A framework for creating assessment criteria

[Dale Emery]

Bloom's Taxonomy is a framework for classifying educational objectives. It's widely known in the academic world, and somewhat known in the world of instruction design. I have no idea how widely used it is in practice. The chunk that's best known is what Bloom called the Cognitive Domain. You can read about it here: http://en.wikipedia.org/wiki/Bloom%27s_Taxonomy

If we're going to assess people, we will want criteria for assessing. At a general level, the criteria will include (1) what a person knows about agility and (2) how the person is able to use that knowledge. The more specific criteria we create will be similar enough to learning objectives that I suspect Bloom's Taxonomy may yield some insight about what we might assess and how.

A gaggle of authors developed a refinement of the cognitive section of Bloom's Taxonomy, extending its one-dimensional classification system to two dimensions: the Knowledge dimension and the Cognitive Process dimension.

The Knowledge dimension is about the kind of knowledge involved in the learning objective. It has four major categories (listed from least complex to most):
- Factual Knowledge (basic facts)
- Conceptual Knowledge (relationships among facts, such as categories, models, theories, principles, generalizations, and structures)
- Procedural Knowledge (how to do stuff: subject-specific techniques, methods, procedures, algorithms)
- Meta-Cognitive Knowledge (strategies for learning, thinking, problem solving; self-knowledge; when and why to apply facts, concepts, procedures)

The Cognitive Process dimension is about what we want people to be able to do with the knowledge. It has six major categories (listed from least complex to most):
- Remember (retrieve relevant knowledge from long-term memory)
- Understand (construct meaning)
- Apply (carry out a procedure in a given situation)
- Analyze (break material into parts to identify how the parts relate to one another and to an overall structure or purpose)
- Evaluate (make judgments based on criteria and standards)
- Create (put elements together to form a coherent or functional whole; reorganize elements into new patterns or structures)

The two dimensions form a grid. According to the model, a good learning objective involves both a key noun that identifies a specific element of knowledge, and a key verb that identifies what we want the person to be able to do with that knowledge. The noun falls into one of the Knowledge dimension categories, and the verb falls into a Cognitive Process category.

For example, we might want someone to be able to recite the rules of simple design. "Recite" falls into the "remember" category of the Cognitive Process dimension. The rules of simple design seem to me like Conceptual Knowledge (more complex than basic facts). So I'd put that objective into the Remember Conceptual Knowledge cell of the grid.

We might also want someone to be able to apply the rules of simple design to refactor some code that has obvious duplication. I'd put that into the Apply Conceptual Knowledge cell.

I wouldn't necessarily want to wield this framework in a rigorous way, but think it will helpful for checking whether we've covered all of the bases we want to cover, and for filling in the some of the holes. One approach would be to (1) list the concepts that make up agility, (2) place each in the appropriate the Knowledge dimension category, (3) construct assessment criteria for each relevant Cognitive Process category.

Here's a link to a book about the revised, two-dimensional framework:
http://www.amazon.com/gp/product/080131903X?ie=UTF8&tag=dalehemer-20

Dale

--
Dale Emery
Consultant to software teams and leaders
Web: http://dhemery.com
Weblog: http://cwd.dhemery.com