In general, unbiased estimators are preferred for inference because:
A) bias in inference is always bad and should be avoided at all costs, so using unbiased estimators is the only way to conduct objective inference.
B) unbiased estimators always have the least amount of variance of all possible estimators.
C) a biased estimator will almost always give very little insight into the parameter being estimated and, therefore, serves as a poor guess of the parameter's true value.
D) in many cases the most probable value of an unbiased estimator is the value of the parameter being estimated; thus, unbiased estimators typically converge on values near the parameter.

Question

Quizplus · Accepted Answer

Unbiased estimators are preferred because they, on average, hit the true parameter value, making them reliable for inference over repeated samples. This does not mean they always have the least variance or that bias is always detrimental in every context, but rather that their expected value equals the parameter they estimate, which is a desirable property for making inferences.

In General, Unbiased Estimators Are Preferred for Inference Because