Previously, I talked about some exciting work that was being done in the area of dynamic assessment. This work together is looking at how to apply dynamic assessment to ELLs. We have our own contribution to make as well. The results give SLPs another potential strategy to use to determine whether children have language difference or language impairment. What’s really cool about this is that it works with children who know just a little English.
Our study (Peña, Gillam & Bedore, 2014), focuses on data from our Diagnostic Markers of Language Impermanent in Bilingual Children, funded by the NIH. In this analysis, we selected the 18 children with language impairment and matched them to 18 other children who had typical development. Our matching criteria included age, sex, non-verbal IQ, age of first exposure to English, and current amount of exposure. We added another set of typical controls, but this time we matched only on amount of exposure and age. We wanted to test whether the cut-offs for the well-matched groups also worked for another set of children who were not as carefully matched. We reasoned that this would tell us more about how DA might work in a clinical setting.
We used our Dynamic Assessment of Narratives Intervention (DANI), which is published by Pro-Ed. Children first told a story (Two Friends) using the wordless picture book found in the DANI kit. Then, we did 2-30 minute interventions with each child in English. We did make some changes to the intervention, and the script is included as a supplement to the article.Basically, we used a comprehensive intervention– one in which we address learning story grammar and story components. Then, children told a second story (Bird and His Ring) from the DANI.
The pretest and posttest stories were rated on story grammar, story ideas and language (complexity and creativity), and story components (setting time & place, character information, temporal markers, etc.). After the intervention, the examiner rated modifiability (affective and cognitive strategies used during intervention).
We ran 4 discriminant analyses to see which set of predictors best differentiated children. Our best classification was more than 95% accurate 100% sensitivity and 94.4% specificity. For this classification we found that a combination of modifiability (compliance, metacognition, and task orientation), DA story scores (setting, dialogue, and complexity of vocabulary) and ungrammaticality (based on the posttest narrative sample) had the highest accuracy. In the paper, we provide a formula that can be used to predict from this combination of scores whether children likely fall in the impaired range or in the normal range.
Based on this paper, and our clinical experience I want to highlight things to remember if you use DA with ELLs:
- There were some kids who just couldn’t tell the story independently in English. We eliminated these children from the study. At this point we don’t know whether DA in English would work for them or not (certainly the story scores couldn’t be used).
- Observation of modifiability contributed a lot of the variance to the model. What this means is that observations that clinicians make can really help us to distinguish language impairment and language difference. Kids who were responsive, flexible, aware of what was going on– even though they might not be completely accurate– were those who more likely had language differences. Kids who clearly had a hard time paying attention, persisted in using unproductive strategies, or who didn’t seem to engage were more likely to have language impairment. Note that I don’t say anything about accuracy here– only about interaction during the intervention.
- With respect to accuracy– the story telling situation for this group of ELL kindergartners was very very challenging. The stories were short and they made lots of grammatical errors. But, the children with LI made EVEN more errors than those without LI. So, once you account for ESL errors, are sentences short? are elements omitted? is vocabulary highly restricted (given their experience with English)?
So, here’s another piece to the puzzle and another way that SLPs can assess children who are second language learners. While this study was exclusively about Spanish-English bilinguals, we think that this approach can also be used for children who speak other language pairs.