E-Prime Challenge 2019 Winner!
Experiment Author: Roland Pfister, University of Wuerzburg
1 | Research question
This experiment compares rule-violation behavior between positively formulated rules of the type "When X then do Y" and negatively formulated rules of the type "When X then do not do Y". Because negations are notoriously difficult to represent – try not to think of a white bear during the next five minutes! – we expect participants to violate rules more readily if they are formulated negatively rather than positively.
2 | Design
Participants navigate through a 2D city maze with a virtual bicycle courier using the arrow keys of the computer keyboard.
Some lanes are explicitly forbidden while participants are free to use others. Each trial contains one lane that participants are allowed to take and one lane that they are not allowed to take. Taking a forbidden lane can speed up the delivery at times and the program tracks when participants choose to commit such a rule violation. The factor Rule Formulation (positive vs. negative) is implemented as a between-subjects manipulation and it is tied to the subject number (see the attribute zRuleType in the List CounterbalanceList).
3| Dependent variables and analyses
The main dependent variable is the choice to violate a rule or not (attribute “Violation”). This choice is modeled as a function of the participant’s rule condition (attribute “zRuleType”) the gain of the current trial (i.e., the number of tiles saved by choosing the forbidden lane; attribute “Gain”).
In-depth analyses of performance data (inter-keystroke intervals) are available via custom logfiles as described in section “5 | Features”. Variable coding:
- Violation: 0 = rule-conform trial, 1 = rule violation
- zRuleType: 1 = positively formulated rule, 2 = negatively formulated rule
- Gain: Potential gain in tiles when using the forbidden lane (negative numbers indicate that the forbidden path is actually longer than the allowed one, positive numbers indicate a shortcut)
4| Settings and requirements
The program is run at a resolution of 1024 x 768. Other resolutions are also supported (e.g., when enabling the "Match Desktop Resolution" feature). If using other resolutions than 1024 x 768 it might be useful to tweak the visual appearance of the instruction screens a little. No external input devices other than a standard keyboard are required.
For demonstration purposes, the experiment has been cut down to two blocks of 5 trials each.
- Maps are generated randomly on the fly! At the same time, the program is written in a way that it can be easily amended for users who would like to run the program with fixed maps that are read in from the hard drive.
- Instructions are repeated if requested by the participant.
- Press “p” during the trial to take a screenshot! The image is saved as “screenAll.bmp” in the experiment folder.
- Custom logfiles: In addition to the usual .edat3 logfile (which can be used to run the main analyses), the program writes a number of additional logfiles to simplify the analyses. (1) “ZodiacZity_Raw.dat” features the main variables/attributes in a compact format and includes data of all subjects run on a given computer. (2) “ZodiacZity_S[SubjectNumber]_Park.dat” is a copy of the compact file for each individual participant. (3) “ZodiacZity_Sample.dat” contains sample characteristics at a glance. (4) “MData[...].dat” files show detailed statistics for individual keypresses within a trial. Main measures derived from this keystroke-specific data are written online to the first two files described above. (5) Each map of each subject is logged in the “maps” subfolder.