Measurement-error

Ray Carroll and David Ruppert are well known research statisticians who have published many joint articles on regression, weighted regression and transformation and they have also written an excellent book together on this research topic. Stefanski has recently published several papers on measurement error models with Carroll. Here they have teamed up to write a statistics text on a unique topic. Measurement error models are common and practical when dealing with covariates that have measurement error. Least squares estimation in linear regression is based on the assumption that the predictor variables are measured without error. There are many articles and an excellent text by Fuller "Measurement Error Models", published by Wiley in 1988 that deals with the linear case. Also look at a section in Chapter 5 of Miller's "Beyond ANOVA, Basics of Applied Statistics" that refers to the problem as the error in variables problem. For the nonlinear case this is the first treatment. Well written and well documented, this text provides an up-to-date account of the theory and methods and provides real applications (e.g. the Framingham Heart Study). This is a great reference as are many of the other monographs in this series by Chapman and Hall/CRC Press. Includes bootstrap approaches in the chapter on fitting methods and models.

Disclaimer: I was a student of Dr. Van Ness and assisted with some minor research for the early chapters in the book.

The text deals with measurement error, ie, situations where there may be error in the measurement of the independent variable as well as the dependent. Focus is given to problem-solving and illustration of key points rather than the most rigorous mathematical proofs. One must be strongly conversant with at least undergraduate mathematical statistics to grasp most of the text.

However, it's an intriguing field with a number of real-world applications that become apparent in the text.

As a professional engineer with a recurring need to crunch large amounts of statistical data, I find that this book is the perfect quick guide to things that forget and don't use that often. If has easy to follow language, and the best part about it is that I don't need to reread the whole thing to get a good explanation of a topic in the last chapter.

I had to knock it down a star because it is a touch out of date. The math is fine, but I wish that there was a companion that explained how to do some of the more uncommon operations using common spreadsheeting or data analysis software. Sometimes, figuring out how to get MS Excel to do what Taylor recommends that I do can be more cumbersome than anything else.

If nothing else, it has a great picture on the cover.

Many undergraduate students in sciences and engineering must have encountered this experience: You conduct an experiment and collect the relevant data. You are asked to fit your data into a straight line by performing one or multiple linear regression. You are also to present any uncertainty and error in your data as well as calculation. You panic and scratch your head and don't know what's the appropriate procedure to carry out these analysis.

Here comes John Taylor's "An Introduction to Error Analysis", which introduces the study of uncertainties to students. The book assumes no prior knowledge and uses a plethora of pertinent examples (drawn from chemistry, physics, and engineering) to illustrate topics like propagation of uncertainties, random uncertainties, rejection of data, least-squares fitting, and distribution.

This book will save hours of studying and researching on error analysis method. It is very well-written and reader-friendly that lower division students will find it useful.

I can't believe I'm the first person to review this. Everyone I work with loves this book, it's a classic.

If you are into measurement errors, you will find this book excellent value. The author goes into a lot of detail spelling everything out. The organisation and English is a bit difficult. I also enjoyed a book called Measurement Uncertainty by RH dieck, but that has many errors. Also check out "An introduction to error analysis" by Taylor...

This is a major piece in the uncertainty hunting. The Rabinovich's book is needed in every national calibration laboratory.

As an experimental physicist by education I picked up this book to brush up my measurement knowledge. I was really surprised to discover that this book is about metrology and not experimental physics measurements. In fact it was very educating for me to read about the many differences between those two closely related trades.

Especially I liked the first chapter it has a philosophical depth. Most of it can be read like poetry: enjoying every day a few pages. The deep and thoughtful understanding of the author is stunning.

The following chapters are of a strong varying quality: some trivial, some educational. The mathematics can be followed (or looked up) by people with a physics background. It is not always clear to me why at some points the details of the calculations are spelled out and at others not.

This book is definitely not a textbook or an introduction to the field. You should already have a basic knowledge of the subject.