With probability set and statistical significance set confidence

SAMPLING

• Researchers use samples to infer characteristics of a population

The entire universe or population of
interest. The ‘population of Bond
students’ is easy to define; but what if I am interested in motorsports fans, how do I define that population? Or something that sounds simple, like ‘Australian
adults’?

A subset of that population.

DEFINING THE TARGET POPULATION

What is the relevant population? Not always easy to define…

Sampling frame: A list of elements (e.g. a
membership list, a student email list) from which a sample may be drawn

– If a researcher was interested in surveying doctors in
Australia, a sampling frame might be a list of all members of the Australian Medical Association.

Nonprobability sampling: A sampling technique in which members are selected on the basis of
personal judgement or convenience.

– Results appear to be random if there is no other systematic pattern to the list (ex: you are standing by the men’s bathroom…)

PROBABILITY SAMPLING METHODS (CONTINUED)

PROBABILITY SAMPLING METHODS (CONTINUED)

Cluster sampling

-0

Probability sampling

– It’s … convenient! A lecturer who uses students for a study; you ask your friends to complete a survey or participate in a focus group.

– Convenient, but potential respondents may be unwilling or unrepresentative. Are my friends representative of technology trends or political opinions?!

NONPROBABILITY SAMPLING (CONTINUED)

Quota sampling

– Respondents are obtained from information
provided by the initial respondents, they refer other individuals.

– Common if dealing with rare populations, e.g., a teddy bear collector would provide contact details for other teddy bear collectors.

– Nonprobability methods are better for projects with financial and human resource constraints. If we are thoughtful (any reason to believe respondents are not representative?!), non-probability samples are often “good enough”.

EVALUATING SAMPLE DESIGN CRITERIA (CONTINUED)

Web surveys are increasing in popularity

– Option to reach a large sample rapidly (but verify that the survey you show formats properly on different devices).

• The confidence level (e.g. 90%, 95% or 99% CL) – Higher confidence requires a larger sample.

There are math formulas to determine an appropriate sample size – the ‘power’ of the study.

DETERMINING SAMPLE SIZE ON THE BASIS OF JUDGEMENT (COMMON)

• Sample size may be determined on the basis of managerial judgements.

• Sampling involves selecting a subset of the target population for the purposes of being able to draw general conclusions about that
population.

• Why sample?

Findings will be reasonably accurate (there are formulas to determine the ‘confidence interval’).