Tributes have been paid to the educationalist Harvey Goldstein, who died from coronavirus at the age of 80 over the weekend.

Mr Goldstein was a statistician and social science researcher whose research had questioned the use of school league tables in education, and who had produced renowned work on multi-level modelling. His work also challenged the use of secondary school league tables as an accountability measure, the government’s use of baseline tests in early years, and the use of the Programme for International Student Assessment (Pisa) rankings as a means of comparing educational systems.

His career in statistics spanned many disparate fields, from public health and epidemiology to education. Former colleagues spoke of his warmth and kindness in mentoring younger academics, as well as his keen sense that statistics was not an abstract discipline but one with significant real-world policy applications in the fields of education and public health. 

While he said he always knew that he would attend university, and went on to become a hugely respected researcher in the field of education, Mr Goldstein’s parents had a limited education themselves.

He was born on 30 October 1939 in Whitechapel, in a Jewish part of East London, “within the sound of Bow Bells”, as he described in an interview in 2017. His father worked with his uncle as a button manufacturer, and later became an unskilled engineer, although Mr Goldstein said he “had no real qualifications of any kind”, while his mother, who died when he was a young child, worked as a hat-maker. 

Mr Goldstein lived with his grandparents when his father enlisted in the army in the Second World War, and the family later moved to Edmonton, North London. Aged 11, he passed the exam to attend Hendon County School, a grammar. He later said that he had always assumed he would attend university - his maternal uncle had won a scholarship, later becoming a headteacher, so there was “great emphasis” on education within his family. 

Having attained four A grades at A level in pure maths, applied maths, physics and chemistry, Mr Goldstein studied at the University of Manchester, where he pursued additional classes in statistics with Toby Lewis, an anarchist lecturer who marched for nuclear disarmament. Former colleagues have spoken of Mr Goldstein’s strong interest in how data could be applied to real issues of social equity. While he was less interested in the pure maths components of his degree, he was fascinated by how statistics could be used to improve society.

Some of his first studies, as a research assistant at University College London, included surveys of living conditions in then poverty-stricken Notting Hill, West London, in the 1960s. Mr Goldstein later said of collecting data: “I saw that here was a way of putting into practice social ideals.” He would frequently remind colleagues that the Royal Statistical Society, of which he was a member, had been founded in 1832 by a group concerned with the alleviation of poverty.

Having completed a diploma in Statistics at UCL, Mr Goldstein worked as a lecturer at the Institute of Child Health on child development studies, such as statistical modelling of children’s height. At UCL, he had worked for two years with Sir Richard Doll, the pioneering researcher who linked smoking with cancer.

Mr Goldstein developed this work in finding links between smoking in pregnancy and perinatal mortality rates, despite the fact that population-level studies in Finland had shown no link. His research showed that the mortality rate for infants whose mothers had smoked was most pronounced for disadvantaged children, as these babies would have had poor nutrition and hence a lower birth weight, which was strongly correlated with infant mortality. The experience taught him that evidence must not simply be derived from large groups but from heterogeneous, diverse groups as well, which was also to influence his work in education. 

In the 1970s, he became a senior research officer at the National Children’s Bureau, continuing to work on child development. He was interested in the underlying theory behind data sets, which could simply be based on prejudice - a concern which informed his research across his career. At the NCB, he found that questionnaires assessing children’s behaviour were based on no particular theory at all apart from one clinician’s judgement, and therefore could be biased.

Criticism of school league tables

His concern regarding nuances that could be lost in simplistic data sets made him critical of both the use of school league tables as an accountability measure, and the use of Pisa as an international ranking through which to compare the effectiveness of education systems in different countries.

Of Pisa, he said comparisons were partly motivated by “big bucks” and that “governments use it to either justify what they’re doing or to criticise what the previous lot did”. In his view, education systems are so fundamentally a part of their country’s culture that the problems of translation, as well as the cultural assumptions of the tests, make “absolute comparisons” impossible. 

He was also highly critical of the use of school league tables as a performance measure. While he worked closely with Hampshire council on school improvement, he felt that the “statistical uncertainty” of school league tables in measuring performance was far too high, even if the contextual value added of schools’ social contexts was taken into account, and especially if it was not.

The raw data of GCSE results as a means of producing rankings was meaningless, he argued.

One former colleague, Pamela Sammons, a professor of education at the University of Oxford, recalled: “Harvey always argued that statistical models need to be complex in order to be accurate - they should model the complexity of reality. If you use a simple model and get it wrong, that is very dangerous.”

Anna Vignoles, a professor of education at the University of Cambridge, added: “What he did was he mounted a robust challenge to inappropriate use of school league tables. Lots of people have criticised them but he demonstrated accurately and scientifically how they could mislead and what their weaknesses were.”

In 1977, Mr Goldstein moved to the Institute of Education as a professor of statistical methods, a post he held until his retirement in 2005 at the age of 65. His work on school effectiveness and assessment sparked his interest in multi-level modelling - essentially exploring the ways that data sets are clustered together. He argued that pupils in a school will share certain characteristics through attending the same school, but statistical modelling needs to reflect that a subset of these pupils may share other characteristics by belonging to the same neighbourhood or area. He founded the Centre for Multilevel Modelling whilst at the IOE. 

“It [multi-level modelling] recognises that students are not individuals floating around on their own, lots of common things will happen to one class, one school - you can’t treat those individuals as though they are completely independent of one another,” said Ms Vignoles. 

Mr Goldstein was awarded the Royal Statistical Society’s Guy Medal in Silver in 1988 and was made a Fellow of the British Academy in 1996.

Former friends have spoken of Mr Goldstein’s remarkable activity and productivity into old age, describing how, while he never appeared harried or stressed and had “time for everyone”, he also seemed to produce more work than most people. After retirement, he worked part-time as a professor of statistics at both the University of Bristol and the Institute of Child Health for 15 years. 

He also held visiting professorships at the University of East Anglia, the London School of Hygiene and Tropical Medicine, the Institute of Fiscal Studies and the Australian Catholic University in Brisbane.

George Leckie, who worked with Mr Goldstein on school league tables as a professor of social statistics at the University of Bristol, described both his extraordinary energy and his kindness to younger colleagues. 

“He was great - very unusually approachable for a senior academic, very generous with his time, with very good intuition on tricky research questions.”

Mr Leckie said Mr Goldstein was actively involved in teaching, sharing his modelling with peers and lobbying government. “Where some academics might shy away from that, he enjoyed using statistics to challenge policy, to formulate policy. He spanned the social sciences, medical statistics - he could turn his hand to many things,” he said.

And, at the end of days teaching modelling courses, he would play flute in an orchestra, and was also teaching himself French. 

Many recalled his generosity with his time and his willingness to help students, even if they were not his own. Ms Sammons said, “He was very generous. He wouldn’t always be looking to be first place [the first-named academic of a research paper]. He was generous with authorship and generous with his time. His interest in social justice shone through his work.”

Writing on Twitter, Ms Vignoles described Mr Goldstein as a “legend” who was “so very kind to me when I first started at IOE”.

Katie Harron, who worked with Mr Goldstein at the London School of Hygiene and Tropical Medicine at UCL, said: “When I started my Phd and I knew that I would be working with Harvey, he was such a big name that I was quite daunted and expected him to be stereotypical scary professor, but he was very approachable, friendly and hands-on.

“He always had time for me and for many other junior colleagues and colleagues across the board, and he was generally very supportive and encouraging. I have very fond memories of my early work with him. He started out as a supervisor and then became a friend, colleague and general inspiration.”

On a personal level, Mr Goldstein was deeply concerned by social inequity and, more recently, the environment. He had shifted to a vegetarian diet, and his interests included cycling in Norfolk with his wife Barbara, with whom he published a book on the subject. Ms Sammons commented that he was “good at a catchy title” and had also helped to come up with the name for their recent publication for the British Educational Research Association on baseline tests, A baseline without basis.

“We will greatly miss Harvey as a colleague and friend.  He made an enormous contribution to social science and education,” she said.