Professor Harvey Goldstein
Harvey Goldstein, FBA who was Professor of Social Statistics at the School of Education, University of Bristol has sadly passed away on the 9th April 2020, aged 80 years. He is survived by his wife Barbara and son, Tom.
Harvey was an internationally renowned statistician and social science researcher whose work had huge impact in many different academic disciplines. He was a hard-working colleague, generous manager and collaborator who influenced many younger researchers and whose loss to statistics and academia more generally is immense.
Harvey was born in the East End of London on 30th October 1939 and grew up in Edmonton, North London. Harvey’s father originally worked as a button manufacturer but became a semi-skilled engineer after the second world war. Harvey’s mother was a milliner who worked from home but died when Harvey was just five and so he was brought up by his grandparents until his father returned from the war to join them. He went to his local primary school before moving up after his 11 plus to the local grammar school, Hendon County School.
Here he first showed an aptitude for mathematics and took A levels in Pure and Applied Maths, Physics and Chemistry gaining straight As and going to the University of Manchester to study Pure Mathematics. Here he achieved his BSc whilst taking statistics courses on the side given by Toby Lewis who would become a lifelong friend and mentor. Having completed his BSc, Harvey moved back to London where he studied for a Postgraduate diploma at UCL and spent 2 years as a research assistant. He also published his first journal article while at UCL in Nature no less with eminent co-authors including Richard Doll.
His first lecturing post was at the Institute of Child Health (ICH) in 1964 where he was to lecture for the next eight years. At ICH he worked primarily on child development studies alongside Jim Tanner producing ground-breaking work on modelling growth trajectories in children. Here he also began his long collaboration with Neville Butler who ran first the 1958 and then the 1970 cohort studies at the time. One example of early work he carried out at ICH and published in the British Medical Journal in 1972, is his definitive study demonstrating the negative effects of smoking during pregnancy on birthweight and perinatal mortality using data from the 1958 cohort. From ICH Harvey moved to the National Children’s Bureau to work with the director, Mia Kellmer Pringle as Senior Research Officer in 1972. He was still very much working on child development using the 1958 birth cohort study and had established another long-term working relationship, this time with Michael Healy, which led to a seminal paper on ‘bone maturity’ in 1976.
In 1977 Harvey moved to the Institute of Education (IOE), London as Professor of Statistical Methods, a post he was to hold to his (first) retirement aged 65 in 2005. At the IOE Harvey began to work more in social statistics, focussing on school effectiveness and educational assessment which led on to his interest in multilevel modelling. Whilst at the Institute Harvey was awarded the Royal Statistical Society’s Guy Medal in Silver in 1988 and made a Fellow of the British Academy in 1996.
At the IOE much of Harvey’s seminal work on statistical methodology, statistical software and applications was achieved with colleagues in the Centre for Multilevel Modelling (CMM) that he founded and led. Although Harvey retired from the IOE in 2005 he spent the next 15 years working part-time in several institutions. Upon his retirement, CMM moved to Bristol and Harvey continued in a part-time role working there until his death. For all that time he has also held a second part-time role as Professor of Statistics back at UCL Institute of Child Health as well as holding at various times visiting professorships at University of East Anglia, the London School of Hygiene and Tropical Medicine, the Institute of Fiscal Studies and the Australian Catholic University in Brisbane.
Harvey’s research contributions cover an incredibly broad range of topic areas including work in both the social and medical sciences primarily around statistical software and methodology. He has published well over 350 journal articles and many of these have been written since his first retirement.
He is perhaps best known in terms of statistical methodology for his original work into multilevel modelling which started back in the 1980s and has continued for over 35 years. He has written a seminal book entitled Multilevel Statistical Methods which has had four editions and many articles with colleagues from the multilevel centre. More recently with Bill Browne he has written work on Bayesian versions of multilevel models, with James Carpenter methodological work on handling datasets with missing data and with Katie Harron and colleagues at UCL on methods to allow data linkage of different administrative data sources. Alongside this methodological research is his work on statistical software as early on he realised that software would be important to ensure that the methodology was accessible to applied researchers. His vision for multilevel software became reality in his work with Jon Rasbash on the MLn and MLwiN series of software packages and he continued working with Chris Charlton and Bill Browne at CMM on more recent software developments supporting a large user community.
Although Harvey is incredibly respected in both methodological and applied research spheres he is perhaps equally well known for how he has brought statistical expertise into the public realm. He has written at length on the uses and abuses of league tables both in education and other fields including his seminal 1996 paper with David Spiegelhalter and his more recent work with George Leckie. He has advised many government departments and public bodies including recently the Home Office and NHS Digital and previously the Office for National Statistics and HEFCE.
Harvey also promoted the statistics discipline through many roles at the Royal Statistical Society, most recently (and for the second time) as joint Editor of its flagship journal Series A (“Statistics in Society”).
Harvey will be hugely missed by his many younger colleagues who he continued to support right up to his death. For them he actively encouraged not accepting the status quo and thinking outside the box and he was always extremely approachable and generous with his time. Through them, we hope his legacy will live on.
William Browne and George Leckie
The Royal Statistical Society hosted an online celebration of Harvey in May 2021 and have made a video of the event available.
The early parts of this obituary draw from an interview that Harvey did in 2017 as part of the Pioneers of Social Research project. The full interview transcript is available and offers a rich and detailed insight into Harvey's career and life, presented in his own words.
John Gray, Fiona Steele and others authored a 20 page memoir published in the British Academy.
Peter Mortimore wrote the Guardian obituary.
An overview of Harvey's academic life appeard in a Tes article on 15th April 2020 and provides another interesting read.
An article also appeared in the Bristol Post.
A Tribute and Appreication article by Therese Hopfenback and Sally Thomas appeared in the journal Assessment in Education: Principles, Policy & Practice.
A short obituary also appeared in StatsLife, maganize of the Royal Statistical Society.
Gemma Moss has written a piece on the IOE London blog.
He was also remembered by Institute of Social Science Research, University of Queensland.
A series of academics reflect on their professional and personal memories of Harvey in the Radical Statistics group.
Readers may also find the following 2019 podcast of interest. Harvey gave this postcast as part of the UCL GOS Institute of Child Health Podcast series. In it he again reflects on his varied career.
Professor of Social Statistics
- Fellow of the British Academy (FBA)
- publications list: 1963 - present (PDF, 426kB)
- Publications 1963-present (downloadable)
- Brief CV (PDF, 133kB)
- Follow me on Twitter
- Please note that from December 2018 I will no longer be maintaining this web site regularly. My publications from 2017 onwards and blog commentaries can be found at harveygoldstein.co.uk
Professor Goldstein is a chartered statistician, is currently joint editor of the Royal Statistical Society's Journal, Series A, has been a member of the Society's Council and was awarded the Society's Guy medal on silver in 1998. He was elected a member of the International Statistical Institute in 1987, and a fellow of the British Academy in 1996. He was awarded an honorary doctorate by the Open University in 2001.
There are four main foci of my research interests.
- The first is the use of statistical modelling techniques in the construction and analysis of educational tests. The implications of adopting such models have been explored in a series of papers since 1977. In a number of papers he has also explored the ideas of criterion referenced assessment, comparability of assessments and the interaction of assessment modes and the purposes of assessment.
- The second interest lies in the area of 'educational (school) effectiveness'. He has been involved in a number of longitudinal studies of 'value added indicators' for comparing institutions and the use of such indicators for school improvement purposes.
- The third research interest is in the methodology of multilevel modelling. He has had research funding for this since 1986 and has supervised the production (with Bill Browne and the late Jon Rasbash) of a widely used software package (MLwiN) and made a number of theoretical developments. The major text on multilevel modelling is his book 'Multilevel Statistical Models': (2011) 4th Edition, Wiley, Chichester.
- Finally, in recent years, largely in collaboration with Bill Browne, James Carpenter and Mike kenward I have been dveloping metods for handling missing data values and measurement errors using Bayesian modelling and latent normal transformations.
- I have started a new BLOG commenting on current educational issues. This is at HGeduc.blogspot.com. If you follow me on Twitter (@GoldsteinHarvey) then you will be told when a new commentary is posted.
- The Department for Education has, for 2012, 2013 and 2014 produced more extensive information in the annual school league tables. Unfortunately, despite concerns being raised with them, there are some serious problems with these which can easily lead to misinterpretations. Click here (PDF, 407kB) for a note on this.
- A report produced for the British Academy provides a comprehensive discussion of league tables in the public sector.
- .Recent work with George Leckie has demonstrated how the uncertainty associated with using school league tables for choosing schools, can be illustrated graphically in ways that make it very accessible to users. Click here (PDF, 201kB). In this connection, some recent work by Allen and Burgess appears to provide evidence that league tables can make useful predictions, but on close examination this evidence turns out to be weak and based upon a questionable analysis. Click here for a critique (PDF, 283kB).
- A 2010 paper in the British Educational research journal seeks to resurrect the Rasch (item response) model for educational testing, using rather dubious arguments. A detailed response to this, which the journal has refused to publish without even going through a refereeing process, can be viewed here (PDF, 414kB). It is, however now due to be published in early 2015.
- New research on the uses of school league tables (PDF, 161kB) shows that for purposes of school choice they convey very little useful information.
- One development of the REALCOM project has been to develop a full multilevel multiple imputation procedure that will handle normal as well as categorical data. This procedure has been implemented using an interface between REALCOM and MLwiN. The software and a manual can be downloaded. The methodology is described in a paper on mixed multivariate responses (PDF, 342kB) that has been published in Statistical Modelling. The latest version will now write out chains for parameters selected for monitoring. This may be useful for studying functions of parameters. See the new manual for details.
- Multilevel Statistical Models Third Edition (2003) - latest updates and datasets. An updated and expanded fourth edition, published by Wiley, was published at the end of 2010.
- LGROW is freely available software for calculating longitudinal child growth norms.
- Commentaries on current issues in education 1996-2005. - Using understandings based upon research, these commentaries are responses to official documents and pronouncements about educational policy.
- My published papers 1963-2014
- Brief curriculum vitae. - Basic personal details, professional activities, awards and research projects.
Multilevel Statistical Models- Second Edition (1995)
Multilevel Statistical Models - Third Edition (2003)
- latest updates and datasets. An updated and expanded fourth edition, published by Wiley, was published at the end of 2010.
Some useful web sites
- Education-line electronic database - This is a database for the grey literature in educational research. It has been set up by the British Educational Research Association to make available speedily unpublished papers, read at conferences or work in progress, including critical evaluations of current issues. Serious researchers are invited to contribute. To access it (as part of a database devoted to a wider effort to disseminate social research findings).
- Department for Education (formerly known as the Department for children, schools and families).
- Department for Business Innovation and Skills (formerly known as Department for Innovation, Universities and Skills).
- Office for Standards in Education - OFSTED.
- Qualifications and Curriculum Development Agency.