5

In Section 9.2.3 of Casella's Statistical Inference, they base their confidence interval construction for a parameter $\theta$ on a real-valued statistic $T$ with cdf $F_T(t| \theta)$.

  1. They first assume that $T$ is a continuous random variable. Then by the Probability Integral Transformation, the continuous random variable $F (T | \theta)$ is uniformly distributed over $(0, 1)$, and therefore is a pivot which can be used to construct a confidence region. enter image description here

  2. They say the situation where $T$ is discrete is similar but has a few additional technical details to consider.

    enter image description here

    My question is when $T$ has a discrete distribution. In such case, $F (T | \theta)$ takes countably many values and is not uniformly distributed over $(0, 1)$. Does the distribution of $F (T | \theta)$ depend on $\theta$, and therefore it is not a pivot? In such case, why can Casella still think it as a pivot, or am I missing something?

Thanks and regards!

Community
  • 1
Tim
  • 1
  • 29
  • 102
  • 189

1 Answers1

2

It is not a pivot. However the resulting interval is guaranteed to have at least the stated coverage. For the discrete distribution the coverage could be a lot higher than stated.

Glen_b
  • 257,508
  • 32
  • 553
  • 939
user62243
  • 21
  • 2
  • 2
    I'm sure you did not mean to use the word "diapered" ([look it up](http://dictionary.reference.com/browse/diaper)). What do you mean? – whuber Dec 07 '14 at 02:38