Happy with Perfection?
http://blog.igenoukan.com/2008/09/happy-with-perfection.html
I see several problems with a 100% mentality to grading. This post covers some of them, and a possible solution or two. Part of the problem is parents, but part of it is that a higher difficulty can lead to lower grades, when there is more learning in the higher difficulty. Homework, testing and grade representation all have their parts in this. Expanding our assessments to include a variety of good qualities can give us a better result.
Iteration and Prototyping
http://blog.igenoukan.com/2008/10/iteration-and-prototyping.html
Perhaps not the first thing to come to mind for assessments, it is relevant. First is that each prototype is different, like a person's skills and approach differ after learning.
Pointless Obstacles are Bad
http://blog.igenoukan.com/2008/10/pointless-obstacles-are-bad.html
This is a little more directly related in that assessments shouldn't be limited by classes. Testing into and out of classes is an example, but not the limit. If a person has taken the course once, why should they be forced to take the course again just to retest on the curriculum. Perhaps this would make colleges and universities seem more like testing and accreditation institutions, but having to pay for the class and redo all of it just to prove capability doesn't make sense except as a business.
Testing Before Teaching, Why?
http://blog.igenoukan.com/2008/12/i-find-following-article-to-be.html
Defending pre-tests really wasn't my initial intent, but it was hard not to in response to a particular article. One example that even the author conceded was testing out of a class, but there is so much more to a well given and used pre-test. It can help teachers and students. One of the big points is it not counting towards the course grade. Are the results important? Yes, but a pre-test that counts for the grade isn't over material covered in the course. It has other uses.
Logical Content Breakdown
http://blog.igenoukan.com/2009/01/logical-content-breakdown.html
This one is about the content of the assessments, and the order. In the example covered in the post, splitting up the content differently and using a slightly altered explanation would make learning the curriculum very different from the current norms, for that topic. To further facilitate that kind of content breakdown and the learning that comes from it, assessments and self-assessments should be given. Yes, I just said that self-assessments should be given. It's a matter of getting students to see their mistakes, shortcomings and weaknesses. The content breakdown can make learning the material easier, but assessments can help learners see how well they have learned the different parts.
Metalearning: Learning How You Learn
http://blog.igenoukan.com/2009/04/metalearning-learning-how-you-learn.html
The big idea is that it's good to know how you learn, and so it is good to learn how you learn. To do that requires assessing the results of learning attempts, and comparing them; assessing how well and learning method works for you. Trial and error is the default method for this. Helping students understand themselves as a learner has many possible positive effects. One is giving students the knowledge to craft good learning environments and select good learning resources for themselves. If you don't know how you learn, experimenting while being graded can seem like a very bad idea.
Skill Based Grading
http://blog.igenoukan.com/2009/04/skill-based-grading.html
I'm a game designer, and this is one of the points I like in gameplay when there is to be improvement; skill based assessment. With skill based gameplay, difficulty is matched with skill levels to provide a suitable difficulty. Well, what if the debriefing of results and skill based results found in some games was used in assessment? I think it would be interesting to have achievements, multiple measurements and the ability to retry after learning from previous mistakes.
Automated Tests
No, I didn't miss a link. This isn't a post I've made before, but rather a record of my thoughts after going through those posts. I think one of the keys to improving assessments is to make them less costly to give and take. Automated tests are a possible solution, if done well. Randomly generated problems will always be able to be solved by the right software, but they can be used as non-graded self-assessments and graded assessments when other tools are not available.
I know that many proponents of improving education are for opening up the resources normally available to students, but the keys to automated tests are providing the resources and not allowing alternative resources. Basically it becomes a game. How well you do at the game shows what you can do, but if you have a tool that solves the problems for you, it's pointless to use. Yet it is also not logical to test a person in a way that removes information that would be available to them in the real deal merely because they don't remember it. So, if the assessment was on derivative calculus as a whole, it might start with picking the correct names of the rules, putting the equations together correctly, using them in problems and correctly identifying the abnormal cases. It would probably take a while, but everything needed is there and requires the learned skills and knowledge to succeed. I think most people would enjoy assessments more if it was less like being put on trial and more like exploratory play. Isn't education supposed to be a safe time to make mistakes before going out into the real world?
Have fun, spread the word and tell me what you think,
Igen Oukan
Read more!