📊 Workplace Accidents in Poland (2021) - Econometric Analysis

Project Overview

This project examines the factors influencing the number of people injured in workplace accidents across subregions in Poland in 2021. Using econometric modeling, we explore how various socio-economic and infrastructural variables impact workplace accident rates.

Research Hypotheses

1. Higher registered unemployment rates lead to fewer workplace accident victims.
2. An increase in the number of boiler houses significantly contributes to a higher number of workplace accidents.
3. A higher number of marriages correlates with an increased number of workplace accident victims.
4. Higher investment expenditures per capita in enterprises result in a greater number of workplace accidents.

Findings & Conclusions

• The model meets all estimation method assumptions but has a high forecasting error, making it unsuitable for accurate predictions.
• Hypothesis 1: ❌ False – Unemployment rate was not statistically significant in the model.
• Hypothesis 2: ❌ False – The number of boiler houses was not a significant variable.
• Hypothesis 3: ✅ True – A higher number of marriages led to a statistically significant increase in workplace accidents (+0.41 victims per unit increase).
• Hypothesis 4: ✅ True – Higher investment expenditures per capita were associated with more workplace accidents (+0.08 victims per PLN increase).

Tools & Methods

• Programming Languages: Python, Gretl
• Statistical Methods:
  • Descriptive statistics
  • Stepwise regression
  • Hellwig’s method
  • Model diagnostics (heteroskedasticity, multicollinearity, residual normality tests)
  • Prediction with confidence intervals