Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Understanding Rasch measurement: estimation methods for Rasch measures.

J M Linacre1

  • 1MESA Psychometric Laboratory, University of Chicago, IL 60637, USA. mesa@uchicago.edu

Journal of Outcome Measurement
|November 26, 1999
PubMed
Summary

Rasch parameter estimation methods, both non-iterative and iterative, yield statistically equivalent results. Careful comparison of estimates across different software programs is essential for accurate analysis.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

New approaches to determining reliability and validity.

Research quarterly for exercise and sport·2015
Same author

Capturing the true burden of dystonia on patients: the Cervical Dystonia Impact Profile (CDIP-58).

Neurology·2004
Same author

FIM levels as ordinal categories.

Journal of outcome measurement·2001
Same author

New approaches to determining reliability and validity.

Research quarterly for exercise and sport·2000
Same author

Investigating rating scale category utility.

Journal of outcome measurement·1999
Same author

Detecting multidimensionality: which residual data-type works best?

Journal of outcome measurement·1998

Area of Science:

  • Psychometrics
  • Statistical Modeling

Background:

  • Rasch model parameter estimation involves various iterative and non-iterative techniques.
  • Understanding these methods is crucial for accurate measurement in psychometrics.

Purpose of the Study:

  • To classify and describe different Rasch parameter estimation methods.
  • To highlight the characteristics of parameter estimates, including standard errors and fit statistics.

Main Methods:

  • Non-iterative methods: normal approximation algorithm (PROX).
  • Iterative methods: datum-by-datum (e.g., PAIR), marginal (e.g., CMLE, JMLE, MMLE).
  • Classification based on distributional assumptions and data handling (complete vs. missing).

Main Results:

  • Five computer programs implementing diverse estimation methods produce statistically equivalent estimates.
  • Parameter estimates are characterized by standard errors (local/general, ideal/inflated) and fit statistics.
  • Different estimation approaches exist, including those for complete and missing data.

Conclusions:

  • Despite methodological diversity, Rasch parameter estimation methods converge to equivalent results.
  • Careful consideration is needed when comparing estimates derived from different software packages.

Related Experiment Videos