Experiment investigating the effect of training order on
category learning performance reported in
Edmunds, Wills, Milton (2017), Experiment 1.
This DAU contains stimulus image files, raw data, and analysis
scripts, all in open cross-platform formats (see the
file format notes).
PLY22longData.csv (845
KB) - Trial-level raw data. This is the file against which
the md5 checksum reported in the journal article is
calculated. Column headings are as follows:
DAU: Experiment name
Participant: Participant ID (unique across entire
experiment)
Condition: Experimental condition (either EasyMed or
HardMed)
Block: Block number
SubBlock: Values from 1 to 5 that change every 30
trials to aid with learning curve graphs
Trial: Trial number
StimulusNo: Unique ID for each stimulus (number
corresponds to numbered PNG files in PLY22stimuli.zip)
Frequency: The value of each stimulus on the spatial
frequency dimension (in arbitrary units)
Orientation: The value of each stimulus on the
orientation dimension (in arbitrary units)
Category: The category assignment for each stimulus
(Category A=0 and Category B=1)
Difficulty: The difficulty of that stimulus (Easy,
Med, Hard)
Response: The response the participant gave (Category
A=68, Category B=75; numbers refer to ASCII code of
response letters on keyboard i.e. 68 = key 'D', 75 = key
'K')
Correct: Whether the participant was correct or not
(0 = incorrect, 1 = correct)
RT: Reaction time in seconds
PLY22verbalReports.pdf
(1.7 MB) - Scanned copies of the strategy reports produced by
each participant. Prefix participant numbers with '22' to
match up to the other files in this repository
(e.g. participant 103 in this file is participant 22103 in the
other files).
PLY22verbalReports.csv
(1 KB) - Participant level strategy information from verbal
reports given by participants. Column headings are as follows:
CERE: Classification of participant's strategy by
Charlotte E.R. Edmunds.
GW: Classification of participant's strategy by Gemma
Williams
Agreed: Classification of participant's strategy
agreed by CERE and GW after discussion. For this and the
previous two columns the codes are:
2D = Generic two-dimensional rule
CJ=Conjunction rule
CJ2 = A double conjunction
NA = No relevant strategy information
RuleX = Rule plus exception
SUR = Sequential unidimensional rules
Udexp = Unidimensional rule plus exception
UDF = Unidimensional rule based on frequency
UDO = Unidimensional rule based on
orientation
CEREyes: Whether CERE thought the participant reported
a strategy or not (1 = yes, 0 = no)
GWyes: Whether GW thought participants reported a
strategy or not (1 = yes, 0 = no)
AgreedYes: Presence of straetgy agreed by CERE and GW
after discussion (1 = yes, 0 - no)
simpStrat: Simplified strategies. All complex
rule-based strategies are given the label "Complex", and
both unidimensional strategies are defined as "UD"
R scripts for analysis of PLY22. You will
need R, plus the
following packages that are available on
CRAN:
effsize, ggplot2, irr, rmisc
PLY22 Functions.R (6
KB) - functions required by the other scripts, below.