Clinical Tutorial Throwback (2018): How to measure syntax in early intervention (hint: MLU alone won’t cut it)

"Describing sentences in terms of utterance length may be appropriate for communicating expectations with non- language specialists, but we believe it is crucial for speech–language pathologists to differentiate word combinations from sentences in order to recognize early difficulty with the acquisition of sentence structure…and to develop intervention plans to facilitate grammar for children with or at-risk for language disorders." 

(Hadley et al., 2018)

Picture it: You’re an EI therapist, and you’ve recently evaluated two two-year-olds. In addition to a standardized assessment, you also collected a language sample to evaluate their expressive language skills. Here’s a sample utterance produced by each child:

Child 1:  I see you!

Child 2:  Dog run.

At first glance, it seems logical to assume that Child 1 has more advanced language skills, right? He or she is producing 3 word utterances, while Child 2 is only producing two. And, utterance length is something that we as SLPs should be looking at when we evaluate young children’s language development. But, focusing only on utterance length will most likely result in an overestimation of a child’s language abilitiesThis recent article serves as a good reminder that we should be extending our language analysis beyond utterance length to include syntax. And, the best part is, the authors explain a quick and easy way to analyze grammatical structures in young children, along with developmental expectations that can be used to identify intervention targets.
The method of analysis that the authors recommend is based first on distinguishing between a word combination and a sentence. What’s the difference? Well, we know that the hallmark of age two is 2-word combinations, right? These are utterances that have an MLU of between 1.5 and 2. Shortly after toddlers begin producing these word combinations, they begin to produce true sentences which include Subject-Verb-Object (SVO) (“I want that.”) and Subject-Verb (SV) constructions (“baby sleep”).
Now, what about “diversity” of those sentences? Sentence diversity has to do with how common or frequently used the words are in a child’s utterance. Put another way, you’re counting the number of different or unique SV combinations that a child produces. So, if we go back to the example of Child 1 and Child 2, the first child’s use of the subject “I” is much more common than the subject “dog”. And, “see you”, although a two-word combination, is most likely a phrase that the child hears and produces often, whereas the verb “run” is most likely less commonly noted in the child’s typical utterances. The authors provide fantastic supplementary materials here that include references and specific procedures for calculating sentence diversity.*
And, while this is all great information, it kinda seems like it’s adding time to the already lengthy process of language sample analysis, right? Well, here’s some good news: the authors provide evidence to suggest that a pared down, quicker version of LSA (what they refer to as, “structure-specific language sampling”) is the way to go if you want to measure sentence diversity. Basically, you’re focusing on and transcribing only one specific grammatical structure within a language sample. And, as an EI therapist, the methods that they recommend for collecting the sample make a lot of sense. For instance, the authors suggest using a parent–child language sample because unstructured conversations provide the most authentic measure of a child’s language skills. (*More information regarding how to collect the sample is included in the supplementary materials.) The best part? As long as you’re consistent with the amount of time that you use to collect the samples, repeated measures of sentence diversity can be used to monitor progress and determine targets for intervention.

All in all, the information in this article seems like it could really be a game-changer for EI therapists who are working with children transitioning from word combinations to more complex utterances. Beyond providing a more specific description of young children’s language skills, this method of analysis may allow SLPs to identify children who will have trouble moving from use of more basic, routine sentence constructions to use of more complex grammatical forms—i.e., kids who may eventually fit the diagnosis of Developmental Language Disorder (DLD).

*The authors are careful to explain that while an assessment of sentence diversity is useful for selecting intervention targets and monitoring progress, a more comprehensive analysis of grammar (such as the Index of Productive Syntax, or IPSN) is needed to identify the severity of a language disorder. One of the 3 supplemental materials included in this article is a table that lists items from the IPSN and how they relate to understanding of sentence structure.

Hadley, P. A., McKenna, M. M., & Rispoli, M. (2018). Sentence diversity in early language development: Recommendations for target selection and progress monitoring. American Journal of Speech-Language Pathology. doi: 10.1044/2017_AJSLP-17-0098.