Hi,
At University College London we have been running a system for a long time now which provides self-paced learning. The key to how good the software is, is that it includes "confidence assessment" - whenever you answer a question, you indicate your level of confidence in your answer (low/med/high), and this influences the marks you receive.
We want to use this "confidence assessment" feature within a VLE, so I've been working to implement it in a test installation of Moodle. I think I've managed it!
A couple of questions:
1) Could this feature be included in the "official" distribution of Moodle in the nearish future? I've set it up so that it can be turned on/off very easily (just a Yes/No option in the quiz settings), and it has no effect on a quiz if you opt not to use the feature.
2) What's the best way of communicating my modified code to you? I've very carefully marked every single change to the code, but I have never used CVS and to be honest I'm a little bit frightened by it.
3) What is required before a modification can be included? I've added a few "Help" files to explain the feature to teachers/students, but I haven't yet been able to translate them into 90 languages.... Nor have I internationalised the labels on the buttons etc.
Dan
How to use this concept of confidence assessment?
I think that it is good to publish also this link. Maybe there are even better ones? http://www.ucl.ac.uk/~ucgbarg/tea/eltpa6.htm
That link is a good one. The author (Prof Tony Gardner-Medwin) is one of the main champions of the technique. If you read his writings on the subject (see http://www.ucl.ac.uk/~ucgbarg/pubteach.htm for a list) you will find better descriptions of how to apply the concept than I can provide!
Hello Dan,
I think this is a really good idea! The explanation of this approach as presented in the article attached to Ger's response to your post is also very interesting. It looks like it is helpful from the student's point of view. Therefore I would add my voice to yours that if possible, confidence assessment be available as an option in the quiz module.
On a related issue, until recently I used the quiz module both for student practice and for examination testing. The student practice quizzes were set up to provide 5 questions at random, with randomized presentation of the answers and no feedback after submission. Students could take the practice quizzes as often as desired, and their score on the practice quiz was the last one taken. My rationale was based on the (probably simple) logic that if they did not achieve a 5/5 grade, they would keep trying it again and again until they were successful. I did not give them feedback or answers in order to stimulate them to think about what question(s) they may have gotten wrong and to think about them more if they got them again in another trial.
My strategy was not successful, since students were able to systematically figure out the answers to questions by answering the first question and leaving all of the other four blank. It was more or less a "brute force" technique of systematically working on only one question until they got it right, and then going on to the next one. Of course, once one person found the right answer to a question, he/she passed it around to other members of class. I figured this out after looking at the log files of quiz attempts. Some students took the quiz 30-40 times before getting the quiz completely correct and others who took the quiz at a later date got a 5/5 on their first trial.
I think the "confidence assessment" approach would help me to get students to think about their answers a bit more. However I would also like to be able to design the situation so that all of the questions are presented with answers already checked (randomly) for each of the questions. That way, it would be impossible to submit the quiz for grading with only one question answered. They would have to think about whether the answer already checked is correct or incorrect, and change it according to their own decision about which answer is really correct. I suppose they could still "try it out" in hopes of several being correct, but if they are also required to provide their confidence assessment of each question, it might stimulate them to think more about the question.
Rick Danielson
I did some of the same with practice quizzes and then a test quiz. The way I accomplished it was to have a bank of questions, four practice quizzes with some random questions and some constant, then opened these quizzes the night before. They had time to practice if they chose and then I took them to the computer lab for the final quiz which counted, which was a random from the entire set of questions. I only opened this one up for a twenty minute period, and allowed them a total of three chances to make their 100%. The only time this would break down is if there is an internet problem. I thought it worked rather well, and the material seemed to be well retained, based on future testing.