Unlike clinical and preclinical experimentation, in laboratory studies (e.g. studies that use primary cultures from different donors), a sample size calculation is not used/reported. It is generally reported that 3 to 5 samples from different donors are used, as rule of thumb, but they do not explain why they use this number.
I wonder if there is a study or an argument validating the lack of sample size calculation and the use of 3-5 samples in this type of studies
I don't believe so.. please look at a beginning statistics text to see what sample size is all about.. Best wishes, David Booth
Oh, I know what sample size is all about. Please look at a beginning reading comprehension text
The use of (typically) 6 as a group size in animal experiments is generally justified by the sheer number of potential variables that are screened. So many findings in animal studies never go anywhere, so the animal people want to focus on results that pretty much pass the Mark I eyeball test. With 6 and 6, data are only significantly different when there is no overlap between the groups, or maybe just one observation.
The small sample sizes for in vitro is interesting. I guess that you are concerned with demonstrating a process or a phenomenon, and that if you can make it happen three or four times it counts as a credible demonstration. But this isn't hypothesis testing, so the notion of sample size and power to detect effects may not apply.
I'd be interested to hear from colleagues in the in vitro area, though.
Interesting contribution here : Article CRIS Guidelines (Checklist for Reporting In-vitro Studies): ...
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs