What Factors Determine Digital Privacy Concerns?

This repository contains the underlying code for my analysis that I shared on medium.

Contents of this Repository

1. The jupyter notebook contains the code and explanations of my design choices.
2. Europinions.pdf is the file that you can find the survey report and the questionnaire that I used in this project.
3. plots is the directory for the plots
4. sem_report is the report that is automatically created by semopy package.
5. functions contains the helper functions that used in the analysis.

Research Questions

There are three questions that I wanted to answer in this study

1. Do personality traits impact digital privacy concerns?

2. Does trust in political institutions have any impact on digital privacy concerns?

3. Do demographics impact digital privacy concerns?

Data

Data is available GESIS archive. The access to the data is free of charge for members. You can get free membership and download the data

Methodology

I used 3 different frameworks.

1. Structural Equation Modelling
2. Multivariate Regression Modelling
3. Causal Inference via Directed Acyclic Graphs

Causal inference was my final goal, however I wanted to build the first two models for comparison purposes.

The following packages are used in this analysis:

pre-processing:

pandas

plotting

matplotlib, plotnine

modelling:

Semopy, Statsmodels, DoWhy

Results

My Provisional causal model indicate that:

Question 1:

Neither of personality traits have significant impact on the target.

Question 2:

trust in institutions have negative impact on digital privacy concerns.

Question 3:

Among demographics, Income has negative impact, wheras age has the contrary. Gender has no impact on the matter