Survey Data and Analysis

This repository contains the survey data, and its analysis for our developer survey to learn about principles of ethical AI.

• Excel Files Note: all excel files that contain survey data we can index by column contain response distributions on the bottom row for each column.

---

Survey Data (/data/)

Survey Responses

• Files:
  • survey_results.xlsx
  • survey_results.pdf
  • ai_study_finalized.csv
• Description:
  These files contain the full accepted responses from the survey.

Quantitative Analysis By Question

• Files:
  • quantitative_analysis.docx
    Figures and textual breakdown of response distribution for multiple choice (quantiative) questions.
  • quantitative_results.xlsx
    Full response distribution for all quantitative questions.

Figures (/data/figures/)

• Description:
  Contains figures generated during analysis, including those for raw data analysis and familiarity analysis.
  Graphs produced from the ethical AI familiarity analysis are saved in data/figures/familiarity_analysis.

Annotations

Annotation Files (/data/annotations/)

• Formats:
  • JSON: Annotator JSON Books (/data/annotations/json/)
  • Excel: Annotations in Excel (/data/annotations/xcel/)
• Description:
  Annotations provide responses with themes and codes, aside from those questions about risk and mitigation methods,
  • Themes refer to commonalities in the text which may be mentioned in the paper.
  • Codes are used to highlight interesting responses.

---

Notebooks (/notebooks/)

Correlation Tests

• File: /notebooks/correlation_tests.ipynb
• Description:
  Contains correlation tests on the survey data.

Familiarity Analysis Notebook

• File: /notebooks/familiarity_analysis.ipynb
• Description:
  Analyzes data related to familiarity with the principles of ethical AI and current regulatory initiatives.
  Relevant graphs are saved to data/figures/familiarity_analysis.

Quantitative Analysis Notebook

• File: /notebooks/quantitative.ipynb
• Description:
  Builds graphs and generates a DOCX file summarizing all quantitative questions, and the demographics table.

Dataset Exploration

• File: data_view.ipynb
• Description:
  F(x)'s to search dataset, i.e., ranking responses by demographics.

---