From mathematics and science to classroom management: an IRT validation of a teaching confidence scale in a sample of primary special education teachers

Introduction Teaching confidence is a critical determinant of instructional quality, teacher attitudes, and student outcomes, particularly in inclusive and special education contexts. This study validated a multidimensional Teaching Confidence Scale for primary special education teachers, encompassing Mathematics/Science Instruction, Instructional Innovation, and Classroom Management. Methods Using an Item Response Theory (IRT) framework, we examined the scale’s dimensional structure and psychometric properties in a sample of 305 primary special education teachers working in inclusive settings. Results Confirmatory factor analysis supported a robust three-factor model with excellent fit ( RMSEA = 0.027, CFI = 0.994, TLI = 0.956). Standardized factor loadings ranged from 0.62 to 0.91 across domains, indicating strong item–factor relationships. Reliability was consistently high, with Cronbach’s α values ranging from 0.83 to 0.97 and McDonald’s ω from 0.92 to 0.98 (AVE = 0.50–0.77). IRT analyses using the Generalized Partial Credit Model indicated satisfactory item fit (RMSEA < 0.05), moderate to high discrimination parameters particularly for Mathematics/Science Instruction and Instructional Innovation and minimal differential item functioning across teacher groups. Discussion Overall, the findings support the Teaching Confidence Scale as a reliable and valid instrument for assessing domain-specific teaching confidence in primary special education, with implications for teacher professional development, inclusive practice, and future research.