CSE 4441/5351 Human-Computer Interaction

Winter 2009

Assignment 1

Due: Thursday, April 9, 2009 @ midnight

Design and administer a questionnaire to a group of people ("participants").  The questionnaire should collect data on characteristics of the participants and aspects of their interaction with computers.  Analyse the data and write a report outlining the topic and related work (Introduction), the methodology (Method) and the results (Results and Discussion). 

Participants are drawn (ideally, at random) from a population.  You may use the York University community as the population or some other population conveniently available to you, such as people at a local shopping mall or subway station or in the neighborhood where you live.

The population can be narrowed as appropriate.  For example, if you are only interested in Apple Mac users, mobile phone users, or people over the age of 50, you can screen candidate participants and use only those from the desired population.  Use at least 25 participants.  Hopefully, you can keep participant involvement to 5-10 minutes, but this depends on the data collected.

Will you ask participants to complete the questionnaire themselves or will you interact with them and complete the questionnaire together?  In the latter case, you are more likely to get complete and accurate information since participants can ask for clarification on questions they do not fully understand.

So, what sort of data are you to collect?  A questionnaire usually begins by asking simple demographic information such as age and gender.  You can also get more specific information, as relevant to the topic, such as first-language spoken, number of hours per day using a computer, hours per day playing video games, preferred browser (IE vs. Firefox), preferred computer type (Mac vs PC), preferred texting method (multitap vs. T9), number of text messages sent per day, estimated typing speed, and so on.  Can you see possibilities for interesting relationships or ways to summarize, group, and graph the data? 

Try to venture beyond the examples above.  Remember, this is your assignment, not mine!  Can you think of other relationships?  Do Faculty of Arts students vs. Faculty of Science students differ in…? Does age or gender make a difference in…?  Do left-handed people prefer…?  Is the number of text messages sent per day related to…? Are mobile phone users more likely to…?  Which pointing device do users of notebook computers prefer?  Do Mac users really dress like that guy in the TV ad?  Are Starbucks devotees more likely to use an iPod than clients at Tim Hortons?

Sounds like fun, right?  Sure, but relationships you find or seek to find need a plausible explanation.   For example, I doubt people who wear glasses send more or less text messages than people who don't wear glasses, but I could be wrong.  If you want to look for such a relationship, that's fine, but make sure you have a plausible reason (and include it in your report).  If, in the end, you don't find the relationship sought, that's fine too. But, an explanation is necessary.

Think about the way information is gathered in the questionnaire.  Do not ask participants their name.  Instead, each is given a code (P1, P2, P3, …). Simple nominal-scale, or categorical, data are gathered as follows:

      What is your gender?    ____ male    ____ female

If you are interested in knowing how many hours per day participants use a computer, the questionnaire could include

      Please indicate your computer usage:
      ___ 0-2 hours/day     ___ 2-5 hours/day     ___ 5+ hours/day

Gathered in this manner, the responses are tabulated as counts, or number of respondents, per category.  These data are useful to show categorical relationships, such as computer usage (low, medium, high) vs. gender (male, female).

You could also solicit the same information as follows:

      How many hours per day do you use a computer? __________

Responses framed in this manner are examples of ratio-scale data.  Provided you have a second ratio-scale characteristic, such as age, you can show the degree of correlation (r) between the two variables.  You can even build a prediction equation where one variable is predicted from the other.

You can also use a 5-point Likert scale with a preamble such as "Please indicate your level of agreement to the following statements".  An example statement is

      There should be a law prohibiting the use of mobile phones while driving.

                 1   2      3         4 5
            strongly      mildly      neutral      mildly      strongly
            disagree      disagree           agree agree

Are males or females more likely to agree with this statement?  I'm not sure, but this assignment provides a vehicle to find out. (Pun intended.)

In the Introduction, discuss related work in the literature.  Google Scholar is a good tool to find related work.  Include citations to at least two papers in the literature (not authored by your professor).  The papers you cite must appear the end of your report in a section called References.  Format citations and references, as per the examples in the template file (see below).

For the report, please use this template.  Save the template on your computer and rename it a1.doc.   Read the template for information on sections, sub-sections, formatting, etc.  The final formatted report should be 3-4 pages.

Submission.  Two files must be submitted.  Submit your report and questionnaire in Prism using the commands

submit 4441 a1 a1.pdf

submit 4441 a1 questionnaire.pdf 

Please include your original files, as well:

submit 4441 a1 a1.doc

submit 4441 a1 questionnaire.doc 

Please retain your completed questionnaires for future reference, if necessary.

 

Good luck.