Statistical Intervals, 2nd Edition

Epilogue

The concept of this book was born about 40 years ago. After agreeing to go for it with Wiley, our publishers, it took two of us 12 years to put together the first edition, published in 1991. After a 14-year hiatus, it took another 12 years for the three of us to get this second edition into print. Why so long? Admittedly, our day jobs had to be our first priority. But another reason is that there is so much to say and limited space to say it precisely and understandably. Statistical intervals, in fact, illustrate the proverbial “bottomless pit” with seemingly endless opportunities for adding to the exposition. We have tried to provide both practitioners and researchers the specific tools they need to quantify uncertainty in their data via statistical intervals for commonly occurring situations and to provide added guidance in constructing such intervals in more complex and less frequently occurring situations. Thus, we will conclude by reemphasizing some of our major ideas, and adding a few anecdotes about:

• The importance of calculating the “right” statistical interval.
• The role of statistical intervals versus other forms of inference.
• The limitations of statistical inference.

Statistical Intervals: Vive La Différence!

We have presented a wide variety of statistical intervals, and tried to explain the situations for which each is appropriate, how each is calculated, and the underlying assumptions. Some of these—such as confidence intervals for the population or ...