i would like to determine the sample size to use in collection of data among children admitted with a certain condition. how do i go about getting the sample size so as to determine the prevalence of the condition in the admitted children, and later proceed to get details from the ones with the condition? please help.
The formula is Z*(stnddev)/sqrt(n)=0.5*(length of the confidence interval) for the infinite population case, for finite populations see p 89 of Scheaffer et al, Elementary Survey Sampling 6th ed. Thomson. Sample size calculators can also be found on the WWW. One such is at
https://www.surveysystem.com/sscalc.htm
Best Wishes, D. Booth
You can use the attached formal on image. Considering 95% confidence level with reflected 5% merging of error and considering sample size is normally distributed (50%probability value). z=1.96 . N= Population
Hi Rose,
Sample size estimation is thoroughly a subjective procedure that requires you to have some level of information before proceeding.
1. First of all, you should get a rough estimate of all children (within the age range you intend to enroll in your study) from the head nurse or doctor at the hospital(s) you intend to recruit patients from. This would allow you to have an idea of the source population (N), which is needed for accurate sample size estimation.
2. Search for literature on similar studies from scientific databases to have an idea of previous researches done in the area. Here you would have an idea of sample sizes and methods previously considered in your intended area of study. You could also estimate effect size from relevant literature. Effect size is needed for sample estimation, especially if you want to use GPower Software to estimate your sample size.
3. Download GPower (simple Google search) and estimate your needed sample size, and possibly over-sample the number to control for non-response and/or attrition.
Good luck!
Since you want to discuss the prevalence of the condition, it appears that your sample will include both children with the condition and children without the condition. Depending on the rarity of the condition, you may need a much larger sample size to discuss children with the condition than you need just to discuss prevalence. Thus, you need to start with some estimate of the rarity of the condition, and then make sure that the number with the condition will be large enough to provide reliable statistics.
Use confidence interval. Note it is easy to do the calculation using build in programs such as the one found on
http://epitools.ausvet.com.au/content.php?page=1Proportion&Proportion=0.5&Conf=0.95&Precision=0.05&Population=457
Search for some hospital based studies /literatures which tells you the prevalence. Use free soft wares like open epi in Google select precision level and confidence level you can easily calculate the required sample size
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size