Rapid Online Assessment of Reading ability (ROAR)

Yeatman, Jason D., Kenny An Tang, Patrick M. Donnelly, Maya Yablonski, Mahalakshmi Ramamurthy, Iliana I. Karipidis, Sendy Caffarra, et al. 2021. “Rapid Online Assessment of Reading Ability.” Scientific Reports 11 (1): 1–11.

Summary

This repo contains the Javascript code for the ROAR and analysis code to reproduce all the figures in the manuscript. All the data reported in the manuscript is included in this repo.

Repo Organization

• data_allsubs contains data organized in .csv files that has been concatenated across subjects and organized as tidy data tables.
  • LDT_alldata_long.csv All the data in long format. Each row is a trial of the LDT task. Raction time (rt), Accuracy (acc), stimulus (word), word length (worrdLength), whether it's a real/pseudoword (realpseuudo), and subject ID (subj) are coded for each trial.
  • LDT_alldata_wide.csv Accuracy data in wide format. Each row is a subject. Each column is a stimulus (word). Each entry is correct/incorrect for that stimulus/subject. Subject ID (subj) is coded in the first column. This data is also displayed in figs/WordMatrix.eps
  • LDT_alldata_wide_sorted.csv Same data as above but the rows are order based on each subject's percent correct and the columns are ordered based on each columns percent correct. See figs/WordMatrix.eps
  • metadata_all.csv Subject metadata including ages and reading scores. This needs to be filtered based on subject ID to match rows to the LDT data.
  • wordStatistics.csv Statistics on each stimulus such as word frequency, bigram and trigram statistics, etc. For definitions of word stats see: https://www.neuro.mcw.edu/mcword/definitions.html#UN1
• analysis Directory of analysis code
  • LDT_Run_All.sh This shell script runs the whole analysis pipeline. In a shell type "source LDT_Run_All.sh" and it will run each step of the analysis pipeline and make figures.
  • Figure1_WordListProperties.R R script that plots statistics about the stimuli and does some other data organization and analysis of response time distributions. The Figures generated by the script were moved to the suppliment.
  • Figure2_PercentCorrect.R Generates the figure relating percent correct on the ROAR to the Woodcock Johnson
  • Figure3_LDT_IRT_Models.R Performs the IRT analysis to choose stimuli that went into the final version of the ROAR
  • Figure4_LDT_AnalyzeWordStatistics.R Creates Figure 3 analyzing properties of individual lists
• Study2 Directory containing data and analysis code for the Study 2 validation. This was the validation of the optimized version of the ROAR going down to first grade
  • LDT_Study2_IRT.R Analysis of Study 2 validation data and generation of figure 4.
  • data Data from study 2