semantic-similarity-stm
This repository contains scripts for review on the semantic similarity effect on short-term memory. Our review article was publised as online first.
Ishiguro, S., & Saito, S. (2020). The Detrimental Effect of Semantic Similarity in Short-Term Memory Tasks: A Meta-Regression Approach. Psychonomic Bulletin & Review doi: 10.3758/s13423-020-01815-7
Workflow
OPTIONAL Modify Summary Table of previous studies
In case you want to include/exclude studies. It should contain effect sizes and file paths for the materials.
Prepare for Materials and Norms
English affective norms Downloda 13428_2012_314_MOESM1_ESM.zip and place a file, BRM-emot-submit.csv in AffectiveNorms folder.
Free association norms Download SWOW-EN2008 assoc. strengths (R123) [8Mb] and place a file, strength.SWOW-EN.R123.csv in AssociationNorms folder.
French affective norms Download ESM 2 (XLSX 284 kb) and place 13428_2013_431_MOESM2_ESM.xlsx in AffectiveNormsFrench folder.
Materials in previous studies
Please contact the aurthor (ishiguro.sho.grocio@gmail.com). I will not distribute files of materials for copyright protection. Alternatively, manually fetch words from articles and create xlsx files.
Do data Cleansing
Remove " and ' in strength.SWOW-EN.R123.csv (the original file)
cd Scripts/
bash data_cleansing.sh
Note. Locate where you save semantic-similarity-stm folder (e.g., ~/Downloads/semantic-similarity-stm).
Create Association Matrix
Create association_matrix.csv.
python free_association_matrix_creator.py
Run Main program
Main program loads the summary table and calculate Similarity and Connectivity.
python main_cal.py
Do preprocess for analysis
cd ../StatisticalAnalysis/
python preprocess_forR.py
Analyse data
Rscript meta_analysis.R
Create tables of results with Similarity and Connectivity indexes (optional)
python table_construction.py
Note
Below, mean_dist_from_centroid calculated (dis)similarity value for an example list (i.e., 'diamond, ... ,sapphire') and connectivity_calc calculated connectivity (or association stregnth) for another example list (i.e., 'apple, banana, orange'). These examples are used in my manuscript. Note that you should run free_association_matrix_creator.py and get a cue-response matrix containing words that you target before using connectivity_calc function.
I hope that these functions facilitate psychological studies!
>>> from grocio_utils import *
>>> mean_dist_from_centroid(['diamond', 'emerald', 'opal','pearl','ruby','sapphire'])
1.0449535299761725
>>> connectivity_calc(['apple','banana','orange'])
0.03786311822026108
Dependencies
Python libraries: scipy, pandas, numpy, matplotlib, seaborn, xlrd, gensim, tqdm
R libraries: meta, metafor, dmetar, grid