This repository contains all data and code necessary to reproduce the results of Flexibly biased learning rates in social learning (Witt, Palminteri & Wu, 2025), in which we investigated learning rate biases in individual and social learning. The code is designed to be run from the top directory.

has all code used for simulating

• modelSim.py has simulation and parameter generation
  • normativeSims.py uses modelSim.py to generate data used in normative analyses -- needs to be run to reproduce those plots as the outputs are above github's file size limit

and everything related to model fitting and recovery

• modelFit.py has the functions to calculate negative log likelihoods of parameter sets given data
  • separate functions for individual and social conditions
  • pilotFits.py is used for fitting experimental data with the biased model
  • pilotFits_plain.py is used for fitting experimental data with the unbiased model
  • fitting_synthesis.py converts this script's output for further analysis

analysis scripts

• CogSci.R analysis script to generate results and figures in the CogSci paper

data from experiments, fitting data, simulations

all the plots used in the paper