Data and Analysis Unit: PLYM5
Last update: 2014-12-04 by Andy Wills
Description
Category learning task, reported in Experiment 4A of Wills, Inkster, and Milton (submitted).
This DAU contains stimulus files, raw data, and an analysis script, all in open
cross-platform formats (see the file format notes).
Citation
The APA-format citation for this resource is:
Wills, A.J. (2014). Data and Analysis Unit: PLYM5. Retrieved from http://www.willslab.co.uk/plym5
If you make use of these resources, please drop me an email: andy@willslab.co.uk.
Resources
- plym5stim.tbz (218 KB) - Stimulus image files
- plym5instruct.txt (3 KB) - Instructions presented to participants.
- plym5data.txt (498 KB) - Trial-level raw data.
Column headings are as follows:
- subj: Participant ID (unique within PLYM5)
- cond: Experimental condition (1 = intentional, 2 = incidental)
- cb: Stimulus randomization condition (1-8, see plym5cb.txt)
- keyb: Response key randomization condition (1 = Policeman on left, 2 = Doctor on left )
- phase: Experiment phase (1 = training, 2 = test)
- blk: Block (Ranges from 1 to 3 in phase 1, and 1 to 8 in phase 2)
- trial: Trial number (resets for each block)
- stim: Stimulus ID (see plym5code.txt)
- resp: Phase-specific response code
- Training phase, Intentional condition: Accuracy ( 1 = right, 0 = wrong )
- Training phase, Incidental condition: Choice ( 1 = seen before, 0 = not seen )
- Test phase (either condition): Choice ( 1 = the 1111 category, 0 = the 0000 category )
- rt: Reaction time in milliseconds
- conf: Confidence rating (1-3: 1 = 'not confident', 2 = 'neither confident nor unconfident',
3 = 'confident'). Only used in the training phase of the incidental condition. Set to zero otherwise.
- confrt: Reaction time in milliseconds for the confidence rating.
Only used in the training phase of the incidental condition. Set to zero otherwise.
- plym5analysis.R (1 KB) - R script for analyses
reported in Wills, Inkster, Milton (submitted). This script can be run by the command
source('plym5analysis.R'), or can be run interactively. Requires plym5cap.R
- plym5cap.R (6 KB) - R script containing a single function that
performs all analyses required in PLYM5. This allows variant analyses (e.g. with
or without participant exclusions) to be run without code duplication. This is
NOT a portable function, it makes use of global variables and will only work unmodified in the context of these
specific experiments.
- plym5logistic.R (3 KB) - R script for logistic regression analyses
reported in the Supplementary Materials of Wills, Inkster, Milton (submitted). This script can be run by the command
source('plym5analysis.R'), or can be run interactively.
- plym5code.txt (1 KB) - The main purpose of this file is to facilitate
automated analysis of plym5data.txt by plym5analysis.R. Column headings
are as follows:
- stim.number: Stimulus ID (as used in plym5data.txt)
- stim.type: Logical stimulus code. The four digits are the values of the four stimulus dimensions
(left to right, D1-D4). The final 's' has no meaning, it's there to ensure the four digits are parsed by R scripts as
a string rather than as a number. In all stimulus randomization conditions,
the stim.type code (and thus also the stimulus ID) represents the stimulus
in canonical form (i.e. with D1 as criterial, D2-D4 as noncriterial, and with 1111 and 0000 as
the category prototypes). Thus stimulus ID 1-4 are category 1111 training stimuli, ID 5-8 are
category 0000 training stimuli, ID 9-10 are the 'critical' test stimuli, and ID 11-16 are the other
test stimuli.
- fr: The response predicted by an Overall Similarity response model (1 = category 1111,
0 = category 0000, 0.5 = no prediction).
- ca: The response predicted by a Critierial Attribute response model.
- nca1: The response predicted by a Noncriterial Attribute response model acting on D2
- nca2: The response predicted by a Noncriterial Attribute response model acting on D3
- nca3: The response predicted by a Noncriterial Attribute response model acting on D4
- lk: The response predicted by an 'alway press the left key' response model.
- rk: The response predicted by an 'alway press the right key' response model.
- plym5cb.txt (1 KB) - Specification of stimulus randomization
conditions. Each column represents one of the eight conditions (cb in plym5data.txt). Each
row represents a stimulus ID (stim in plym5data.txt). Each cell gives the filename of the
physical stimulus presented for a particular cb condition and stimulus ID. For example,
1010 indicates stimulus file 1010.png in plym5stim.tbz.
Note: The column
titles (e.g. "cb5:R2D34") give the value of cb (5 in this example) and then describe the operation
required to get from the canonical structure (cb = 1) to this structure. "R" means rotate. Specifically,
R1 means rotate left once i.e. all four digits in the stimulus code move one place to the left, with the leftmost digit
becoming the rightmost digit. R2 means two left-rotate operations. After the rotations one or
more of the dimensions are flipped (all 1s become 0s and all 0s become 1s). For example, D3D4 means
flip D3 and D4, after rotation.