Scribd
Upload
64 views

Rasch Model: January 2010

The Rasch model is used to create interval-level measurement scales and assess how well data fits the model's criteria for measurement. It estimates the amount of the underlying trait revealed by each person's ability and item difficulty on a logit scale. The model requires that item difficulties and person abilities be estimated independently and produce invariant measures across samples. It models the probability of a correct response as a logistic function of the difference between ability and difficulty. Extensions include the rating scale model for Likert scales and the many-faceted Rasch model for performances judged by raters.

Uploaded by

Raj'z Kingzter
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
64 views

Rasch Model: January 2010

The Rasch model is used to create interval-level measurement scales and assess how well data fits the model's criteria for measurement. It estimates the amount of the underlying trait revealed by each person's ability and item difficulty on a logit scale. The model requires that item difficulties and person abilities be estimated independently and produce invariant measures across samples. It models the probability of a correct response as a logistic function of the difference between ability and difficulty. Extensions include the rating scale model for Likert scales and the many-faceted Rasch model for performances judged by raters.

Uploaded by

Raj'z Kingzter
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/314045679

Rasch Model

Chapter · January 2010


DOI: 10.1002/9780470479216.corpsy0773

CITATION READS

1 3,377

1 author:

Trevor G Bond
James Cook University
94 PUBLICATIONS   7,373 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Piagetian research in early educational settings. View project

Piaget: Formal Operations; Theory, Practice & Empirical Evidence View project

All content following this page was uploaded by Trevor G Bond on 15 October 2017.

The user has requested enhancement of the downloaded file.


Rasch Model : 499 words; 3 references; 3 suggested readings; 15 key words.
The Rasch Model
Proponents of the Rasch Model (RM, Rasch, 1960) model claim that it is distinctive in terms
of its focus on the production of interval-level measurement (Andrich, 1988). There is an
important emphasis that tests and questionnaires should produce data that fit the model as the
RM sets out the criteria for successful measurement. From that perspective, the Rasch model
is indispensable for the construction of invariant measurement scales and the monitoring of
their performance across samples and over time; e.g., the calibration of large item banks. RM
is widely adopted in educational and psychological testing as well as for rehabilitation and
quality of life indicators in health settings.

For all items and persons in a data matrix the Rasch model estimates how much of the
underlying latent trait is revealed in each person ability and item difficulty, estimated along a
logit (log odds unit) scale that is common to items and persons. The total score (N correct
responses for ability; N persons correct for item difficulty) is the sufficient statistic for
estimating Rasch measures; i.e., total score is the sufficient statistic for person ability as it
contains the complete information about that ability. Additionally, fit statistics are used as
quality control mechanisms to determine which test items should be added together to
produce total scores; misfitting items/persons should be put to one side for later
consideration. Indeed, the requirements of the Rasch model are often seen as the explicit
statement of the implied conditions of any technique which uses total N correct as a summary
statistic, based on those performances which should/not be counted, and the transformation of
those counts to an interval measurement scale.

The RM feature of parameter separation supports the model’s goal of specific objectivity:
item/person estimates are calculated independently of the distribution of those difficulties /
abilities in the item / person sample. The consequence of specific objectivity is the
requirement for invariant measures: item difficulties should remain the same (within error)
across all appropriate samples, and person estimates should not vary according to the choice
of items in a test. Lack of invariance, say, revealed as DIF (differential item functioning)
should prompt diagnostic consideration of item/person performances.

In the RM for dichotomous data, the probability of any correct response is modeled as a
logistic function of the difference between person ability and item difficulty, each expressed
in logits (log odds units): higher ability persons are more likely to succeed on all items; less
difficult questions will be more easily responded to by all persons; and the order of the item
difficulties remains the same for all persons. For polytomous data (Wright & Masters, 1982),
the Rating Scale Model (RSM) predominates in the analysis of Likert-style data, while Partial
Credit Rasch Model (PCM) allows response options to vary across items. The many-facets
Rasch model (MFRM) provides for the estimation of additional facet(s), such as rater severity
when judges are used to score persons on items according to graded criteria (e.g., essay
marking, performance certification.)

References
Andrich, D. (1988). Rasch models for measurement. Newbury Park, CA: Sage.
Rasch, G. (1960/1980). Probabilistic models for some intelligence and attainment tests
(Expanded ed.). Chicago: University of Chicago Press.
Wright, B.D. & Masters, G.N. (1982). Rating Scale Analysis: Rasch Measurement. Chicago:
MESA press.

Key Words: Rasch model, measurement, invariance, probability, test, DIF, parameter, scores,
item response theory (IRT), scaling, unidimensional, rating scale, Likert scale, judge.

Some suggested readings:


Online: http://en.wikipedia.org/wiki/Rasch_model
Article:
Pallant, J. F. & Tennant, A. (2007) An introduction to the Rasch measurement model: An
example using the Hospital Anxiety and Depression Scale (HADS), British Journal of
Clinical Psychology, 46(1), pp. 1-18.
Book:
Bond, T.G. & Fox, C. M. (2007). Applying the Rasch model: Fundamental measurement in
the human sciences. (2nd ed.) Mahwah, N.J.: Erlbaum.

View publication stats

You might also like