Script and data for assessing interdisciplinarity in cognitive science

Data

The data folder

All the data used to produce the results is contained in this repository. However, because of copyright restrictions, we only included the processed data that results from the r_scripts/data_processing.R. This processed objects can be found in the saved_objects folder in .rds format. For raw data questions, please contact the authors at pc684 at cornell dot edu.

The theories_data folder

This folder includes the data and scripts for the LSA theory analysis part of the paper. The .py scripts take a catalog and transform it into a lemmatized document by term matrix. The scripts have to be run in the following order:

1. clean_database.py
2. lemmatizer_stemmer.py
3. document_by_term_generator.py

To produce the data for the study, all 3 scripts must be run on new_abstracts.csv and catalog.scientometrics.csv. The output of these scripts generate clean_database.csv (Scientometrics paper data) and clean_all_abstracts.csv (this paper's new data) through scripts 1 and 2, and dbt_matrix_scientometrics.csv (Scientometrics data) and all_dbt.csv (New data, compressed into all_dbt.zip). all_dbt_reduced.zip contains a compressed version of the reduced document by term matrix produced in the theories.R script.

R scripts

The r_scripts folder contains the main analysis scripts for the paper. The saved_objects folder contains .Rds files that save the intermediate objects generated in these scripts to save time. data_processing.R takes the files in the data folder and produces a usable version for the rest of the analysis. This script takes a very long time to run as each author file has to be incorporated. The output of this script is saved in journalData.Rds and authorByJournal.rds. JSD.R calculates the Jensen-Shannon divergence and Gini Index for all papers in the journals data folder. The output of this script is contained in complete_frame.rds. theories.R builds semantic space with each theory included in the Scientometrics paper, and applies them to the journal papers in Cognitive Science journals through a random forest. The predictions and scores for each theory in each article of the CogSci journals are saved in cogsci_theories.rds. main_script.R takes the objects produced by the other scripts and generates the figures and models reported in the paper. Optionally, it can be commented/uncommented to run all scripts and not rely on the saved_objects folder. functions.R contains most of the general purpose functions used in the rest of the scripts.

Figures

• Folder author_figures includes the theory distribution for each author who (co-)authored a paper in the three cognitive science journals being considered.
• Folder main_figures includes the figures used in the main text of the paper.
• Folder supp_figures includes the figures used as supplementary material in the paper.