I've read that statistical significance rises with sample size, but with cell line DNA we work with DNA from thousands of nearly genetically identical cells in each sample. Does this mean that even though the p-value for my correlation for each cell line is higher than 0.05, the result is still statistically relevant because we can count each of the cell's genetic material as its own sort of replicate?
Nope. The question is: what is your unit of randomization? If you randomize a culture plate to a treatment, than the whole culture plate is your unit, not a single cell. Only if you randomize single cells (what you almost surely don't do), the cell would be the unit of randomization.
it's essential to consider other factors such as experimental design, potential confounding variables, and the overall context of your research. While a higher sample size can increase statistical power, it's still important to interpret results cautiously and consider the broader implications of your findings.
Yes, Jochen Wilhelm is right. That would be similar to measuring some property in a certain volume of a given solution and stating that the sample size is of the order of the Avogadro number because you consider individual molecules.
- Badges
- Science topic