Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

How can i establish that for each $$p \in [0,1] $$ if $$X_{1},X_{2}...$$ are a coin runs to parameter p, with the propability P to cover up the confidence region $$R:=\left[\overline{X}_n-\frac{\sqrt{5}}{\sqrt{n}}, \overline{X}_n+\frac{\sqrt{5}}{\sqrt{n}}\right]$$ with Chebyshev inequality: $$P((\overline{X}_n-EX) \ge c\sigma) \le \frac{1}{c^2}$$ not less than .95!

I start with

$$=> $$ $$P((\overline{X}_n-EX) \le c\sigma) \ge 1 - \frac{1}{c^2} \ge 0.95 $$

$$P(\overline{X}_n-\frac{\sqrt{5}}{\sqrt{n}}\le p \le \overline{X}_n+\frac{\sqrt{5}}{\sqrt{n}}) = 0.95$$

$$1-\frac{1}{c^2} = 0.95$$ $$c=2\sqrt{5}$$

From here I am stuck...

share|improve this question

1 Answer 1

up vote 1 down vote accepted

Let $S = \frac{1}{n}\sum_{i=1}^n X_i$. Then, $E[S] = p$, $\text{var}(S) = \frac{p(1-p)}{n}$ and so by the Chebyshev inequality $$P\biggr\{p - c\sqrt{\text{var}(S)} < S < p + c\sqrt{\text{var}(S)}\biggr\} \geq 1 - \frac{1}{c^2}$$ where the right side has value $0.95$ when $c = \sqrt{20}$. But if $S$ does lie in the interval $$\left(p-\sqrt{20\cdot\text{var}(S)}, p+\sqrt{20\cdot\text{var}(S)}\right),$$ then $p$ must be in the interval $$\left(S - \sqrt{20\cdot\text{var}(S)}, S+\sqrt{20\cdot\text{var}(S)}\right),$$ that is, this is a $95\%$ confidence interval for the unknown parameter $p$. Since $\text{var}(S)$ has maximum value $\frac{1}{4n}$ when $p = \frac{1}{2}$, we have that $$\left(S - \sqrt{20\cdot\text{var}(S)}, S+\sqrt{20\cdot\text{var}(S)}\right) \subset \left(S - \sqrt{\frac{5}{n}}, S+\sqrt{\frac{5}{n}}\right),$$ and this superset is also a $95\%$ confidence interval for $p$.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.