MKTG966 Measurement Data Analysis in Marketing Spring 2005 Prof. Eric T. Bradlow

1
MKTG966 Measurement Data Analysis in Marketing
Spring 2005Prof. Eric T. Bradlow

• Paper Presentation by
• Matthew Conger Sergio Luiz Fernandes Ka-Chuen
  Hui

2
Latent Variable Models for Teratogenesis using
Multiple Binary OutcomesJulie M. Legler
Louise M. Ryan Journal of the American
Statistical Association, 92(437) 13-20, March
1997
Julie M. Legler Director, Statistics Department
of Mathematics, Statistics, and Computer
Science St. Olaf College Northfield, MN
Louise M. Ryan Professor of Biostatistics Departme
nt of Biostatistics Harvard School of Public
Health Boston, MA
3
Epileptic Women Require Anticonvulsant Drug
Therapy During PregnancyHowever
Broad Nasal Bridge BNB
Clinodactyly CLINRL
Hypoplastic toenails and fingernails TNHYP, FNHP
4
Item Response Theory (IRT) orLatent Trait Theory
(LTT) or Probabilistic Test Theory (PTT)

• A visual guide to item response theory
• Ivailo Partchev (Friedrich-Schiller-Universitaet
  Jena)
• http//www2.uni-jena.de/svw/metheval/irt.php
• One-parameter Logistic (1-PL) Model or Rasch
  Model (1960)
• http//www2.uni-jena.de/svw/metheval/irt/ptb.html
• http//www2.uni-jena.de/svw/metheval/irt/trf1pl.ht
  ml
• Two-parameter Logistic (2-PL) Model or Birnbaum
  Model (1968)
• http//www2.uni-jena.de/svw/metheval/irt/irf3pl.ht
  ml
• http//www2.uni-jena.de/svw/metheval/irt/trf2pl.ht
  ml
• Three-parameter Logistic (3-PL) Model
• http//www2.uni-jena.de/svw/metheval/irt/irf3pl.ht
  ml
• http//www2.uni-jena.de/svw/metheval/irt/trf3pl.ht
  ml

5
Latent Variable Model for Multiple Outcomes
6
Conclusions