Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
MARGIN OF ERROR
Message posted by Jim on May 9, 2000 at 12:00 AM (ET)
In calculating the margin of error in a given sample, is the population size ever relevant? If I draw two random samples of 1,000, one from a large population, say 1m, and another from a small population of, say, 3,000, do I apply the same formula to determine the margin of error?
Common sense would seem to suggest that when the sample and population sizes are so similar, the margin of error must be lower. Is this true and is there a differemt formula for calculating margin of error in such cases?
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: MARGIN OF ERROR
Message posted by William Gray on May 12, 2000 at 12:00 AM (ET)
It depends more on what sub-groups you want to break out of the main group, than on what the size of larger population is.If you want to look at men and women among the population, for example, you need to sample enough men and women to give you the margin of error you consider acceptible. That would increase your overall sample (and decrease its margin of error, according.)
Re: MARGIN OF ERROR
Message posted by Doug Balson on May 17, 2000 at 12:00 AM (ET)
there are two formulas for calculating the margin of error (one for a finite population and one for an infinate population). For all practical purposes when a population is over 1 million the second formula is used but I think it makes little practical differnce. The interesting point here is that your sample size realitive to the population makes no difference/impact on the margin of error. It is the ablsolute size of the sample and its charteristcs (Homogeneity) that do this.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.