Specs Grading Update #3: On Partial Credit, Following Directions, and the Awesomeness of Tokens.

I really thought I’d have little to write about on this subject until the midterm, but it turns out this specs grading experiment is requiring all kinds of reflection and micro-adjustments on my part, and thus I have plenty to share this week.

See my previous posts on converting my research methods course to specs grading, and insights from the first two weeks. Up this week are three reflections following the first few weeks of class and the grading of the first few assignments.

For those new to specs grading, see more general posts here, here, here, and here.

Realization #1: In the past, I was really dependent on partial credit in my grading process. One of the hallmarks of specs grading is no partial credit: student work is either satisfactory or unsatisfactory, and I’m following that policy at both the macro and micro level. That means that not only is an individual assignment graded this way, but each question within an assignment, too. But by not giving partial credit, many students can end up getting an unsatisfactory on the entire first assignment because they don’t completely answer every question. This is in part due to poor assignment design on my part.

See, for this experiment I essentially retro-fitted my pre-existing problem sets into ‘homework modules’, individual concept-based assignments due regularly throughout the course that cover the key concepts and skills in a research methods course. In most cases, I kept the old points system, and just made created thresholds for satisfactory and excellent work. So for example, if a student earns 120/150 points, their grade is ‘Satisfactory’, and 135/150 points is ‘Excellent’. That should work fine—except, without partial credit available, a student has to get the entire question correct to get any credit at all.

For example: my first homework module, Source Analysis I: Basis of Evidence, asks students to assess three pieces on the same subject that vary in their scholarly-ness, quality, and suitability for use in papers. For each source they are given five tasks:

1. Produce an APA citation for the source
2. Identify the source’s central claim
3. Identify the types of evidence used by the author(s)
4. Defend answers in b and c
5. Argue whether or not this source is appropriate to cite in a paper

See the problem? In part d, if they successfully defend the central claim, but either forget to defend their points about the types of evidence, or got c incorrect and therefore would struggle to get d correct…or, or, or–they get zero points! And since this part of the assignment had the highest weight (20/50 points for each source), students might end up with 90/150 points across the three sources even if everything else was correct. With partial credit this isn’t an issue. I could give them 10 points for the part they got right, leading to an overall score of 120/150—a Satisfactory grade. But when using an all-or-nothing scale, I must be very, very careful with question wording and point allocation. Multi-part questions must be broken up into smaller chunks, so students can earn full points on the parts they do attempt or get right. These are good lessons for me, but I regret the potential growing pains for my students.

Realization #2: I also have to consider whether I am really willing to fail a student’s assignment if they do not follow all the directions/specifications. The specifications in every assignment say to follow all directions, but if a student’s substantive work is perfect, can I deny credit because they failed on procedural matter? For example: on the same assignment above, I specifically ask that students embed their answers into the document rather than starting a fresh one (so I can see the questions as we go, making it easier to grade, and to minimize the chances that they miss part or all of a question) and that they list their answers by source, rather than by question. If students fail to do one or the other, then I must decide whether or not to grade it as ‘unsatisfactory’. Since I realized this during the first assignment, I decided to point it out in a class announcement on canvas and in the next class session, but not to fail any assignments on this basis. But this has really tested me in how rigidly I will follow my own rules.

Realization #3: The token system is a lifesaver. Last time I noted that I was having trouble deciding how generous to be with granting tokens for participation. I worried about token hoarding, or using them to forgive a ton of absences. I realized in this first batch of grading, though, that students are going to NEED their tokens, as issues like the ones discussed above could result in students regularly turning in tokens to re-do their assignments. Since every student needs to get a satisfactory on every required homework assignment (21 of them), I can’t risk students failing the course because they don’t have enough tokens to turn in. While my preferred method of awarding tokens is for attending campus events and talks and turning in a summary report, my students—all hardworking but overcommitted to jobs, family, classes, and for the lucky ones, activities)—are realistically only going to be able to earn one or two tokens that way, max. I need to have an influx of tokens into the economy just to ensure they won’t fail the course.

So my adjustment in week 3 is to award two tokens per class session. They can be for any reason I want—a student who said something clever during discussion, or a student who was attentive the entire time, or who helped a classmate with a tricky problem during our in-class exercises. I think erring on the side of generosity is the way to go here, at the very least to reduce the potential for student stress and anxiety over whether they will have enough tokens to be able to re-attempt their assignments. My announcement of the two-per-day rule went over very well (unsurprisingly) and there already seems to be a friendly competition to see who earns the days’ prizes.

Moving forward, my main concern is how much time grading is going to take. The first assignment went very slowly for me, as I still felt the need to provide extensive feedback (in anticipation of re-dos) and justifications for all points awarded/removed. My next report will focus on grading, once I have a few more assignments under my belt.