Chapter 7: Sampling Flashcards
Population
Entire set of people or products in which you are interested
Sample
Smaller set, taken from the population
Just because a sample comes from a population doesn’t mean it generalized to the population
Census
Testing every person/product in the population
Population of Interest
Population to which researchers want to generalize to
E.g. men with dementia or undergraduate women
Biased Sample
Unrepresentative sample
Sampling bias
Some members of population of interest have a higher probability than other members of being included in the sample
Unbiased Sample
Representative sample
All members of the population have an equal chance of being included in the sample
Only unbiased samples allow us to make inferences about the population of interest (generalize)
Sampling Bias
When the sample does not accurately represent the population (some members have a higher probability of being included than others)
Could contain too many of the unusual people
E.g. sampling a chip from very bottom of bag or quiche from variety pack
E.g. people who respond to rate professors
Unrepresentative Sampling Techniques
Convenience sampling
Self selection sampling
Purposive sampling
Snowball sampling
Quota sampling
Convenience Sampling
Using a sample of people who are easy to contact and readily available to participate
E.g. college professors sampling college students
E.g. go to depression clinic to interview willing people (not all people with depression get treatment)
E.g. sampling those who invite themselves (Sona) (doesn’t include people who don’t want to participate) - might not be a problem with deception in terms of topic of study
E.g. people who complete online surveys
Self Selection
Sample only contains people who volunteer to participate
People who take the time to complete online surveys tend to have stronger opinions
Not all internet polls have this bias - can be random selection invitation which rules out this bias
Representative Sampling Techniques
Probability Sampling (random sampling)
Simple random sampling
Systematic sampling
Cluster sampling
Multistage sampling
Stratified random sampling
Oversampling
Combining techniques
Purposive Sampling
Want to study only certain kinds of people
Targeting a very specific group for recruitment
Nonrandom selection
E.g. studying only smokers for a smoking intervention study
Snowball Sampling
Variation on purposive sampling
Relying on participants to help find other participants
Participants asked to recommend a few acquaintances for the study
E.g. asking people to contact others from their support group for depression
E.g. truck drivers
Quota Sampling
Researcher identifies subsets of population of interest and sets target number for each category in the sample
Sample from population of interest non randomly until quotas are filled
External Validity
How well would the claim generalize beyond the current sample
Determined by how sample was selected (e.g. is sample representative of population)
Most important for frequency claims
Probability Sampling (random sampling)
Every member of the population of interest has an equal and known chance of being elected for the sample
Regardless of whether they are convenient or motivated to volunteer
Have excellent external validity, can generalize
Simple Probability Sampling
Generate random numbers, each participant number generated is included
Systematic Sampling
Select two random numbers
Start with first number member (e.g. number is four, start counting from member #4)
Use second number to select participants (e.g. number is 7, every 7th participant starting from participant 4 is selected)
Can be difficult to find and enumerate each member
Cluster Sampling
People are already divided into arbitrary groups
Clusters of participants within population are randomly selected
All individuals in each selected cluster are used
E.g. randomly select high schools out of all high schools in population
Multistage Sampling
Two random samples are selected
Random sample of clusters and a random sample of people within the clusters
E.g. select high schools at random and sample certain students from each school
Stratified Random Sampling
Researcher purposefully selects particular demographic categories or strata
Retain important characteristic of population in sample (e.g. male:female ratio)
Randomly select individuals within each category, proportionate to assumed membership in population
E.g. ensuring sample of Canadians has representation of South Asian descent equivalent to in population
Two categories/strata in their population (South Asian Canadians and other Canadians)
All members of both categories selected at random
Differences Between Cluster Sampling & Stratified Random Sampling
Strata are meaningful categories (ethic or religious groups), clusters are arbitrary (any set of high schools)
Final sample sizes of the strata reflect their proportion in the population, clusters are not selected with proportions in mind
Oversampling
Variation of stratified random sampling
Oversampling: researcher intentionally overrepresents one or more groups
E.g. oversampling the South Asian Canadian population to include more than what is represented in the population
Adjust the final results so members in oversampled group are weighted to actual proportion of the population