New target for "unexpected grade"

Most of the targets we currently support are intended to help improve success rates for all students, and define "success" as variations on course completion (continued activity, completion criteria, competencies, etc.) However, there are other legitimate priorities held by some institutions. In general, we can categorise institutions as having one or (usually) more of the following educational priorities: Academic, Pragmatic, Individualist, Idealist. To date, our models have fallen into the "Pragmatic" category.

The "Academic" model of education emphasises that the most academically qualified students should be identified and encouraged to pursue advanced study and careers in scholarship and teaching. This model accepts that student "success" is likely to fall on a normal curve, and that in general, student academic ability is a latent construct that does not change radically over a short period of time. In this kind of model, a rapid change in student performance, for better or for worse, is likely to be evidence of anomaly. A student who suddenly does much worse in academic work may be suffering from unusual circumstances that call for intervention; a student who suddenly exceeds expectations by a wide margin may result in additional scrutiny to verify that the grade accurately reflects the student's own work.

The analysable for this target is the student, and the sample is the graded activity. The prior history of the student over all graded activities is used to predict the grade of an upcoming activity. This prediction is not displayed before the activity is actually graded. Instead, after the activity is graded, a comparison is made between the actual grade and the predicted grade. If the two values differ by a significant amount, an insight is generated to the grader.