Dear all,
I need some help deciding on the minimum sample numbers for meaningful statistical analysis. I will be working with blood samples and so far I have planned 3 experiments;
1) I have one independent variable (2-levels) (eg. 2 different pH values), and control and treatment groups. So I would like to see how the pH affects the measurements in my control/treatment groups. At the same time, I would like to divide my samples into categories (eg. blood groups A,B,O,AB +) and see if there exist a difference regarding the measurements done for control/treatment
2) Next, I would like to measure a quantifiable value (eg. hemoglobin content) for the same group of people and then see if there is a difference between blood groups
3) Lastly, I want to measure another quantifiable value (eg. Ca2+ content) at the same 2 different pH levels (like above) and at last check if there exists a difference at Ca2+ levels for the blood groups as well
As I will be working with human material I am trying to keep the number of donors I need to a minimum. I also think that I will not be able to have the same amount of donors per blood group.
I am not sure if I can lower the number of donors by doing repeated measures, or if I should just get a population of eg. 20 people to do the experiments and assign the categorical value (blood groups) randomly. I know the donor size is also dependent on the type of statistical analysis but I am having a hard time calculating the effect sizes.
I appreciate all the answers/ideas/comments :)
Thank you!!
Hi Ipek
Am I correct that your study 1 and 3 are in-vitro (ex-vivo) laboratory experiments but 2 is a straight forward observational study? Let's start with the study 2 (observational). I am afraid you are going to need a large cohort to get any meaningful answer to your question because the differences you are going to see are small. Furthermore, there should be sufficient number of each blood group type in your study population. On the other hand, being experiments, you may not need very many subjects for your study 1 and 3 (in fact, you can combine both studies). The problem is that you need to have equal number of A, B, AB and O for your experiments. Your suggested number of 20 may work (5/blood group) for a pilot study. If you are getting discernible differences, you can then do a larger study.
Good luck.
Siva
Dear Muttuswamy Sivakumaran,
thank you for your answer. You are indeed correct; the experiments 1&3 are planned in-vitro. Your suggestion is very valuable and helpful, I would definitely keep it in mind while planning further.
Thank you!
Have a nice day:)
Hi,
go through
How to choose a sample size (for the statistically challenged)
https://tools4dev.org › resources › how-to-choose-a-sampl...
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Enzymology
- Enzymes