Forum archive 2000-2006

Michael Gage - slides for NeXT discussion

Michael Gage - slides for NeXT discussion

by Arnold Pizer -
Number of replies: 0
inactiveTopicslides for NeXT discussion topic started 4/5/2001; 6:29:58 PM
last post 4/5/2001; 6:29:58 PM
userMichael Gage - slides for NeXT discussion  blueArrow
4/5/2001; 6:29:58 PM (reads: 664, responses: 0)
Topics for NeXT discussion To view this in slide mode go to http://webhost.math.rochester.edu/webworkdocs/slides/373/1 Click next to begin: Abstract Abstract: Arnie Pizer and I designed a web-based homework checker called WeBWorK which we have used at the University of Rochester in one form or another for five years now. Students enter answers to calculus homework problems through a web-browser and are told immediately whether or not their answer is right, but NOT what is wrong with the answer (if there is something wrong). The student gets to try again on the same problem. This is like doing homework with someone looking over your shoulder, checking your work, but not helping any more than that. The immediate feedback keeps most students at the homework problems until they get them right, and since every single problem is graded I can count homework heavily toward the final grade. Students like this. They feel in control about the amount and quality of the homework they do. They also get frustrated since the computer can't tell them what they are doing wrong. Every new form of instruction, such as WeBWorK or any other web-based instruction package (Blackboard, WebCT, etc.), provides new opportunities, but also limitations, and requires new strategies which take advantage of the strengths, while minimizing the effect of the weaknesses of the new medium. After giving a very brief idea of how we have used WeBWorK, I'd like to lead a discussion (1) on how to write homework problems that take advantage of web-based capabilities, (2) on what other support is needed to help students effectively learn from homework(workshops? recitations?..) (3) on how to minimize the effect of the constraints imposed by using the web and computers, (4) when to seek to improve the current software (5) and when to utilize some other media entirely. These topics are clearly relevant to many other web and computer based instructional modules as well as to WeBWorK. I have some ideas on these subjects and some experience, but not many answers. I'm looking forward to the workshop. Mike Gage P.S. For those wishing to get a feel for WeBWorK in advance: You can find general information beginning at http://webwork.math.rochester.edu. To see how problems are written try http://webwork.math.rochester.edu./tutorialCourse. To see the widest possible variety of WeBWorK problems look at http://webwork.math.rochester.edu/demoCourse. Or contact me: gage@math.rochester.edu Main points (1) How to write homework problems that take advantage of web-based capabilities, (2) What other support is needed to help students effectively learn from homework(workshops? recitations?..) (3) How to minimize the effect of the constraints imposed by using the web and computers, (4) When to seek to improve the current software (5) When to utilize some other media entirely. Observations on web homework Students keep at the homework until they get it right (probably due to instant feedback about incorrectness) Automatic, universal grading makes it more defensible to count homework heavily. Students feel more in control of how much the amount and quality of their work will count. No direct help E-mail feedback helps a little to remedy this, but it's time consuming - in my opinion it's more effective than office hours. Workshops? recitations? Anecdotal evidence seems to indicate that with web homework is more focused in recitations. Can it get too focused? Some students have a WeBWorK note book with printouts of the problems, their work in solving the problems. Others are more disorganized. It's easier to help the former. Questions on web homework Is having the homework also available in hard copy form important (essential, useful?) in making the homework process effective and efficient? How can you redirect frustration when students are not getting the problem right. Build in contextual hints, depending on which wrong answer is given. Providing workshops or recitations? something else? how do you convince students that even though the computer "knows" their answer isn't right, it doesn't "know" what they did wrong? Is having different problem versions a good idea? It makes it somewhat harder to write (or at least debug) problems Some versions of problems may have differing difficulty Does this discourage co-operative learning? Homework model vs Gateway quiz model -- how do you compare these? Is one better or more appropriate? Homework model Student works on the same problem until time is up (or some number of attempts is exceeded) No answers provided, just checking correctness Problems are each done independently (Usually) no time limit (in hours) for working on the problem -- just a due date Problems are moderately different for each student--but each student gets only one version of the problem to work on until they get it. Gateway quiz Student submits an entire quiz The quiz is graded and returned If the student does not get a high enough score they take a new version of the quiz. What are the inherent limitations on web homework questions? One should ask the questions one should, rather than the questions one can! Proofs of theorems can't be effectively checked Narrative descriptions are even less likely Number and function answers can be checked fairly easily Multiple choice, matching, etc. etc. Graphical input is at least a possibility WeBWorK, WebAssign, CAPA, Blackboard, Prometheus, WebCT WeBWorK and CAPA, and to some extent WebAssign use a "programming" or algorithmic paradigm for developing problems. More flexible -- higher learning curve. In particular WeBWorK allows one to program the answer checking as well as the delivery of each problem. Others use a "database" paradigm. Fill in the parameters in one of several problem types. WeBWorK and CAPA have facility for TeX based hard copy output. I don't think the others do. Things to consider Extensibility Ease of use -- particularly for students, but also for instructors Availability -- especially for students. WeBWorK won't work with a class that does not have easy access to the internet.

<| Post or View Comments |>