Estimation of appropriate sample size is one of the vital part for clinical as well as non clinical researchers. The available literature has suggested different methods to estimate sample size. Which method did you preferred for your researchers and why?
Dear @Ratan, from statistical point of view: the sample size for clinical studies should be as large as possible; since this will cost more time and money, so, the researcher should find it out how large the size should be. Usually, grouping the elements of the population of the study in clusters and take representative cases from each cluster will lead to a proper conclusion of the results of the study.
Dear @Ratan, from statistical point of view: the sample size for clinical studies should be as large as possible; since this will cost more time and money, so, the researcher should find it out how large the size should be. Usually, grouping the elements of the population of the study in clusters and take representative cases from each cluster will lead to a proper conclusion of the results of the study.
The sample should be adjusted to the question/hypothesis and scale of analysis (e.g. age or gender effects, environmental effects, genetic effects; ecological versus evolutionary scales of analysis). We should not assume that a local sample from a study population (e.g. people from one city) is representative for a target population (e.g. local versus regional versus continental versus global scale). Very large, but local samples, are not necessarily representative, perhaps complicating data interpretation because of biased sampling (E.g. Hurlbert 1984, Ecology).
Dear Ratan,
The cost, should be minimized, however the minimum sample size should be statistically meaningful as well.
Following are the basic considerations for estimating sample size for a CT
1) Study Design
2) Hypothesis Testing
3) Primary Study End Point
4) Expected Response Test vs. Control
5) Clinical Important/Meaningful Difference/Margin
6) Level of Significance
7) Power
8) Drop-out rate
You may find the attached article useful
Sample Size Estimation in Clinical Trial
Perspect Clin Res. 2010 Apr-Jun; 1(2): 67–69.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3148614/
Dear Ratan
I fully agree with, and endorse the point made by dear Abedallah and Kamal about the sample size in clinical studies so as to lead to more reliable results. Of course, since this has resource and cost implications, the time constraints and funds constraints should also be kept in view in this regard.
To me, there is absolutely no difference between the recommended experimental design for tests in birds versus humans, from a scientific point of view!
Arguments for discussion:
Biology of humans and non-human vertebrates are likely at similar levels of phenotypic and genetic complexity. Investigations indicate that a non-domesticated bird and a human have many biological characteristics in common. Humans and wildlife share many scientifically defined terminologies, including how organs, tissues, cells and their phenotypic, physiological or chemical compounds are named. Research in wildlife might reveal phenomena not investigated in humans. Alternatively, some research conducted with humans cannot be done with wildlife.
Larger sample sizes usually yield more power and accuracy to clinical research and trials by dimishing the error variance/margin of error, through a more effective accommodation of variables that are significantly different, yet limiting the influence of exceptions and outliers. That said, the definition of an 'adequate' sample size is quite variable as well, depending on the objective and context.
Sample size depends mainly on variance of the population, which can be concluded from a pilot study. The larger the variance among the pilot study responses, the larger the sample size will be.
Dear Mahfuz,
the relationship between sample size and variance will depend both on the size of the sample (5 versus 20 versus 100 versus 2000 etc..) and the scale of sampling (within-individual, between individuals within study populations, between study populations, etc)...
Dear @ Abedallah, Kamal, Debi and Asmat. I agree that larger sample size will always increased the cost of research in term of time and money.
Dear @ Behrouz, I agree with your steps and thanks for the valuable attachment. It is very helpfu to me...
Dear @ Monica, I agree with you point "Larger sample sizes usually yield more power and accuracy to clinical research". This may be another point why we requir appropriate sample size.
Dear @ Mahfuz, I agree with the point that conducting a pilot study is one of the common method use to determine sample size...
Dear @ Marcel, I agree to the point " possibility of variability in sample size with different variences...
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size