Going Deeper into the CLASS Measure: Language Modeling

In my last post, I shared some of my personal struggles to master the CLASS measure and promised to take you on a “deeper dive” into some of the trickier CLASS concepts I encountered in my CLASS journey.

We’ll start with conversations—what we might consider the vehicle for all of the other Instructional Support dimensions.

Overall, the focus of Language Modeling (LM) is for teachers to model more complex language and to increase the amount of talking kids do. So when observing and coding for LM, I always ask myself these key questions first:

1. How successful was the teacher at getting kids talking? Not that much? Kind of? Very?
2. Was there a balance of teacher and child talk?
3. Was it obvious that the teacher was intentionally modeling more advanced, more complex language for the children?
4. Were the back-and-forth exchanges I observed really conversations or just strings of questions and answers (i.e., question-answer-question-answer-question-answer)?

It is important to remember that conversations are dialogues—they include not only questions and responses but actual sharing of information between two persons. The “contingent responding” behavioral marker underscores that teachers’ responses and follow-up questions should be relevant to what the child has said. Conversationally, children often don’t give us a lot to go on, so we can easily fall into the trap of skipping the responding part and moving straight on to another follow-up question, which results in the question-answer-question-answer type exchange, not an authentic conversation. I call it “peppering a child with questions”– something no child really enjoys. Often times, children will clam up or walk away rather than engage; they feel like they’re being quizzed.

One coach I recently spoke with said she encourages teachers to converse with children “the same way you would while getting your nails done with a friend.” Different content, obviously, but the natural, dialogic flow of the conversation should feel just as comfortable. This means that teachers need to be encouraged to “just talk with” children, rather than talking “to” them or “at” them.

When coding the dimension of Language Modeling, coders need to analyze and evaluate the form, frequency, and the effectiveness of teacher-child conversations. By effectiveness, I mean “how much talking did the children do?” It should be at least as much as the teacher did.

Remember that a question-answer string will not count as a high-range conversation, as the sharing of information (contingent responding) is missing, often resulting in very short responses on the part of the child. With high-range conversations, “there is a natural flow of information during center times or other open periods of time that encourages students to converse and makes them feel as if they are valued conversational partners” (Pre-K CLASS Manual, p. 79).

Stay tuned, in my next blog I will invite you to dig deeper to consider another Instructional Support indicator, this time from the Quality of Feedback dimension.

We are excited to have Sara Beach guest blog. As a former Teachstone Staff Trainer, she frequently presented on topics such as Helping Teachers with the Instructional Supports, through active, adult-learning approaches. She has been an Infant-toddler teacher, center director, education specialist, coach-mentor, and early childhood college instructor, and her highest honor has been supporting teachers.