quota Ensures various subgroups of a population is represented in a specific proportion Left to the discretion of researcher Digital and online ease Relies on in-depth sampling frame
Probability sampling procedures
Every population element is known and has the opportunity to be sampled
Sampling error is reduced and more objective Inferences can be made about the population, theoretically without sampling error Includes: - Random Each respondent has a known and equal chance of being selected Every combination or population elements is a possibility Ensures sampling error is not added to by the researchers - Systematic Randomly selects the first respondent then develop a system for subsequent sampling Once the first is selected, all population elements no longer have an equal chance E.g. every 30th unique customer on a customer database - Stratified Break population down into mutually exclusive strata (sub groups), then sample within these strata - Cluster