Common questions about standards-based grading

A FAQ and advice about SBG

Photo by Camylla Battani on Unsplash

I often talk with people who are learning about alternative grading for the first time. This is a lot of fun: I get to see their eyes light up as they realize they can solve problems they didn’t even know were fixable.

I use standards-based grading (SBG) as my go-to example of alternative grading. While SBG is far from the only approach, it’s easy to explain and illustrates many key ideas in alternative grading. As a result I get lots of questions about how SBG works. Today, I’ve put together the most common questions I get from people who have just been exposed to standards-based grading.

Recap: Standards-Based Grading

Standards-based grading is one particular flavor of alternative grading – I think of it as an “elemental” approach that can be mixed, matched, and reworked to fit many different contexts.

At its core, SBG is grading at the level of individual skills or topics. SBG classes include a list of standards that describe the key things that students will learn in class. These are generally discrete skills – skills that can be practiced and demonstrated separately from each other.¹ More specifically:

A standard is a clear and observable description of an action that a student can take to demonstrate their learning of some specific topic.

For example, here is a standard from the Calculus 2 class I’m currently teaching. We’ll be learning about this next week:

D.3: I can solve separable differential equations and initial value problems.

Quizzes, problem sets, etc. include questions that address various standards. In my classes, I usually write one page of related questions (with room for answers) for each standard, and write the standard at the top of that page. Instead of earning points or letter grades on each assignment, students earn a single mark on each standard – typically Successful or Not Yet – based on how well they’ve shown their understanding of that standard on that assignment. Students have multiple opportunities to earn Successful on each standard by completing new questions that cover the same standards on future assessments. Final grades are based on which, or how many, standards the student has successfully completed – not on how many attempts were required.

Now, on to the questions!

Q: How many standards should I have?

A: This is easily the most common question, and also the hardest to answer because it’s entirely context-dependent. But that won’t stop me from trying to answer it!

I usually aim for 12–15 standards – about 1 per week in a typical 14 week semester, with a bit of wiggle room. That also tells you a bit about the level of detail at which I write my standards: Enough to encompass a topic that takes 1 or 2 classes to introduce and practice.

Going into greater detail with standards (and therefore adding more of them) makes for an unwieldy amount of assessment and grading. Robert and I both have horror stories about what happens when you have 25 or 30 or 68 standards, so don’t try to assess every single lesson-level objective.

On the flip side, broader – and therefore less-specific – standards tend to mush together multiple skills, making them hard to assess. For example:

I can solve quadratic equations using factoring, the quadratic formula, completing the square, and a calculator.

That’s a lot of options. I could easily write a question that tests one of those methods. Or I could write a question that allows students to select and demonstrate one such approach. But I’d have to write a lot of questions to make sure they can really use all of those methods, and suddenly my exam is 20 pages long.²

Creating a list of standards that works for you is an iterative process. My bottom-line advice is to work hard to create 12–15 standards, take detailed notes, and edit them carefully for the future. Also take a look at the many alternative grading syllabus repositories: Let those who’ve come before you do some of the heavy lifting of creating a list of standards!

Q: Doesn’t labeling questions with a standard give away the answer?

A: To help students know what to focus on and what constitutes “success”, it’s important to clearly label questions or assignments with the standards they address. But, does that tell students how to solve a question?

This all depends on what the skill is that you’re assessing, and the basis for this question usually conflates two separate skills: choosing an appropriate method to solve a problem, and actually applying that method. Sometimes you have a collection of methods and want to see if students can select the appropriate method for a situation. In that case, the skill is selecting the method. Other times you might want to see if students can apply a specific method, in which case telling students to use that method is just part of the question itself.

To illustrate, here are two standards I’m using in a Calculus 2 class this semester:

I.3: I can choose an appropriate integration technique to evaluate an integral and explain my choice.

I.4: I can evaluate an integral using elementary functions, substitution, and integration by parts.

These illustrate exactly this difference. To assess I.3, I give students several integration questions and ask them to identify the technique and explain their choice. But to assess I.4, I directly label a question with the appropriate technique (“Use substitution to evaluate this integral…”). If a student identifies a technique correctly and demonstrates that they can also apply it correctly, they can earn credit for both I.3 and I.4. If they identify an appropriate technique with good reasoning, but then make an error in using it, that means they haven’t shown proficiency with just I.4.

Or, you might genuinely want to see if students can put everything together, rather than testing these skills separately. In that case, consider using a different flavor of alternative grading: Specifications grading, which focuses more on synthesis and communication.

Each approach is valid and makes sense in different contexts. A great thing about alternative grading is that you can mix and match approaches to create a system that works best for you!

Q: If a student meets a standard in Week 1, how do I know they can still do it at the end of the semester?

A: This is based on the assumption that students need to earn Successful on each standard just once. That’s one option, but far from the only one.

I often answer this by pointing out that the question applies equally well to traditional grading: If a student answers a <fill in the blank with your favorite topic> question on one quiz or exam, how do you know they can still do that at the end of the semester? The answers all apply to SBG as well.

For example, you might say that a student needs to answer multiple questions on a topic. In SBG, that means requiring multiple Successful marks to earn credit for a standard. I often require two Successful marks and space out the attempts so that students end up demonstrating their understanding across several weeks (or even months).

Another answer might be that the final exam is a chance to check students’ continuing understanding of key topics. I often designate a small number of “core” standards (typically 4 or 5) that must be reattempted or “recertified” on the final exam, even if a student has already successfully completed them. Of course, the trick with a final exam is that it’s, well, final and therefore resists reassessment. To address the reassessment problem, a student’s work on the final exam only contributes to a “+” or “-” modifier on their final grade. Other items, such as the total number of standards completed during the semester, determine the letter grade itself.

For example, with 5 core standards, here’s how I might distribute grade modifiers based on the final exam:

5 core targets completed: Add a “+”
3 – 4 core targets completed: No change
2 or fewer core targets completed: Add a “-”.

This has proved to give a good balance between ensuring students retain knowledge, and not overwhelming them with an anxiety-inducing final exam that doesn’t really let them show what they know.

Q: If you ask students the same questions over and over, won’t they just remember how to solve them the next time?

A: Sure. So… don’t do that! Again, this is a problem with traditional grades just as much as with alternative grading, and the same solutions apply. Ask students different questions that come at the same standards from different angles. Change the setup or situation, but ask the same questions about it. Come up with creative ways to get at key ideas. Do all the same things you would do when writing a new question for a final exam or make-up quiz.³

Q: OK wise guy, so how do you write that many new versions of a question?

A: This is a much better question! Sometimes it’s possible to “change the numbers” or replace one graph/function/model with another and ask the same questions about it.⁴ If you go this route, be kind to yourself and pre-write as many questions as you can before starting to teach the class. There are discipline-specific tools that can help with this process – ask your colleagues.

But in general, I recommend not getting yourself into this situation. Don’t offer students unlimited attempts, so that reassessments become mindless repetition. Plan ahead and set a clear expectation for how many attempts will be offered. When I use the “1 attempt plus recertify core standards on the final exam” approach described above, I usually offer 2 or 3 attempts at each standard during the semester plus one on the final – so, at most 4 total questions to write per standard. I offer these attempts on a regular basis, usually on pre-scheduled quizzes or exams, so that I can plan ahead on writing new questions.

Also carefully consider how many times students must earn Successful on each standard – do they really need to do that three times per standard (therefore leading to 5 or 6 or 7 separate attempts)? Or will 1 or 2 do?

Finally, think carefully about what Successful means. It should not require perfection. Which errors are important enough and relevant to the specific standard you’re assessing so that a student really needs to make a new attempt from scratch? If an error isn’t that important, then maybe the student still succeeded. I often add a third mark, Revisable, for minor issues that still need to be corrected. Rather than making a new attempt, a student can come to an office hour, explain the error and how to fix it, and then earn Successful if they convince me. This can reduce the need to write extra questions.

Q: Reassessments add work. How do you manage the extra grading? What about extra workload for students?

A: This is important to think about, because adding reassessments on top of an already full pile of grading is a recipe for being locked away in grading jail. For students, adding reassessments on top of a full schedule of assignments leads to being overwhelmed and can build frustration that works against learning.

The solution in both cases is to reduce assignments to the bare essentials. My rule of thumb is that reassessments add another 50% above the time required to complete or grade the original assignment. Use that to estimate workload for both you and students, and then start figuring out which assignments can be removed to open up that extra room.

Take a look at your assignments and ask what their purpose is, and vigorously reduce or even remove anything that doesn’t require a full cycle of work, feedback, and reattempt. What’s left will probably be a lot less than you expect. Perhaps a major project is essential to building student understanding, in which case you can focus on giving feedback on it, while cutting out other items. An exam can probably be reduced to a focused set of the most important and relevant questions. A homework set assigned for practice could be given peer feedback, or you could post a key and grade only for completion. Other items meant to encourage engagement or for practice can usually be graded only for completion and effort, if at all.

Along the way, make sure – as I mentioned above – that you have a reasonably small list of standards. Keeping the number of standards low means fewer assignments to assess them. We all have favorite topics that aren’t actually required by the course description. Those can usually be addressed entirely in-class, without any associated assessment or standard.

Q: SBG looks nice for intro or skills-based classes, but what about upper-level classes?

A: SBG is great for classes that are primarily skills-based, which often means introductory classes. But most instructors want students to be able to show synthesis or “putting it all together”, and that’s even more true for most upper-level classes. For assessing synthesis, communication, and other higher-order skills, Specifications Grading is an excellent choice. And the choice isn’t a binary one: you can mix together elements of SBG (for discrete skills), Specifications (for synthesis or communication), and other approaches (such as ungraded practice work) in the same class. Since this is a post about SBG, I won’t go into more details here, but if you’re interested in these differences or how to mix and match different approaches, check out my post on Standards or Specifications, and Chapter 7 of our book, also conveniently called Grading for Growth.

A final question

These are some of the most common questions I’ve heard about standards-based grading, but they certainly are everything. What’s a question you’ve always had about standards-based grading (but were – maybe – afraid to ask)? Post it in the comments and I’ll do my best to answer it!

I’ve collected these questions and many others in my (frequently updated) Alternative Grading FAQ.

1. Not all skills can be easily separated from each other, and not all situations benefit from trying to separate them. That’s why other approaches to alternative grading, like specifications grading, are more appropriate in some circumstances.

2. We have a whole blog post on how to write standards that includes lots of advice about how to hit the right balance of specificity and generality.

3. I’m honestly astonished at how often I get this question. I’ve worked hard to expunge any suggestion that you should re-ask the exact same question each time, but there’s always somebody who latches on to that idea. Shrug.

4. One of my favorite approaches is to give students a scenario that includes an incorrect answer to a question, based on a a common misconception. Their job is to explain what’s wrong, why, and how to fix it.