undefined | Better HN

0 pointsgerminalphrase8y ago0 comments

Disrupt Standardized Testing

We spend ~2 billion on standardized testing in the U.S.; however, perceived drawbacks in how these tests are structured and administered prevent more widespread adoption (though expansion would provide valuable data that is currently unavailable to front line educational staff). The need for standardized testing is largely predicated on a lack of trust in the assessments that classroom instructors are creating and administering, but these untrustworthy assessments are being created in the millions every year. It is a classic 'business that still uses filing cabinets' as the collective value of these materials is largely silo'ed within the binders/file cabinets/'05 Dell Inspiron of the teachers who made them.

Provide a platform that is structured around the creation and administration of assessments BUT the test contents are made available for use by other instructors. The value to the instructor centers on four attributes: immediacy, authenticity, accessibility, and personalization. Immediacy in that an assessment can be created simply by building an appropriate request (e.g. 'comprehension' 'to kill a mockingbird' 'grade 9' 'ten questions'). Authenticity in that this generative model pushes the instructor to rapidly curate their provided assessment elements as appropriate to their lessons (accepted questions remain, rejected questions are replaced: much faster than creating from scratch). Accessibility in that as instructors modify assessment elements (e.g. for an ELL student) those modifications become available to all instructors in the future. Personalization in that the platform could provide any number of assessments that the teacher distributes according to the needs of their individual students, this distribution could likewise be automated to react to past assessment results and further extend the teacher's capabilities.

At the metal, it's one sliver of a Content Management System paired with a Cognitive Tutor. You would need a process for vetting the validity of individual assessment elements, but the teacher curation may offer an initial path (as judging the validity of assessment items would be an implicit element of the workflow). On the back end, you would be in a place to sell assessment data to school districts that is poorly provided by current testing services. You could also sell certified results for higher education, talent identification, and so on. Further on, this platform would be well positioned to create a marketplace for educational content, games, and domain specific tools.

0 comments

2 comments · 1 top-level

throwawayjava8y ago· 1 in thread

This is a bit like saying "disrupt the federal contracting business, so much money for such crap software!"

Standardized tests are a highly political product, and in most states the creation of a standardized test has very little to do with creating an optimal (or even good) assessment.

The idea of using teachers as part of the vetting process for questions is a nice one, but your actual customer sit in state capitols, not classrooms.

germinalphraseOP8y ago

I don't think you're entirely wrong, there is political pressure behind standardized testing; however, this is largely because the alternative was no trustable testing. Making teachers' assesments trustable is politically palatable to teachers while also fulfilling the administrative goals that standardized testing has been asked (and largely failed) to resolve.

j / k navigate · click thread line to collapse