I attended the 30th Annual HCIL Symposium on Wednesday and Thursday. Unfortunately, I missed a few of the sessions because of work responsibilities, but I’ve included my raw notes below. I don’t have time to do all the linking and such that I usually do, but you can find most of the papers and other material on the HCIL site.
Welcome/Intro
Jen Golbeck
– provost intro: get a taste but then can’t stay; just here since October
– awards for student research (sponsored by Yahoo!): Cody Dunne – DataVis award; Megan Monroe – DataVis honorable mention; Beth Bonsignore – HCI award; Jason Yipp – honorable mention; conference travel award of $900 each; Cody gets $500 check since he’s graduated with his PhD
Visualization: Turning complex big data into meaningful small data
Ben Shneiderman
– in 1983, visualization wasn’t a big thing; it’s a big thing now including Obama’s “Big Data” initiative started in 2012; goal is “visual reasoning”
– Solving Problems with Visual Analytics (2010)
– photos of multi-screen applications; 10 million to 100+ million pixels
– photos of small screens on mobile devices; a few meaningful pixels
– visualization mantra: overview, zoom and filter, details on demand
– users make the decisions; humans not machines; machines support humans
– SciViz – 1,2,3D visualization; InfoVis – multi-var, temporal, tree, network
– early projects: visible human, faceted menus, american memory, hypertext link, ACM communications hypertext influenced Tim Berners Lee; called embedded menus, Lee coined hot links
– zoomable, dynamic user interfaces
– SpotFire examples; gene discovery, etc.
– Treemap examples: smartmoney marketmap, etc.;Treeversity: comparing trees over time
– “Visualization gives answers to questions you didn’t know you had.”
– NodeXL network visualization examples: political votes, twitter use, grants
– medical visualization: LifeLines, LifeLines2, EventFlow; Oracle Health Sciences supported; hope to incorporate into an Oracle product some day [Hyperion likely]
– “One picture is worth 1,000 words. We think one interface is worth 1,000 pictures.”
– [very similar materials to what he presented at CTO Summit the other night; quicker with more jargon]
– Q&A: most exciting new think in visualization today? spreading ecosystem – vis embedded in many other systems leading to visual literacy and explore by themselves, the broadening audience; scaling up to larger and larger data sets, how do you get billions and trillions of items lead to a specific decision
The Power of Simplicity
Ben Bederson
– fundamental tension between simplicity and power, richness and focus; tension in design and communications for a long time
– example: directive to make government data open and available digitally; used youtube to introduce the importance of this initiative; 60 seconds to get the message across
– researchers tend to focus on power; most folks are looking for simplicity
– PhotoMesa – zoomable photo library UI; convinced this would take over the world; what succeeded? one dimensional scrolling like Picasa and most other photo systems (even file browsers)
– zoomable web browsing – research showed most browsing was revisiting pages they’ve already visited, experimentation showed that it was more effective for people; firefox implemented panorama view which was zoomable in one version; what succeeded? the back button
– zoomable animated presentations – KidPad, CounterPoint, PPTPlex, Prezi (people dabble with it, but don’t often reuse it; too complex)
– even Apple tried 2D grid in Spaces, went to 1D; MS did so in OneNote
– What happened? “eye candy” that doesn’t add much; don’t scale well; spatial memory not as good; 2D harder to scan than 1D; where to zoom?
– What happened (similar)? complex-cognitive load; too much screen space; no focus on common tasks?
– What should we do? Build systems that are simple, small, and support common tasks
– successful zoomable: Google Maps (inherently 2D – new version removed UI even though added features); animated transitions on iPhone; pinch to zoom (natural interaction)
– which InfoVis are simple, small, and support common tasks (and are readily available and and web based)? Google Charts and D3 are a start, but not enough
– Simplicity is Power
– recent paper reviewing 15 years of zoomable UI
– Q&A: new Honda Accord UI is not simple; any hope that the complexity of these things has peaked? not connected to automotive design research; in my home, some people love the power and others who despise the complexity; society is grappling with this problem of power vs. simplicity; these are not new ideas, but they are still huge challenges; new user -> intermediates -> experts -> intermediates; how does simplicity apply to voice interfaces like Siri? HCIL has traditionally been against too much use of voice interfaces; Google seems to be developing the technology to where the recognition quality is high enough it may work in some areas; Dick Bolt – “put that there!” 30 years ago; Google showed a new commercial implementation that may finally see that work; please keep making crazy zoomable interface, researchers need to keep pushing on the power edge, over time they get more usable and embedded in system, simplicity comes later? we shouldn’t avoid power and only go for simplicity; but how do we think about the balance; my own research would have been more successful if I had considered simplicity
Designing with and for Children
Allison Druin
– all of her work has been in teams, including many children
– why design with children? today’s children are not the same people we were as children; same questions, e.g. “what does it mean to be brave?” – bullying in physical world vs. digital; “what does it mean to be different?” – today many more dimensions of individual difference (obesity, autism spectrum, international adoption, etc); “what is my future?” – today with the media can see people excelling at young ages, have I missed my chance?
– in only 15 years, computers have gone from being magic, to ubiquity, to mobile computers used by children, to using all kinds of computers anywhere- social; the social and technology ecosystem is much more important today
– must always work with children to understand who they are and what they experience
– children in the design process: user, tester, informant, design partner
– users can absolutely contribute to the design process at all points, even if they aren’t designers or developers; co-design, cooperative inquiry, participatory design
– designing with kids – lots of craft materials for prototyping – is messy; edutopia video of kids co-creating
– reach of children: 15 years, 47 child-partners, 10 dissertations, 38 students-staff-faculty, 98 publications, 26 partnerships with companies, universities, and others (not just sponsors, but partners)
– Do Not Touch button created with a partner (Nickelodeon) – press it, slime or creatures appear; nothing to do with the application, but very successful for the partner
– lessons of children – mobile technology, robots, what does it mean to connect
– the blue sky: SCALE – processes aren’t good enough for co-design with dozens to hundreds to more people; diversity – include children from everywhere; ROI – can’t prove it yet
– challenge: when creating your new technologies think reflectively, voice change, partner
– Q&A: do any of the children come back after 20 years and say what happened? from 7–11 can work with us; stay away for at least year; come back as an intern; high school kids teams with some partners; longitudinal study looking at outcomes – many wrote college application essays based on their experiences at the lab
Social Media
Jen Golbeck
– citizen science -Â biotrackers.net
– single language translation – human machine collaboration
– social media – how are communities created
– example of spam: first recorded spam was on USENET; community was livid and there were no filtering mechanisms; community took revenge in many ways after they found who had posted it
– we’ve hit a point where technology has surpassed our ability to understand how it’s being used
– Jen’s research is creepy; it shows what can be done with the information on the networks; she loves the research but hopes no one actually implements some of this stuff
– algorithms for computing trust between people in social networks
– predict political preferences based on who they follow on Twitter; method predicted 98% of who someone would vote for – in two elections
– computed personality scores based on twitter; within 10% of other methods which is lower than normal variance in other methods between two tests
– another study of 65000 people’s “likes” on Facebook; could predict race, religion, sexuality, behaviors such as drinking alcohol or eating habits; can build models predicting attributes from seemingly unrelated data from social networks
– example: Target sent offers for pregnancy items to his teenage daughter before father knew it; their data can even accurately pinpoint due date; really creeped people out so they started adding random other items in the catalog to hide their focus
– what if we started using this commercially? the outcomes are not always accurate, and there’s no way for an individual to fix an erroneous predictions about them; most people don’t understand the issue that the things they share are leading to predictions way beyond that
– challenge: how to inform people what sort of risk and use that providing a particular piece of information may entail; what is the value
– Q&A: as people learn more about what these algorithms can do, could some people learn to manipulate by gaming the system? the simpler techniques can be easily gamed, but the more complex stuff is hard to understand what’s really happening in the  algorithms; the behaviors and predictions are emergent; e.g. people who “liked” and “enjoy being a mom” correlated with lower intelligence, and liking curly fries correlated with higher intelligence; can I learn things about myself or what information an organization knows about me? that’s a dissertation hard problem to reverse engineer that, really fascinating, great research problem; with these kinds of algorithms, it may be that an organization could know more about your deepest desires and beliefs than you do, what about the ethics of this, there’s a responsibility for the people creating these technologies to serve a public good and make it transparent? there are people researching the ethics of these sorts of things, I know little about the ethics of this so I wouldn’t want to make ethical statements, whole research area in how to communicate the output of algorithms
https://www.faychildrensclinic.com/clomid/
Touchscreen Accessibility: Supporting Individual motor abilities
Leah Findlater
– touchscreen mobile and tablets will outnumber PCs in 2013
– videos of people with motor impairments using touchscreens [iPads]
– research: what are the mainstream touchscreen devices are in use on a daily basis? what are they being used for? what kind of adaptations are being used?
– most studies on motor impairment use 5-20 participants; how do we do a study with many more participants? find and analyze user generated content (youtube videos) showing people with physical disabilities using touch devices
– 187 videos in dataset; resulted 60 disability related search terms * 9 techology-related terms; 101 different users
– coded videos on 21 dimensions; e.g. age, emotion, direct/indirect interaction, touch details, use of external objects
– medical conditions: cerebral palsy in 45+ videos; almost half were children, almost even split male/female; 78% ipad, 17% iphone, 5% other
– Finding 1: interaction styles – 92% direct touch (29% index finger – fingernail doesn’t trigger touch; 16% hand – precise control difficult; 5 videos with nose – single point of input; a couple with feet – ipad is a stimulus; sling support in 13% of videos), 8% indirect (stylus, head stick, mouth stick);long dwell times, fingernail, accidental touches, dragging and sliding, multitouch
– Finding 2: adaptations – screen protection; gloves and sleeves; homemade headsticks and mouthsticks; physical barriers to mask inputs
– Finding 3: sentiments expressed – mostly positive or neutral (e.g. “gives me freedom and indpendence”); 6 videos with negative
– how can we design more accessible interactions? allow control over sensitivity of device; provide alternatives to multitouch; ignore long touches; support DIY physical guides; Apple’s assistive touch was not shown being used in the video data set
– reflect on the method: user generated content is effective for design research
-Q&A: did you create a playlist to provide a set to others? haven’t decided if it’s appropriate to aggregate other people’s videos even though they are publicly available
Gamifying Green: Persuasive Techology and Gamification to Promote Proenvironmental Behavior
Jon Froehlich
– is it making sports teams green? not today
– gamification – the use of game design elements in non-game context
– examples: fitness, finance
– examples: bottle recycling game (VW); used way more than nearby conventional one, Nest leaves, ecodriving trophies
– reward for good green behaviors
– gamification ingredients: points, levels, goals (of different scales), leader boards, feedback, collections, playfulness, narrative, customization, self-expression, unpredictable reinforcement
– feedback is key – informational and motivational processes; assess and direct
– eco-feedback
– not easy to combine into a good game; requires good design and execution
– opower (arlington) – power bills provide comparison, rewards, levels, goals, feedback; loss aversion, feedback, descriptive social norm (if everyone is doing it, I should be; drive to average), injective motive (what direction is good)
– ecodriving: link between driver behavior and fuel efficiency – 26% variation; if 1/3 of drivers paid attention 33 million tons of CO2 per year; constant mileage measurement changes behavior; Nissan carwings with trophies etc.; not enough research yet; Fiat study found 6% improvement
– gamification is faddish and hyped; it’s an inadvertent con; intrinsic vs. extrinsic motivation; when sensors are wrong feedback is wrong and behavior incorrectly influenced, and may not be able to fix it
– we’re at the Atari 8-bit level in this today; long way to go, but lots of potential
– Q&A: how does runkeeper work; is this personal or social? running is personal to me; other folks may care; problem is particularly damaging because of this
Crisp Answers to Fuzzy Questions: Designing lessons for crowdsourcing decision inputs
Alex Quinn
– buying a car is difficult; 80% of buyers spend 18 hours or more researching
– crowdsourcing for data-driven decisions: how do you design the tasks to make this effective
– mechanical turk design is a barrier to workers using it
[had to leave for conference call]
Combining Crowdsourcing and Google Street View to Identify Street-level Accessibility Problems
Kotaro Hara
– there are few ways to find accessibility problems a priori
– can street-view with crowdworkers help do this for us?
– traditionally street audits are performed by governments and other orgs; expensive and not kept current
– categories: missing curb ramp; object in path; surface problem; ending sidewalk; other
– can motivated workers identify problems? can workers perform the task?
– 229 images, manually curated for testing; average age of photo 3.1 years
– observed moderate to substantial agreement between researchers
– does it reflect what wheelchair users would see as problems? 75 images with think out loud with wheelchair users
– strong agreement between researchers and wheelchair users
– ground truth method: majority vote of researchers = a problem
– how accurate are turkers? 185 turkers – 78% accuracy with 1 turker; 88% with majority vote
– some false positives; overlabeling; some turkers low quality; majority vote helps removes some of the noise
– more turkers leads to more accuracy; not much more benefit after 5 turkers
– future research: scaling, automatic curb detection, accessibility aware navigation system
– Q&A: what happens when images change on Google? ask turkers to recode new picture; any work with governments or others to rate relative severity of problems? found some agencies interested; potholes? we’re not working on street problems, but it could be done; how reliable is “looking” at finding all the problems?
Challenges and Potential of End-User Gesture Customization
Uran Oh
– gestures designed by experts; some people want to define their own for personal use
– potential benefits: memorability, efficiency, accessibility
– research has shown people who deign them remember them better; shortcuts are more efficient; individual differences lead to some gestures easier than others
– how do typical users create gestures? what challenges to they encounter? how can we support this process?
– 20 participants with touch screen experience
– open ended gesture creation – 12.2 gestures created on average, number of gestures correlates with arbitrariness, tended toward familiar gestures, gesture recognizer was opaque, finding: difficult and need to know how recognizer works
– action-specific gesture creation – 12 specific actions, most not satisfied with their first set, inability to improve, 76-88% accuracy after training
– saliency of gesture features – rank distinguishability of 9 features, objective features are more distinguishable, speed and scale are less distinctive, order count and finger count are more distinctive
– benefits of customization cannot be achieved without system support
– Q&A: any suggestions for the recognizers? recognizer needs to understand mental model of the user; password gestures might be better if the recognizers are better
Effects of Scientist’s Feedback and Peer Teamwork on Novice Citizen Scientists’ Motivation and Contribution
Yuron He
– citizen science – citizens can contribute to scientific projects
– challenges: more volunteers and data
– previous studies, existing volunteers; this study, potential volunteers
– does cultivating citizen scientists work
– 74 freshman from scholar program; collect tree and bird data; paper form and digital photos
– independent variables: working as a pair or alone; positive (nice job!) vs positive directive feedback (including instructions for improvement)
– dependent: volunteer motivation; data quantity
– motivation: intrinsic, identified regulation, external regulation, amotivation
– pairs higher in intrinsic and identified regulation
– directive positive feedback -> lower amotivation
– same quantity alone or in pairs
– pairs collect more photos; individuals are more sensitive to positive directive feedback as to quaniity in one type, but lower in other; in pairs, having more fun; alone, more serious; directive destroys fun but increases seriousness
– performance varies by ways of working, different difficulty of tasks, and different types of feedback
– Q&A: what’s the next step? study a real citizen science project to see if non-class context has different effects
Of Natures and Gamers: Lessons from Designing a Mobile App for Citizen Science
Ann Bowser, Derek Hansen,
– floracaching – based on geocaching using citizens to photograph specimen plants
– challenge: limited budgets lead to poor technical solutions which is a barrier to citizen participation
– natures: people who do citizen science no matter what, motivated intrisicly; gamers: motivated by the game; geocachers, a blend
– which activities to natures and gamers find engaging; discover something engaging for both
– co-design sessions at 2 universities with 58 participants
– four different tasks with different difficulty and motivations
– surveys and focus groups for quantitative and qualitative data
– similarities: like game for citizen science, social interactions
– differences: perceptions of gamification (gamers like the competition; saw it as a game; want more; subset of natures didn’t like the competition; see the difference between the gaming vs science aspects), guidance vs autonomy (gamers like contributing to a bigger picture, but game must give clear assignments; natures liked everything given the time, like open endedness), integration and enrichment (gamers would do it if it fit with their lives, but not otherwis; background running with notification – passive until notified; natures like to be outside and is one of their hobbies; help bring friends into it)
– implications: gamification is effective with both groups; guidance and integration for gamers; autonomy and enrichment for nature group
– explore the use of quests and guilds for gamers; allow natures to opt out of game aspects and allow deeper engagement in the domain by follow on requests from intrinsic activities
– Q.&A: would you add prizes for going to more sparser populated places, harder to get to? depends on the scientific campaign, maybe give more points or badges for obscure tasks; current app encourages visiting same plant over and over; how will you engage the scientists? biotracker app is for a specific scientific campaign; do the citizens get to see the results and purpose of their participation? currently provide access to all the data, but an active area of research to figure out how to engage participants in the results and purpose; is this a separate community from geocaching, on purpose or accidental? would love to have that community involved; current split is 50/50 citizen scientists/gamers+geocachers
Enhancing Cyberinfrastructure with Cognitive Principles
Tim Clausner
– dealing with genomics data at scale is hard
– iPlant – 11,000 users, 380TB of data, 103 million files; people (integrators) contribute apps to the environment to deal with different aspects of the data and work
– discovery window: apps window, data window, analyses data
– goal: enhance user performance via cognitive science
– first phase: observational study – 4 expert users and 6 nominal users; video recorded using system and heuristic analysis of videos
– users don’t realize there is drag-and-drop in system to tie data file to app
– 3 windows present usability problems; purpose of each window not clear; design problems, not user problems; more problems from scale of cyberinfrastructure; no clear relationship between inputs, apps, and outputs especially when a chain of apps is used
– meaningful elements of the workflow are left implicit; where are things stored; flow; succession of files and apps; ad hoc naming is insufficient; retesting of apps is required for successful results; user must rely on memory to reconstruct prior workflows
– are there cognitive principles which offer guidance toward enhancing the interface? key: attention
– visualize the flow with input, analysis, and output and the order the apps are run, associate the location of where each role is stored
– create a workflow that consists of data->app->analysis, with recursive encapsulation
– shift metaphor from container to object-path
– future: study performance improvement by applying cognitive priciples
– Q&A:
Techniques on Temporal Event Sequence Simplification
Rongjian Lan
– temporal event: event that happens at a specific time or over a period of time
– Â intra-record understanding vs. inter-record understanding
– EventFlow – timeline view with triangle for point event and bar for interval
– large complicated data sets are too complicated with EventFlow visualization; need simplification techniques
– intra-record simplification – remove a subset of events in a record
– inter-record simplification – remove a subset of records
– three levels: 1) filter-based 2) transformation-based 3) temporal search and replace
– filter-based: intra-record – removed subset of event categories or remove events outside temporal window; inter- remove a subset of records or a subset based on attributes
– transformation: [missed first type]; interval merging – remove small gaps and overlaps; marker event – ignore all events but marker event
– temporal search and replace: based on graphical query system (search by drawing on timeline and define constraints – wildcard, repetition, permutation); search for target sequence, replace with different representation
– Q&A: can I save the manipulations and reuse it on another set? yes, the simplifications are stored in an XML format; query by example? this builds on that, a graphical representation with constraint definitions
Basketball Play-by-Play Analysis Using EventFlow
Megan Monroe
– always wanted to apply EventFlow to other domains, to make it easier to explain and less depressing
– basketball games have similar event sequences
– ESPN provides real-time play by play stats of all events in all NCAA and NBA games
– basketball game is an alternating series of possessions; events happen within a possession
– dual datasets: offense to defense data set; defense to offense set; align the sets at the possession change
– seems like both teams were similar in this game; but by filtering down to the outcome of the possession after a made shot; shows that UNC had a cleaner, more efficient offense than Maryland
– EventFlow is up and running; talk to Megan
– Q&A: are you letting other schools have access to this? not yet; Maryland coaches ignored it; you realize you can pay for the next 30 years of this lab with this? yep, and other schools don’t have Megan to present it; how long does this take? only the manual data cleaning for mis-sequenced event
Expanding and Evaluating Network Motif Simplification
Cody Dunne, Ben Shneiderman
– network visualizations get very complex; need simplification
– better layouts, alternate visualizations, graph summarization
– motifs are repeating patterns in the network; patterns are functionally equivalent and can be replace via glyphs
– glyphs: fan, 2-connector, D-connector (more than 2 nodes), cliques
– [using orange to purple instead of red to green now]
– crescent glyph replaced diamond, users were confused as an occluded fan,now a tapered diamond
– interactivity: mouse over + drill down to all data
– implemented in NodeXL
– developed motif detection algorithms
– how effective are these representations? simplified vs. node-link representation; 31 tasks; users were faster in all cases using the simplifications; more accurate in many cases; often better at estimating the sizes; when asked to count, took longer, but less error; finding the cut point, no time difference but more accurate; find an individual node, faster and more accurately in the larger sets; even with hidden labels, no slower slower but just as accurate; comparisons were slower and less accurate
– valuable in reducing complexity and understanding larger or hidden relationships; however, frequent motifs may not be covered, glyph design has tradeoffs, requires some training
– Q&A: have applied this to healthcare delivery networks, and see how much money is flowing between providers, and what the outcomes there are? not specifically, but some medical cases, care about the attributes, and this problem space has some interesting stuff; dartmouth health atlas has found interesting patterns in outcomes and costs across counties, this would really help because it takes years to get the results out (4 years), data is available? may be other visual analytics that are better for this; any application of temporal motifs? some people studying it now, not a lot of solid network temporal models yet, have a paper on dynamic changes on networks in revision for journal; what about users that found the labels despite being hidden? found them by luck, hidden took a little longer, but no difference in accuracy
User Interface Techniques to Reduce Wrong Patient Selection Error
Awalin Sopan
– errors where wrong patient is selected for a treatment is a big problem
– computers don’t necessarily help
– multitasking, interruptions, urgency, fatigue, etc
– classified errors, analyzed tasks, recommended 27 specific UI techniques, prototype techniques
– errors: mistakes, slips, failure to recognize
– task analysis: recall patient -> select patient-> verify selection -> place order-> confirm order; wrong patient at three points (verify, place, and confirm)
– reduce mistakes in recall stage: facilitate recall (with more metadata and photo), avoid confusion (more attributes and sorting, filtering, and grouping; highlighting possible confusions)
– reduce slips in select stage: improve selection, text readability, highlight target under cursor, 2D grid instead of list
– increase recognition in verification and confirmation stages: photo made recognition go from 7% to 43%; decision support systems; animation between stages
– small changes in UI can make big difference in patient safety; include clinicians and HCI researchers in the design process; to err is human, the systems should make up for it
– Q&A: in our work, showing the complete order and history on the screen all the time as well as the patient information? good; responsiveness can increase or decrease errors, did you find that? the prototype didn’t have this problem, lit review suggests usability testing to see actual performance, shorter response times increase errors but slowing them down causes frustration; do you see differences between systems within a single hospital vs across multiple hospitals? across hospitals, have the same problems, but the same solutions apply
Visualizing Changes Over Time in Datasets Using Dynamic Hierarchies
John Alexis Guerra Gomez
– trees represent hiearchies; we know how to visualize, navigate, store
– how do we compare hierarchies
– tree: set of nodes and links that represent parent-to-child, each node is labeled, one or more numeric and categorical attributes
– types of trees: fixed (containment), dynamic (change order of tree is okay), mixed (some parts changeable, others static)
– comparisons: type 0 – topology change but no value change; types 1 and 2 – no topology changes but value changes; type 3 and 4 – topology and value changes; type 1 and 2 extensively worked on, 3 a bit, treeverse does all 5
– showing change in visualizations: color-coded lists, treemaps, bullet in treeverse 1, stem view treeverse 2
– StemView – icicle view with actual change (color), relative change (height), ending value (width), direction of change (above or below line), created/removed nodes (white vs black border)
[didn’t see the end because I had to leave]
Video Game Controllers: Negative and positive transfer across games and genres
Ken Norman
– current game controllers are pretty standardized; the equivalent of the querty keyboard; likely to be around for the next 100 years without much change
– controlled dimensionality – complexity of the game controller to the game; pong=1, half-life=7
– controllers are more complex, up to 9 CD; fun and frustrating; modes and combos can increase options; CD of up to 30+ in Tomb Raider Underworld
– students coded 5 games of their choice; 323 games of over 4,500 mappings; 37 games were coded by more than one coder
– categorized games by genre (using wikipedia as reference); music had lowest dimensionality, adventure games had highest
– frequently used functions: start button=menus and pause; x button=action and jumping; left-stick=move; right-stick=camera move; D-pad used for a lot of different things
– what happens when going to a new game with different mapping? worst case, everything swapped; negative and positive transfer (proactive interference, retroactive interference)
– develop a measure of commonality; sequels have high positive transfer but add new modes; mixed transfer within a genre; negative transfer between genres
– gamers abilities vary: highly proficient at one mapping; highly adaptable; know their preferences and customize the controller
– if designing a game, stick to the common mappings; database is valuable
– if getting ready to play a new game, think about how its mapping compares to what you know
– Q&A: what about differences in manual dexterity? yes, vast differences, especially between genres, other research maps skills required; what about things like kinect? not mature enough yet, but will definitely develop common gestures; does database have games based on CD? yes, out DB has all that
When Face-to-Face Fails: Opportunities for social media to foster collaborative learning
Tamara Clegg
– goal: find science relevant and think scientifically
– need social interactions around science
– observing, critiquing, share insights and ideas
– challenges: stigmatize low achievers, dysfunctional interactions
– Computer Supported Collaborative Learning has focused on the cognitive, not the social
– how can technology support face-to-face interaction
– SINQ = social media for supporting collaborative learning; micro-contributions, vote on quality, aggregate into projects
– how can SINQ help? used in kitchen chemistry with other tools on iPads
– compared with and without SINQ: unproductive interactions without SINQ (competing foci, easily derailed, arguments, talking out of turn; more successful with SINQ (more focused, shifting social landscape, idea seeds, facilitate conversations with facilitators, authorship and recognition, physical separation)
– key: factor in the context, other: facilitate communication, community repository of contributions, multiple entry points
– Q&A: anonymous or by name in SINQ? they were aware, but SINQ made it less charged and could deal with it asynchronously
Clinky the Robot: Preliminary programming for preschoolers
Mona Leigh Guha
– motivation: harness potential of computers, debugging skills are benefition, CS not very diverse
– can children (3-5 years old) program? comfortable with computers, independent in exploration, developmentally appropriate, want to program
– Clinky the Robot – iPad app to teach young children programming skills
– cooperative inquiry techniques; participatory designs, wireframes, implementation, formative evaluation with both teachers and children
– kidsteams developed idea, then did layered elaboration, then facilitator prototyped it, formative evaluation (interview teachers then watched kids interact)
– Clinky teaches programming concepts in increasingly complex exercises
– kids liked programming, but need to iterate, evaluate learning process, and expand levels
– consider scaffolding support for technologies, kids can learn programming
– Q&A: is programming the best word for what you’re teaching? originally that was the intent, even though it is generally problem solving; did kids view this differently than playing a game? they did have fun, and did feel it was a game, still have to test their learning
Using Mobile Touchscreen Devices as Peepholes to Support Children’s Second Language Learning
Brenna McNally
– goal: envision a future in which all students are proficient in English and at least one other language
– struggle: attitudinal factors, learning styles; opportunity: mobile, movement in learning
– how can we incorporate movement into mobile language learning
– peepholes: window into a virtual world; static – panning, dynamic – move device
– Scenic Worlds: virtual world, categorize vocabulary; chose German because it is largely unfamiliar at these ages (7-11)
– formative design session (how do kids envision language learning, how would they use mobile; combination of realistic and imaginary elements, small bits of language, variety of movements, gamification); evaluation session (sit, touch, tilt, move – needed direction to stand up and move arms; simplification); iterate (dynamic and static conditions); pilot test (how do they react? in pairs, much more conversation and interactions, more response to surveys; application interactions like zoom; some got frustration; hold the world still with one finger, and move things with other)
– ongoing work: learn more about how they interact with the peepholes, are they learning – word recall?
– what interactions can stimulate learning? this is a boring task (categorizing words), but the kids were excited and engaged; told parents and grandparents what words they learned
– Q&A: how many words? each condition had a different set of words, 9 clouds and 3 jars (12 words in each condition); this doesn’t just have to be for kids? 8 and up
Designing Alternate Reality Games for Learning
Beth Bonsignore
– world without oil: alternate reality game (ARGs); 32 days simulated 32 week escalating oil crisis; 1700 participants, 1500 story fragments
– cathy’s book: best sellers for children and young adult books
– arcane gallery of gadgetry: detective through history
– not video games; pervasive, transmedia interface
– not games at all: participatory narrative, storytelling archeology, real world imagination boundaries
– pervasive, authentic, story; collaborative sense-making; personal agency; critical thinking (counterfactual reasoning)
– studies find many ARGs promote this kind of learning naturally; most don’t look at design attributes, promoting learning, or look at younger learners
– AGOG: scavenger hunts, cryptography, electronics; “play the past”; 60 students 13-14 years old, 50% “FARM” students (Free and Reduced Meal); missions, roles, interact with in-game characters (protagonist by proxy); final mission with participation of all the players; narrative content embedded in multiple media, scoped the amount and number of sites for student audience; used real historical artifacts and data, but used actual data gaps for parts of game
– results: engagement, critical thinking, interactive/interlocking tasks made kids feel part of the story and that their participation mattered
– Q&A: what’s the next step? put a proposal together to take this further to more schools and more kids, how to include players in the design of the game, doesn’t scale well, lots of investment, capture participants energy to scale it
The Roles and Challenges of Technology in Supporting Learners’ Ownership in Science Learning
Jason Yip
– traditional science learning is alien, boring, and disconnected
– ownership of science learning: control, personalization, investment, territory
– science learning is hard
– case study of kitchen chemistry – 4 focal learners
– storykit (collaborative storytelling), zydeco (collaborative museum exploration), SINQ (social network for learning)
– theme 1: personal documentation – personal elements, audio messages, ownership and control but can be distracting; sciencekit – personal story elements but structured for science learning
– theme 2: role switching – physical activity and recording in parallel – switch back and forth, physical activity alone – have to put the iPad down or ask facilitator; ipads don’t mix with water; children are gross and sticky (covers don’t work); can wearables and sensors support storytelling in messy environments
– theme 3: collaborations – frustrating input for collaboration; typing vs. audio – paper and pen may have worked better; future – separate personal and group data, but use group data in personal narrative
– Q&A: is storykit part of sciencekit? more the inspiration
FACIT PD: A Framework for Analysis and Creation of Intergenerational Techniques for Participatory Design
Beth Foss, Jason Yip
– participatory design: overarching method philosophy; people should be included in the co-design of the things that surround them; cooperative inquiry; techniques for collaboration between adults and children
– interactions with technology are very different than 30 years ago; design for gestures, mobile; interactions will be different in the future
– does an existing technique work for this design situation
– FACIT PD: 8 dimensions in 3 categories, design goals, techniques
– dimensions: partner experience (“big paper” technique), need for accomodation (“mixing ideas” technique), design space (“bags of stuff” technique), maturity of design (“stickies” technique), cost (“paper prototyping” technique), technology level (need for digital tools and recording devices), portability (“pig paper” technique), physical interaction (“layered elaboration” technique)
– does an existing technique work for this design? look at the dimensions of the design problem vs the techniques
– how can I develop new techniques? look at dimensions the design problem needs and generate techniques that fit those dimensions (e.g. “Team Construction” technique)
– Q&A: do you have a library of techniques coded, and when will you have a card set? read the paper, lots of ours and others, and the card set idea is awesome; how does this system help deal with technology changes? prioritize the dimensions and build your technique for the most important
Domain Expertise in Cooperative Inquiry with Designers and Learners
Tamara Clegg
– challenge: growing into design domain experts because do it many time; working only with design expert children has limitations
– challenge: growing into subject domain experts; may not be familiar with the  design technique
– use both kinds of experts
– what are the differences in expertise among children? that’s the study
– 3 techniques; kidsteam are the design experts; subject experts were from the kitchen chemistry program
– similar design themes with specific differences: mobility, personaliztion, games, tagging, social, narrative
[results went by too fast to document]
– KC experts wanted contextual details, references, unobtrusive devices; design experts had wild ideas, specific technology features, aesthetics, opinionated
– ideally work with both types of experts; what if you don’t have access to both? choose appropriate techniques, ask co-designers to criticize freely, focus designers on observation, build relationships
– Q&A: if you can have both, should you put them together or separate sessions? typically, separate sessions, need time to ramp up in their expertise, but speculate that bringing them together at times
Cooperative Inquiry Extended: Creating technology with students with learning differences
Mona Leigh Guha
– cooperative inquiry: adult: “There’s no wrong answers.” child: “But is there a right answer?”
– CI developed with children with typical development
– over two semesters, at a school for children with learning disabilities: first semester, normal CI; second semester, modified CI
– kids with learning disabilityes, anxiety, ADHD, speech or language disabilities, mild to moderate spectrum; many with more than one condition
– first semester, more older kids; designed a game with no CI modifications
– changes: informal time (like snack time at lab), higher adult to child ratio (1 to 1 in lab), communicate in multiple modalities, plan for high levels of engagement
– why informal time? kids curious about facilitators and what they did; facilitators curious about the kids
– why more adults? adults are more like floaters than design partners with fewer adults
– multiple modalities? always right things down and have students repeat back
– why high engagement? kids really engaged the process and wanted to influence outcomes
– second semester: implement changes and see how it worked
– informal time: social time at beginning of each session, personally focused question of the day (instead of task focused); informal conversations prompted design ideas, adult design partner more defined, adults better understand children’s needs
– more adults: extra adult per small group; improved support for individual child partners; adult free to provide leadership
– multiple modalities: verbal and written activities; one child repeats; less hesitation with unfamiliar activities, less deviation from activities (positive and negative)
– high engagement: all design decisions as a team; multiple activities per session; more prep before each session; children contribute ideas according to preferred method; leadership roles available to child partners; not as much of a problem in 2nd semester
– CI can lead to positive experiences; researchers can use CI in these populations
– Q&A: did you record any informal sessions? no, didn’t have the cameras out, would have been great data
Enhancing the Web with End-User Programming
Tak Yeon Lee
– goal: expressive and easy programming environment to allow non-programmers to create browser extensions
– programming environment loaded as a floating window at the bottom of any web page
– inefficiencies of the web: missing relevant information, removing unwanted information, repetitive operations, finding needle in a haystack
– building extensions is difficult; programming knowledge + time and efforts
– automate repetitive jobs, integrate data from multiple sources, custom filtering, automated summaries
– “pick” operations, “change” operations (arithmetic, string manipulation, filter, sort), “add” operations (new elements, change style, trigger events, hide, etc)
– Â programming by example: in traditional programming need to know syntax; by example just provide sample input and output; computer figures out the right operation through trying many; then apply the program to other data sets
– OurUMD site shows grade distribution in classes at UMD
– future: usability improvements, evaluation of system learnability
– Q&A: would you be willing to have a kidsteam session, what would happen? would be a bit challenging, but would work; would it make it better for adults or just children? should be similar, web site complexity is a big factor; what types of longer term research questions, specifically about learning within the system? what can people learn about abstract computation without programming, what is the minimum amount of information for programming must a user provide, some operations can not be represented as input and output, how to collect user’s intention; could this be used for grading other things besides programming exercises, would also like to use this system to extract data from google scholar for my own research? trying to generalize to other kinds of grading but string matching works best for code; when you do the “add” operation, how do you ensure that the entire page is messed up? no heuristics right now, user needs to keep things stable, security and robustness issues, extensions need to be reviewed by community and improved, make it social
Missed Sessions
– Designing Decision Aids to Help Patients Choose Treatment Options, Lyndsey Franklin
– Twinlist for Medication Reconciliation: Evaluation with clinicians, Catherine Plaisant
– Designing Tangible Computing for Creativity, Tim Clausner
– A Tangible Interactive Shirt for Teaching Anatomy and Physiology to Children, Leyla Norooz
– Sharing Intimacy through Huggable Bears, Zahra Ashktorab
– Recruiting and Retaining Young Participants: Strategies from five years of field research, Beth Foss
