census - information from the entire population population - the entire data set subset (sampling frame) - a ‘slice’ of the population
random sampling
simple random number sampling
- done by index
- easy and fast
- removes bias
- requires indexed sampling frame
- may be unrepresentative
systematic
- randomly sorted list, selected by stepping (population size / sample size)
- easy and fast
- removes bias
- requires indexed sampling frame
- may be unrepresentative
- is more tedious than simple random number sampling
non-random sampling
stratasfied
- groups as allocated into stratas, each strata has a set size
- no bias within stratas
- more representative of the population
- takes longer to do
- can be harder to identify groups for the population
- biased
quote
- ask people as you find them, if they fit your quota
- does not require a defined sampling frame
- inherently biased
opportunistic
- randomly ask people when given the opportunity (ie coming across people whilst standing in a location)
- does not require a defined sampling frame
- inherently biased