Question

Are all good samples random? This is an opportunity to bring up opinion polling, which typically tries to obtain views from particular groups (e.g., men, women, older, younger, employed, unemployed, Democrat, Republican, etc.) and then “weights” the results by the prevalence in the population.

Magazines often report surveys giving statistics such as “63% of women expect the man to pay on the first date.” Are these random samples? These surveys are most definitely not random – they are typically click-through from the magazine website – and so can provide an opportunity to discuss the sort of biases that can result from lack of random sampling.

Answer 1

1. No, not necessarily all good samples are random.

The given example of opinion polling clearly justifies this statement. In order to reflect the actual population opinion, opinion polling must be done by selective sampling through particular groups.

2. No, if the surveys are obtained as a click-through from magazine website then it cant be said random. This is because the sampled data will only be obtained from the people who visit the magazine website. So we can certainly say that here it will be the case of insufficient representativeness of the overall population. As a result, the conclusion will also be biased only reflecting opinion of a subset of the population. This is also called sampling bias.