CSc 280

Project 4: Go Forth and Redesign

Due dates:

Approval deadline

Mon, March 3, 2008

Experiment Design
Presentations

Thurs, March 6, 2008

Paper Due

Fri, March 14, 2008, at noon

Final Presentation

Mon, March 17, 2008, at 1pm

Objectives

Summary

In this final project, your team (which can again be different if you choose) will plan an entire usability study from start to finish. You will make a proposal for a redesign of an existing artifact, create at least two alternative designs, implement them, develop a hypothesis related to some aspect of your designs, plan an experiment to test the hypothesis, carry out the experiment, and then analyze, write-up, and present your results.

The Details

Your final project is to redesign an interface of your choosing. Some examples might be a part of a web site, a cell phone, a PDA, a game, a piece of hardware (like a mouse), or just about any computer interface you can think of. Note that I said part of in that last sentence. The temptation here will be to pick something way too big: (I hate the way eBay does its auctions. Let's redo it!) But you must refrain from trying to solve all of a software's usability problems all at once. In academic HCI research, the goal is to pinpoint one piece of the design that you can test with an experiment to see if that piece has a real usability effect or not. So you must keep your redesign small and focused. You may feel that Travelocity is sorely in need of a complete revamp, but that's too much. Redesigning the way a customer inputs her trip details on the reservation screen is of a more appropriate scope.

The only real requirement of your redesign is that it must be a comparative design that leads to a controlled experiment. So you must have at least two alternative designs. Remember to keep the number of IVs to a minimum (one would be good) and to minimize the number of uncontrolled variables. The designs should differ from each other by one small aspect, though they may both differ from the original artifact by as much as you like.

The Proposal

Your choice of what you redesign should not be based on preference (or hatred) of the existing system, but rather on solid heuristical evidence that its usability can be improved. Your entire team needs to have an in-person meeting with me by the end of the day on Friday, Feb. 29 to have me sign off on the project. In that meeting, you should give heuristical evidence of why an artifact's current usability is faulty, and how your redesigns will address it. You do not need to have all of the experiment design worked out (that's for the following week: see below), but you should have enough of the interface design mapped out so that all of its functionality is known, and so that I can see that your idea is of appropriate scope. Feel free to have informal discussions with me before this group meeting to vet possible ideas so that you don't go down wrong paths for too long.

Experiment Design presentation

In class on Thurs, March 6, your team will present your plan for how you will conduct your final experiment. All team members may present as a group. The purpose of this presentation is to obtain constructive feedback from your peers and myself as to how your experiment design can be improved (e.g. validity threats you may have overlooked, quality of measurements taken, quality of operationalized variables, etc.) Each team may take up to 20 minutes to present with ample time for comments and questions to follow. Remember, you'll need to explain what you're redesigning too! Be sure to include as much information as possible including tasks, subjects you wish to recruit, protocol, IVs, DVs, questionnaires, etc. Handouts may be useful. The whole point is to make sure each team ends up with useful data that can actually confirm/deny the hypothesis you are posing.

And the rest is up to you

Just about everything else is in your hands. You have just over three weeks for this project, so it's possible to do two rounds of testing if you move quickly enough. A test with a low-fi or Wizard of Oz prototype done in the early stages may keep you from wasting time implementing something unusable in the later stages. Alternatively, your team may arrive at a more usable design via a thorough set of task analysis interviews and ethnographic observation sessions. It's up to you to choose the tools that will serve you best. But choose wisely. Your grade will depend, not just on the quality of the final report, but also on what tools you decided to use during these three weeks to arrive at the best design you could. For example, if your team decides not to use coding or event logging just because it takes a long time to analyze, even though you have chosen an experiment design that could benefit from that data, then your overall grade will be adversely affected, even if what you did, you did well.

How to turn this in

As always, the result of all this will be a written and an oral report from each group. The audience for the paper is still an HCI-knowledgable reader who is not familiar with this assignment. This time, however, the audience for the final presentation will be the same as the audience for the paper. There will be invited guests at the final presentation, so it will be necessary for you to include enough background (not HCI background, but background on your design and your experiment) so that everyone in the room can follow. The final presentation can be conducted by your entire team. Clips, as always, are welcome. Each group has 25 minutes to present, with 15 minutes of Q & A to follow. Note that the paper is due on the Friday of the last day of classes, while the presentation occurs three days later.

Please turn in an electronic copy of your written and oral report as well as the implementations of your redesigns on Blackboard. In addition, turn in a paper copy of your report on the due date.

CSc 280 Projects