I have this kind of problem:
The market share $Z$ of a company is estimated in two independent survey polls. The sample sizes are $n_1=500$ and $n_2=2000$ with corresponding observed shares $p_1$ and $p_2$, respectively. These two results are to be combined. Two alternative estimators of $Z$ are presented as follows:
$Z_a=\frac{p_1+p_2}{2}$
$Z_b=\frac{p_1+4p_1}{5}$
- Are these estimators unbiased?
- Which one of these estimators is more efficient?
I know that
$E(Z)=0.2p_1+0.8p_2=E(Z_b)\rightarrow Z_b$
is unbiased and
$E(Z_a)=0.5p_1+0.5p_2$
so $Z_b$ is biased. Still I can't calculate MSE of these estimator so I don't know which one of these estimators is more efficient