Wednesday, June 23, 2010

Final exams in a standards-based system

Programming notes: If you actually visit the site instead of in your Reader, you'll notice there's a tab for the standards-based grading implementation series. You'll also notice I've changed the blog roll titles to Twitter names. I did it in my Reader because I had trouble mentally linking people on twitter to their blog. If you're on there and DON'T want your twitter name connected to your blog, please let me know ASAP and I'll change it. On the other hand, if you're not on there and should be, let me know that too. Don't be shy. I've got about 150 blogs in Reader so I probably just missed it looking over the list.
Shawn told me he's posting on this as well and I'll update this link when it comes out, but it's definitely in the comments here. Edit: As promised.

Note to Shawn and Matt: I'm going to need a list of all your upcoming posts. I hate that you say everything I have to say, only better. Oh wait. I love that.

Just to warn you, I'm a middle school teacher and we don't give finals at my school. Be more skeptical of anything I have to say than you usually are.

Before we move on, here's a reminder: Assessment is not just tests.

So what do we do about  finals in a standards-based system? It will require some compromises.

In the Ideal World you probably don't give a final. You might have some culminating project of awesomeness (edit:fixed link) but if you've really done your job you don't even need a final. Actually, if you've really done your job your students can just assign themselves a grade because they're such master self-assessors. Some would argue that the ideal world doesn't even have grades. I understand that argument but I don't fall into the all-grades-are-the-root-of-all-evil camp.

In the Almost Ideal World you still need to give grades and you still operate in a fairly traditional school setting. You need to give a written test, not some sort of showcase.

Assuming you've broken your curriculum into topics or skills lists, you've got specific things you want the students to know and do. You also assess them regularly. For me, when I start off a topic I'll give a quiz about once a week and the last week I'll give two. Then I'll move on and re-quiz1 every few weeks and slowly bring it back more frequently as the grading period closes.

During the last two(ish) weeks of the trimester, I will fully reassess each student on each topic as needed. A student who has been banging out review quizzes all trimester doesn't need to be reassessed. It would be hoop jumping, which is what we're avoiding with all this SBG business. Think of your student that has 102% in your class. There is nothing positive that can come from the final. At best, she stays the same. I've had teachers with "fail the final, fail the class" policies. It shocks me they're willing to invalidate an entire semester of hard work because of a single test.

On the other hand, some students will need to reassess on everything. They want to take your final.

Those students might be taking something that might look like a final. The difference is that the final is based on need and is transparent. They know what they need to take and can focus accordingly. It's not a guessing game of, "Is this cumulative? How much of the first month is going to be there? Is Mr. Buell going to ask anything about calculating acceleration?" A student should never fail your class because they guessed wrong about what would be on the final.

Just in as I'm writing this: Relevant tweet by @jerridkruse

Compromise 1: The final will be given as needed. Students will only take the parts you need more evidence for or they want another chance to show you something. Goals should be clear going in.

Ok, you probably do need to give everyone a midterm or final. You might be able to format it by topic and let certain students just cross out the ones they've passed out of. On the other hand, you might also be giving a common final and have no choice in the matter.
So what's a good SBGer to do?

Remember this: Finals are just one more piece of evidence. It is up to you to decide if that evidence overrides all of your previous evidence.

So what happens? Students take the final like normal. Now you can sit with your excellent department and pour over testing data.

When it actually comes to the grade, you need to make a decision here. Does the evidence from the final override what you previously believed about the student's level of learning?

What does this mean? Assuming you've got some fairly recent assessment data on your student, you are within your rights to 100% totally and completely disregard something on the final. You have a section on your final about states of matter. Your student has aced every states of matter problem you've given her, she leads tutorial groups on it, she performed an interpretive dance on the states of matter, and has predicted the existence and properties of the Bose-Einstein Condensate through a pure thought experiment. On your final, she bombs it. Why does she bomb it? No idea. Almost Ideal World you can go back and talk with her. In Your World, she's on the beach somewhere because it's summer vacation and you're the only sucker still in school. Maybe she broke up with her boyfriend that day. Maybe she didn't get enough sleep because she was studying for her history final. Maybe she just read the question wrong.2 In this case, you, as a teacher and a human being and not a scantron machine, is able to required to morally obligated to make a decision. You need to decide if all of your previous evidence were wrong or if your final was wrong (or something in the middle).

This is where organizing your gradebook into separate topics is clutch. If you have an entry that says "Final exam" and it averages everything in, you're going to have to massage the points somewhere to make it work out. That usually means randomly giving points. That distorts the record of learning. Not to mention it gets...uncomfortable... and isn't very CYA friendly.

On the other hand, in a topic-based system, you just update the other scores as needed while keeping intact the current topic score.

Compromise 2: Give everyone the final. Use it as a single piece of evidence and decide for yourself if it is relevant and meaningful.

Before you decide what to do with your final, take a second to remember how angry you get when you hand out the All Important State Test to your kids and you realized you guessed wrong about what standards they would be testing. Sarcasymptote told me his recent NY test had 39 questions on 104 standards. Your final should NEVER be a game of chance. Take another second to remember how outraged we all get because someone out there is trying to measure our worth as a teacher based on a single test, on a single day.3

All of this is a way of saying that your tests should never override your judgment. Never let the points make a decision for you. You are the teacher. Own your grades.

1: Usually, mid-course that's not a full assessment. I'll lump all the previous topics together on one quiz and only include one or two big ideas per topic. The rest of the period or the next day, they'll work on the topic they had the most trouble with.
2: I once almost failed a class because I didn't turn over the page and see there were questions on the back of the final. My professor let me come in and just talked with me about what we learned. Take that "standards-based grading doesn't prepare them for college" people! Thank you Dr. Ebbesen.
3: If you want an assessment guiding principle, you couldn't go wrong with, "Don't do anything your state does for high stakes testing." or WWMSTD? What would my state test do?


  1. I like the sound of this, especially for lower grades where they are getting the hang of a comprehensive style test.

    However, I remember my freshman year of college meeting up with old classmates who told me, "I wish my parents had made me take finals (comprehensive finals, by semester)" The school had a policy where you could get out of the final by having no absences and an A, or one absence and a B, etc. My parents made me take the finals even though I had an A and no absences. In college/uni, I had some idea of how to study for a final and did decent on those the first semester, whereas many of my classmates struggled with figuring out how to study. College was a shock of freedom and independence as it was, and this was one less adjustment I had to make. In that sense, taking seniors and, perhaps, juniors through studying for and taking a semester comprehensive final could be very useful for at least some of them.

  2. Two responses:
    1. In California they take long, high stakes tests ALL THE TIME. I think they start in third grade. In 8th grade they take 4 state tests (math, science, ELA, history) plus various district mandated benchmarks, writing prompts, and a math placement test for high school. Include the CAHSEE, ACT, SAT, AP tests and I think they're pretty well covered in that area.

    2. I think what you're getting at is study skills + testing stamina. You don't necessarily need to combine those two things into a final grade for the class. There's a diff between what's worth teaching and what's worth assessing, although there's substantial overlap. Study skills are certainly worth teaching and might be worth directly assessing. On the other hand, there are certainly better ways to teach testing stamina than giving really long tests (strategies to help re-focus, how to take quick mental breaks, ways to deal with frustration, building persistence, etc). A problem with most finals is that they're often as much a test of stamina as it is what they've learned.

  3. We had a system like the one Younger Rachael mentioned above but linking to attendance. Any student who missed more than three days per term was required to take final exams and they are required to be 25% of the final grade.
    Then, we began to reports back from graduates who felt that exemption from finals meant they were will-prepared for college, teachers started complaining that we weren't holding students accountable for learning, etc. so they ADDED another End-of-Course assessment that all students had to take worth 10% of the pre-final exam grade. That became convoluted and crazy so next year, we are going to a single final exam, which all students MUST take, but I've not yet heard what they will require its weight to be. It makes defining final exams with regard to sbg a bit difficult. :(

  4. I'm writing from the future here (Dec. 2011), leaving the promised trail of breadcrumbs as I go from beginning to end.

    This is all really helpful. I am about to give a final exam (ironically) during which I will be working out some ideas for doing more justice to these ideas in my college level teaching than I have been able to do so far.

    But how exactly does this help tighten the table legs?