How can I calculate sample size based on confidence interval? and when it is recommended to use this approach rather than Cohen's power approach?
I depends on the assumed distribution.
If the distribution is normal with (quite well) known sigma, then you get a good approximation with
n = ((z*sigma)/w)²
where z is the normal quantile for the given confidence level and w is the desired half-width of the confidence interval.
Example:
Given the standard deviation (sigma) of the variable is 4. The 95% confidence interval should be plusminus 0.5. The z-score for 95% confidence is 1.96.
n = ((1.96*4)/0.5)² = 245.8
So you will need a sample size of n = 246 to get an expected width of 95% confidence intervals of plusminus 0.5.
EDIT: I did not see the previos answer while I was writing mine. Nice to see that it shows exactly the same solution (on the same example that the assumed distribution is normal!).
Hello,
I'm not sure about what you mean but if I good understand your question then this answer can be helpful. If you want to calculate a sample size based on confidence interval you can use this formula (i.e. for estimation average value):
n=(za2s2)/e2
where s is a standard deviation, e is a half-length of confidence interval and za is a value of Standard Normal Distribution for 1-a, and a is alpha - level of significance.
I depends on the assumed distribution.
If the distribution is normal with (quite well) known sigma, then you get a good approximation with
n = ((z*sigma)/w)²
where z is the normal quantile for the given confidence level and w is the desired half-width of the confidence interval.
Example:
Given the standard deviation (sigma) of the variable is 4. The 95% confidence interval should be plusminus 0.5. The z-score for 95% confidence is 1.96.
n = ((1.96*4)/0.5)² = 245.8
So you will need a sample size of n = 246 to get an expected width of 95% confidence intervals of plusminus 0.5.
EDIT: I did not see the previos answer while I was writing mine. Nice to see that it shows exactly the same solution (on the same example that the assumed distribution is normal!).
Dear Professor Muayyad Ahmad
hoping you are in a good health,
Kindly find the attached file which contains "The recommended sample size for a given population size, level of confidence, and margin of error appears in the body of the table". Based on the research advisor 2006.
In addition to a website for research advisor 2006
the issue of using this approach rather than Cohen's power approach depends on the researcher experience. I think if the researcher didn't know Cohen's power approach well, he will use like such that table.
http://www.research-advisors.com/tools/SampleSize.htm
Thank you all colleagues for your contributions and answers. That was helpfull
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size