Tom Sgouros keeps dragging me into the psychometric weeds and was trying to explain the concept of "theta" in psychometrics last week. I ended up reading some of this paper on Psychometric Principles in Student Assessment, including this paragraph:
The assessment design framework provides a way of thinking about psychometrics that relates what we observe to what we infer. The models of the evidence-centered design framework are illustrated in Figure 1. The student model, at the far left, concerns what we want to say about what a student knows or can do—aspects of their knowledge or skill. Following a tradition in psychometrics, we label this “θ” (theta). This label may stand for something rather simple, like a single category of knowledge such as vocabulary usage, or something much more complex, like a set of variables that concern which strategies a student can bring to bear on mixed-number subtraction problems and under what conditions she uses which ones. The task model, at the far right, concerns the situations we can set up in the world, in which we will observe the student say or do something that gives us clues about the knowledge or skill we’ve built into the student model. Between the student and task model are the scoring model and the measurement model, through which we reason from what we observe in performances to what we infer about a student.
This gives me a little more language to describe my most basic reaction to the Common Core ELA. I think the most fundamental design principle in the CC is to collapse the task model and student model as much as possible.
Picking one example quickly, here's 2.2b from the English (that is for England) 2007 Programme of Study for Key Stage 4:
Students should be able to understand how meaning is constructed within sentences and across texts as a whole.
You can see that a broad range of tasks could address this student model. You can also see that it would be easy to end up with a task model which does not completely cover to the theoretical student model.
The equivalent in the Common Core would be:
Analyze how an author’s choices concerning how to structure a text, order events within it (e.g., parallel plots), and manipulate time (e.g., pacing, flashbacks) create such effects as mystery, tension, or surprise.
If this is the student model, and the task model is "the situations we can set up in the world, in which we will observe the student say or do something that gives us clues about the knowledge or skill we’ve built into the student model," how much difference is there between the two here?
To be clear, I don't think my observation here is controversial. However, I haven't seen it discussed -- because there has been little serious analysis of the structure or design of the Common Core ELA. My complaint is that as a result, the student model is way to narrow, specific and incomplete to represent what we really want students to know and be able to do.
No comments:
Post a Comment