Propensity Score Matching: An Application on Observational Data

The study aimed to determine the survival rate of first-class passengers using the Titanic dataset from Kaggle. Descriptive statistics revealed that first class passengers had way more chance to survive as compared to other classes, which underscores the role of socioeconomic status in determining chances of survival. Evaluation metrics, which assess model performance independently for male and female cohorts, shed light on gender specific projected accuracy. The analysis of propensity scores matching data for male and female passengers separately ensured that each gender category had control groups and treatments that were equally distributed. It was discovered that women had higher survival rates compared to men and these findings also identified disparities in the levels of surviving among genders. Improvements in covariate balance were indicated by post-matching statistics for both the male and female cohorts, indicating that the matching process was successful for both genders. The treatment effect estimates for male and female passengers were computed independently, and the findings showed that a number of characteristics significantly improved the survival rates for each gender group. The overall results of the study emphasized how important it is to include gender when analyzing survival outcomes using the Titanic dataset. In addition, age was suggested as an important factor whereby young people had higher chances of being saved.