Original questions written at exam standard. Work each one before you open the solution.
Answer. The population is the entire set under study; a sample is a subset of it.
Conclusions are drawn about the population using data from the sample, which is why the sample must represent it.
Answer. Complete and free of sampling error; but expensive and slow.
A census is also impossible where testing destroys the item — you cannot crash-test every car.
Answer.
, leaving 20 females.
Answer.
, taken from a random start between 1 and 40.
Answer. It is not random, so the probability of any individual being selected is unknown.
Statistical inference requires known selection probabilities. Quota sampling lets the interviewer choose within each quota, which introduces unmeasurable bias.
Answer. It does not control other variables, so a lurking variable may explain the link.
Only a controlled experiment — with randomisation and a control group — isolates the effect of the variable being changed.