BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Statistics Reading Group
SUMMARY:David Cox's 1972 proportional hazards paper - Phil
Dawid\, University of Cambridge
DTSTART;TZID=Europe/London:20081029T163000
DTEND;TZID=Europe/London:20081029T173000
UID:TALK14653AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/14653
DESCRIPTION:In March 1972\, David Cox presented his paper "Reg
ression Models and Life Tables" to a meeting of th
e Royal Statistical Society. The paper and the en
suing lively discussion were published in the Jour
nal of the Royal Statistical Society\, Series B (M
ethodological)\, Vol. 34\, No. 2 (1972)\, pp. 187-
220 . Accord
ing to the Web of Science\, this paper has had ove
r 23300 citations\, which is almost certainly a gr
oss underestimate. Most of those citations are fr
om the medical literature\, since what Cox did was
to provide a simple yet extremely flexible soluti
on\, based on his formulation of a "proportional h
azards model"\, to the problem of comparing surviv
al across different individuals while taking full
account of differing treatments\, baseline and tim
e-varying personal characteristics\, withdrawal fr
om follow-up\, etc. In 1990\, Cox won the prestig
ious Kettering Prize and Gold Medal for Cancer Res
earch for "the development of the Proportional Haz
ard Regression Model." The importance of the pape
r as a contribution to Science is thus clear. But
its contributions to statistical methodology were
of equal originality and importance\, and it is t
hose I propose to discuss. \n\n\n\nCox's model was
perhaps one of the earliest examples of a "semipa
rametric model"\, and his method of eliminating th
e nonparametric part through the formation of a "p
artial likelihood" was certainly imaginative. But
his argument for the validity of the method was h
euristic in the extreme\, and it was not until muc
h later that it was set on a firm mathematical bas
is through the application of martingale theory.
It could be argued that its philosophical basis\,
which has much in common with the prequential appr
oach to statistical inference\, is less secure. I
hope others will join me in a thorough-going disc
ussion of the contents and ramifications of this p
athbreaking paper.
LOCATION:MR11\, CMS
CONTACT:Richard Samworth
END:VEVENT
END:VCALENDAR