ASA Connect

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

Going off your preference for "motivation and examples", have you considered compiling a series a articles that address the curriculum?

E.g. This article by Gershman and Blei described the motivation of non-parametric method for modeling.

https://www.cs.princeton.edu/~blei/papers/GershmanBlei2012.pdf

This could be a helpful illustration that the value of non-parametric methods go far beyond "How to perform a hypothesis test when you data doesn't meet assumptions A, B, C, D."

------------------------------
Glen Wright Colopy
DPhil Student
University of Oxford

Original Message:
Sent: 07-07-2016 16:41
From: Malcolm Sherman
Subject: non-parametric course

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

My personal favorite non-parametric book is Hollander & Wolfe "Non-parametric Statistical Methods".  I still use it when I forget the nuances of different methods.  I have the first edition from 1973,  but a 3rd edition is available from 2014 with Eric Chicken added as coauthor. Nonparametric Statistical Methods, Third Edition

Wiley		remove preview
Nonparametric Statistical Methods, Third Edition Praise for the Second Edition "This book should be an essential part of the personal library of every practicing statistician."- Technometrics Thoroughly revised and updated, the new edition of Nonparametric Statistical Methods includes additional modern topics and procedures, more practical data sets, and new problems from real-life situations. View this on Wiley >

------------------------------
Susan Spruill
Statistical Consultant

Original Message:
Sent: 07-07-2016 16:41
From: Malcolm Sherman
Subject: non-parametric course

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

I use Higgens.  I do traditional rank-based and permutation-bsed material for 7 weeks or so, and then 6 weeks on bootstrapping and smoothing,  There's a SAS companion, but as I recall no R.  You'll probably want to write your own R material in any event, however, since R tools develop so quickly.

------------------------------
John Kolassa
Rutgers University

Original Message:
Sent: 07-07-2016 16:41
From: Malcolm Sherman
Subject: non-parametric course

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

I second Hollander, Wolfe, and Chicken's Nonparametric Statistical Methods. The initial descriptions are inuitive with many examples including R code in the most recent edition. If you want more detailed explanations, each chapter has an appendix immediately following with more complex proofs and extensions. I come back to this book frequently for personal consultations.

------------------------------
Cynthia Bell
Senior Statistician
McGovern Medical School at UTHealth

Original Message:
Sent: 07-07-2016 16:41
From: Malcolm Sherman
Subject: non-parametric course

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

The best current nonparametric book is "An Introduction to Modern Nonparametrics" by Higgins. Very complete and better than some of the older books with its current perspectives on bootstrapping, permutation tests, randomization tests, and more. Use it as a reference in consulting, in building libraries for graduate students, and in one-on-one independent studies. I have used several of the older books by Gibbons and others for previous classes in non-parametrics. Hope that helps!

William L. Seaver
Associate Professor in Business Analytics and Statistics
2007 Allen Keally Outstanding Teacher Award
2001 University's Excellence in Teaching Award

Business Analytics and Statistics
Haslam College of Business
University of Tennessee
Knoxville, TN 37996-0532
865-974-6862


Original Message------

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

Original Message------

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

Original Message------

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

I highly recommend Nonparametric Statistical Methods, 3rd edition by Myles Hollander, Douglas Wolfe, and Eric Church. I've used earlier editions of this book many times. Here is a link with more information: http://www.wiley.com/WileyCDA/WileyTitle/productCd-1118553314.html

Patti Costello
Eastern Kentucky University


Sent from my iPad


Original Message------

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman

I am still using the text by William Conover for its useful examples and overall completeness in material coverage. The short "theory" pages are listed at the end of each section, so a course can be taught without the theory. I simply add separately whatever statistical programs and new developments in the nonparametric contributions in the literature. I have extracted several ideas from this text, and I built on these ideas research articles that I have published. I realize that in this thread few have recommended so far this text.

------------------------------
Raid Amin
Professor
University of West Florida

Original Message:
Sent: 07-07-2016 16:41
From: Malcolm Sherman
Subject: non-parametric course

I've been asked to develop a course in non-parametric statistics for
upper division math majors and computer engineering grad students.
Students will have had some calculus and a standard calculus based stat
course covering topics like t-tests for normal samples and least-squares
regression. My immediate problem is to find a text. A 40 year old
out-of-print book (Sturdy Statistics by Mosteller and Rourke) comes
closest to what I am looking for, which is a somewhat informal text that
emphasizes motivation and examples instead of proofs. But M&R makes no
mention of bootstrap methods and doesn't use R or any other statistical
package. Suggestions for a suitable text would be appreciated. thanks.
Malcolm Sherman