<p>Test-retest reliability of the EEAS total score.</p>

EFA, guided by eigenvalues and the scree plot, provided initial support for a unidimensional structure, with a single factor accounting for 32.7% of the variance. Rasch analysis supported essential unidimensionality (first residual eigenvalue = 1.90). Nine of 10 items showed acceptable infit MNSQ values (0.64–1.34); Q8 (telephone) showed elevated infit (1.92) and outfit (1.83) values, indicating misfit, which may reflect the unique and multidimensional cognitive demands of technology-mediated communication in this population.

Person-item targeting remained within acceptable clinical limits, though items were slightly easier for the average person’s ability (person mean = 0.49 logits). The EEAS-K demonstrated excellent internal consistency (α = 0.886), moderate to strong convergent validity with the VAS (r = 0.611), and excellent test-reliability (ICC = 0.91). DIF analysis identified significant gender-based DIF for Q7 and age-based DIF for Q7 and Q8, suggesting raw scores for these specific items should be interpreted with caution across these demographics.

No significant DIF was found for hearing-loss-severity.</p> <p>Conclusion</p><p>The EEAS-K is a reliable and valid tool for assessing listening effort in Kannada-speaking adults with hearing loss. The Wright map demonstrates adequate person-item targeting, and the dual-method psychometric approach provides satisfactory measurement evidence supporting clinical and research use.</p></div>