How to calculate the effect size for clinical trial sample size calculation? and how to calculate the standard deviation in cases of educational interventions for diabetic patients with hba1c as the primary outcome
Please read this article:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3148614/
Hi Ehab Mudher Mikhael
The links below highlights how to calculate the effect size for clinical trial sample size calculation as well as the standard deviations
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2876926/#!po=86.0000
https://www.researchgate.net/publication/11190202_Determining_the_sample_size_in_a_clinical_trial
Here you have different effect size calculators:
https://www.polyu.edu.hk/mm/effectsizefaqs/calculator/calculator.html
QUESTIONS ASKED There are three questions asked. These three questions do not deal with the same issue: (i) effect size, (ii) sample size, and (iii) standard deviation.
1. EFFECT SIZE is the magnitude of phenomenon. There are several methods of calculating effect size. Cohen's d is one method:
d = (X^1 - X^2) / Spool
where X^1 is the mean of group 1 and X^2 is mean of group 2; and Spool is the pool standard deviation.
See this link: https://en.wikipedia.org/wiki/Effect_size
2. SAMPLE SIZE. if the population size (N) us known, use the Yamane equation to determine sample size:
ny = N / (1 + N(e2)
where N = population size and e = error level, i.e. 0.05 is confidence interval is 95%. If the population size is not known (non-finite), then take test sample to determine minimum sample size requirement for continuous data by:
n = (Z2σ2) / SE2
where Z is the critical value at at given confidence interval (see Z table), σ is the expected standard deviation, and SE = σ / sqrt(n). For discrete data, use the following equation:
n = (Z2PQ) / e2
3. STANDARD DEVIATION. Given series of 10 observations (x1, x2, ..., x10) where the expected value is the mean (x^), the variance is the average mean-difference squared of the series: var = SUM[SQRT((1/n)(xi - x^)2)]. the standard deviation is the square root of the variance: SQRT(var).
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size