Chapter 7 - Case and Comparative Studies Flashcards
Sample
A subset of the population
Sample Statistic v Population Parameter
Sample statistic is the number one gathers from a population, the population parameter is the actual statistic (that we’ll never know).
Sampling Error
The difference between the sample statistic and the population parameter.
Three Factors in Representative Sampling
Sample Frame
Sample Selection
Sample Size
Sample Frame
Everyone whom the sample is trying to represent.
- List of all units of the target population
o Every single UofS student for example
- Considerations:
o Does it even exist?
o If it does, would it be shared?
o Is it complete / accurate?
Sample Selection
Can either be probability based or non probability based
Probability Based Sample Selection
o Everyone has equal chance of being selected
Based on probability theory
o Can estimate likelihood of representativeness
o Can estimate accuracy of sample statistics (sampling error)
o Appropriate for statistical analysis
Non-Probability Sampling Designs
Accidental Sample:
Convenience sample - Stopping people on the street, taking a survey of people walking through arts
Self-selected / voluntary sample
Purposive Sample - Researcher selection and judgment (they use their understanding of the population to determine who they should be surveying – experts, lived experience)
Snowball sampling – going to one person, getting another name, another name
Probability Sampling Designs
Simple Random - Random selecting from a list
Systematic selection - A selection interval is calculated based on the sample needed
Stratified - Selection units within it - A random sample of Canada? But your sample of Canada will be mostly Ontario and Quebec. So you need specific strata.
Cluster - Based on geographic units – and then those are randomly selected
Sample Size
Negative relationship between sample size and sampling error
o As sample size increases, the margin of error/sampling error decreases
Criticizing Probability Sampling
o What are possible gaps?
o How could the sampling frame possibly introduce bias?
Criticizing Non-Probability Sampling
Convenience or purposive sample?
Purposive = knowledge as research for specific people
o Potential for ‘closed loops’?
Some communities that are too similar or are a part of one particular network wouldn’t be representative
Non-Probability Sampling Is Used in These Four Contexts:
o Exploratory research
o Extremely small populations
o Unavailable/inadequate sampling frames
o High refusal populations
Sample Sizes and Non-Probability Sampling
- Focus on quality of information obtained, rather than quantity of cases
- Looking for data saturation: point at which data collection no longer provides the researcher with new information
Data Saturation
Point at which data collection no longer provides the researcher with new information
