BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Optimal Design for Item Response Theory Models - Professor Heinz H
 olling\, University of Muenster\, Germany
DTSTART:20181018T150000Z
DTEND:20181018T160000Z
UID:TALK111412@talks.cam.ac.uk
CONTACT:Professor John Rust
DESCRIPTION:Optimal design allows for estimating parameters of statistical
  models according to important optimality criteria\, e. g.\, minimizing st
 andard errors of estimators. Thus\, optimal designs may considerably reduc
 e the number of experimental units\, such as respondents or items in empir
 ical studies. For a long time\, optimal design has not received much atten
 tion within psychology\, but meanwhile interest for this subject is rapidl
 y increasing as such designs are needed\, e. g.\, in large scale assessmen
 t\, adaptive testing or automatic item generation.\nIn this presentation\,
  first\, fundamental principles of optimal design are introduced using wel
 l-known linear models\, e. g. analysis of variance or simple regression. T
 he rationale of adaptive\, Bayesian\, and minimax designs needed for nonli
 near models will then be outlined. Such designs are presented for Item Res
 ponse Theory (IRT) models\, e.g.\, 1Pl and 2PL model or linear logistic mo
 del. Finally\, two R packages for deriving Bayesian and minimax designs ba
 sed on recently developed algorithms will briefly be demonstrated.\n
LOCATION:Room W4.05\, Cambridge Judge Business School\, Trumpington Street
END:VEVENT
END:VCALENDAR