Quantifying phonological knowledge in children with phonological disorder.
Philip N Combiths, Jessica A Barlow, Emilie Sanchez
PMID 31379215WHAT IT FOUND
For preschoolers with phonological disorder, a consonant accuracy cutoff of 20.4% on a large 293-word probe correctly classifies 90.0% of sounds as present or absent in the child’s phonemic inventory.
This allows clinicians to estimate functional phonological knowledge from standard accuracy scores without manual minimal-pair analysis.
Key findings
01A 20.4% consonant accuracy cutoff on the 293-item Phonological Knowledge Probe correctly classified the phonemic status of 90.0% of English consonants across all 275 children.
02Consonants that were 'in' the children's phonemic inventories had a mean accuracy of 74.0%, whereas consonants that were 'out' had a mean accuracy of 12.0%.
03The accuracy cutoff was less effective when calculated from shorter standardized tests (44-53 words), with classification accuracy dropping compared to the larger probe.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used archival data, which may have introduced some inconsistencies in transcription or sampling. The findings are most robust for preschool-aged children; there were only 57 school-aged children in the sample, limiting conclusions for older groups. The method works best for later-acquired sounds. For early-acquired sounds, most children with phonological disorder already produce them accurately and use them, creating a 'ceiling effect' that makes accuracy scores less predictive of functional use. The high classification accuracy (90.0%) relied on a large 293-word probe. Using shorter, standardised tests (44-53 words) resulted in lower accuracy and reliability. The cutoff of 20.4% had higher sensitivity than specificity, meaning it was better at identifying sounds that *are* in the inventory than at identifying sounds that are *not*.
Declared interests
Funded by NIH NIDCD grant F31 DC017697-01. The authors declared no other conflicts of interest.
The easy way to misread this
Do not apply the 20.4% accuracy cutoff to short standardised tests like the GFTA-2. The high classification accuracy was only achieved with the large 293-word research probe; using the shorter test significantly reduces reliability. Additionally, do not use this method to assess early-acquired sounds (like /m/, /b/, /d/) or older school-aged children, as the relationship between accuracy and phonemic use is much weaker in these groups.