My undergrad roommate Kristen (now Dr. Dorsey, after earning her PhD in ECE from CMU) emailed me about an article on specs-based grading, asking what effect it might have on intrinsic motivation (which we'd been discussing with some of our former suitemates over an extended email thread. I love my suitemates).
My reply was that I've also heard the technique called "contract grading, " and it has pluses and minuses. This Chronicle of Higher Ed article has a decent discussion of the minuses, which mostly consist of "watch out for loopholes and students trying to game your system to do minimal work."
Contract or specs-based grading is exactly what it sounds like: writing out detailed instructions as to what students must do to earn a certain grade in class. And I mean detailed. Turn-your-class-into-a-videogame detailed. The kind of contract you'd write out when specifying a technical component you're outsourcing to a subcontractor. "If you submit 4 of these 10 assignments and are absent fewer than 3 times, you get a B." "To earn an A, your essay must answer the following questions in grammatically correct English..."
There's been a limited amount of empirical research on its effects. Via the POD mailing list, here's a study on contract grading's effects on a science class (psychology) and a humanities class (composition). Spoiler: contract grading was "more effective" at student retention and higher grades than a traditionally-graded control group.
Now: what about intrinsic motivation -- the sort of thing most teachers wish their students had? You know, the students who want to learn about nanoelectronics because it's so beautiful! and they love love love electronics! just like you do.
Here's where it gets tricky. Intrinsic motivation can be fragile, and extrinsic rewards can destroy it. If a kid loves playing the violin, and you start rewarding her with ice cream every time she plays, she may learn to play in order to get ice cream -- and will stop playing the violin as soon as the ice cream ceases.
This means (in my opinion) that contract grading contracts should be written so that students who are on fervent fire can keep on running without needing to stop to puzzle out bean-counting. Your expectations should be clear and flexible enough that students who do have a project in mind can see how they would do those things anyway if they were doing the project well -- the goal here is minimal re-routing of an intrinsically motivated student who's already running full-tilt down a path. Also, the contract should explicitly state that students can talk with you about renegotiating the contract to fit a project they really want to do.
Depending on your student population, you may or may not not have a lot of intrinsically motivated students from the first day. Hopefully you won't have many amotivated ones who just don't care at all. If so, the contracts can help by turning amotivated students into extrinsically motivated students. Extrinsic motivation means that they are motivated, but by something other than an inner love for the subject.
Extrinsic motivation has a bunch of sub-categories, but it's not necessarily "bad." Heck, we try to extrinsically motivate students: "you should do well in this class because it'll help you get a job." (People often confuse intrinsic with extrinsic motivation. "He's so motivated to do well because he wants to keep his scholarship!" is still extrinsic motivation -- you may not need to keep prodding this student to do his work, but the scholarship is what's driving him, not necessarily a deep-seated love for circuit theory.)
Basically, contracts can turn all students into extrinsically motivated students -- which is great for amotivated students, but not so good for intrinsically motivated ones. So be careful when writing your contracts so that the amotivated students can't find loopholes -- and the intrinsically motivated students won't get distracted by having to worry about "playing the game" in order to get points.