I am aware that with standard error and population's standard deviation available from previous studies one can calculate an appropriate sample size. But how do I locate the standard deviation from studies on disease prevalence (such studies where one is counting for individuals infected and those uninfected)? Please shed some light. Thank you.
You need to have an expected infection rate, and you need to know the required accuracy.
1) I want to prove that disease X is not in Nigeria. The only way to do this is to test all individuals. The equivalent is that I want to prove that everyone has disease X (100% infection).
2) I want to show that disease X is probably not in Nigeria. Here I need a sample size large enough such that the 95% confidence interval about zero is less than some value.
3) I want to show that half the individuals are infected with a 95% confidence interval of 30%.
4) I want to show that half the individuals carry disease X with a 95% confidence interval of 0.00005%. This would take a very large sample because the 95% confidence interval is so small.
An option for reducing testing costs is to pool samples. However, this only works if test sensitivity is sufficient to detect one infected individual out of the pool size. It also only works if the infection rate is low. Depending on the number of pooled samples, "low" is in the range of 20 to 30%. Pooling becomes more effective if infection rates are
Thank you very much for your detailed answer @ Timothy Ebert. How exactly do I then determine the sample size needed in case of option 3 and 4?
The fastest and easiest approach is to run the BinGroup program and enter different values.
If you want to read more, the user manual for BinGroup might help, also see
https://www.researchgate.net/publication/233576917_Reexamining_the_Pooled_Sampling_Approach_for_Estimating_Prevalence_of_Infected_Insect_Vectors/overview
and citations therein.
I don't know of a user friendly pre-packaged program for calculating sample size of pooled samples given current infection rate, pool size, and desired confidence interval.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size