I'm troubling over the role of number of observations in a within subjects design on power. Let's say I have a 2x2, or even a single factor, completely within subjects design. A power analysis might tell me I need a relatively small sample, say <50 (depending on the size of effect I'm looking for, etc.). However, what if I only have 10 trials, versus 100 trials? How do I add that as a parameter? Surely the number of observations per participant will impact the stability of my results.

Now, let's say I have a large number of observations per participant; each participant receives 100 trials, 50 in each condition. But what if I only care about a small number of trials - for example, items rated as "very confident" on a memory task? I may obtain very few (and potentially 0) trials per participant with the response I'm interested in examining. How do I plan for this, and what should this mean if I'm trying to figure out my appropriate sample size?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.