Statistical Analysis with Missing Data by Roderick J. A. Little

By Roderick J. A. Little

Compliment for the 1st version of Statistical research with lacking Data

"An very important contribution to the utilized data literature.... I supply the publication excessive marks for unifying and making obtainable a lot of the previous and present paintings during this vital area."
—William E. Strawderman, Rutgers University

"This book...provide[s] fascinating real-life examples, stimulating end-of-chapter workouts, and up to date references. it's going to be on each utilized statistician’s bookshelf."
The Statistician

"The e-book may be studied within the statistical tools division in each statistical agency."
Journal of professional Statistics

Statistical research of knowledge units with lacking values is a pervasive challenge for which typical tools are of restricted price. the 1st version of Statistical research with lacking information has been a typical reference on missing-data tools. Now, reflecting wide advancements in Bayesian tools for simulating posterior distributions, this moment variation via said specialists at the topic bargains a completely up to date, reorganized survey of present method for dealing with missing-data problems.

Blending conception and alertness, authors Roderick Little and Donald Rubin evaluation old methods to the topic and describe rigorous but basic tools for multivariate research with lacking values. They then supply a coherent thought for research of difficulties in accordance with likelihoods derived from statistical versions for the information and the missing-data mechanism and practice the idea to a variety of vital missing-data problems.

The re-creation now enlarges its assurance to include:

  • Expanded assurance of Bayesian technique, either theoretical and computational, and of a number of imputation
  • Analysis of knowledge with lacking values the place inferences are in accordance with likelihoods derived from formal statistical versions for the data-generating and missing-data mechanisms
  • Applications of the strategy in quite a few contexts together with regression, issue research, contingency desk research, time sequence, and pattern survey inference
  • Extensive references, examples, and exercises

Amstat News requested 3 evaluate editors to expense their best 5 favourite books within the September 2003 factor. Statistical research With lacking Data was once between these selected.

Show description

Read Online or Download Statistical Analysis with Missing Data PDF

Best biostatistics books

Introductory Biostatistics for the Health Sciences: Modern Applications Including Bootstrap (Wiley Series in Probability and Statistics)

Available to drugs- and/or public policy-related audiences, in addition to such a lot statisticians.
Emphasis on outliers is mentioned in terms of detection and treatment.
Resampling facts software program is integrated throughout.
Motivating functions are offered in gentle of sincere theory.
Plentiful routines are sprinkled throughout.


“…readers will come away with a greater realizing of biostatistics. .. ” (Statistical equipment in scientific learn, Vol. 14, No. 1, 2005)

"…a strong textbook for an introductory path in biostatistics on the complicated undergraduate/beginning graduate level…covers a broader number of themes that's frequently present in related textbooks…" (The American Statistician, Vol. fifty eight, No. 2, may possibly 2004)

". .. lucid writing kind and various real-world examples upload to the book's allure and price. .. complicated topic uniquely available. .. " (Zentralblatt Math, Vol. 1028, 2004)

“. .. the authors’ lucid writing kind and various fascinating real-world examples make a fancy topic rather available to a various viewers. .. i like to recommend this publication to undergraduate information scholars, clinical group of workers, public overall healthiness trainees, or nursing and scientific scholars. .. might additionally make an outstanding textbook for an introductory. .. biostatistics path. ” (Mathematical experiences, factor 2004b)

". .. the booklet is such a lot profitable in assembly the wishes of scholars. .. " (Technometrics, Vol. forty five, No. four, November 2003)

Analyzing rater agreement

Contract between raters is of significant significance in lots of domain names. for instance, in medication, diagnoses are frequently supplied by means of a couple of physician to ensure the proposed remedy is perfect. In legal trials, sentencing relies, between different issues, at the entire contract one of the jurors. In observational experiences, researchers elevate reliability by means of analyzing discrepant rankings.

Parametrische Statistik: Verteilungen, maximum likelihood und GLM in R

Beispielreich baut das Buch Schritt für Schritt die statistischen Grundlagen moderner Datenanalysen für Anwender auf. Dabei wird besonderer Wert auf einen roten Faden gelegt, der alle Methoden zusammenführt. Ausgehend von den Grundlagen in beschreibender Statistik, Verteilungen und greatest chance, werden alle anderen Verfahren als Spezialfälle des GLM entwickelt (ANOVA, a number of Regression).

Design and Analysis of Cross-Over Trials, Third Edition

The 1st version of layout and research of Cross-Over Trials fast turned the normal reference at the topic and has remained so for greater than 12 years. In that point, although, using cross-over trials has grown swiftly, fairly within the pharmaceutical enviornment, and researchers have made a couple of advances in either the idea and strategies appropriate to those trials.

Extra info for Statistical Analysis with Missing Data

Sample text

On the other hand, if we have little knowledge about the form of the uncensored distribution, we cannot say whether the data are a censored sample from a symmetric distribution or a random subsample from an asymmetric distribution. In the former case, the sample mean is biased for the population mean; in the latter case it is not. 10. Historical Heights. Wachter and Trussell (1982) present an interesting illustration of stochastic censoring, involving the estimation of historical heights. The distribution of heights in historical populations is of considerable interest in the biomedical and social sciences, because of the information it provides about nutrition, and hence indirectly about living standards.

It is applied to the case of multivariate normal data in Chapter 11. The resulting algorithm is particularly instructive, because it is closely related to an iterative version of a method that imputes estimates of the missing values by regression. Thus even in this complex problem, a link can be established between efficient model-based methods and more traditional pragmatic approaches based on substituting reasonable estimates for missing values. Chapter 11 also presents more esoteric uses of the EM algorithm to handle problems such as variance components models, factor analysis, and time series, which can be viewed as missing-data problems for multivariate normal data with specialized parametric structure.

Approaches to measuring and incorporating imputation uncertainty are discussed in Chapter 5, including multiple imputation, which reappears in Parts II and III in the context of model-based methods. 4. Model-Based Procedures. A broad class of procedures is generated by defining a model for the observed data and basing inferences on the likelihood or posterior distribution under that model, with parameters estimated by procedures such as maximum likelihood. Advantages of this approach are flexibility; the avoidance of ad hoc methods, in that model assumptions underlying the resulting methods can be displayed and evaluated; and the availability of estimates of variance that take into account incompleteness in the data.

Download PDF sample

Rated 4.78 of 5 – based on 5 votes