Professional Documents
Culture Documents
testing
Item Response Theory (IRT)
Prepared By :
1. Logamalar a/p Chegaran 823545
2. Tilagawathy a/p Palasupmaniam 823565
Overview
What we will cover today
-Test bias
-Test Fairness
-Test Accommodations
- Assumptions of Classical Test Theory (CTT).
- Item Response Theory
- Similarities and differences between IRT and CTT
Assumptions of Classical Test Theory (CTT)
1. The error and the true scores from the same test have a
correlation of zero.Hence, the variance of the observed
score is expected to be equal to the sum of the variances of
the true and error score (Lord,1980)
ie Ɣ Te = 0
Sumber : Prof. ‘Dibu Ojerinde, OON ,Joint Admissions and Matriculation Board
(JAMB), Abuja,
Nigeria
Assumptions of Classical Test Theory (CTT)
Sumber : Prof. ‘Dibu Ojerinde, OON ,Joint Admissions and Matriculation Board (JAMB), Abuja,
Nigeria
Assumptions of Classical Test Theory (CTT)
X ║ X1 if X1 = X2 = Ti + Ei
Sumber : Prof. ‘Dibu Ojerinde, OON ,Joint Admissions and Matriculation Board (JAMB), Abuja,
Nigeria
Descriptions of IRT
“IRT refers to a set of This latent variable is
mathematical models that
describe, in probabilistic usually a hypothetical
terms, the relationship construct [trait/domain or
between a person’s response ability] which is postulated
to a survey question/test item
and his or her level of the to exist but cannot be
‘latent variable’ being measured by a single
measured by the scale” observable variable/item.
Fayers and Hays p55
Assessing Quality of Life in
Clinical Trials. Oxford Univ
Press: Instead it is indirectly
Chapter on Applying IRT for measured by using
evaluating questionnaire item multiple items or questions
and scale properties.
in a multi-item test/scale.
7
Assumptions in IRT
• Unidimensionality
– Examinee performance is a single
ability
• Response Dichotomous
– The relationship of examinee
performance on each item and the
ability measured by the test is
described as monotonically
increasing.
• Monotonicity of item performance
and ability is typified in an item
characteristic curve (ICC).
• Examinees with more ability have
higher probabilities for giving
correct answers to items than
lower ability students
(Hambleton, 1989).
• Mathematical model
linking the observable
dichotomously scored
data (item performance)
a b
to the unobservable data
(ability)
c
• P(θ)
i gives the probability
of a correct response
to item i as a function
if ability (θ)
• b is the probability of
b=item difficulty a=item
a correct answer
discrimination (1+c)/2
c=psuedoguessing parameter
• Three items
showing
different item
difficulties (b)
• Two-parameter
model: c=0
• One-parameter
a model: c=0, a=1
b
• Different levels
of item
discrimination
IRT has almost completely replaced CTT as method of choice.
IRT has many advantages ove CTT that have brought IRT into
more frequent use.
IRT allows for greater reliability.
IRT can be used in CAT
IRT allows for difficulty and ability to be on the same scale.
IRT can be analyzed using multi-level modeling.
Arsaythamby Veloo/Rosna Awang Hashim,Teori Ujian dan Pentaksiran Pendidikan,UUM ,Sintok.2016 pg.28
3 basic Compenents of IRT
3. Invariance –
Position on the latent trait can be estimated by the items with know
IRF’s and item characteristic are population independen within
linear tranformation.
Sumber : Psy 427 Cal State Northridge, Andrew Ainsworth, PhD, slides.
Differences between IRT and CTT
Dimension CTT IRT
Definition CTT is a theory about test scores IRT is a general statical theory
that introduces 3 concepts. about examinee item and test
Test score(often calld observed performance and how
score),true score and error score. performance relate to the abilities
that are measured by the items in
the test.
Cttirt1-150715175719-1val-app6891.pdf
Test Bias
Test Fairness
Test Accomodations
TEST BIAS
DEFINITION
http://www.academia.edu/9336249/Academic_Achievement_Test_Bias_and_Fairness
SOURCES OF TEST BIAS
Types of Test Bias
Construct bias
occurs when the construct measured yields significantly different
results for test-takers from the original culture for which the test was
developed and test-takers from a new culture.
DEFINITION
http://www.academia.edu/9336249/Academic_Achievement_Test_Bias_and_Fairness
Test Fairness
About fairness
means the test item should not have any biases. It should not be
offensive to any examinee subgroup.
http://www.academia.edu/9336249/Academic_Achievement_Test_Bias_and_Fairness
Test Accomodations
DEFINITION
Students with Disabilities: Guidelines for Special Test Accommodations, August 2015, p.3
TEST
ACCOMODATIONS
Timing/ Response
setting Presentation
scheduling