Significant figures
Sir: In defending herself against Stephen Glover's criticism of her report in the Observer on the recently reported relation- ship between the consumption of hamburg- ers and the occurrence of non-Hodgkins lymphoma, Caroline Richmond (Letters, 31 August) seems to confuse statistical signifi- cance with practical or theoretical signifi- cance.
The statistical significance of the researchers' evidence indicates that there is only a 5 per cent likelihood that the find- ings are a fortuitous, chance happenstance, but it tells us absolutely nothing about the importance of the findings, contrary to Miss Richmond's interpretation. The more frequent occurrence of the disease among hamburger eaters might be only slightly and trivially greater than that among the non- hamburger eaters and be of no practical significance to anyone, but yet could be sta- tistically significant. Such a state of affairs is not unlikely in large-scale studies such as this one, in which some 35,000 people are said to have participated.
(I cannot comment on Miss Richmond's contention that 'doctors are no longer san- guine about giving blood', but I have noticed lately that mine seems to be pissed off at taking urine samples.) Max Prola
The Croft, Ditchfield Lane, High Legh, Knutsford, Cheshire
Previous page
