Statistical Modeling

Now Released!!

	ISBN 978-1448642397	Statistical Modeling: A Fresh Approach Daniel Kaplan Read the preface and sample chapters on line If we statistician/teachers can follow Danny's example of teaching modeling in a first stat course, in such an accessible, thoughtful, statistically sophisticated way, the consequences for the future of our profession could be profound. --- Prof. George Cobb, Mt. Holyoke College

NEW! in January 2010
Exercises and Activities. Exercises are distributed on line, available from anywhere at any time. Cut and paste the link into your syllabus. (AcroScore instructors: request a customized link for your access code.)

Getting the book:

Software:

Datasets and additional operators (as an R workspace): ISM.Rdata
Reference sheet for R commands
R statistical package

Resources for Instructors:

An annotated list of Exercises and Activities. (AcroScore instructors: request a customized link for your access code.)
Data sets in CSV spreadsheet format. (They are already in the R workspace ISM.Rdata.)
alder : birthweight : congress : courses : cps : galton : gestation : grade-to-number : grades : hdd-minneapolis.csv : I95citations : kidsfeet : marriage : running-longitudinal.csv : sat : speeders : swim100m : ten-mile-race : utilities : WeightLoss : whickham : zebra-mussels :

Seminars and Workshops:

UPCOMING:
- Computing in R workshop for faculty to be held at Dennison University in late May 2010. (Details to be posted.)
- Project MOSAIC kick-off workshop at the Institute for Mathematics and its Applications in late June 2010.
- "Models of the Alternative Hypothesis" at the Joint Statistical Meetings in August 2010.
PREVIOUS Workshops/Seminars
- "Remodeling Data Analysis!" an MAA mini-course at the Joint Mathematics Meetings, San Francisco, January 2010,
- Teaching Statistical Modeling Workshop at USCOTS: June 23-25, 2009
- 1-day instructors' workshop January 4, 2009 at the Joint Mathematics Meetings in Washington, DC
- CAUSE-sponsored seminar on Simulation and Resampling

Reviewers' comments on the book:

This is an unconventional, modern, and compelling book that turns statistics on its head by introducing modeling from day one. While standard methods are introduced later (to help students understand them when they see them in context), the focus of the book is on approaches that quantify differences while controlling for other factors. It's readable, has a raft of online exercises that can be used before, during and after class, and communicates the excitement of statistics.
There remains a compelling need to better connect the science, mathematics, statistics and modeling courses at the introductory level, and this book helps to make those links.
This book puts conceptual tools into the hands of students with which they can build an understanding of statistics that goes well beyond the end of the course. I think Kaplan achieves in the book what he has set out to do.
This text is not simply a bunch of tools. Students learn bigger statistical concepts and they should be more prepared for future study in statistics than those taking a standard methods course.
The text approaches statistics from a modeling perspective. Rather than starting with the usual t-tests and onesample z-tests, it jumps right into multiple regression. Through multiple regression, it derives t-tests, ANOVA, ANCOVA, and in the process uses linear models to provide a unifying theory for all of the methods usually taught in an introductory statistics course.
This text gets the first statistics course right in that, by emphasizing models, it teaches the students to see elementary statistics as a consistent whole rather than as a collection of unrelated techniques and tests.
In my opinion, every introductory statistics teacher should have this book on his or her shelf and be familiar with the material in it.
Different approach, current statistical usage, open source software and one you must read!
It is the best text I've seen that communicates the excitement of statistics by pruning away many of the cookbook like checklists of techniques and focuses on models.
Perhaps in my next job, I can convince the statistics curriculum committee to adopt this text.
It is unlikely to appeal to traditionalists (it can't be just plopped down into last year's conventional syllabus), but there is a generation of statistics teachers coming along who are likely to embrace it. It is also being recognized that future independent thinkers in diverse fields must have computational and modeling skills that transcend the common familiarity with mice and menus.
The Statistical Geometry chapter and related material is brilliant, and offers an insight into correlation that will really resonate with mathematics teachers; ditto for the Logic of Hypothesis Testing chapter.
The exercises are a particular strength of the project. These are wonderfully diverse and effective ways of engaging students at multiple levels. The range of uses for these is very attractive. Students can be assigned reading questions for before lecture, straightforward drill concepts to ensure that they have a basic understanding as well as elaborations that extend this learning to a new level.
[The textbook we have been using] mistakenly assumes that students are oafs and that they need to be told but not taught. It is terrible to the core that puts off the readers. This manuscript is different. It builds a good conceptual foundation, and then motivates and propels the reader forward. This manuscript rightly assumes that students are intelligent and with a proper guidance, direction and tone they can learn, understand and appreciate the beauty and usefulness of statistics in real life. I personally would like to thank the author for presenting the material the way he did. He makes the book a compelling and enjoyable read.
The integration of R into the text was the main reason I chose this text. I love how he has a computational section at the end of each chapter, and I think his coverage of R is extremely organized, with not too much, but not too little. R is my preferred stats package since it is versatile, powerful, and free!
I have a couple of shelves full of introductory statistics books, but for various reasons none have ever seemed suitable for my undergraduate biology/ecology students. Daniel Kaplan has now given me a book I can confidently recommend.
This book, as the title suggest, focusses on a modeling/regression approach but still includes an appropriate discussion on null hypothesis testing and anova. This discussion comes near the end of the book once the modeling approach is clearly understood. It could easily be used as a first statistics book (which is how I would use it), but it could also follow a more traditional null hypothesis based first course.
The writing style is clear and easily followed, with just enough explanation to understand why something is being done, but without the mathematical notation that scares so many students away. A geometrical approach is taken to explain some of the more theoretical aspects. It's also just long enough to be useful, without intimidating students with an enormous number of pages. Chapters are generally short and there is a nice logical flow as you work through the book. The practical examples use R, are well thought out, and great care is taken to explain the output.
Probably not the book for an undergraduate statistician, but for other disciplines that need a good understanding of not only statistical practice, but also statistical thinking, I have yet to come across anything better. Indeed I am so impressed with this book that I felt compelled to write my first ever Amazon review.