# Which Active Grading Scheme is Best?

Active SBG means:

1. Emphasizing the learning that grades represent, and trying to avoid holding grades as the final product of education.
2. Allowing students to react to their grades. Grades are the beginning of a conversation, not the end.
3. Helping students to understand their grades by organizing them into topics (vanilla SBG).
4. Actively keeping students informed by assessing their skills often and giving them feedback as soon as possible.

These ideals are great, but we have to be really careful about how we implement them – there are some major traps to avoid. For example, if we implement #2 by throwing out old grades, do we inadvertently overemphasize tests and contradict #1? If we break our whole course down into discrete topics to implement #3, do we run the risk of trivializing the broader connections _between_ concepts in our curriculum?

I’m creating grading software, and trying to decide what the default grading scheme should be.  Here’s my current favorite.  What do you think?

## Rubric for each assessment:

• 0. Has not demonstrated any skill or understanding
• 1. Has demonstrated the beginnings of understanding, but still makes conceptual errors
• 2. Has demonstrated understanding of the standard, but still makes mechanical errors
• 3. Has demonstrated basic mastery of the skill
• 4. Has demonstrated mastery of the skill, can connect this skill to other skills, and can creatively apply it to new situations.

It may not be possible for a single assessment to accurately measure a 1 and a 4, here, since any assessment that can measure a 4 would probably overwhelm a student who is at a 1 level.  An assessment that could measure a 4 would probably include many different standards simultaneously – its score could be entered in multiple standards at once.

## Reassessment policy:

Students may ask for a reassessment on a given standard whenever they like.  Reassessments might take the form of problems, conversations, a project, etc.  If the student wants the chance to earn a 4, he should ask ahead of time, since these assessments take time to create.

## Method of calculating a standard score:

A decaying average that counts the most recent score as 60% of the total grade, recursively. Whenever a student gets a new score for a standard, that single score is worth 60%, and the combination of all the old scores is worth 40%.  Examples:

First assessment Second assessment Third assessment Final score The math
2 3 4 3.4 $0.6\cdot 4 + 0.4(0.6\cdot 3 + 0.4(2))$
4 3 2 2.6 $0.6\cdot 2 + 0.4(0.6\cdot 3 + 0.4(4))$

.

First assessment Second assessment Third assessment Fourth assessment Final score The math
2 3 4 4 3.8 $0.6\cdot 4 + 0.4(0.6\cdot 4 + 0.4(0.6\cdot 3+0.4(2)))$

## Method of Calculating the Overall Grade

To find a student’s overall grade, I evaluate the list below from the top down.  The highest qualifying grade is assigned.

An average of at least 3.6 with a minimum score of at least 3: A

An average of at least 3.3 with a minimum score of at least 2.5: B

An average of at least 3.0 with a minimum score of at least 2:  C

An average of at least 2.7 with a minimum score of at least 1.5: D

F

Examples, with 6 standards in the course:

S1 S2 S3 S4 S5 S6 Average Minimum Final Grade
Student 1 3 4 4 4 3 4 3.7 3 A (on average, shows high-level understanding.  Has no gaping areas of weakness)
Student 2 3 4 4 3 3 4 3.5 3 B (on average, shows somewhere between mechanical and conceptual mastery)
Student 3 2 4 4 3 3 4 3.3 2 C (shows a mix of low- and high-level understanding, is good with the mechanics, but still can’t really get standard 1)
Student 4 3 3 3 3 3 3 3 3 C (always shows mechanical mastery and has never shown higher-level understanding. Has no areas of bigger weakness)
Student 5 4 4 1 4 4 4 3.5 1 F (usually shows a very high level of understanding, but really doesn’t understand standard 3.  Note that a single assessment with a score of 3 or higher on standard 3 would launch this student to a B)

This answers a lot of my worries, as long as I can really keep up with the demand for level 4 assessments in a way that connects multiple standards.  Students must show a lot of high-level understanding to get an A, and they must be able to connect concepts.  Students who just want to pass the course can easily get a C by learning the rote, independent skills of the course.  This seems about right to me.

Worries I still have:

• Under this method, if I give a final exam and a student with straight 4s gets a 1 on one of the covered topics, is it really right to drop that student to a C?  His grades would be something like 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 2.2 4 4 4 4.  A C?  Maybe it’s not the minimum that I should be testing, but some measure of deviation?  Maybe I need to leave 3 days for emergency reassessment after any “final” exam?
• This system is complex.   Is it so complex that students would feel completely confused about their grade?
• All standards are “weighted” equally here.  Do I need to reserve some topics as more important?

What do you think?  I’m trying to make this software for anyone who wants to use Active SBG in their class.  What do you need to effect your favorite grading system?

## 15 thoughts on “Which Active Grading Scheme is Best?”

1. Riley, I like this a lot. One thing that would be nice is an option to do a slight variation in weighting the standards. The way I teach, I classify some standards as core—you can’t pass if you can’t do this. Some are intermediate—no A without these, and some are adanced—do everything + these=100. Would it be possible to do something like this with your gradebook?

1. Anything is possible, John! It’s fun being on this side of the software, where if it should do something, I can *make* it do something.

The trick, of course, would be to present the configuration of your standard classification in an understandable way. How do you tell your kids which are which? Are there asterisks by the core standards, or do they come on different lists?

Honestly, I don’t think I’ll have this implemented by the time the beta opens (November!), but it could be in there by semester 2. I hope you’ll sign up and voice the need for this again. Thanks for the comment!

2. Nice idea. My immediate concerns, because I’m doing SBG now and these jumped out at me, for my kids, in my class, at my school. Maybe not useful for your kids at your school…

1. Writing, coordinating, and grading each of these individual reassessments takes me TONS of time. Your system encourages a lot of reassessing — which I get — but maybe too much? Like, since the last grade isn’t just “taken,” then kids who are motivated will be reassessing a lot to “up” their prior grades’ averages.

2. Will this grading be too hard for the kids to understand? Does it detract them from understanding “I get this” vs. “I sorta get this” vs. “I don’t get this.”? I suppose they’ll catch on after a quarter or so how to calculate their grades, but will they really be able to interpret them? My kids are just getting the “last grade counts” thing now, at the end of the quarter.

3. The opposite of (1): With the complexity of the grading scheme, will kids really see reassessing as in their favor? A 4, 3, 2 gives a kid a 2.6. Will it be a detractor for kids trying again?

Sam

1. 1. Totally. Some limit on reassessing might be required – once a week? Major downside.

2. Yeah, maybe. I think they could get it. Maybe you have to make recursive function definitions the first standard? 😉

3. It does put some of the pressure back on, doesn’t it. On the other hand, do you feel good about a kid who scores a 4, 3, 2 having a 4?

1. I think any number crunching is fine, as long as they’re just suggestions. I remember demoing a system (maybe Global Scholar) that gave like 4 “final scores” based on most recent, mean, median, and this power law formula. In the end though you still manually chose a letter grade.

1. I like it. I worry that kids can’t know what their grade is until you decide, but I guess that’s where you sit down with the student and discuss his or her grade!

2. I’m with you on your points Sam, one benefit I gave the kids when I was spelling it out to them earlier in the year is that “it takes the mystery out of where your grade comes from, especially since they have copy of the rubric. To point #3 though, I think it strikes a balance between #1 and #3. They won’t just retest willy-nilly because if they come flying in unprepared hoping to up their ‘number’ but instead decrease it, they (should) be making conscious decisions about when they are retesting and putting in meaningful work before retesting to guard against the ‘backslide”. I suppose your calc students can understand the inner-workings of the recursive relationships, but my ‘blue-students”? Unh-Uh. My only gripe then, if they are analyzing the system that way, is that they are still ‘playing the game’ instead of “I need to learn this”. I’m talking myself in a circle again, so I’ll quit here.

3. Riley, it looks great! My thoughts:
– Is it possible (and maybe you’ve already done it – and that’s what you mean by describing the “default” structure above) that the teacher could select from a set of possible methods of grade calculation, or even adapt them and/or create their own?
– I also worry about complexity – I think one important ingredient for students and families is a set of really clear examples for each grading scheme.
– Re: reassessing – could the teacher specify reassessment limits? Or have the ability to weight the various assessments for a given standard?

1. Yeah, totally. Options abound. But only options I like! Normal averaging will be an option for calculating the final, overall grade, but not for individual standard grades.

I think automatically generated examples are a great idea – I worked them into the app today. Already, the feedback loop is working!

Weighting assessments is complicated, and I haven’t seen a great reason to do it. I mean, I get the final exam thing, I guess. But I already feel so shady about the accuracy of assessments that I’m hesitant to say, “hey, this assessment is about twice as accurate as this one, so it counts for twice as much, word?”

Hope you jump in on the beta, Sarah! Great ideas are flying between us like SPARKS!

4. I so like the ‘decaying average’ for accumulating the assessments! I also like John’s idea of weighting standards.

I don’t exactly weight the standards per se, but one thing I do with my children’s grade system is that knowledge of content presented is only one factor of the grade… timeliness and quality of assignments is the other factor. This may help with some ‘timeliness of learning’ comments on previous posts. So for example, if I have it set at 60% content, 40% timeliness/quality, then even though a child masters the content as exemplified by tests, he still can’t get an A unless he does well on the reports/activities/discussions/etc. (AND does them in the assigned time period).

In my case I had the ‘luxury’ of shifting the percentages as needed by class (or even by child). For example, I had a child who struggled doing the in-class and extra-class assignments, but who could ace a test like you wouldn’t believe. So for certain classes, I shifted the percent allocations a bit to give more weight to the non-test activities, which in turn helped him focus more on his time management and getting his work done on time (and legibly)!

1. Shifting percentages to change the focus of a course has always seemed dangerous to me. I liked grades in my classes to represent a student’s skill level, and would be hesitant to change that for a student with a high skill level, just so I could focus him on something else.

How did your student with the time management problems respond to your shifting of the percentages?

1. Like I said, it was a ‘true luxury’, if you will, because we homeschool…so I made that change in weighting across ALL of his classes. I saw mixed results – it greatly improved his timeliness and organization in English and Science related courses, and I am seeing that fruit as he is taking some college-level science & lit courses.

But with math it was less dramatic, there was some improvement, but he still more or less continued to scrape by on test grades. He’s always been ‘lucky’ that he picks up math concepts quickly & easily – but if you don’t actually do some practice, you don’t learn to pay attention to details… and that we are also seeing in his college level math courses! Details matter.

5. Yaacov says:

I’m currently using a four level rubric that matches up pretty much exactly with yours. I write it:
Unfamiliar Concepts – 0
Beginning Progress – 1
Progressing Towards Comptence -2
Competence Achieved – 3
Exceeds Expectations – 4
I don’t put the numbers, but students are quick to connect them. I will probably reword 1 to Beginning Understanding and 4 to Makes Further Connections.

I like it but I’m finding that it’s pretty much impossible to get a level four on smaller pieces of work.

As for methods of calculating an overall grade, I think you really need to ask what the purpose of an overall grade is. If it’s an indicator of the readiness of the student for post-secondary studies, having the lowest mark be the most influential seems harsh. Student 5 will likely do just fine at PSE, and almost certainly will do better than Students 3 and 4.

If the final grade is a certification that the student understands the entirety of a specific body of knowledge, you have a stronger argument for your method.

If it’s an indication of what percentage of a body of knowledge a student understands, then you should use a straight up average.

etc.

1. Great points. My thought with the minimum score is that certain skills are simply vital. If I have a kid in my algebra class that can simplify exponential expressions like a whiz but can’t graph a function, I’m not going to pass him.

Of course, that’s an unlikely scenario. Maybe we needn’t worry about the case where a kid can do everything except a single skill.

At my school, we did not all agree about what the overall grade meant. I’m not a big fan of standardization in general, but it seems like if we do standardize (by agreeing on a set of grades) we should agree on what those grades _mean._ Thanks for the comment!