Sampling Framework!!!!!

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Sampling Framework!!!!!
Message posted by Andrea on October 6, 1999 at 12:00 AM (ET)

I am trying to conduct a survey of local companies. I want to take a sample of these companies to survey....Is there any rule of thumb as to how many I should survey? I saw somewher that it should be at least 15 for each breakdown you use (ie if i analysed for a certain type of business employing 10 to 20 people). But the same article said I should have at least 30% of the complete number of employees!!!!! I don't understand! The complete population for some of the breakdowns may be between 25 and 2000.....if that's any help! I hope someone can give me some....any advice on this PROBLEM!!!!!!!!!

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Sampling Framework!!!!!
Message posted by JG on October 6, 1999 at 12:00 AM (ET)

To do meaningful random sampling you must know or be willing to guess a great deal of information about your population. Also, only if your sampling procedure is random or probabilistic does sompling theory apply.

You have to guess the population size, any strata in the population that are relavent, the variability of the population or of each strata, the required accuracy and reliability of the statistics that you plan to calculate from your sample, if things change with time, etc.

There are web sites that discuss 'statistical sampling procedures, etc.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.