Missing Data: A Gentle Introduction by Patrick E. McKnight PhD

, , Comments Off on Missing Data: A Gentle Introduction by Patrick E. McKnight PhD

By Patrick E. McKnight PhD

While such a lot books on lacking facts concentrate on utilizing subtle statistical recommendations to accommodate the matter after it has happened, this quantity offers a technique for the keep an eye on and prevention of lacking information. In transparent, nontechnical language, the authors support the reader comprehend the differing kinds of lacking information and their implications for the reliability, validity, and generalizability of a study’s conclusions. they supply sensible innovations for designing reviews that reduce the chance of lacking facts, and for addressing this significant factor whilst reporting examine effects. while statistical treatments are needed--such as deletion strategies, augmentation tools, and unmarried imputation and a number of imputation procedures--the booklet additionally explains tips to make sound judgements approximately their use. Patrick E. McKnight's web site bargains a periodically up-to-date annotated bibliography on lacking info and hyperlinks to different net assets that handle lacking facts.

Show description

Read Online or Download Missing Data: A Gentle Introduction PDF

Best methodology & statistics books

Advanced ESR Methods in Polymer Research

A definitive paintings on ESR and polymer technological know-how via cutting-edge major authoritiesThe prior two decades have visible remarkable advances in electron spin resonance (ESR) thoughts, really as they observe to polymeric fabrics. With contributions from over a dozen of the world's most sensible polymer scientists, complicated ESR equipment in Polymer learn is the 1st publication to assemble all of the present tendencies during this intriguing box into one accomplished reference.

Bioinformatics for Geneticists

This well timed e-book illustrates the worth of bioinformatics, no longer easily as a suite of instruments yet fairly as a technology more and more necessary to navigate and deal with the host of knowledge generated through genomics and the provision of thoroughly sequenced genomes. Bioinformatics can be utilized in any respect levels of genetics learn: to enhance research layout, to aid in candidate gene identity, to help facts interpretation and administration and to make clear the molecular pathology of disease-causing mutations.

Missing Data: A Gentle Introduction

Whereas such a lot books on lacking info specialize in employing refined statistical suggestions to accommodate the matter after it has happened, this quantity offers a strategy for the regulate and prevention of lacking facts. In transparent, nontechnical language, the authors aid the reader comprehend the different sorts of lacking facts and their implications for the reliability, validity, and generalizability of a study’s conclusions.

Compendium of Organic Synthetic Methods Volume 5

Presents artificial chemists with a mode for fast retrieval of data from the literature, directory fabric via response style instead of by means of writer identify or ebook date. every one up to date quantity will current the most recent artificial tools for coaching of monofunctional and difunctional compounds.

Additional info for Missing Data: A Gentle Introduction

Example text

The two groups may react differently throughout the study. Disappointed individuals may drop out of the study (Bradley, 1993; McPherson & Consequences of Missing Data 31 Britton, 2001). If groups are nonequivalent with respect to satisfaction with group assignment, that would result in differential attrition and might influence study outcomes. Differential attrition also results in unequal group size. As noted previously, missing data reduce the sample size and thus the power to detect significant treatment effects.

The goal is for these measures to converge (or triangulate) on the construct, thereby increasing the likelihood that it has been adequately captured by the measurement process. In many ways, the use of multiple measures protects researchers from problems stemming from missing data that occur when only a single measure is used. With multiple measures of the same construct, if data are missing for one or two measures, the others can still provide information about that construct, and therefore all is not lost.

Unbalanced cell sizes may limit the use of some statistical tests (if deletion methods are used; see Chapter 7) and may reduce statistical power to detect significant treatment effects. Unequal cell sizes may result in the violation of two assumptions, normality and homogeneity of variance, that are the basis of the t-test and the F-test. Under these conditions, alternative formulas for these tests should be used to minimize the error of inference (Keppel & Wickens, 2004). A mismatch between participants’ preferences for treatment and their assignment to study groups could contribute to differential attrition bias as well.

Download PDF sample

Rated 4.93 of 5 – based on 37 votes