Sampling

population

whole set of items that are of interest

census

observes every member of a population

sample

observations taken from a subset of the population

sampling units

individual units of a population

sampling frame

list of sampling units that are either named or numbered

advantages and disadvantages of census

ad: completely accurate result

dis: time consuming, expensive, hard to process all data, can’t be used when testing process destroys item

advantages and disadvantages of sample:

ad: cheaper, quicker, less to process

dis: not as accurate, not large enough for info about subgroups

random sampling

each member of population has equal chance of being selected
simple random, stratified, systematic

simple random sampling

all sampling units are numbered/named in a sampling frame and chosen at random by either a random number generator or a lottery

ad: free of bias, easy and cheap, equal chance for each sampling unit

dis: need sampling frame, becomes expensive and time consuming when sample is big

systematic sampling

required elements are chosen at regular intervals from an ordered list eg every 5 people are chosen from a list

ad: simple and quick, suitable for large samples and population

dis: sampling frame needed, can have bias if sampling frame not random

stratified sampling

population divided into mutually exclusive stata (eg males and females) and random sample is taken from each, each strata has same proportion

ad: accurately reflects population structure, guarantees proportional representation

dis: population must have clear strata, same disadvantages as random

non-random sampling

quota, opportunity

quota sampling

population divided into strata, proportion of strata determines number in each quota, interview people until all quotas are filled up

ad: allows small sample to be representative of population, no sampling frame needed, quick, easy, cheap, easy comparision between groups of a population

dis: bias, dividing into strata can be inaccurate and expensive, non-responses not recorded

opportunity sampling

taking sample from people who are available at time study is carried out and who fit criteria you are looking for eg first 20 people who have brown hair outside a hair salon

ad: easy to do, cheap

disadvantage: unlikely to be representative, highly dependent on researcher