Harvey Goldstein

ORCID: 0000-0003-3878-4122
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • School Choice and Performance
  • Statistical Methods and Bayesian Inference
  • Statistical Methods and Inference
  • Educational Assessment and Improvement
  • Urban, Neighborhood, and Segregation Studies
  • Education Systems and Policy
  • Advanced Causal Inference Techniques
  • Data Quality and Management
  • Optimal Experimental Design Methods
  • Higher Education Governance and Development
  • Birth, Development, and Health
  • Evaluation and Performance Assessment
  • Regional Economics and Spatial Analysis
  • Parental Involvement in Education
  • Survey Methodology and Nonresponse
  • Spatial and Panel Data Analysis
  • Student Assessment and Feedback
  • demographic modeling and climate adaptation
  • Health disparities and outcomes
  • Census and Population Estimation
  • Data-Driven Disease Surveillance
  • Psychometric Methodologies and Testing
  • Early Childhood Education and Development
  • Regional Development and Policy
  • Genetic and phenotypic traits in livestock

Witten/Herdecke University
2025

Vestische Caritas-Kliniken
2025

University College London
2009-2021

University of Bristol
2012-2021

At Bristol
2021

Great Ormond Street Hospital
2019-2021

Royal Statistical Society
2015-2020

British Academy
2020

Cabot (United States)
2006-2020

Lund University
2020

Contents Dedication Preface Acknowledgements Notation A general classification notation and diagram Glossary Chapter 1 An introduction to multilevel models 1.1 Hierarchically structured data 1.2 School effectiveness 1.3 Sample survey methods 1.4 Repeated measures 1.5 Event history survival 1.6 Discrete response 1.7 Multivariate 1.8 Nonlinear 1.9 Measurement errors 1.10 Cross classifications multiple membership structures. 1.11 Factor analysis structural equation 1.12 Levels of aggregation...

10.1198/tech.2006.s417 article EN Technometrics 2006-08-01

SUMMARY In the light of an increasing interest in accountability public institutions, this paper sets out statistical issues involved making quantitative comparisons between institutions areas health and education. We deal detail with need to take account model-based uncertainty comparisons. discuss establish appropriate measures institutional 'outcomes' base-line exercise care sensitivity when interpreting apparent differences. The emphasizes that methods exist which can contribute...

10.2307/2983325 article EN Journal of the Royal Statistical Society Series A (Statistics in Society) 1996-01-01

This general methodological discussion presents the essential features of multilevel modelling. The models described solve a number problems in different fields, including social and biological sciences, where data have hierarchical structure. Practical examples are used to illustrate topics such as school effectiveness, progress reading achievement, child growth, attitudes. A basic familiarity with structure application linear or multiple regression is assumed.

10.2307/1164754 article EN Journal of Educational Statistics 1988-01-01

10.2307/2344367 article EN Journal of the Royal Statistical Society Series A (General) 1976-01-01

Abstract In multilevel modeling the residual variation in a response variable is split into component parts that are attributed to various levels. applied work, much use made of percentage attributable higher level sources variation. Such measure, however, makes sense only simple variance components, Normal response, models where it often referred as intra-unit correlation. this article we describe how similar measures can be found for both more complex random and with discrete responses....

10.1207/s15328031us0104_02 article EN Understanding Statistics 2002-12-02

Charts are presented which give centile standards for boys9 and girls9 heights at ages 2 to 9 when parents9 height is allowed for. Mid-parent used (i.e. the average of father9s mother9s height). A comparison made with results from existing `parent-unknown9 British standard charts. child 3rd on parent-unknown charts (i) 20th new if his parents small enough adults, (ii) about 1st 97th centile. Conversely a has only be 25th population in conventional limit normal parental Thus result...

10.1136/adc.45.244.755 article EN Archives of Disease in Childhood 1970-12-01

Journal Article Multilevel mixed linear model analysis using iterative generalized least squares Get access H. GOLDSTEIN Department of Mathematics, Statistics & Computing, University London Institute EducationLondon WC1HOAL, U.K. Search for other works by this author on: Oxford Academic Google Scholar Biometrika, Volume 73, Issue 1, April 1986, Pages 43–56, https://doi.org/10.1093/biomet/73.1.43 Published: 01 1986 history Revision received: March Received: July

10.1093/biomet/73.1.43 article EN Biometrika 1986-01-01

SummaryAn updated system for estimating dental maturity is presented. It extends the original (Demirjian et al., 1973) based on radiographs of 7 teeth by including two extra stages, and enlarging standardizing sample to include 2407 boys 2349 girls. Percentile standards from ages 2·5 17·0 years are presented separately girls.Scoring systems percentile different sets 4 a comparison all three made. suggested that these may measure somewhat aspects maturity.

10.1080/03014467600001671 article EN Annals of Human Biology 1976-01-01

SUMMARY When a study produces estimates for many units or categories common problem is that end-users will wish to make their own comparisons among subset of these units. This occur, example, when school performance are produced all schools. The paper proposes procedure, based on the graphical presentation confidence intervals, which enables such be carried out while maintaining an average required type I error rate. means two independent samples presented graphically, it practice accompany...

10.2307/2983411 article EN Journal of the Royal Statistical Society Series A (Statistics in Society) 1995-01-01

SUMMARY This paper discusses the use of improved approximations for estimation generalized linear multilevel models where response is a proportion. Simulation studies by Rodriguez and Goldman have shown that in extreme situations large biases can occur, most notably when binary, number level 1 units per 2 unit small underlying random parameter values are large. An approximation introduced which largely eliminates situation described Goldman. Keywortis: �BINARY RESPONSE; GENERALIZED LINEAR...

10.2307/2983328 article EN Journal of the Royal Statistical Society Series A (Statistics in Society) 1996-01-01

A national sample of several thousand children has been followed longitudinally from birth. At the ages 7 and 11 years physical mental retardation due to smoking in pregnancy found, this deficit increases with number cigarettes smoked after fourth month pregnancy. Children mothers who 10 or more a day are on average 1.0 cm shorter between three five months retarded reading, mathematics, general ability compared offspring non-smokers, allowing for associated social biological factors.

10.1136/bmj.4.5892.573 article EN BMJ 1973-12-08

Summary When multilevel models are estimated from survey data derived using multistage sampling, unequal selection probabilities at any stage of sampling may induce bias in standard estimators, unless the sources fully controlled for covariates. This paper proposes alternative ways weighting estimation a two-level model by reciprocals each sampling. Consistent estimators obtained when both sample number level 2 units and 1 within sampled increase. Scaling weights is proposed to improve...

10.1111/1467-9868.00106 article EN Journal of the Royal Statistical Society Series B (Statistical Methodology) 1998-01-01

Summary A common application of multilevel models is to apportion the variance in response according different levels data. Whereas partitioning variances straightforward with a continuous variable normal error distribution at each level, extension this binary responses or proportions counts less obvious. We describe methodology due Goldstein and co-workers for apportioning that attributable higher binomial logistic models. This they referred as partition coefficient. consider extending...

10.1111/j.1467-985x.2004.00365.x article EN Journal of the Royal Statistical Society Series A (Statistics in Society) 2005-03-10

A procedure is proposed for the analysis of multilevel nonlinear models using a linearization. The case log linear discrete response data studied in detail.

10.1093/biomet/78.1.45 article EN Biometrika 1991-01-01

Multiple imputation is becoming increasingly established as the leading practical approach to modelling partially observed data, under assumption that data are missing at random. However, many medical and social datasets multilevel, this structure should be reflected not only in model of interest, but also model. In particular, reflect differences between level 1 variables 2 (which constant across units). This led us develop <b>REALCOM-IMPUTE</b> software, which we describe article. software...

10.18637/jss.v045.i05 article EN cc-by Journal of Statistical Software 2011-01-01

Linkage of population-based administrative data is a valuable tool for combining detailed individual-level information from different sources research. While not substitute classical studies based on primary collection, analyses linked can answer questions that require large sample sizes or hard-to-reach populations, and generate evidence with high level external validity applicability policy making. There are unique challenges in the appropriate research use data, example respect to bias...

10.1177/2053951717745678 article EN cc-by-nc-nd Big Data & Society 2017-12-01

In the social and other sciences many data are collected with a known but complex underlying structure. Over past two decades there has been an increase in use of multilevel modelling techniques that account for nested structures. Often however structures more cannot be fitted into First, cross-classified models where classifications not nested. Secondly, we consider multiple membership observation does belong simply to one member classification. These extensions when combined allow us fit...

10.1177/1471082x0100100202 article EN Statistical Modelling 2001-07-01
Coming Soon ...