Libby Hemphill research and posts on social media, collaboration, and related technologies

Seminar notes 2/6/06

Guests: John Laird (EECS), Chris Quintana (Ed)

John: 7 principles of game design
Games - rules + goal (default goal, not just one you make up; i.e. winning at basketball)
Addictive nature - from intermediate goals; short-term, long-term, overarching goals
Interactivity - concentrate on activity rather than story, character, artwork, etc.
Sidenote: gaming and surgeons; surgeons who played a lot of game (something like Smash Brothers) were better (faster, fewer errors)
Feedback - players need to know what goals they need to achieve, where they are in the goals, etc.
Variety
Limit meaningless repetition - skip parts already seen, goal-oriented repetition OK
Consistency - know what to expect, visible reason for failure
Fairness & balance - no single dominant strategy (rock/paper/scissors)

John's editorial - games don't really teach you anything about the real world; some exceptions

1. training - second-language learning example; game where you have to speak Arabic to achieve goals in the game; inflexibility of the game important in helping you pronounce appropriately

Chris - asks about Schank's case simulation stuff
Note: John's making distinction between training and education, skills vs. concepts

Do games help us learn? Priti and John - maybe games help with component skills (expanding working memory and visual attention) but not domain knowledge development

John - good game design and learning at odds - good games don't allow time for reflection; good games move to new goal quickly

Gender and games - maybe these are design principles for "games for boys"
Jude - maybe role playing's involved some how
LINK: purple moon - games for girls, different goals

Design vs. design research

John - generalization, taking findings from one place and using them elsewhere makes it research
Chris - James Paul Gee's group at Wisconsin
John - 1/10 games makes money; we don't have a science of entertainment (movies, games don't always make money)Priti - John, are you a learning scientist?
John - no, I'm not a learning scientist; interested in how people learn, but it's too difficult in how to understand that; not interested in sociological aspects of learning
Priti - we'd probably claim that's learning sciences research

Research Methods (LS class)

Guests: Judy Olson, some other guy in a green fleece

Notes about PIM research: Forks poetry faculty

Ann Brown – article in learning sciences
Cobb – Ed Researcher 2003

Intervention – for psych like treatment

Eric’s question: isn’t this like HCI? Barry’s answer: yeah, pretty much

Priti: figuring out what counts as a small theory
Barry: how would you test it? – falsifiability is key

Frameworks, theories, models –
Judy: frameworks are early, say what the variables might be; models and theories are predictable and formal, models more mathematical;
Priti: same in cogsci, models are just more specific
Barry: model of teaching built on theories
Question: how does this fit with 722? Theories inform models, etc.

Priti: for Anderson, ACT-R is the theory

Conjecture –
Maybe the “why” of hypotheses (conjecture: why A leads to B, where “A leads to B” is your hypothesis)
For Barab – the scheme or mechanism of how things are going to relate
Ex. Chi’s article –
Hypothesis: self-explanation improves learning
Conjectures: when and how self-explanation would lead to learning

Question: how do conjecture, model, formalism, coding scheme, design study fit together?
First pass: conjecture and model tell you what aspects you think are important and how they fit together, formalism is boxes for coding scheme that fit that model

Appropriate comparisons – not did they learn but would they have learned more in some other way?

Design vs. design-based research –
{diagrams on the board}
design-based research has a lot of variables, a lot of interactions, explanations for all of those, iterations on those variables and interactions
design – variables but not necessarily explanations of interactions
question: how is design-based research different from other kinds of research?
Answers: iteration, less control over variables (design-based research makes explanations but doesn’t necessarily control variables so much as report them)

How DBR is different –
More akin to engineering, turn theory and understanding into a difference in the world
Don’t just describe the world but perturb it

Question: where do design-based research studies get published, and what do those publications look like?
Answers:
ijCSCL
question is “who cares about this topic”
Cognitive Science Society – emerging learning sciences theme this year (implicit and explicit learning)

Ed Research –
Epidemiological – get a lot of data, just grab a lot of data (lots of variables) from all over, large statistical methods (public v. parochial)
Small-scale intervention – tardy slips, points for detention, reduce class size; small single change and examine what changes occur
History – Dewey lost, Benet won (psychological testing, quantifying educational success)

Judy:
Anderson automatic post office study
State transition diagrams about design meeting – size of bubble corresponds to total time in discussion
Go research (replicate Chase & Simon chess memory stuff)
Recall orders reconstruct patterns – name your family members, play Go or chess, give someone a place to start, regularities in the recall order indicate a tree structure
Acquiring structure of a field

Design-based research in collaboratories – go in thinking something’s going to work, try it out, tweak it, then see what happens and iterate

DBR and the field – don’t have to be in the field to be doing DBR

Barry asks Judy, “how do you relate to learning sciences”?
Judy: nope, I don’t study learning; maybe adoption

Barry: isn’t improving group performance a learning task?