"How are cut scores for proficiency determined?”I am assuming that he is making reference to cut-off scores such as the 40th percentile cut-off mark for "proficient" on standardized tests for the purposes of reporting AYP.
Mark's response follows.
"....But…..the dirty little secret to setting proficiency levels is still rooted in the question, “How many do we want to fail?'"I think Mark would agree that we (educators) don't want any students to fail. His point is obviously to answer with a question that continues the conversation in his classroom. I'd like to take it one step further as it relates to standards-based grading/reporting. As I mentioned in an earlier post, I am in the midst of transitioning to a 1-4 point scale for the purpose of reporting out student understanding of learning targets. My current scale looks like this:
4 – demonstrates thorough understandingScale meets letter grade
3.5 – high level of understanding, but with small errors
3 – demonstrates understanding, but with significant gaps
2 – shows some understanding, but insufficient for a passing grade
1 – Attempts the problem
Until now, I had not thought about the implications of this quasi-rubric matching up with my overall grading scale of A: 90 and up; B: 80 and up; C: 70 and up; D: 60 and up, F 59 and below. Breaking the 1-4 scale down into percentages and then assigning a letter grade makes for an interesting discussion.Let's assume a student is being assessed on two learning targets. On one they demonstrate a high level of understanding (4) and on the second they show some understanding, but insufficient for a passing grade (2). Assuming all learning targets are weighted equally, as they are now in my grade book, this student has earned a "C" or 6/8 learning target points. I subsequently report out that they have a "75%" level of understanding. By mastering one learning target and failing another, should a student receive a "C"?
I wonder if a better way of reporting out student learning might be by the number of learning targets they've mastered. I'm thinking about a hybrid system that might require students to have a "4" level of understanding on at least 90% of the learning targets in order to get an "A." The percentage here is up for discussion, but if that system is a step in the right direction, what would an appropriate break down for various letter grades look like? Should there also be a required number of "3" or "3.5" learning targets in order to earn a given grade? In other words, should a student who has demonstrated a high level of understanding (4) on 90% of the learning targets, but has significant gaps on the rest (3) earn a different letter grade than a student who has also mastered 90% of the learning targets, but has some small errors (most likely computation in my math course; 3.5) on the remainder earn the same overall letter grade?
I can think of many more scenarios similar to the one I just described. A quick fix is eliminating letter grades, but unfortunately that's not an option for me as public school teacher who is required to use a student information system. The bigger question worth discussing is what makes an "A"?