fit_clmm: Fit Cumulative Link Mixed-Effects Model to Simulated Ratings

Description

Fit Cumulative Link Mixed-Effects Model to Simulated Ratings

Usage

fit_clmm(.data, main_effect = FALSE)

Arguments

.data

Data frame, with the format as resulting from a call to gen_data.

main_effect

Whether to test the main effect of repetition (TRUE) or the repetition-by-interval interaction (FALSE; the default).

Value

A vector, with the following elements.

R: Fixed-effects estimate of the main effect of repetition.
I1: Fixed-effects estimate of the main effect of interval (1).
I2: Fixed-effects estimate of the main effect of interval (2).
I3: Fixed-effects estimate of the main effect of interval (3).
R:I1: Fixed-effects estimate of the interaction (1).
R:I2: Fixed-effects estimate of the interaction (2).
R:I3: Fixed-effects estimate of the interaction (3).
dev1: Deviance for the model including the effect(s) of interest.
dev2: Deviance for the model excluding the effect(s) of interest.
chisq_RI: Chi-square value for the likelihood ratio test.
p_RI: Associated p-value.
thresh.1|2: First cut-point (threshold).
thresh.2|3: Second cut-point.
thresh.3|4: Third cut-point.
thresh.4|5: Fourth cut-point.
thresh.5|6: Fifth cut-point.
thresh.6|7: Sixth cut-point.

Details

Fits a cumulative link mixed-effects model to the data and tests the specified effect (interaction or main effect) using a likelihood-ratio test using ordinal::clmm(). The function's main purpose is to be used in power simulation.

If the interaction is to be tested, the following two models are compared:

trating ~ R * (I1 + I2 + I3) + (1 + R:I1 + R:I2 + R:I3 | subj_id) + (1 + R:I1 + R:I2 + R:I3 | stim_id)

trating ~ R + I1 + I2 + I3 + (1 + R:I1 + R:I2 + R:I3 | subj_id) + (1 + R:I1 + R:I2 + R:I3 | stim_id).

If the main effect is to be tested, then the following two models are compared.

trating ~ R * (I1 + I2 + I3) + (1 + R | subj_id) + (1 + R | stim_id)

trating ~ I1 + I2 + I3 + R:I1 + R:I2 + R:I3) + (1 + R | subj_id) + (1 + R | stim_id).

Examples

Run this code

# NOT RUN {
set.seed(62)
dat <- gen_data(24) # test main effect
# }
# NOT RUN {
fit_clmm(dat, TRUE) # takes a few minutes
# }
# NOT RUN {
# }