** Using WeBWorK **

A colleague and I are in the middle of a study to determine how effective these two different approaches are in helping students retain information from early in the semester. We hope to have results this summer.

Each approach has its problems. For (1), we like the focus on the topics of the week and the ability to ask challenging problems on WebWork, but we miss the review. For (2), faculty find that students focus so much on the WebWork assignments which are all review that they do not work on problems from the "current" material each week.

I anticipate our department looking for a middle ground between these approaches, pending the results of our study. What do other schools/faculty do? How are you using webwork? How do you design your sets?

Thank you in advance for your help!

Jeff

<| Post or View Comments |>

Here at the University of Michigan we've essentially done both of the homework modes you've described, albeit mostly because of pragmatic reasoning. In some classes we have daily or weekly homework that follows the material that is being covered in class very closely. This is, essentially, transferring the homework sets we might otherwise want to assign so that students will work the material that we're covering into WeBWorK homework sets. In other classes, where we haven't been as comfortable instituting that amount of homework, we've had sets that are due every week or so that cover much longer sections of material---going back a week or more---which is closer to your Saxon model.

I'm happy to respond to additional questions if you have any.

Gavin

<| Post or View Comments |>

At Rochester, some professors use a variant of your first approach: create sets that contain material from each week of a course, have this material covered in recitations the following week, and then have the WeBWorK assignment due early the next week (e.g. Sunday night). In recitation the TA's will go over and/or have students work in groups on e.g. the TA's WeBWorK problems. This has the effect of turning recitations into quasi workshops.

We haven't done any studies to determine the benefits of this method or any other.

I'm sure other schools will be interested in the results from your study.

Arnie

<| Post or View Comments |>

Thank you for your comments. Carolyn Yackel and I will keep you posted on the results of our study.

In the meantime, I am anticipating that the faculty here are going to want to change the homework sets to consist of 1/2 review and 1/2 new problems.

Jeff

<| Post or View Comments |>

I just discovered Webwork, so I'm exploring what is possible.

I'm designing a Math for Elementary Teachers course, and would like to incorporate review/practice exercises. In contrast to Jeff's approach, I would like the exercises to be generated at the student's request.

This reflects a philosophy that encourages/requires a student to take more responsibility for their own learning. Some students really don't need any review questions on, say, adding fractions. Others need a LOT of practice. They shouldn't need the professor to tell them this, by giving everyone the same quantity of review questions.

I would like to be able to give an optional problem set like this:

Practice questions on adding fractions

Number of problems [pull down menu, 0 to 20] Level of difficulty [easy, medium, hard, mixed] [then Webwork creates a set of problems with these properties]

Susan Addington Cal State San Bernardino

<| Post or View Comments |>

I don't believe that anyone has implemented a capability such as the one you describe. It is possible to assign different problems to different students (although the grading modules will probably get confused).

I think you could also write a problem with an extra check box which tells the problem to reseed the next time it is viewed. This would give the student the same problem but with new numbers. We might have to write a pg module that allows you to save the new seed, but otherwise not many changes would be required.

To get a new version of the problem the student would check "new version" of problem before hitting the "submit button".

Has anyone else tried something of this sort?

--Mike

<| Post or View Comments |>

Though it wasn't really what it started out to do, the versioning in the gateway testing module might be able to do something like this, possibly with a little modification. Briefly, the way the versioning works is that students are assigned a template homework (or gateway) set, and when it's accessed the system creates a new version of the homework set for the student. I haven't tested this for non-gateway tests, but the theory is that it should be able to do general problem sets as well as tests.

I think this could do something similar to what Susan described if one created a number of template sets, say an 'intro/beginning' set, an 'intermediate' set, 'hard', etc. Then a student could take as many versions of each set as s/he felt s/he needed.

It might also be that it might be useful to use problem/topic groups for this type of assignment, as well. In gateway testing each problem on the test is drawn from a group of possible problems.

The two things that I can think may need to be changed to make this work for what Susan is trying are (1) I believe currently all gateway tests are timed, which is probably not what she wants, and (2) I haven't carefully tested versions of assignments that allow more than one attempt per version.

I'd be happy to talk about how the gateway testing module would better address what Susan wants if it sounds as if this is possibly a good solution. I'm going to be doing some development on the system at some point this summer, and could try and wrap that in.

Gavin

<| Post or View Comments |>

I have put together a mechanism for doing something like what you are requesting that does not require changes to the system. You create a set that contains a single problem, but that problem lets the student select a topic and difficulty level, and then a problem of that type is loaded. The student can step through additional problems in that category via a "get a new problem" button.

You can get a copy of the code needed to do this from the Union College CVS repository at

  http://cvs.webwork.rochester.edu/viewcvs.cgi/union_problib/examples/?cvsroot=Union+College

Some caveats:

The statistics at the bottom of the page (number of attempts, percentage correct) are not reset when a new problem is requested by the student (that would require changes at the system level). So these will be cumulative over ALL the problems the student has attempted. (And the "Show past answers" will show all the past answers.)

Hardcopy will always print the first problem in the first topic and difficulty setting. (The hardcopy versions of the problem don't have access to the data that tells which problem the student selected.)

The code tries to clear the previous answers when a new problem is selected. There are situations when this might not get them all (though I think they will be pretty rare).

I haven't done a lot of testing with this, so other issues might arise with certain combinations of problems. For example, there might be troubles if two problems use static images with the same names. (I think WW will give them the same name, and the browser will use the cached version of the first one in the second problem.)

Let me know if there are problem with this.

Davide

<| Post or View Comments |>

Is there an example publicly available I can just try out, to see how it works? I am a bad programmer, and in the middle of writing a textbook, so any help is very welcome.

Susan

<| Post or View Comments |>