Speech intelligibility tested by the Pediatric Matrix Sentence test in 3–6 year old children

ObjectiveThe present study was aimed at the development and application of the Polish Pediatric Matrix Sentence Test (PPMST) for testing speech intelligibility in normal-hearing (NH) and hearing-impaired (HI) children aged 3–6.Methods & ProceduresThe test was based on sentences of the subject-verb-object pattern. Pictures illustrating PPMST utterances were prepared and the picture-point (PP) method was used for administering the 1-up/1-down adaptive procedure converging the signal to noise ratio (SNR) to the speech reception threshold (SRT). The correctness of verbal responses (VR), preceding PP responses, was also judged.Outcomes & ResultsThe normative SRT for the PP method was shown to decrease with age. The guessing rate (γ) turned out to be close to the theoretical value for forced-choice procedures, γ = 1/n, where n = 6 for the six-alternative PP method (γ ≈ 0.166) and n = 4 for the four-alternative PP method (γ ≈ 0.25). Test optimization resulted in minimizing the lapse rate (λ) (ratio γ/λ ≈ 8.0 for n = 4 and γ/λ ≈ 5.6 for n = 6, both for NH and HI children). Significantly higher SRTs were observed for HI children than for the NH group.Conclusions & ImplicationsFor children aged 3–6, tested by the developed PPMST, speech intelligibility performance, for both the VR and PP method, increases with age. For each age group, significantly worse intelligibility was observed for HI children than for NH children. The PPMST combined with the PP method is a reliable tool for pediatric speech intelligibility measurements.

► Test is based on sentences of the subject-verb-object pattern.
► Correctness of verbal and picture-point responses was judged.
► For children aged 3–6, normative speech reception thresholds were shown to decrease with age.
► Results obtained for hearing impaired children were significantly higher than the normative data.