Specs Grading Update #4: On poor syllabus design, late work, and the woes of grading

It’s time for another specifications grading update! These posts are my attempt to be highly transparent about the benefits and challenges of using this method in my research methods class this semester, with all the trials and tribulations and the reflection and changes they prompt here on display.  Check out Posts 1, 2, and 3 if you haven’t already, or for a deep dive into the ins-and-outs of specifications grading as a system, take a look at our extensive discussions on the subject over the last year.

Today’s topics: requiring so much that I set up my students to fail; dealing with late work; and how all that grading is going. In other words, let’s talk about how even extensive reflection and consideration can result in basic syllabus mistakes that pose unacceptable challenges to students.

Let’s start with this basic question: Why in the world did I require 21 assignments? Yes, pedagogically speaking, this made sense: these assignments, collectively, added up the basic knowledge required of an introductory methods course. They covered topics such as writing research questions and hypotheses to measurement, ethics, sampling, correlation v. causation, and everything else. Back in the summer, I spent a lot of time considering whether to require everything, or allow students to complete some smaller number or percentage of the total.

This student just failed my class due to poor syllabus-making on my part.
This student just failed my class due to poor syllabus-making on my part.

Here’s the problem: if a student completes twenty assignments with a satisfactory or higher score, but misses assignment #21 because they forgot or overslept or had a family emergency, then according to my syllabus they fail the course outright. Sure, the tokens help with this, letting students get a 24 hour extension per token, up to 72 hours—but what if they don’t have enough tokens in the bank, or they completely forget for a week or more? These students FAIL THE COURSE.

Upon reflection, this is of course unacceptable. I realized by week four that I needed more flexibility in both the short and long term. In the short term, I reduced the requirement from 21/21 assignments to 18/21 of all students at the B bundle level and below. A bundle students still had to complete 21/21. All students regardless of bundle are also now allowed to use a token to eliminate a missing assignment from their record. They can do this once during the semester. To say my students were grateful for this change is a massive understatement.

I also made all the optional homework assignments due by the last day of the course, so that students can make up what they need to after evaluating where they stand on all homework assignments, rather than having to do them throughout the course.

In the long term, for the next version of the course, I need further changes. There are a few options. One is to reduce the number of required assignments so there are fewer duedates. Having work constantly due for the class makes sense pedagogically, but is difficult for students to keep track of. A few days of illness or distraction can spell disaster for even the most hardworking students, and the current set up is, I think, too stressful for them. By moving the optional assignments to the end of the semester, I hope to help with this in the short term, but next semester I may combine the homework modules and have them due weekly on the same day so they are easier to track

Option two is to let them use as many tokens as necessary to make up for the lateness of the assignment, as opposed to the three-day rule. This would at the very least alleviate the panic of forgetting an assignment.

A third option is a somewhat more dramatic version of option 2: make the homework module due dates be ‘suggested’ but not actually due until the end of the class. After this semester, every assignment and its specs will be complete, so I could actually upload them on the first day of class and let students complete them at their own pace. There are obvious downsides here: some students may procrastinate and not do any assignments until the end of the semester, for example, and there is a risk of plagiarism if I turn back graded work for some students when others have not turned theirs in. I’m not sure if those concerns outweigh the potential benefits, though.

What about late work? Here’s another issue: students that do turn their assignment in on time but get an unsatisfactory have just one shot to try again, by turning in a token for a revision attempt. If they get it wrong a second time, the rule is that they have to complete an optional assignment to make up for the original required assignment. This means that they do not have to keep working on mastering the original concept. The gaps in the knowledge I originally deemed ‘required’ will continue. One way to solve this is to provide optional assignments that synthesize concepts together—so they have more chances to practice and master the original skills.

Finally, let’s talk grading. The first few weeks of the class, grading still took a lot of time. I left each student substantial individualized comments justifying their grade. Now, things are a lot smoother. Some specifications require less time on my part to assess—such as those asking students to correctly answer an objective question, or to make a credible attempt at completing answering a subjective question. The lack of partial grading on individual items helps a lot. Each part o the question is graded separately, so I don’t have to think about how many points out of 10, for example, a student earns. Instead, they earn x number of points for a correct answer, and zero for an incorrect answer.

For example, take this recent assignment on hypotheses. Here are the specifications:

To earn a ‘Satisfactory’ grade on this module, you must do the following:

  1. Turn the assignment in on time (10am CST 9/27/16) into World Classroom assignment marked ‘Hypotheses” in a readable .doc, .docx or .pdf or audio format. Re-Upload this document with answers embedded in the existing text, or if submitted as an audio response, clearly state the question as written before giving your answer.
  • Follow all course policies as well as all directions in the assignment.
  • Completely answer at least three of the below questions (that is, attempt to identify the independent and dependent variables and to write a hypothesis)
  • Earn 12 points (80% of required)

To earn an ‘Excellent’ grade on this module, you must complete all the requirements for Satisfactory, plus:

  • Completely answer two additional questions (total of five)
  • Earn 22 points (88% of required)

And here is the assignment. They were given six research questions to tackle:

For each of the following research questions,

  1. Identify the dependent variable (1 point)
  2. Generate a possible independent variable. (1)

Note: Your IV must logically be able to impact the dependent variable without a long explanation. You do not have to state your logic as part of your answer, but it may help you develop your own thoughts and let me see what you are thinking.

3. Write a hypothesis suggesting a relationship between the two variables, following all the rules of good hypotheses. (3)

An incorrect answer on (c) does not eliminate the points a student might earn on (a) and (b). In the past—meaning earlier in this semester—I would have said that each question was worth five points, and if they got even one part wrong, they got 0/5. Now, they can earn points for the parts they get right, which is both a fairer system and one that makes grading easier.

I can’t really express the joy of not worrying about points and how to allocate them throughout the semester. In the past I made my courses worth 1000 or 10000 points, and then allocated points to different assignments. I then struggled with how to allocate points within assignments, as the points really mattered. I would ask myself: Should this question be worth 1 point? 5? 10? What if the points I use add up to some weird number like 132, because I can only think of 33 multiple-choice questions? What if on one assignment the point total ends up being 200 and another 500 due to arbitrary reasons? These questions no longer plague me.

On 'Whose Line is it Anyway?
On ‘Whose Line is it Anyway?” as in my class, everything is made up and the points don’t matter.

With specs grading, the points just don’t matter. I can make one assignment worth 20 points and the next 1000 points—but all points are INTERNAL, and the only thing that matters is the percent of available points that a student earns within that assignment. So 18/20 on assignment 1 and 900/1000 on assignment two mean exactly the same thing: an Excellent. Same with 16/20 and 800/1000. That’s satisfactory. It could be out of 300,000 points and it would be the same—it goes into the gradebook as satisfactory, unsatisfactory, or excellent. So the only balancing is internal to the assignment and relating each question to each other. In the past, I would have to remember that writing a solid hypothesis was worth 4 points, no matter what assignment it appeared in, as making it worth more or less when it was basically the same skill felt wrong. I don’t have to do that anymore—it can be worth different amounts in a given assignment, no problem.

Next week is fall break, and after that my students will be taking their oral exam midterms, attending a writing workshop, and collecting data for their group project (the Best X in Town). I’m hopeful for—but not expecting—smooth sailing for the next few weeks in the class.