The four-parameter normal ogive model with response times

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In recent years, interest in the four-parameter logistic (4PL) model (Barton and Lord, ETS Res Rep Ser 198(1):i-8, 1981), and its normal ogive equivalent, has been renewed (Culpepper, Psychometrika, 81(4):1142–1163, 2016; Feuerstahler and Waller (Multivar Behav Res 49(3):285–285, 2014)). The defining feature of this model is the inclusion of an upper asymptote parameter, in addition to those included in the more common three-parameter logistic (3PL) model. The use of the slipping parameter has come into contact with many assessment applications, such as high-stakes testing (Loken and Rulison, Br J Math Stat Psychol 63(3):509–525, 2010), low-stakes testing (Culpepper, Psychometrika, 81(4):1142–1163, 2016), and measuring psychopathology (Waller and Reise, Measuring psychopathology with nonstandard item response theory models: Fitting the four-parameter model to the Minnesota Multiphasic Personality Inventory, 2010). Yet as mentioned in Culpepper (Psychometrika, 81(4):1142–1163, 2016), the recovery of the slipping parameter also requires larger sample sizes and longer iterations for the sampling algorithm to converge. Response time (RT), which has already been widely utilized to study student behaviors, such as rapid-guessing, was included in our model to help recover the slipping parameter and the overall measurement accuracy. Based on the hierarchical framework of response and RT (van der Linden, Psychometrika 72(3):287–308, 2007), we extended the four-parameter normal ogive model by incorporating RT into the model formulation. A Gibbs sampling approach to estimation was developed and investigated.

Original languageEnglish (US)
Title of host publicationQuantitative Psychology - 84th Annual Meeting of the Psychometric Society, IMPS 2019
EditorsMarie Wiberg, Dylan Molenaar, Jorge González, Ulf Böckenholt, Jee-Seon Kim
PublisherSpringer
Pages55-67
Number of pages13
ISBN (Print)9783030434687
DOIs
StatePublished - 2020
Event84th Annual Meeting of the Psychometric Society, IMPS 2019 - Santiago, Chile
Duration: Jul 15 2019Jul 19 2019

Publication series

NameSpringer Proceedings in Mathematics and Statistics
Volume322
ISSN (Print)2194-1009
ISSN (Electronic)2194-1017

Conference

Conference84th Annual Meeting of the Psychometric Society, IMPS 2019
CountryChile
CitySantiago
Period7/15/197/19/19

Keywords

  • Four-parameter normal ogive model
  • Hierarchical framework of response and response time
  • Response times
  • Slipping

ASJC Scopus subject areas

  • Mathematics(all)

Fingerprint Dive into the research topics of 'The four-parameter normal ogive model with response times'. Together they form a unique fingerprint.

Cite this