Abstract
In the present study, the item response theory (IRT) based psychometric validation technique was used to assess
the measurement properties of Teacher Attitude towards Teaching (TAT) scale for 153 secondary school teachers
of the district SBS Nagar, Punjab. This instrument began with 15 Likert-type items with five ordered response
categories, which were analysed by Samejima's Graded Response Model (GRM). The original unconstrained
GRM resulted in scores between 0.082 and 1.901, and four items (I1, I4, I7 and I13) were eliminated for poor
discrimination, leaving 11 items. The retained items gave a range of discrimination values from 0.356 to 1.999
(M ≈ 0.907) with I10, I9, and I15 being the best discriminators. Likelihood-ratio test (LRT(10)) was applied to
determine whether the unconstrained model was more likely than the constrained model, and the result was that
the unconstrained model was more likely, LRT(10) = 28.52, p = .001, and item discrimination was found not to
be uniform. The Item Information Curves illustrated complementary coverage and the Test Information Function
demonstrated the greatest precision in the lower to middle portion of the trait continuum with less precision at the
high end of the trait continuum. The reliability of this instrument was measured using ordinal alpha (0.689) and
McDonald's omega (0.690) which resulted in moderate but acceptable reliability for a first validation step. The
results provide some initial validation of the 11-item TAT instrument and serve to identify the need for additional
validation of the instrument with larger, independent samples.
the measurement properties of Teacher Attitude towards Teaching (TAT) scale for 153 secondary school teachers
of the district SBS Nagar, Punjab. This instrument began with 15 Likert-type items with five ordered response
categories, which were analysed by Samejima's Graded Response Model (GRM). The original unconstrained
GRM resulted in scores between 0.082 and 1.901, and four items (I1, I4, I7 and I13) were eliminated for poor
discrimination, leaving 11 items. The retained items gave a range of discrimination values from 0.356 to 1.999
(M ≈ 0.907) with I10, I9, and I15 being the best discriminators. Likelihood-ratio test (LRT(10)) was applied to
determine whether the unconstrained model was more likely than the constrained model, and the result was that
the unconstrained model was more likely, LRT(10) = 28.52, p = .001, and item discrimination was found not to
be uniform. The Item Information Curves illustrated complementary coverage and the Test Information Function
demonstrated the greatest precision in the lower to middle portion of the trait continuum with less precision at the
high end of the trait continuum. The reliability of this instrument was measured using ordinal alpha (0.689) and
McDonald's omega (0.690) which resulted in moderate but acceptable reliability for a first validation step. The
results provide some initial validation of the 11-item TAT instrument and serve to identify the need for additional
validation of the instrument with larger, independent samples.
Keywords
Attitude Scale towards Teaching, Teaching
References
Mamta, Vijay Kumar Chechi. (2026). VALIDATION OF THE ATTITUDE TOWARDS TEACHING SCALE USING ITEM RESPONSE THEORY: A STUDY AMONG SECONDARY SCHOOL TEACHERS OF SBS NAGAR DISTRICT (PUNJAB). https://doi.org/10.5281/zenodo.23417849