Toby Lewis at 100
On 19 October 2018, in a West London care home, Toby Lewis turned 100. Physically fit, if a bit frail, he looks and sounds as anyone who had not seen him for 40 years would recognise instantly. He greets you with the same polite, gentlemanly air that always masked a certain eccentricity, even anarchism, that made him such an interesting and lively companion.
Together with a sharp intelligence and a total lack of deference towards authority, he inspired many of us eagerly to await his contributions to statistical seminars where he would often ask pertinent questions that illuminated otherwise opaque deliveries.
I first met Toby in 1959 while studying maths as an undergraduate in Manchester and wondering what I really wanted to do with my life. I remember the day we were due for a statistics lecture and Toby appeared, sporting a campaign for nuclear disarmament badge, which instantly decided me. We became friends and for a short time were colleagues at University College London, before he went to a chair at Hull in 1968.
Toby’s family – his wife, Catherine, and two daughters, Rae and Gina – all enjoyed the change from Greenwich, and here Toby developed his lifetime interest (fittingly) in statistical outliers and fully pursued his fruitful collaboration with Vic Barnett, which had started in Manchester when Vic was a PhD student, and resulted in what remains one of the standard texts in this area.1
Toby was highly gifted from the start. Immediately before the outbreak of World War II, he left Oxford with a first-class degree in mathematics. During the war, he spent time attached to the Australian army, partly in Borneo, and at the end, having become “Major Lewis”, he spent three years working on what he described as “unexciting operational research”. It was during a spell at the National Physical Laboratory with Edgar Feiller after the war, followed by work in the statistical advisory unit of the Ministry of Supply, that he transformed himself into a statistician. He had married by then and his big career opportunity came in 1957 when he was recruited by Maurice Bartlett as a lecturer in statistics in Manchester. Then Maurice moved to UCL to replace Egon Pearson, and Toby followed him in 1962.
When he moved to Hull in 1968 to take charge of a new department of statistics, he quickly joined the cultural scene by moving into a grand house next door to Philip Larkin, whom he would consistently defend as “not really being a misogynist”.
Moving to the Open University in 1979 to establish a new statistics group, which quickly gained a high academic reputation, was an exciting time for Toby and he really enjoyed the whole concept of distance teaching and making video programmes. His anarchist tendency made him a natural enemy of bureaucracy, and he found a splendid opportunity to mock it when, having already been well established in his post, the University Senate was then asked to approve his appointment: Toby voted against!
It was during this time that he took up the case of a Bristol maths student who he regarded as having been treated unfairly by the university establishment on suspicion of cheating. What really outraged him was the way they had utilised statistical evidence from the pattern of his examination answers in order to downgrade his degree, which Toby demonstrated was wholly unjustifiable. The case he put forward on behalf of the student was eventually lost, at least partly – as he pointed out – because the student had been assumed guilty and was then required to prove his innocence! The episode testifies to his keen sense of justice and his determination never to give up on a cause he believed in.
He retired in 1986 and with Catherine moved to Diss in Norfolk, quite near Great Yarmouth – his place of birth. He remained active by teaching and researching at the University of East Anglia, where he had an honorary professorship that lasted until he was in his early nineties. He worked with the Royal Statistical Society (RSS) in organising public outreach conferences, and actively campaigned to get the British Association for the Advancement of Science (BAAS; now the British Science Association) to understand that statistics could be a great motivating factor in persuading young people to take an interest in maths. Together with Debra Hurcomb at the RSS, he did achieve some success in getting statistics featured at the BAAS’s annual conferences.
Toby also began to spend some time in our multilevel modelling centre at the Institute of Education, one result of which was a jointly edited book on quantitative educational assessment in 1996,2 as well as a paper about engine design in 2006 that introduced multilevel modelling to the engineering profession.3 A particularly noteworthy achievement was his prowess at croquet, where often he belonged to the winning partnership at the Centre’s annual tournament (see below)!
ABOVE Toby Lewis and Michael Healy: multilevel croquet champions, 2004.
A few years after Catherine died in 2007, Toby decided to move back to London to be nearer to his daughters and, until very recently, he lived independently in a sheltered apartment.
Toby isn’t very active now. His life revolves around the daily routine of a friendly care home, and regular visits from Gina and Rae. Just as he always did, he enjoys the life he has, and welcomes visitors with all his familiar courtesy. It has been a privilege to know him, to have shared many of the things he cared about and enjoyed, and to have had an example to try and emulate.
About the author
Harvey Goldstein is professor of social statistics at the University of Bristol Centre for Multilevel Modelling. He 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.
I am grateful to the following for supplying helpful materials and useful comments: John Bibby, Barbara Goldstein, Rae Lewis, Gina Lewis and Kevin McConway.
- Barnett, V. and Lewis, T. (1994). Outliers in statistical data. Chichester, Wiley. 3rd edition. ^
- Goldstein H & Lewis T. (1996). Assessment: Problems Developments and Statistical Issues. April, J Wiley & Sons, pp. 265. ^
- Goldstein, H & Lewis, T. (2006). Holliday’s engine mapping experiment revisited: a design problem with 2-level repeated measures data. Int. J. Vehicle design, 40, 317-326. ^
Murray Aitkin writes:
I was touched by Harvey Goldstein's stories of Toby Lewis in the December 2018 issue of Significance [published above]. I had several delightful encounters with Toby, which I would like to record for Significance.
Toby was visiting Sydney in 1973 on the tour of distinguished statisticians which Joe Gani had organised following his appointment to the CSIRO. He gave a seminar at UNSW which I attended, titled "Breweries, bottles and other stochastic phenomena", a typically light-hearted title. I introduced myself afterwards and we had a long discussion. I thought, "What a charming man!"
On 4 December 1973, the head of statistics at Macquarie University, where I then worked, called the statistics discipline staff together for a meeting. He had a major announcement: Maurice Bartlett was being considered for a special five-year appointment to a chair of statistics in the school.
Bartlett was retiring from the chair of biomathematics at the University of Oxford, and had previously taught at Cambridge, Manchester and University College London. There had already been discussions about a new chair of statistics, and the need for restructuring of our courses.
Bartlett arrived, and met the statistics staff. We were told later that the university was not appointing him. Later in December, the university advertised formally a chair of statistics in the School of Economic and Financial Studies. The wording of the advertisement was quite unusual. It explicitly ruled out candidates interested in other areas (like theoretical or applied probability); they had to fit into the existing structure:
"It is not intended that the establishment of this Chair should result in the University entering new areas of study, but rather that the professor should provide further leadership in the areas at present being developed. These areas are: Actuarial Studies; Demography; General Applied Statistics, Operational Research, Simulation; Mathematical Statistics. It is intended that the successful applicant should be able to guide teaching and research in at least one of the above applied areas.
"In addition to formal qualifications in mathematical statistics, the successful applicant must have a background in the application of statistics, which will enable him [no "or her"] to lead the teaching along present lines, which are based on the idea of equal emphasis on theory and applications. Applicants should indicate the extent of their training and experience in the above-mentioned areas of study, or in any other fields which they regard as relevant to the present teaching and research activities within the statistics area. The University reserves the right to fill any position by invitation."
Before the closing date for applications, I had an unexpected overseas call from Toby Lewis. The conversation went something like this:
"Toby, how nice to hear from you!"
"Murray, I saw the advertisement from your university for the chair of statistics. Are you applying?"
"Toby, I could not get promotion to Associate Professor. There was no point in my applying."
"Oh good! Well, I'm applying!"
I was amazed. "That's wonderful news," I said.
Events moved slowly after this. Toby was flown over in July 1975 for interview, and met the staff at a staff meeting. I was not present. Before the interview he asked me to tell him what was happening in the discipline. I went over to his hotel and spent more than an hour describing the situation. Toby listened without comment.
He was offered the post. He took some time to negotiate details with the university. He called me again, and said: "Murray, I have decided not to accept the Macquarie post."
I was devastated. "Oh God, why?" I asked. "I hope you weren't put off by my description of the situation."
"No," Toby said. "We could not reach agreement on my leave and travel."
This was a great disappointment. The chair was not filled for another year.
A year later I arrived in Lancaster for a three-year SSRC Professorial Fellowship. I contacted Toby, and soon after gave a seminar at Hull and enjoyed again Toby's stimulating discussions and contributions. We met from time to time at RSS meetings, always with happiness and laughter.
When I left the UK I lost contact with Toby, and was delighted to hear from Harvey's article in Significance that he was alive and well, and unchanged in personality.