CMSC 434/828S - Fall 2003
Prof Guimbretiere

Introduction to Human-Computer Interaction


Home Contact Syllabus Schedule HW & Project Links

Project phase #4 (due 12/04/03)

Evaluation

Overview

Now that you have a running prototype, it is time to have target users evaluate it for real. During phase #4, you will design and conduct a qualitative evaluation of the system you built during phase #3.

What to do

To conduct your qualitative evaluation you should:

  • Starting from your initial list of goals, identify at least 3 major tasks you wish to evaluate,

  • Design a protocol for your qualitative evaluation. Your protocol should describe how you will evaluate the 3 major tasks you picked including the testing technique used, where the test will take place, the testing configuration, the initial state of your system and the final state to be reached by participants. Your protocol must at least include a direct observation evaluation, a post-observation survey and debriefing,

  • Run the protocol for at least 6 participants. For best results, you must be careful on how to pick your participants: as before, it is important to pick participants who are representative of one or more personas and to have at least 4 participants per persona. Don't forget to have each participant sign the consent form before starting the experiment.
    To avoid wasting your participants time (and yours) on fixing the protocol during the experiment, it is very important to debug your protocol on members of your group before running it with participants.  Finally, for your results to be useful you must not influence participants during your evaluation but only take notes.
    In order preserve the privacy of the interviewees you should not gather private information such as name, ID and the like.

  • Analyze the information gathered during the qualitative evaluation, and identify which part of your design should be changed.

Deliverables

Your group will deliver a portfolio written to the imaginary VP of product development in your company. The portfolio will include the following sections.

Section 1 (40 points): Protocol

  1. A detailed explanation of the design rational behind your protocol (~ 2.5 pages), including:

    • The overall goals of your study,

    • Which testing technique you picked and why,

    • How you selected the tasks to be performed and how they are related to the study goals,

    • How you selected your users and why,

    • Your consent form,

    • Your experimental setting including where the test took place, how it was performed, the initial state of your interface and the expected final state,

    • The questionnaire you used for the debriefing and how each question relates to the goals of your study.

Section 2 (40 points): Observations

  1. For each participant in your experiment present a summary of the testing session (1 page per participant), including:

    • A summary of your group observations,

    • The participant completed consent form and survey,

    • A summary of the debriefing sessions.

  2. The key findings of your evaluation, and for each of them, how you reached this conclusion from the data you collected (1 page).

Section 3 (20 points): New prototype

  1. Pick two major problems and propose a solution to address these problems. Your solution should be compatible with the rest of the design. It might be useful to use a Before/After structure so that changes you made are easy to understand. For each change, do not forget to explain why your changes will address the problem at hand (~2 pages excluding figures).

Section 4 (20 points, 828S students only): Experimental design

For this project you are asked to run the experiment you proposed during phase #3 and produce a report of your results:

  1. Provide a copy of your consent form,

  2. Present possible discrepancies between your proposed design and the design actually used. In particular, state how many participants you ran (the same participants can be used for both studies),

  3. Present your results. Explain how these results agree (or not) with your predictions. Explain possible discrepancies.

  4. Present and explain other insights you gained during the experiment.

For this section it might be useful to look at the book called "Doing Psychology Experiments" by D. Martin, This book is placed on reserve at the CS library.

Notes (for 434 and 828S)

The portfolio is intended to document the progression of your design, which includes your final project. Your portfolio must be neat, well-organized, and visually appealing. Portfolios should be constructed out of a 1" or smaller 3-ring binder (we will not appreciate having to carry around larger binders). Your portfolio should also use titled section separators to separate the major sections. The cover of the portfolio should include the names of the group members, the group number, and the title of the project. The first page should be a table of contents, which will grow over time.

Each group should be prepared to give a 10 minutes presentation of their portfolio during the class when the project is due. The presentation could be made either using PowerPoint slides or by using material provided in the portfolio.

Grading will be based upon the sophistication, logic and maturity of the presentation, and the completeness of the work.