Estimation Criteria

This lecture discusses the criteria for estimating a parameter with an estimator. It covers two types of criteria: asymptotic behavior as n approaches infinity and finite-sample comparison for a fixed n. Consistency of an estimator is essential but not sufficient for being a good estimator. The lecture also explores bias, variance, and efficiency of estimators, emphasizing the importance of having low bias and low variability. Additionally, it delves into the concept of confidence intervals and the interpretation of confidence levels. The lecture concludes with a discussion on the construction of approximate confidence intervals based on estimators and the calculation of standard errors.