[jpereira]

It's kinda crazy that assessment has been coupled with instruction for so long, it pretty obviously creates a warped incentive model that stands on some real shaky ground.

I think the model also doesn't scale well for assessment, which prompts the creation of few, ineffective, but highly scalable assessments upon which the functioning of the entire system rest, which in turn prompts practices like teaching to the test. This creates even more messed up incentives but this time facing students and teachers as opposed to institutional practices.

I think that assessment, when decoupled, can't be done with an institution or institutions. Instead we need social networks that use a consensus process to define knowledge and who has it.

[graniter]

Agreed. I think about the analogy of interchangeable parts and if that might apply to education. Maybe teachers should just teach and assessors should assess. Interchangeable parts made it so equipment could be more efficiently be manufactured. But that didn't preclude there being a lot of different types of equipment. In fact, there's probably more types of equipment because of interchangeable parts. So for education, having an assessor focus on assessments, independent and separate from instruction, might make it so that there could be easier to assess and more types of assessment, so the teacher doesn't have to always do it. And that might help compare results across teachers.

College Board owns SAT, AP, and (essentially) Common Core so there's been a trend to tighter coupling. I like the idea of an AP Test that anyone can and be assessed separate from the teacher. Same for SAT. I think we need more of that model, not less.

[jpereira]

I don't see the advantage of tying assessments to singular institutions especially ones structured like College Board. What this promotes a bottleneck of what can be communicated to others meaningfully. Currently the SAT can be broadly communicated (because CB and others have invested very large amounts of resources into it) and so students optimize for it. Instead they could be putting their efforts to towards learning that is more personally and socially valuable.

I don't think this is just because the SATs are bad tests but fundamentally emerges from the institutional single source of truth model, where the incentive is for the institution to make their test as general/broad as possible and the most broadly accepted , as to capture the most students and hence fees.

[germinalphrase]

I agree that we need more and better assessment - but I don't understand why we would allow a private body to do what should be a utility function of our educational system. Currently, the College Board is able to fill this niche (at great cost) due to a lack of trust in the assessments of individual instructors. Rather than further increasing the cost of assessment by expanding the use of private arbiters, we should seek to make the assessments that are already taking place (in far greater number and scope) more reliable and trustworthy.

I believe this would take the form of a dedicated assessment tool that allows teachers to create individualized assessments JIT based on - or in concert with - their planning and performance workflow. If I'm teaching a structured lesson on polynomials, I should be able to create a valid assessment simply be requesting that one is made. My students should be able to take that assessment on the spot and the results should be immediately available.

I will take assessments with 80% confidence of validity 10x/week over assessments with 100% validity once per year every. single. time.

There's more to be said here - but that's the gist.

[graniter]

Why not have multiple assessors? Let people/parents/whoever pick who they want. Some schools use ACT and others SAT. Having a single assessor seems suboptimal. Alternatives tend to keep things more in check.

[philipps]

I like the concept of community-driven recognition, but I would caution against separating assessment from learning (if you used "instruction" to mean the moment when learning happens). Communities that engage in collaborative work (like open source projects, or guilds) are very good at surfacing and recognizing expertise. Making assessment a separate process increases the risk that we measure the things we can count, rather than the things that really count (butchering the Einstein quote).

[jpereira]

I definitely see where you're coming from, but I think the danger there is not necessarily anything to do with the coupling of assessment and learning but the coupling of assessment with the learned, i.e those who possess a skill/piece of knowledge/trait should be the ones who define the assessments around it. I like that you bring up guilds because that's very much baked into their process I think.

[j45]

Assessing Competency and Outcomes is one thing when done in a learning environment and quite another in the real world.

If students need to be prepared for the real world, this will need to increasingly become reflected in the amount of time spent in real-world learning vs in-classroom.

The question becomes - how will real world learning get to the classroom to reach the student, and how can the classroom meaningfully be with the student when they are with the real-world?