Data from a replication of Douven and Verbrugge (2010)
investigating the relationship between acceptability ratings of conditional
sentences and conditional probabilities (variable rating) across
three different types of conditionals (inductive, abductive, deductive).
Format
A data frame with 4,590 rows and 5 variables:
- submission_id
A unique identifier for each participant.
- item_nr
An identifier for the context-statement pair presented to the participant.
- rating
The kind of rating the participant has given.
- cond_type
The type of conditional statement.
- response
The measured acceptability rating on a 7-point Likert scale.
References
Douven, I., Verbrugge, S. (2010). The Adams Family. Cognition, 117, 302-18. https://doi.org/10.1016/j.cognition.2010.08.015.
Examples
if (FALSE) {
# fit an ordinal model using 'brms'
# regress response as a function of rating and conditional type (with interaction)
fit <- brms::brm(response ~ rating * cond_type,
data = AT,
family = cumulative("logit"))
# check encoding scheme used in the model
get_cell_definitions(fit)
# get draws for acceptability averaged over the type of conditional
extract_cell_draws(fit, rating == "acceptability")
# compare main effects of rating
compare_groups(fit, rating == "acceptability", rating == "cond_probability")
}