First, let me clarify what I mean, since the terminology is inconsistent.
By 'replicate', I'm referring to measurements on an experimental unit subject to the same conditions, within a single experiment. Some sub-types:
- true replicates, often called biological replicates in my field, which are independent experimental units.
- pseudoreplicates, which are replicates that are not statistically independent because of the way the treatment has been applied; there are correlated errors that need to be accounted for.
- technical replicates, a subset of pseudoreplicates, which are typically multiple measurements made at the same time on a single sample, to account for measurement error.
What I am not referring to is repeating an entire experiment at a subsequent point in time (as in 'the replication crisis').
[Note: I'd be interested in knowing other terms that are used for the above concepts]
Till recently, we only had replicability (solely about the replication crisis meaning) and pseudoreplication, but none that I could find relating to the other senses I mentioned earlier.
I created the replication and technical-replicate tags when answering a question last week but realised that the former is going to be confused with the replicability tag and the latter is perhaps narrower than needed.
I propose replacing replication with experimental-replicate (which I just created) and making the former a synonym of the latter. Does this seem clear to you, or is 'experimental replicate' going to be interpreted as repeating a whole experiment?. It may be worth retaining the distinct pseudoreplication tag but we could either ignore the technical-replicate tag and let it get deleted (presently zero tagged questions, no loss) or or make it a synonym of either.
Given how inconsistent the terminology seems to be, there may well be better ways of framing this or other related terms that I'm unaware of. So I'd like to throw this open for discussion.