Sign InThe Forum > Article Comments > Pressuring politicians and populist terrorism > Comments
Pressuring politicians and populist terrorism : Comments
By Geoff Alford, published 26/6/2008Newspaper polls are often just political grandstanding and bring the market research and polling industry into disrepute.
- Pages:
- ‹
- 1
- 2
- Page 3
- ›
- All
It depends how you specify your objectives. If the objective is to identify the “most important issues facing people today” – assuming valid questions – we only need a random sample of 60-70 people to be 95% confident of having identified important issues.
Illustration: define “most important” as any issue important to 20% of the population. To be safe, allow sampling error of +/- 10%. In practice, accept any issue nominated by 10% or more of the sample (some lesser issues may be identified, but unlikely to miss major issues). Then:
N = “1.96^2 x p% x (100-p)% divided by SE^2” , where SE is the allowed sampling error.
N = 3.84 x 20 x 80 / 100 = 62 people
Concerns about non-sampling bias (refusals, non-contacts) might warrant a sample of 200 people as a “safety blanket”.
Your University should have taught you is the bigger the sample, the greater the confidence in “the result”. Or, if the result was derived from invalid questions, then “the bigger the sample, the greater the confidence in the same rubbish results”. It is that simple.
“Elections” are “the outcome”, and sample-based polls are trying to predict that outcome. Hence, elections can produce surprises.
Regarding “a bigger sample being better than a smaller one?”, it depends. A bigger sample will give you more confidence in “the result”, as defined above. But why waste $100K on a survey poll, when $50K will adequately answer your questions?
If you have experienced two competing surveys with vastly different results, and the actual election result also quite different, I would suspect something awry, if not incompetence.
Regarding optional voting systems, people select themselves to vote. We have no way of knowing whether the election result truly reflects the population’s views (a key argument for compulsory voting).
An exception is if 80% of people volunteered to vote, and they all voted the same way. But, you would need to know the context. Is it a referendum for independence in a newly freed country – OK. However, if you were in Zimbabwe, you might be suspicious.