MARTIN GARDNER could take the terror out of statistics. It was this quality combined with his outstanding mathematical ability that made him one of the most distinguished medical statisticians of his generation.
Doctors have recognised increasingly that many of the scientific studies published in their journals are statistically flawed. And this matters - because the treatments offered to patients are based on the results of those studies. If the statistics are misguided then the treatment may seem to be effective when it is not. Most medical studies depend on statistical interpretation because of the great variability in disease and in treatments: results may or may not be due to chance and only statistical analysis can tell.
Gardner, who was appointed Professor of Medical Statistics at Southampton University in 1985, helped at every possible stage in the process of raising the quality of statistics in medical studies. He taught medical undergraduates and postgraduates and could show them what was really important. Too many doctors imagined that statistics was a matter of finding the right test and using a computer to apply it. Gardner emphasised constantly that it was the statistical design of the study that was most important.
Hundreds of doctors and medical students were helped by Gardner with the design and analysis of their studies. It was in his nature to concentrate on the strengths in their studies and build on those while calmly removing the weaknesses. He also played a crucial part in raising the quality of statistics in medical journals by leading the team of statisticians who advised the British Medical Journal. Invited to help the journal over a decade ago, Gardner became more and more central to the process of judging the 4,000 papers a year submitted to the journal.
In the early days the medical editors were in awe of the statisticians and thought wrongly that statistics, being mathematically based, was an exact science. Gardner taught us to be more sophisticated consumers of statistical advice, and eventually he joined us in our weekly discussions on which papers to publish. The outside world might imagine that this meeting would be a serious affair, but in fact it is lively, irreverent, and huge fun. Gardner raised the scientific standards of the meeting and fitted in perfectly.
The broader world knew Gardner best for his work on the cases of leukaemia seen around the Sellafield nuclear plant in west Cumbria. He was invited to join the committee investigating the excess cancers because of his considerable experience in occupational hazards and because of universal respect for his statistical abilities and his integrity.
The investigating committee confirmed in 1984 that there was an excess of leukaemia around Sellafield, and Gardner then compared the histories of those who had leukaemia with the histories of controls to try and identify the cause. The committee knew that it couldn't simply be environmental contamination with radiation because not enough radiation had been released. Gardner's study - the results of which echoed right round the world in 1990 - suggested that at least part of the cause was paternal exposure to radiation before conception of the child who subsequently developed leukaemia. This hypothesis - known to all as the Gardner Hypothesis - is biologically fascinating as well as hugely important for public policy. Sadly, Gardner died in the middle of the scientific debate of his hypothesis.
Although Martin Gardner did work of outstanding international importance, his first commitment was to his family. He met his wife, Linda, when he was a teacher at the University of California in 1966-67, and they had three children after returning to Britain. He loved sport and had enough ability that he seriously contemplated becoming a professional footballer. If he had, then perhaps Southampton would have won more trophies but medical statistics would still be in the dark ages.