Measuring speech sound development: An item response model approach

Research aim: The primary aim of our study is to investigate if there is an ordering in the speech sound development of children aged 3-6, similar to the ordering in general language development. Method: The speech sound development of 1035 children was tested with a revised version of Logo-Articulation Assessment. The data were analyzed with the Mokken Scale Program (MSP) in order to construct scales with satisfactory scalability (H-coefficient) and sufficient reliability (rho). Results: The majority of children over 4.3 years of age turned out to have mastered most speech sounds. An ordering was only found in the youngest age group (3.8-4.3 years of age), for the sounds of /r/ in initial and final position and /s/ in initial position. This resulted in a set of scales. The scales developed for /r/ (in initial and final position) and /s/ were moderately scalable (H > 0.43) and reliable (rho > 0.83), and independent of gender. Moreover, we found variation in the judgment of speech sound development, which may perhaps have been due to where exactly the examiner was positioned during the assessment procedure: in front of the child, or sitting beside the child. Conclusions: We could not detect an ordering for all speech sounds. We only found an ordering for In in initial and final position and /s/ in initial position. In the Mokken analysis we conducted, these scales turned out to be moderately strong and reliable. Our research also underlines that speech sound development is judged not only in an auditory sense, but judgment also depends on the visual interpretation of the listener. (C) 2013 Elsevier Ireland Ltd. All rights reserved