Remember SUGAR? It’s the new, alternative language sample analysis protocol meant to work within the realities of a busy SLP’s workload. It’s been a while, so here’s a quick recap: SUGAR involves calculating four metrics on a 50-utterance sample where you only transcribe child utterances:
Mean length of utteranceSUGAR (MLUS)*
Total number of words (TNW)
Clauses per sentence (CPS)
Words per sentence (WPS)
For specifics and examples, check out the complete procedures (including videos) on their website.
While the creators of SUGAR have provided some support for its validity, the diagnostic accuracy of the four measures hasn’t been tested—until now! In this new study, the authors recruited 36 3- to 7-year-old children with DLD (currently receiving or referred to services) and 206 with typical language, and used the SUGAR protocol to sample their language. All four measures showed acceptable sensitivity and specificity (above 80%), using research-based cutoff scores (see the paper for specifics on cutoffs for each measure). The most accurate classification, according to the authors, was achieved with a combination of MLUS and CPS.
One of SUGAR’s big selling points is that it’s quick (like, 20 minutes quick), at least for kids with typical language. Did that still hold for the children with DLD? Actually, in this study they took less time to provide a 50-utterance sample than their typical peers. Bonus!
Language sampling can be daunting for the full-caseload SLP, but we love that research like this is identifying promising LSA measures that have high diagnostic accuracy (higher, we might add, than many commercially-available tests), while addressing our time and resource barriers.
An important note: there are many methodological differences between SUGAR and other LSA procedures, and SUGAR has not been uncontroversial. We’ll be on the lookout for more research on SUGAR’s diagnostic potential or comparing SUGAR to more traditional protocols to help us really understand the pros and cons of the different LSA methods.
*When calculating MLUS, derivational morphemes (-tion) are counted separately and catenatives (hafta, wanna) count as two morphemes.
Pavelko, S. L., & Owens Jr, R. E. (2019). Diagnostic Accuracy of the Sampling Utterances and Grammatical Analysis Revised (SUGAR) Measures for Identifying Children With Language Impairment. Language, Speech, and Hearing Services in Schools. doi:10.1044/2018_LSHSS-18-0050