What do you think?
Rate this book
Design and Analysis: A Researcher's Handbook
The fourth edition of Design and Analysis continues to offer a readily accessible introduction to the designed experiment in research and the statistical analysis of the data from such experiments. Unique because it emphasizes the use of analytical procedures, this book is appropriate for all as it requires knowledge of only the most fundamental mathematical skills and little or no formal statistical background. Topics single- and two-factor designs with independent groups of subjects; corresponding designs with multiple observations; analysis of designs with unequal sample sizes; analysis of covariance; designs with three factors, including all combinations of between-subjects and within-subject factors; random factors and statistical generalization; and nested factors. This book lives up to its name as a handbook, because of its usefulness as a source and guide to researchers who require assistance in both planning a study and analyzing its results.
624 pages, Hardcover
First published December 1, 1972
5 people are currently reading
74 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 9 of 9 reviews
August 30, 2009
I first came across the 3rd edition of this book, written solely by Geoff Keppel, during the spring semester of 2000. I was taking Dawn McBride's "Experimental Design" class, and this was the text. I enjoyed the class, but was only doing B+/A- work during it, as I was not putting in the time the course deserved and going away every single weekend for the first half of the class and, mostly for interviews for Ph.D. programs. As a result, on the final exam, I was in a situation where I had to get a 96% in order to get an A in the class. Luckily, I studied well, set the curve in the class (so high it turned out to be 102%), and received the A.
During the fall of 2004, the person who usually teaches "Statistics for Experimental Design" at Northwestern University received a grant, and wasn't able to teach the course. It had been canceled the previous two years, and they didn't want to have their graduate students, who were often finished with course work by the end of their third year, to take the course. They put out a call for possible teachers in the Chicago-land area. I was finished with my coursework, and the department chair passed the word along to me. I taught the class, and used this updated version of Keppel's text, as it was the usual text used at Northwestern as well. This was one of my hardest teaching experiences -- first, it was my first graduate course to teach. Second, Northwestern had first, second, and third year students in the class. The first year students found the class too fast and the third year students found the class too slow. Third, this was my first experience teaching on the quarter system, as I had been used to the pace of semesters. However, I did my best, and am happy I did it. I learned a lot about the Analysis of Variance (ANOVA) in the class, as well.
In the summer of 2006, I moved to California State University, Fullerton as a faculty member. There, my department chair was Dan Kee. Dan had received his Ph.D. in educational psychology from UC Berkeley, where Geoff Keppel had taught him class. Dawn McBride was doing a master's degree with Dan, but she later transferred to UC, Irvine. Incidentally, she published with Nancy Segal, who I also published with. She also published with Dan Kee, as well as Barbara Cherry, a faculty member in psychology I knew well.
As a teacher, Geoff Keppel would give students his text, tell them to read it, and then sit at the front of class and ask "what questions do you have for me?" This assumed that students had read a stats book (GASP!) and understood enough to articulate good questions. I don't think that even my Northwestern graduate students were sophisticated enough to appreciate that approach. However, Dan Kee stated that he enjoyed his discussions with Keppel.
The text focuses not on the design of experiments part, as the text might lead on to believe, but rather on the statistical analysis of data from experiments. The most common technique for this, the ANOVA, is the topic of 90% of this book, with some detail given to analysis of covariance, power analysis, and multilevel modeling. Keppel's notation is slightly different than the usual statistical notation for ANOVA, and I don't believe that anything is gained by it. In fact, when I teach ANOVA, even when using Keppel, I tell students that I'm going to use a slightly different notation and that they are free to use either (and ideally, would move back and forth between the two). I know of no other statisticians who use Keppel's notation and very few psychologists, all of whom cite Keppel, who use it. Despite that idiosyncracy, the text has thorough discussions of nearly all basic ANOVA topics, and goes into them with more detail than any psychology texts at the time (although there are recent texts that are quite good -- notably texts by (1) Cole & Maxwell's, (2) Judd & McClelland, (3) Kirk, and (4) Anderson). The one topic I would cover more is multiple comparison procedures for within-subjects designs. Supplementing Keppel's text with Toothaker's Sage little green book on "Multiple Comparison Procedures" would provide an ideal coverage of the topic. Dawn McBride, when she taught my class, also added articles on other techniques, such as meta-analysis, though the text could easily stand alone. I should also note that I especially like the book by Braver, MacKinnon (yes, the same MacKinnon who writes so much on mediation analysis), and Page on using SPSS for ANOVA -- SAS is generally a much better system to use for ANOVA, but with the Braver, MacKinnon, and Page text, I was able to make SPSS do nearly everything I wanted to do. In summary, Keppel's text is an excellent foundation for ANOVA. There are ways it might need supplementing, but those are likely specific to researchers or teachers using the text rather than general lack on its part.
During the fall of 2004, the person who usually teaches "Statistics for Experimental Design" at Northwestern University received a grant, and wasn't able to teach the course. It had been canceled the previous two years, and they didn't want to have their graduate students, who were often finished with course work by the end of their third year, to take the course. They put out a call for possible teachers in the Chicago-land area. I was finished with my coursework, and the department chair passed the word along to me. I taught the class, and used this updated version of Keppel's text, as it was the usual text used at Northwestern as well. This was one of my hardest teaching experiences -- first, it was my first graduate course to teach. Second, Northwestern had first, second, and third year students in the class. The first year students found the class too fast and the third year students found the class too slow. Third, this was my first experience teaching on the quarter system, as I had been used to the pace of semesters. However, I did my best, and am happy I did it. I learned a lot about the Analysis of Variance (ANOVA) in the class, as well.
In the summer of 2006, I moved to California State University, Fullerton as a faculty member. There, my department chair was Dan Kee. Dan had received his Ph.D. in educational psychology from UC Berkeley, where Geoff Keppel had taught him class. Dawn McBride was doing a master's degree with Dan, but she later transferred to UC, Irvine. Incidentally, she published with Nancy Segal, who I also published with. She also published with Dan Kee, as well as Barbara Cherry, a faculty member in psychology I knew well.
As a teacher, Geoff Keppel would give students his text, tell them to read it, and then sit at the front of class and ask "what questions do you have for me?" This assumed that students had read a stats book (GASP!) and understood enough to articulate good questions. I don't think that even my Northwestern graduate students were sophisticated enough to appreciate that approach. However, Dan Kee stated that he enjoyed his discussions with Keppel.
The text focuses not on the design of experiments part, as the text might lead on to believe, but rather on the statistical analysis of data from experiments. The most common technique for this, the ANOVA, is the topic of 90% of this book, with some detail given to analysis of covariance, power analysis, and multilevel modeling. Keppel's notation is slightly different than the usual statistical notation for ANOVA, and I don't believe that anything is gained by it. In fact, when I teach ANOVA, even when using Keppel, I tell students that I'm going to use a slightly different notation and that they are free to use either (and ideally, would move back and forth between the two). I know of no other statisticians who use Keppel's notation and very few psychologists, all of whom cite Keppel, who use it. Despite that idiosyncracy, the text has thorough discussions of nearly all basic ANOVA topics, and goes into them with more detail than any psychology texts at the time (although there are recent texts that are quite good -- notably texts by (1) Cole & Maxwell's, (2) Judd & McClelland, (3) Kirk, and (4) Anderson). The one topic I would cover more is multiple comparison procedures for within-subjects designs. Supplementing Keppel's text with Toothaker's Sage little green book on "Multiple Comparison Procedures" would provide an ideal coverage of the topic. Dawn McBride, when she taught my class, also added articles on other techniques, such as meta-analysis, though the text could easily stand alone. I should also note that I especially like the book by Braver, MacKinnon (yes, the same MacKinnon who writes so much on mediation analysis), and Page on using SPSS for ANOVA -- SAS is generally a much better system to use for ANOVA, but with the Braver, MacKinnon, and Page text, I was able to make SPSS do nearly everything I wanted to do. In summary, Keppel's text is an excellent foundation for ANOVA. There are ways it might need supplementing, but those are likely specific to researchers or teachers using the text rather than general lack on its part.
April 16, 2020
It's old, but it's good. It gets confusing sometimes, but I guess that is because Keppel presupposes that the reader knows what he means. Anyway, what I like most about this book is that it explains statistical terms very well. I think these terms are necessary for every statistical test you may carry out (not only for ANOVA, which is the test used in the book). Also, Keppel gives numerical examples that are fully worked out. These examples become handy when you use a statistical program, such as R, for the first time, and you're not sure if the program gives you the results you want. So you can use the examples in Kepple to see if you can get the same results using the new program.
February 12, 2024
Adding this because amazon linked it and because it is the ONLY textbook I can say I have EVER read front to back, every page. Honestly, taught me a lot. Probs won't ever read another textbook front to back so cheers to the only one.
March 5, 2007
A very interesting and easy-to-read book for those who work with different types of univariate analysis of variance (ANOVA). I believe reading and working through this book is a must for those who are doing research in the field of humanities.
Although the authors tried to cover some basic statistical concepts briefly; reading this book requires basic knowledge about hypothesis testing, and experimental design.
Although the authors tried to cover some basic statistical concepts briefly; reading this book requires basic knowledge about hypothesis testing, and experimental design.
February 20, 2008
Essentially this book reviews all relevant ANOVA-based analyses encompassed within the General Linear Model. It's an excellent reference when trying to fit your variables into an appropriate statistical model.
May 29, 2008
My favorite statistical text -- Keppel presents the basic concepts of experimental design and Analysis of Variance in a straightforward, readable, and accessible manner. Unfortunately, this text doesn't branch out much beyond ANOVA models.
May 29, 2010
This is a textbook used in a graduate class in statistics I took at Kansas State University. Book is in like new condition complete with dustcover but does have highlighting.
February 12, 2015
Excellent reference. Sets up the logic for various tests in a comprehensible way that isn't too mathematical.
November 20, 2022
I have the third and fourth edition and I think they're some of the best statistics books I've read on ANOVAs.
Displaying 1 - 9 of 9 reviews