Experimental Design for Combinatorial and High Throughput Materials Development
暂无分享,去创建一个
Each chapter ends with an extensive set of exercises. The problems are satisfyingly diverse and cover the spectrum of skills: theoretical (proving and deriving results given in the text), computational (simulating data and carrying out speci c analyses; verifying results on the data examples), and conceptual (reading and summarizing key points in a paper, expressing opinions about techniques). A few of the questions are rather vague; for example, “Review the literature on missing values in ANOVA from Allan and Wishart (1930) through Dodge (1985)” (2.1, p. 39) and “Review results concerning the convergence of EM” (8.5, p. 188). I am undecided as to whether these and similar exercises are so vague as to be frustrating for students, or just vague enough to be tantalizingly challenging, although I lean toward the latter view. All in all, this book is well written and very readable, and I would not hesitate to use it as a text for a graduate-level class on the analysis of missing data or as a supplementary text in a data analysis course. The authors assume certain minimal knowledge on the part of readers—familiarity with likelihood-based inference (although this is also brie y reviewed in Sec. 6.1), standard models for the analysis of complete data (in particular, linear models and multinomial models for count data), and properties of common distributions, such as the normal. The book should thus be well within the grasp of the intended audience of applied statisticians. Graduate students beyond the rst year of study should also be able to follow the material without too much dif culty. I did nd a few typos and one gure in Chapter 10 that appeared to be incomplete, but these minor aws do not detract from the value of the book as a whole. All in all, Statistical Analysis With Missing Data is a comprehensive, updated treatment of an important topic by two of the leading researchers in the eld. In summary, I highly recommend this book to all statisticians who work with real data.