Summary

This repository contains code written by Farhad Mohsin and Inwon Kang for the moral dilemma project

The directories are as follows:

1_airplane_scenario/2_plant_scenario:

• html files used for survey rendering each scenario to the mturk participant.
• Contains multiple versions, ultimate.html being the final version

All the python files should be ran in python3

Scripts

classes.json

contains seed data to combine and create random stories from.

various storygen.py

uses data in classes.json to create random stories of different types (purely random, focused on one attribute, etc) to randomly generate features for the survey questions

LP_.. files

Deals with constructing and testing the Lexicographic Preference model

Data

All the data is stored under the data folder. Files containing sensitive information such as worker id have been removed, and can be accessed in a private folder. They are divided by each round of batch that we sent them out on, and the cumulative data is in the one outside of the rounds folders.

3_option.csv

• Contains the survey results where participants are asked to rank from 3 possible candidates to give the life jacket to.
• Question ID column denotes each unique questions.
• [AGE, GENDER, PURPOSE OF TRIP, CAREER, HEALTH, SURVIVAL] columns denote the features of each candidates that are being picked.
• Survival denotes the survival chances of each participants with/without the life jacket.
• FOCUS denotes the aspect the survey focused on. It can be blank (random), survival, purpose of trip and career.
• [Age Group, Gender, Education] columns denote the demographic information of the participants.
• Written Response contains an optional written response from the participants that explain their choices.
• [A score, B score, C score] denote the score (out of 10) assigned to each candidate by the participants.
• Ranking contains the ranking parsed based on the scores.

notes.md

Previous notes on the analysis on the data are stored here