Curves and Erstwhile Midterms

Left-shark logic.

I spent far too many hours trying to concoct a fair system for curving your Q2 GAP scores. I played with weighted percentages and threshold mechanics and achievement-based curves — none of it worked. It just didn’t feel right.

I would have just thrown your GAP essays into the air and gone with random chance, but there are no printed copies this quarter — and throwing Chromebooks into the air turned out to be a little more dangerous than I thought.

The problem is obvious, and should have been obvious before the arbitrary +5 you received at the end of Q1: Any curve shifts the focus away from performance and onto numbers. It stops the discussion about your profile and all the traits and skills that profile represents and replaces it with the oblique and damaging language of grades.

But we have to do something. The extrinsic motivation of numbers is so powerful that without a way to soften the blow, some of you won’t develop the honesty that will eventually make this entire post moot. Your stress is often alleviated by a higher number, too, and it’s hard to ignore that.

I considered the data in front of me. I have your requested profile scores, your essay profiles, notes on those essays, and then a slew of harder data — your midterm score, your self-reported midterm score, other writing samples, and tons of discussion notes.

In all that, what struck me most was the range in self-reported scores. Some of you were close; others were 20 or 30 points off. But it was the correlation that mattered. If you were significantly inaccurate — more than four or five questions off — it was almost always that you were too high. And if you were that inaccurate, you were also overwhelmingly likely to overrate yourself for the grade abatement profile.

This is obviously the Dunning-Kruger effect. It’s also obviously the most important metacognitive pillar in our course: the ability to be honest and accurate when evaluating a performance, whether that performance is on a midterm or an essay or in an interstitial conversation.

So the most important piece of data you’ve given me is how far off from your actual performance you were when it came time to self-assess. The closer you were, the more impressive your metacognitive ability is. It’s not a carnival guessing game, either; it’s about memory, reflection, and investment in the course. On an objective test of course fundamentals, the distance reveals to a large extent your investment in the course overall.

Your Q2 curve, then: You earned points derived from how close your self-reported score on the erstwhile midterm was to your actual score. You also earned points derived from how close your self-reported GAP score was to your actual GAP score. Going higher is a little worse than going lower, but the distance matters most.

There isn’t any complicated formula. If you were correct about your grade abatement score, that earned you points. If you were close to your erstwhile midterm’s score, that earned you more points, depending on how close you were.

I considered very high raw scores, too. You ought to be rewarded for knowing the course fundamentals, and with our midterms canceled, this is one way to do it.

Remember, though, that any mechanism that focuses on grades puts us back in shark-infested waters. Curves and boosts are momentary concessions to the system around us — chum in the water to throw the shark off our scent. Do not obsess about them, and do not belabor them. Swim for dry land.

I’m just using that metaphor so I can post this gif, which comes from the comments here. Drew Magary has a less favorable impression of the viral outbreak of “left shark” memes — read it here, but be very aware that his essay (and even title) aren’t exactly safe for school/work.

Addendum: On the Precipice

After posting the scores on 2/8 to beat the winter storm, I immediately received emails about Q2 grades, despite asking for patience and promising you the mountains of feedback that are now available

Some of those emails came from students earning a 5+ this quarter, but only if they also had a +2 curve for the erstwhile midterm’s DDI/DK (as explained in the rest of this post). That becomes a 79 on the report card.

Well, we’re firmly back in shark-infested waters. Or to use a more timely metaphor, we’ve run outside into the middle of a winter storm.

I think many of you would agree that this one is aptly named.

We are fixated on numbers again, not performance or process. But this was the danger in giving you a range of achievement-based curves: Some of you ended up on those old precipices — 79, 89 — where the first digit completely obscures the second one. Your brain freezes up.

So I’m just giving everyone +1. It’s already in Infinite Campus, and it seems to have helped quite a few of you.

Now you need need to sit down and think deeply and honestly about why that one point affects your emotions and ability to process feedback so dramatically. If you weren’t affected, you need to recognize that you also have that reflexive reaction to a 79 or an 89 — and that means it’s your responsibility to think with your peers.