On May 27, 2025, Westminster International University in Tashkent (WIUT) hosted a research colloquium featuring Zarrukh Rakhimov (PhD candidate in "Econometrics and statistics"), who presented his research study titled "Bootstrap Confidence Intervals in Linear Models."

The presentation focused on the reliability of bootstrap confidence intervals in linear regression analysis, especially under conditions where classical assumptions—such as normality, homoscedasticity, or complete data—may be violated. Using Monte Carlo simulations, Rakhimov compared traditional OLS-based intervals with various bootstrap methods across multiple challenging data environments, including small samples, outliers, and missing data not at random (MNAR).

Particular attention was given to the “bootstrap pairs” technique, which showed strong performance in delivering consistent interval coverage when traditional assumptions did not hold. The findings suggest that bootstrap methods can serve as a valuable alternative or complement to standard techniques, especially in empirical research involving imperfect or non-ideal data.

The session generated keen interest among faculty, researchers and students, providing practical insights for improving statistical inference in applied research.