I need to perform several replicates of an experiment, say I have 3 treatments and 4 replicates. After performing the calculation of the sample size, I get n = 100 samples needed. This means that for each replicate I need 100 samples or can I split the samples among my replicates (i.e., 25 samples per replicate)? Thank you.
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Usually, if you want to draw conclusions about a population, we use representative samples which are adequately sized (sample size) and reflects all the features of the population. If calculations say that the sample size should be 100, we need 100 data points to make statistical arguments about the population based on these 100 data points (sample size).
You mentioned there are 4 replicates i.e. you counted 4 separate setups and so, if you divide 100 into 4 groups, each with 25 data points, you have, statistically, only 1 replicate with 100 datapoints, not 4.
Please don't forget to bring with you a pinch of salt.
Furthermore, in statistics, a rule of thumb is more the merrier. In short if you ever get confused in situations like this, choose 100, not 25, in a manner of speaking. Of course your supervisor will wonder why his reagents don't last as long as it used to.
