kb07: Sample visual world eyetracking data

In gmpm: Generalized Multilevel Permutation Models

Description

Sample data from a visual-world eyetracking study included to demonstrate use of gmpm functions on multilevel data with difficult dependencies. Data are from a subset of Experiment 2 in Kronmuller and Barr (2007). A parametric reanalysis of these data can also be found in Barr (2008).

Usage

kb07

Format

A data frame containing 2464 rows, each one an aggregate over multiple trials within the same condition for the same subject.

[,1]	SubjID	factor	experimental subject
[,2]	Speaker	factor	identity of the speaker (same,diff)
[,3]	Load	factor	was listener under cognitive load?
[,4]	AggID	numeric	index of unique speaker*load*subject combinations
[,5]	ms	factor	time in ms from description onset
[,6]	t1	numeric	time in 100 ms increments
[,7]	T	numeric	count of looks to target
[,8]	O	numeric	count of looks to non-target object
[,9]	X	numeric	count of frames including looks to blank regions or blinks
[,10]	N	numeric	total number for frames (T+O+X)
[,11]	NT	numeric	total non-target frames (N-T)

Author(s)

Dale J. Barr <[e-mail]>

References

Kronmuller, E. and Barr, D. J. (2007). Perspective-free pragmatics: Broken precedents and the recovery-from-preemption hypothesis. Journal of Memory and Language, 56, 436–455.

Barr, D. J. (2008). Analyzing 'visual world' eyetracking data using multilevel logistic regression. Journal of Memory and Language, 59, 457–474.

See Also

gmpm

Examples

data(kb07)

# first let's fit the data using a binomial model
#
kb07.binom.gmpm <- gmpmCreate(cbind(T,NT) ~ t1*Speaker*Load | SubjID,
                       "binomial", kb07, ivars=c("Speaker","Load"))

# you will need to increase the number of runs in the command below
# (to, say, 999) to get sensible results

kb07.binom.gmpm <- gmpmEstimate(kb07.binom.gmpm, list(maxruns=19))

summary(kb07.binom.gmpm)

# now let's do a more powerful multinomial analysis
# where we break out looks to blank regions and blinks
# into a separate "junk" category (X)

kb07.mnom.gmpm <- gmpmCreate(cbind(O,T,X) ~ t1*Speaker*Load | SubjID,
                     "multinomial", kb07, ivars=c("Speaker","Load"))

# again, you'll need to increase the number of runs to get something
# sensible
kb07.mnom.gmpm <- gmpmEstimate(kb07.mnom.gmpm, list(maxruns=19))

summary(kb07.mnom.gmpm)

gmpm documentation built on May 2, 2019, 5:27 p.m.