BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Ronald Fisher's 1934 likelihood and inference paper - Peter McCull
 agh\, University of Chicago
DTSTART:20081112T163000Z
DTEND:20081112T173000Z
UID:TALK14654@talks.cam.ac.uk
CONTACT:Richard Samworth
DESCRIPTION:The fiducial argument is Fisher's attempt to make inferential 
 statements\nabout the likely value of parameters without recourse to a spe
 cific prior\ndistribution on the parameter space.\nIn its original form us
 ing pivotal quantities\, the fiducial argument is\nvirtually indistinguish
 able from Neyman's theory of confidence intervals.\nFisher strove to dista
 nce his theory from that of confidence intervals\nby emphasizing correctly
  the importance for inference\nof ancillary statistics and recognizable su
 bsets.\nDespite many efforts by Fisher\, Fraser and Barnard and others ove
 r the\nyears\, the passage from a pivotal statistic to a probability distr
 ibution\non the parameter space remains a conceptual stumbling block.\nIn 
 this talk\, I will address the question of whether\, in any circumstances\
 ,\nparametric inference is possible without a prior distribution.\nThe vie
 w initially taken is similar to that of\nHora\, Buehler\, Geisser and Dawi
 d\,\nthat inference and prediction are indistinguishable activities.\nFrom
  this point of view\, the parameter space is relatively unimportant\,\nand
  the observation space becomes the focus of inferential activity.\n\nThe p
 aper is available at:\n\nhttp://www.jstor.org/stable/2935559
LOCATION:MR11\, CMS
END:VEVENT
END:VCALENDAR