Is putting SUGAR (Sampling utterances of grammatical analysis revised) into language sample analysis a good thing? A response to Pavelko and Owens (2017)

Purpose: In this letter, the authors respond to Pavelko and Owens’ (2017) newly advanced set of procedures for language sample analysis: Sampling Utterances and Grammatical Analysis Revised (SUGAR). Method: The authors contrast some of the new guidelines for transcription, morpheme segmentation, and language sample elicitation in SUGAR with traditional conventions for language sample analysis (LSA). They address the potential impact of the new guidelines on some of the target measures in SUGAR—mean length of utterances in morphemes (MLU_m), words per sentence (WPS), and clauses per sentence (CPS)—and provide their suggestions. Results: Inclusion of partially intelligible utterances in SUGAR may over-or underestimate children’s MLU_m and reduce the reliability of computing WPS. Counting derivational morphemes and the component morphemes of catenatives (e.g., gonna) may result in overestimation of children’s morphosyntactic skills. Conclusion: Further data are needed to determine whether MLU_m including derivational morphemes and the component morphemes of catenatives is a better measure of children’s morphosyntactic skills than MLU_m excluding those morphemes. Pending such data, the authors recommend maintaining traditional LSA conventions and measures. Furthermore, free, fast automated utilities already exist that reduce barriers for clinicians to conduct informative, in-depth LSA.