Time, Year-end Grades and SBG

In an earlier post I wrote about how I felt that I tend to move slowly through curriculum.  One of the things I do that slows things down is frequent quizzing and post-quiz self/group assessment.  Usually once every 5 classes (or less) we will have a quiz that can take anywhere from 10 minutes to 25 minutes.  Once everyone is finished, the quizzes are handed back to the students and we go over the solutions. Students make corrections and ask questions as we go along.  This whole process can take 40 minutes of class time.  Adding up all the quizzes, it’s a big chunk of time.

There’s reason to think that these quizzes are good use of time - it is a retrieval practice and there is timely formative feedback to me and the students.   However, in my Math 8 classes I’ve been doing something a bit different.  Every couple of days I give the students a very short, fast quiz.  There will be 4 to 6 questions covering topics that go back to the beginning of the year. They get about 10 minutes to do the quiz and then we quickly go over it.  We don’t use quiz dividers and we’re not recording the results. The purpose is fast retrieval practice with interleaved topics.  It is very effective.

These two different types of quizzing have their pros and cons.  For pros, the SBG quizzes have more in-depth questions on a particular topic and allow for tracking progress on the topic whereas the quick quizzes interleave topics and give relatively fast feedback.  SBG quizzes also allow students to focus on specific skills that require work.  For cons, the SBG quizzes take up a lot of class time and often do not synthesize topics.  The quick quizzes have a con in that they aren’t useful for tracking progress because each different topic only touches one or two partial aspects of the learning objective.  For example, on a SBG quiz on adding integers I will ask several questions to cover an array of situations or variations but on a quick quiz there will only be one or two questions. From this quick quiz we can’t infer much about student progress on a learning objective of adding integers if the only evidence we have is from 2 questions.

Adding to this mix, I continue to struggle with finding the best way to produce a summative grade.  For several years I’ve mainly been relying on converting SBG progress to an overall grade.  This makes it easy for students to track their progress and gives them many opportunities to improve.  The downside to this is that the grade doesn’t necessarily reflect overall performance because the learning objectives are done in isolation from each other.  To improve on this I’ve used various “performance tasks” that require the students to pull from several different skills, bodies of knowledge and competencies.  Once again though, time has been my bogey man.  A proper performance task requires 1 to 2 days every 4 to 6 weeks.  As a result, I’m often skipping the performance task in order to keep my classes moving along and the student overall grades are mostly made up from the SBG progress.

I believe there are other unintended consequences from pulling an overall grade from SBG and quizzes.  First, the overall grade is not generated in a way that is shared with the rest of the educational community.  In other words, the overall grades I create may have little in common with overall grades of a more traditional classroom and they represent something different from what others, like university admissions people, expect.  Second, I think I’m seeing an increase in anxiety in students in writing tests because their expectations from SBG quizzes are not necessarily predictive on how they will do on tests.  A student can make good progress in an SBG system where things are chunked out nicely for them but they don’t get much practice in synthesizing separate topics in a higher stakes assessment.

My big question at this point is knowing which is more effective or better for student learning on day to day or month to month basis, SBG quizzing or quick quizzes?  If the answer is quick quizzes, does this mean that all grading returns to the performance task? If a grade is created mostly from performance tasks, can we honor the SBG idea of allowing for constant improvement?

Here are some possible schemes for tracking progress and grades.

Formative Summative Comments
1. Quick quizzes, self tracked, 0% of grade 3-5 tests/tasks, 100% of grade, re-testable Fast assessment. Puts more responsibility on students.
2. SBG, self-tracked, 0% of grade 3-5 tests/tasks, 100% of grade, re-testable Students link their formative assessment to specific learning objectives. Puts more responsibility on students.
3. SBG, tracked in gradebook, 50% of grade 3-5 tests/tasks, 50% of grade, not re-testable Provides more structure for learning improvements Higher stakes testing.
4. SBG, tracked in gradebook, 75% of grade 3-ish tests/tasks, 25% of grade, not re-testable Provides more structure for learning improvements Lower stakes testing. Closest to what I do now, takes a lot of time. Final grade has issues with validity.

There is also the idea of using a performance task as an example of student work to include in a student portfolio or report card, as opposed to reducing the task to a grade. I like this idea because it solves the problem of developing a shared meaning of where the student is at.  By including a piece of student work along with an exemplar of similar work, anyone that reads the report card should have an understanding on the student’s progress.  IIn the end though, we are required to report an overall grade whether it’s on a 100 point scale, a 4 point scale, or anything in between.

FWIW, the performance tasks I am using are shifting more towards individual tests or even 2-stage tests (individual + group).  While I have a couple of lab-like performance tasks that are really authentic, interesting and challenging, they cannot be done individually.  Since the performance task is used to make overall/summative inferences, I don’t think it’s fair to draw this information from a task that is almost all group work.  Furthermore, if I want to allow for students to improve their learning and have their grade always reflect their most recent understandings, it is very hard to use one-off lab performance tasks because they cannot be replicated by individuals and I cannot come up with enough variations to allow for re-assessments.