Skip to content

lottybrand/Dallinger_Analysis

Repository files navigation

Dallinger_Analysis

These scripts are for analysing the data of an experiment on prestige-biased social learning, run via the "Dallinger" experimental software. The experimental scripts can be found at www.github.com/lottybrand/lottysBartlett The preprint for the corresponding paper can be found here https://psyarxiv.com/mn9t6/

The file "data_info" includes an explanation of all of the variables in the data files and the full_data file, and where they come from.

The file "data_inputting.R" converts the jSon scripts from all the datafiles in the raw_data_files folder into one large database

The file "dallinger_data_cleaning.R" cleans this file and creates new variables, subsets and dataframes to be used in the analysis scripts - culminating in full_data.csv

The file "analysis_script.R" includes all the analysis models and their corresponding predictions, this can be reproduced if dallinger_data_cleaning.R is run first (full_data.csv is the saved dataframe from line 113 of dallinger_data_cleaning.R) some plots are also produced in this file.

The folder "results" contains a summary of the main results, some descriptives in "results_prelim" and some detailed results from model4.2, this was during confusion over the use of a_bar in these models (see https://twitter.com/LottyBrand/status/1195482057931153408)

The folder "plots" contains many exploratory plots and some plotting script, although some plots are created in the analysis_script.R

The file "score_check.R" includes code that double-checked our score and copying variables were calculated correctly

The folder "Follow_up_experiment" contains code that explores the within and between topic scores and behaviour, in preparation for our second experiment

The folder "reviewer_responding" includes scripts that helped deal with some reviewer comments

About

data cleaning and analysis scripts for upcoming Dallinger experiment

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published