Access and sampling issues

Chapters 9 & 13

• Who do you need to study in order to investigate your research questions?
• You need to confirm at the earliest opportunity that you have the necessary permission to conduct your work. This is sometimes a prerequisite of ethical approval.
• (Also see page 209)
• What kind of sampling strategy will you employ?
• e.g. probability sampling, quota sampling, theoretical sampling, convenience sampling
• Can you justify your choice of sampling method?
• How will you select your sample size?
• With small or targeted samples, how have you selected your participants?
• (Also see pages 195-197, 397-399)

Sample Size and Confidence Interval Calculators:

www.raosoft.com/samplesize.html - from Raosoft

www.surveysystem.com/sscalc.htm - from Creative Research Systems

Exercise: sample size

1

Question

Suggest an appropriate sample size to measure a population with respect to the percentage of persons owing two cars in a family in Huddersfield. We may have a rough idea of the likely percentage and wish to sample to be accurate within 5% points and to be 95% confident of this accuracy.

Generally for percentages, sample size may be calculated using the formula:

Where:

n = sample number

n = estimate proportion

q = non-estimated proportion

SE = standard error of percentage

Consider the standard error of a percentage.

Assume we guess that the percentage ownership of two car families is 30%

Then standard error of a percentage is:

But 2 (SE(p)) must equal 5% (the level of accuracy required)

i.e.:

Rearranging:

i.e.:

Sample size = 336