Quote:
Originally Posted by AnemicOak
I'll take your word for it. 42,000 out of a population of over 33 million seemed really small to me, but I know nothing about these things. 1,000 seems very small.
|
I don't know how large a figure you need to draw reasonable conclusions from a survey, but 1,000 is quite common for basic surveys. The bigger issue is usually the quality of the sampling, i.e how those 1,000 people are selected can have a drastic impact on the accuracy of any conclusions draw from the data.
For example, it'd be utterly silly for anyone to sample 1,000 people asking do you buy/read ebooks, if that sample is drawn from 1,000 mobile read members
Surveys can be useful, but everyone needs to remain skeptical of conclusions drawn unless the raw data is available as well as details of how/when/where surveys were carried out.
Not saying the 42k sample is right or wrong, I've not looked at it and don't plan to, but people shouldn't take things at face value without questioning, just as you're questioning the sample size.