The second Saskatoon Girl Geek Dinner was held in the luxurious (not kidding) basement of Point 2. We ate lasagna, Caesar salad and garlic bread. There was dessert too, but I'm not sure what it was - it looked good though. After some chatting and eating we settled in for our speaker Julita Vassileva, of the MADMUC Lab in the Computer Science Department at U of S. Julita is a a Professor and researcher in online communities and she's also the Cameco NSERC Prairie Chair of Women and in Science and Engineering.
Her presentation was titled "Social Visualization As a Tool for Motivating Participation" or as she said, "what can we do as designers to make our sites more sticky or attractive?". I made some notes which I will share here, but, of course, I wasn't able to type as fast as Julita talks. As they say, "you had to be there".
Social Visualization is based on several areas of related work:
- User Modeling - creating profiles of users,
- Artificial Intelligence in Eduction - mistrust and privacy issues - 10 yrs ago there was the idea to make models inspectable, leading to Open User/Learner Modelling (OLM).
- CSCL (Computer Support and Computer Learning) People learn through collaboration - which interaction is productive vs. non-productive.
- Interaction Analysis (IA) - Computer support and collaborative work - Carl Gutwin, Director of the Human Interaction Lab at U of S was one of the first people to coin the term "awareness" (assume awareness in terms of Interaction Design) - so you know that one person is typing as you are.
- Ensure awareness of progress & reflection
- Learner to correct errors or involve in dialogue
- Provide teacher with a tool to monitor learner progress
- Provide teacher with overview
- Provide model and suggestions
HCI/CSSW
- Communicating knowledge
- Activating social norms (can exert pressure)
- Triggering users to self regulate their behaviour
Comtella, developed by MADMUC Lab, is a peer-to-peer file and bookmark sharing system, allows students and researchers to share papers (either their own, or downloaded from the web). They studied Comtella in order to determine patterns of participation and learn how to improve participation among contributors. One goal was to get people to identify with an elite group of top contributors. Users visualized a star - denotes more or less giving depending on brightness of the star. They competed only by number of papers contributed. Participation jumped (due to cheating to boost their ratings - very motivational) after new views introduced but then dropped again.
Discovered that they needed to eliminate user effort and integrate as much info as possible into one view. Visualization should be intuitive and self-explanatory. Had to look into "Freakonomics" to find out how to deal with cheating. To counter set the rules so it's cheaper to behave honestly (obscure the rules - don't tell them how you are measuring participation), reward early contributions, do not reward excessive contributions, reward depends on the quality (received ratings).
Gave explicit currency for rating - C-points - earned with each act of rating. Can be invested to "sponsor" own links. These decay over time. Want to see you contributions at the top. You can pay for sponsored links with your C-points. In a test group 2x as many ratings compared to the control group. Found they could orchestrate the desired behavior. More contributions at the beginning of the week and no over contributions.
A version with no competition was also introduced with the purpose of encouraging reciprocity in reading postings and encouraging ratings. Came up with a visualization of the symmetry of relationship of people in the community.
Should be a movement to reciprocity, ideally. People are curious. You will be flattered if someone is reading your post and go to read theirs. Immediate reward - if you rate a post it will change colours (based on energy units - goes from blue to yellow as it gains popularity).
Experiment with Philosophy class and CMPT class (they were comparable - but not identical) - to test connections of lurkers. The visualization stimulated participation.
Important to decide which actions you wish to stimulate - social comparison should be centered on them. Pack them in default view. Make visualization attractive and intuitive. Don't overload with too much meaning, be careful with those that don't have too much meaning as they might attach too much meaning where there isn't any.
There is a gamer in everyone (esp. competitive computer science students). People are generally vain they want to see their stuff. Compete to get gold membership, or star brightness. Even the brightest plastic star was desirable to those who could only afford this - different dimensions of competition.
Participation depends on what your purpose is.
