Assessments in virtually all courses means grades. Assignments are graded, projects are graded, tests are graded, and then all those grades are averaged together to yield a final grade. Despite their ubiquitous acceptance and use, grades have many problems.
The critiquing approach to learning addresses these issues by integrating learning and assessment in a completely different way. Instead of a process based on many individual inaccurate, low-information grades, with little opportunity for getting it right, the process is repetition and feedback until mastery.
The main arguments against a critique-intensive approach are that it takes too much time and expertise to give detailed feedback, and takes yet more time at the end of a term to review and evaluate each student's history of interactions for an entire class.
Enter the Critiquer, a web-based system for critiquing process. The current system, the Code Critic, is designed for programming courses in any language. The Critiquer provides two sets of interfaces:
- Pages where students can submit their work, review their critiques, and track their overall progress
- Pages where instructors can critique submitted solutions, and review and get a summary of each students entire history of submissions and review
In order to make critiquing feasible with large classes, the Critiquer makes it easy for instructors to build and use a database of commonly used critiques. An instructor just
- clicks on a line of student code, and attaches to it
- a critique selected from a list of available critiques, or
- a new critique written on the spot and optionally added to the list for future use,
- mark the solution as done or needing re-work and re-submission
When reviewing a re-submission, the instructor can quickly see the previous submission with critiques to see what was pointed out before.
At the end of the course, or at any point when the instructor wants to give students a grade-equivalent for their work so far, the instructor can get a report summarizing for each student
- the number of different assignments completed and marked done
- the total number of submissions sent in by the student
- the critiques repeated most often and how recently
For a number of papers on an early Java version of the Code Critic, see Lin Qiu's home page.