How to estimate Confidence intervals - Documentation

[AndersS]

Some time ago I asked a question about "Confidence Intervals for Quantiles".
Now I have found a very good and recent (2017) publication - Available for students and researcher in electronic form. For free.

Statistical Intervals A Guide for Practitioners and Researchers, Second Edition  Meeker,  Hahn,  Escobar

ISBN 9780471687177 (cloth) | ISBN 9781118595169 (epub) -2017 - 630 pages

Lots and lots of useful information.

Some useful pages for "Confidence Intervals for Quantiles" :

5.2 Distribution-Free Confidence Intervals and One-Sided Confidence Bounds for a Quantile – page 78 (file page 109)

Epilogue - page 395 (file page 419)

Appendix A - Notation and Acronyms - page 395 (file page 419)

Quantile Function – page 426 (file page 449)

Appendix G - Distribution-Free Intervals Based on Order Statistics – page 497 (file page 516)

On the web site:

https://bcs.wiley.com/he-bcs/Books?action=index&bcsId=10767&itemId=0471687170

there is more information, like a list of suitable software package.

I hope that this updated information will be useful.

(Disclaimer - I am not really a statistician. More SAS programmer and problem solver)
/Br AndersS (Anders Skollermo)

[sbxkoenk]

Thanks for update.

I think you are referring to this (solved!) topic:

How to estimate Confidence intervals?
https://communities.sas.com/t5/SAS-Programming/How-to-estimate-Confidence-intervals/m-p/898485

Koen

[AndersS]

Hi! Yes I marked the subject as SOLVED earlier. Then later on I found a much better documentation. So I wanted to tell people about it. /Br AndersS

[AndersS]

Hi! YES - This documentation and the discussion there is really good.

/Br Anders

[Season · Accepted solution]

Thanks for sharing. I read this book while I was looking for formulae of the analogue of standard errors of medians and the first and third quantiles in order to pool my results after multiple imputation in PROC MIANALYZE but found that all of the three statistics don't have a "standard error" as long as their confidence intervals are computed in a distribution-free manner described in this book. In all, it helped me a lot, so it also has my recommendation.

I would like to recommend two monographs on confidence intervals to our friends: Confidence Intervals for Discrete Data in Clinical Research and Confidence Intervals for Proportions and Related Measures of Effect Size. I hope they can help data analysts in this Community.

[AndersS]

Hi!   Note that there is also a Facbook page about confidenc intervals:

https://www.facebook.com/confidenceintervals

[Season]

Thank you for sharing!