My linguistics rating experiment contains 36 stimuli in total for rating. There are 6 themes (e.g. doctor, farm, etc), and each theme has 3 conditions (good, bad, mixed), and each condition has 2 sentences. So there is a total of 3x2x6 = 36 sentences. I had 54 participants, each participant rates one sentence from each theme, with each rating 6 sentences in total.
My goal is to test if the 3 conditions significantly produce different ratings (e.g. condition good produces better ratings than condition bad). I am trying to interpret the data results using the brms package in r, taking participants and themes into consideration (in the case where these two factors contribute to ratings as well).
Currently, my model looks like this:
fit_1 <- brm(
rating ~ 1 + condition + (1|theme) + (1|participant),
data = stimuli_rating, family = cumulative()
)
Is this how the model is supposed to be?
fit_2 <- brm(
rating ~ 1 + condition*theme + (1|participant),
data = stimuli_rating, family = cumulative()
)
What about this?
Thank you!