Average Reviews:
(More customer reviews)I have a Master's degree in statistics and I am currently pursuing a PhD in chemistry. To prepare for a statistics lecture I am giving my research group, I wanted to see how material was being taught to those in the sciences who were unfamiliar with statistics. My hope with this book was to get a grasp of what educators felt were important statistics for scientists to know and to gain ideas I could use to help teach statistics.
Short Overall Recommendation: Buy this book. You will find the book dense with material, yet easy to read with concise, statistically accurate explanations. The book contains most of the important statistical techniques that researchers need to be familiar with and know. At times, the flow of a chapter was illogical, but as a whole the content of that chapter was always beneficial.
The book lacks textbook-style problem sets at the ends of chapters but you will find the use of Excel functions can suffice, especially if you have your own data to work with (there is also an index of "Worked Examples"). But because this is not a textbook, if you are trying to learn statistics without any background knowledge, you may find the dense style without problems and answers difficult to handle. While I expand on the benefits and disadvantages of this book below, I believe that the book is very well written and the benefit gained from reading it and having it as a reference far outweigh any disadvantages.Lengthy breakdown for those who want more details:
### Benefits of this book:
--Excel functions were taught so that the average laboratory researcher can perform statistical tests right on their computer. On page 9, the authors clearly state, "Excel is not a statistical software package." See my note below in Disadvantages for more.
--The book clearly states this is for biologists, and the examples are directly from biology. Bremer and Doerge did a good job reaching out to their target audience.
--The book is hardcover with the inside pages wire bound. This type of binding allows a researcher to open the book and let it sit hands-free open on a bench.
--In section 2.3, Bremer and Doerge describe the friction that can exist between a researcher and a statistical consultant. They offer brief, but very appropriate advice for how a researcher should prepare for a meeting with a statistician. However, they did not emphasize the benefits of meeting with a statistician prior to running an experiment.
--There is an "Excel Function Index," which I thought would be helpful for a researcher. However, they did miss some indexing. For example, on page 30 the "=NORMINV()" function is introduced, but it is not defined in a gray box as all the other Excel functions and the index does not cite this page for that particular function.
--Chapter 8, "Special Topics" included a very brief (~25 pages) overview of Classification, Clustering, Principal Component Analysis, Microarray Data Analysis, Maximum Likelihood and Bayesian statistics. I felt this was a good conclusion to the book, although I would have called this chapter "Advanced Topics" and emphasized in the introduction that Excel cannot be used for these types of analyses.
--Both authors are women. In the world of science/math textbook authorship, I believe women are severely underrepresented. Since this is such an outstanding book, I believe that this subtle detail should not go unnoticed.### Disadvantages (let me emphasize again that I do not believe these outweigh the benefit of reading this book and having it on your shelf):
--The use of Excel functions serves as a good introduction to statistics, but I fear a reader may misinterpret the message of the authors. While Excel is a good tool to learn statistics, it is not a program that should be relied upon for data analysis, especially for data analysis that will result in published manuscripts. I encourage interested individuals to pursue Minitab. I used this program before my formal statistics training and always found it user-friendly. With Bremer and Doerge's book in hand the Minitab functions should be easy to understand.
-- Although rare, there is an occasional overestimation of the reader's knowledge of statistics; one example is on page 44 where the error term of a regression model is introduced. The authors describe the notation: "epsilon ~ Normal(0,1)" as epsilon having a "normal distribution with parameters 0 and 1." I believe it would have been useful here to say that those parameters are the mean and variance, respectively.
--The format of the book was good, but I felt like they were trying to cram the pages. With a benchtop reference guide, I believe a recipe-style approach would have suited their audience better. For example, each new section would have started on a new page (or maybe every time a new term was introduced). Also, adding in tabs for the different chapters would have been nice. However, contradicting what I just said, one great advantage of their current approach is that a researcher may simply add their own tabs using a sticky note.
--The list of popular statistics software packages on page 8 does not include S-Plus, although it is listed later on page 88. The authors could have expanded their list on page 8 (including packages like S-Plus, Maple, Matlab and Statistica) and just referenced the list later in the text. Perhaps mentioning some well-known packages commonly used by biologists, such as GraphPad Prism or Origin would have been useful.
--On page 108, the authors introduce "Correlation and regression" and discuss r, r^2 and their meanings. While correct, the authors missed an opportunity to expand the discussion to emphasize that r^2 is not a measure of linearity. The term coefficient of determination was not used to describe r^2. Finally, neither r^2, r-squared, nor coefficient of determination could be found in the index.
--On page 130, the authors describe that ANOVA and regression models "both assume that the residuals are independent, normally distributed, and have constant variance." Yet, in Section 7.1.6, "Verification of Assumptions" they do not describe the independence requirement. There is also no description of a diagnostic check for independence.
--ANOVA was discussed, but no experimental design was introduced. I found this to be quite disappointing. So much more could be learned in science if experiments were properly designed in order to extract out correct data. The only way to do this is to consult with a statistician before an experiment is run (but let's be honest, this rarely happens in the sciences). Or the scientist themselves can be equipped with basic statistical experimental design.
--There was no discussion of nonlinear regression. Section 7.1.3, "Logistic Regression" explicitly says that Excel cannot perform logistic regression. However, nonlinear regression can be performed in Excel and is a technique used by biologists.
Click Here to see more reviews about: Statistics at the Bench: A Step-by-Step Handbook for Biologists (Handbooks)
Statistics at the Bench is a convenient bench-side companion for biologists, designed as a handy reference guide for elementary and intermediate statistical analyses. The expectations for biologists to have a more complete understanding of statistics are growing rapidly. New technologies and new areas of science, such as microarrays, next-generation sequencing, and proteomics, have dramatically increased the need for quantitative reasoning among biologists when designing experiments and interpreting results. Even the most routine informatics tools rely on statistical assumptions and methods that need to be appreciated if the scientific results are to be correct, understood, and exploited fully. This book is not a textbook. It is a hands-on manual for working scientists. Statistics at the Bench provides a simple refresher for those who have forgotten what they once knew, and an overview for those wishing to use more quantitative reasoning in their research. Statistical methods, as well as guidelines for the interpretation of results, are explained using simple examples. Throughout the book, examples are accompanied by detailed Excel commands for easy reference. Related Titles from the Publisher Lab Math: A Handbook of Measurements, Calculations, and Other Quantitative Skills for Use at the Bench At the Bench: A Laboratory Navigator, Updated ed. Binding and Kinetics for Molecular Biologists Experimental Design for Biologists Bioinformatics: Sequence and Genome Analysis, 2nd ed.
Buy cheap Statistics at the Bench: A Step-by-Step Handbook for Biologists (Handbooks) now.
