Tn=n+1x+1
An estimator Tn of parameter θ is consistent if it converges in probability:
n→∞limPr(∣Tn−θ∣>ε)=0
for all ε>0
For a binomial population: θ=x/n - the proportion of successes.
Then:
n→∞lim(Tn−θ)=n→∞lim(n+1x+1−nx)=0
If
(n+1x+1−nx)=0
then:
Pr(∣n+1x+1−nx∣>ε)=0
because ε>0 .
So:
n→∞limPr(∣n+1x+1−nx∣>ε)=0
Comments