Materials, data, and analysis code for the paper "Maze made easy: Better and easier measurement of incremental processing difficulty".

Note: This study is a replication and comparision with Witzel et al. 2012. The experimental sentences, half the comprehension questions, and the G- and L-maze distractors all come from Witzel et al. 2012. Additionally, all data/conditions described as "lab" come from their experiment.

Materials:

This folder has source for running the experiment in ibex. Copy header, material, and footer sections together into a .js file to go in data_includes part of an ibex experiment. (This uses the Maze and Maze_separator modules described in ibex_maze)

• header.js: top part of file for Maze tasks (comment out G-maze part and uncomment L-maze part to show L-maze instructions)
• footer.js: bottom part of file for Maze tasks
• header_spr.js: top part of file for SPR
• footer_spr.js: bottom part of file for SPR
• g_maze.js: materials for G-maze
• l_maze.js: materials for L-maze
• spr.js: materials for SPR
• gulordava.js: materials for Gulordava A-maze
• one_b.js: materials for Jocefowicz A-maze

Labelled:

This folder contains the sentences labelled by critical word location, used to add labels to results during analysis.

Data:

This folder contains our data files.

• Raw: data files produced from ibex (tsv format, lines starting with # are comments, see ibex documentation for more details; README in ibex_maze may be useful for maze specific results)
• Processed: .rds files produced by read_Results.R from raw data; data is tidy and rectangular.

Witzel:

This folder contains data from Witzel et al. 2012.

• .txt data files are their raw data files
• .txt materials files are used to associate data with materials when parsing
• .rds files are processed versions made by parse_witzel_data.R

Code:

This folder contains R code for parsing results into tidy data format and R code for all analyses and graphs in the paper.

• read_Results.R: takes raw files, tidies the data into .rds files (for our data)
• parse_witzel_data.R: takes raw data, tidies the data into .rds files (for Witzel data)
• Results.Rmd: runs regressions to determine estimated effect sizes, graph of effect sizes (note: models take time to run)
• errors.Rmd: makes graphs of participants errors, can be used to view by item error rates
• power.Rmd: does power analysis (note: models take a lot of time to run)