Spring 2014
This seminar covers methods for conducting research in human-computer interaction (HCI). These topics will be pursued through independent reading, assignments, and class discussion. The seminar comprises four assignments that not just apply HCI research methods but push the envelope. The assignments are designed to be meaningful and have the potential to be widely visible or to be published. We will have readings that teach HCI experimental research methods and provide examples of valuable contributions.
The goal of this course is to provide students with the background necessary to perform research in HCI and the skills required to conduct human-centric research. There will be little or no content in this course about user interfaces, but students will find topics in CSCI 1950i (User Interfaces) relevant. Enthusiastic students who have not taken CSCI 1950i should contact the instructor for a registration override code. The class will meet in room 506, CIT from 1:00pm-1:50pm MWF.
Jeff Huang, 407 CIT, jeff at cs dot brown dot edu
Alexandra Papoutsaki, 507 CIT, alexpap at cs dot brown dot edu
Crowdsourcing assignment Experiment with different crowdsourcing models to generate an accurate database of all computer science professors, including metadata like degrees, subfield, rank, and advisor. You will each be responsible for a handful of universities and we will aggregate the data at the end to make the database public, and publish the lessons learned.
Experiment bughunt Comb through six empirical papers from CHI/UIST and identify experimental errors. We will aggregate this information to find out what are the common mistakes in HCI research, and publish our findings online.
Redesign assignment Create a provocative redesign of a classic interface using the design research approach. For example, you may change the user model from action-object to object-action, or use real-world metaphors to design affordances. Popular online examples: dontclick.it, boarding pass redesign, automotive gesture touchscreens.
Fitts' Law study Run a variant of the classic Fitts' Law experiment, which can be combined as a class for a meta-analysis that can potentially be submitted for publication as a class.
* Best Paper Award winner; we will look at the corresponding reviews in class to see what reviewers liked about them.
** If you have trouble accessing this paper, an alternate link is http://cs.brown.edu/courses/csci2951-l/[lastname].pdf.