Presumed prevalence was 50%, as parent study in the region, but actual prevalence was 16%.
Does this much of difference makes my calculated sample size wrong?
Thanks in advance.
From the sample size you can calculate 95%CI for population proportion using the sample proportion of 16% and the binomial (or normal approx) distribution. Then by inversion you can calculate how likely a proportion of 16% would be in a random sample of size N from the population with proportion 50% . If it is very unlikely you can reject the hypothesis that is was a random sample from this population ie there is something different about it or it has not been randomly sampled.
Alternatively, if you are trying to get a sample with a proportion within a certain range of the population proportion, you can calculate the sample size necessary for this, assuming random sampling.
Hi Rukman,
I do not think this would automatically imply that your sample size was wrong. Alternatively, it could also suggest that your sample population was significantly different from the target population.
Regards,
Prevalence has other contributing factors. Could it be that the number of deaths from outcome has increased? Has the quality of care and management improved such that more people are cured? However this is not true for chronic conditions, other parameters should be considered. It goes in same line with an intervention; we need to findout what happened???
Your assumptions for calculating the sample size was the anticipated prevalence. But the results are the actual prevalence based on many other factors discussed earlier. So, I don't think it's I'll justified or decrease your external validity.
No because you used a sample size that was based on using p=0.5 which gives you the maximum needed sample size. See: https://www.google.com/search?client=firefox-b-1-d&ei=iGHMX9CkCpOC9PwP4set4A0&q=what+sample+size+should+i+use+for+a+binomial+experiment&oq=What+sample+size+should+I+use+for+a+binomia&gs_lcp=CgZwc3ktYWIQARgCMgcIIRAKEKABMgcIIRAKEKABMgcIIRAKEKABMgcIIRAKEKABOgQIABBHOgUIIRCgAToICCEQFhAdEB46BQghEKsCOgkIABDJAxAWEB5Q0cMcWKurHWC7zx1oAHACeACAAegBiAGjD5IBBjAuMTMuMZgBAKABAaoBB2d3cy13aXrIAQjAAQE&sclient=psy-ab
Esp. the videos Best D. Booth
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size