= Solution
{c}
Since the true mean is known to be zero, the <chi-squared distribution> gives $Q/\theta\sim\chi_n^2$. Let $q_\beta$ be the $\beta$ quantile of this law. The probability
$$
\mathbb P\left(q_{\alpha/2}\leq Q/\theta\leq q_{1-\alpha/2}\right)=1-\alpha
$$
rearranges, using positivity, to the exact <confidence interval>
$$
\boxed{\left[\frac{\sum_iX_i^2}{q_{1-\alpha/2}},\ \frac{\sum_iX_i^2}{q_{\alpha/2}}\right].}
$$
There are $n$ degrees of freedom, because no unknown mean was fitted.
Back to article page
