Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Now We're Thinking with Portals

A presentation for the Teach with Portals team

Rob Lee

on 30 December 2012

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Now We're Thinking with Portals

Now We're Thinking
with Portals The effects of playing
the Portal 2 multiplayer on perspective-taking Questions asked
by the experiment Which parts of the Portal 2 multiplayer support cooperation or collaboration? (Cooperation = working on separate tasks towards the same goal) (Collaboration = interacting together on the same task) Does closer interaction in Portal 2 assist perspective-taking
(‘tuning in’ to each other’s way of thinking)
in later tasks? Which interactions best predict this tuning in?
Collaboration, cooperation, one person leading/teaching the other,
or equal contribution? Method 60 people were recruited (30 pairs)
none had previously played the Portal 2 multiplayer 30 Players played the calibration course for 15 minutes
(4 pairs completed the course early) 30 Players played a control level,
created with the Perpetual Testing Initiative level editor then converted to a multiplayer map using Hammer

(this was before the Perpetual Testing Initiative
allowed for the creation of multiplayer levels) The control level was near-identical to the calibration course,
except that elements of each puzzle requiring interaction between players were removed, and replaced with equivalent solo puzzles Players played on a battered PC laptop and a MacBook pro over a wireless network, but frame rate was acceptable on each. Other Parts of the Experiment Before playing Portal 2, each participant completed the ‘Mind in the Eyes’ test,
which assesses people's ability to read others faces and understand their emotions: http://glennrowe.net/baroncohen/faces/eyestest.aspx playful comforting irritated bored This has previously been found to predict success of perspective-taking in a Lego-building task. People sit either side of a table
with a screen places between them,
so they can see each other’s faces
but not their hands or workspace One person becomes the director, and is given a Duplo model made of 6 or 7 abstractly arranged bricks.
The other becomes the builder, and is given the precise bricks needed to build the Duplo model The director must then instruct the builder in how to build the model The partners take it in turns to be director There are six models in total Fewer errors in the completed model indicate more successful ‘tuning-in’ of each participant to the other’s mindset This involves tailoring your language so that it is clear to your partner,
plus being able to recognize when your partner misunderstands. What My study found Simply being in the group that played the
calibration course increased players' ability to tune in
to the other person’s perspective in the Lego task. Players talked to each other just as much in the control level as the calibration course. This is exciting because it suggests that the level design of the calibration course can help to guide conversations between players towards collaborative interaction However, conversations in the control level often consisted of the more able player explaining to the other how to solve each puzzle. It was only in the calibration course that more conversations led to more puzzles being solved with input from both players. More Findings In the calibration course, the proportion of players solving each puzzle with input from both players increased with each successive puzzle completed* *except for puzzle 5 This was not true for the control level This is exciting because it suggests that the level design
of the calibration course not only supports collaboration,
but that the more success a pair has in collaborating as equals,
the more likely they are to collaborate as equals in later puzzles More Findings - Lego Task Pairs made fewer errors in each successive Lego model that they built together as director and builder This suggests that they got better at tuning in
to each other the longer they worked together,
at least for the purposes of the Lego task When a pair managed to solve puzzles jointly in Portal 2, this gave them a head start in tuning in to each other’s perspective in the Lego task, and they made fewer errors on the first model built together This is particularly exciting, because there is very little research exploring whether
collaboration on one task can help people to interact more effectively
in future tasks that require different methods of working together.
These results suggest that it can. More Findings - Collaboration in Portal 2 Several different aspects of interactions between players in Portal 2 predicted high levels of success in the Lego task: 1.) If puzzles were solved with input from both players, they did better in the Lego task 2.) If players gave each other roughly equal numbers of instructions, they did better in the Lego task 3.) The more generous that players were in their assessment of how much their partner contributed to puzzle solving in Portal 2, the better the pair performed in the Lego task These results are exciting because they highlight aspects of interaction
that are not often recognized as key to collaboration: What further research could follow on from this? Smaller-scale research
modifying the current experiment Longer-term research looking at the impact
of collaborating in
Portal 2 over time Smaller-scale research modifying the current experiment In my experiment the use of the in-game ping tool and gestures
had no effects on performance in the Lego task.
This may be because players were new to using the ping tool,
and because most of the puzzles were simple enough to explain using
spoken word, so use of the in-game tools was quite limited. However, we still need to see whether
effective use of the in-game communicative tools
could have a greater impact on later interactions
over longer experimental periods
(for example, a few hours of play) It would also be interesting to explore the use of these tools
as a replacement to spoken communication, rather than a supplement, e.g. by turning off the ability to communicate through the microphone headsets. An in-game equivalent of the Lego task could be constructed
using the Portal 2 level editor, with each player in a separate room,
and one player being required to describe to the other
how to set up their room so that the two are identical (or mirrored) Once several such levels have been constructed, they could be used
as an intervention, to see whether this novel form of in-game interaction
has any particular effects upon actions in future tasks. Finally, whilst widely used, the Mind in the Eyes task
is arguably limited by its use of static photographs
of people’s eyes. In real life, people are not just pictures,
they move around as they speak. It would be useful
to develop an improved version of this test that utilizes short video clips of people’s eyes
as they show particular emotional expressions. An additional modification to the Mind in the Eyes test that relates to the Lego test would be to reverse its structure: In a counterbalanced study, such an ‘Eyes in the Mind’ test could be used to investigate how closely people agree that particular words refer to the same real-world occurrence. Longer-term research looking at
the impact of collaborating in Portal 2 over time Future research into the effects of collaborating in Portal 2 cannot be limited to exploring
the effects of playing it for only 15 minutes A longer-term intervention, with participants
interacting on a weekly basis, would allow for
the exploration of whether collaboration in Portal 2: Has a cumulative impact over time, Peaks after a certain amount of collaboration, Or only yields short-term benefits immediately after each playing session. The impact of such collaborative play could be investigated using multiple measures of perspective-taking,
multiple collaborative challenges, and other social measures, such as relationships between collaborators,
and between each collaborator and their other peers, for example in a primary school setting. The ChaT Lab has extensive experience of performing interventions in primary-school settings, and I have extensive teaching assistant experience at primary school level, in particular
I have supported kids with dyslexia, behavioral difficulties, cerebral palsy and autism. On January 19th I am flying to India to volunteer for 3 months as a team leader with Raleigh International, supporting a team of 6 English and 6 Indian volunteers aged 18-23.
I am currently planning preliminary research that will look into the impact of
collaborating over time on perspective-taking ability between the volunteers. Longer-term interventions also open up the possibility of investigating the educational processes involved in learning to use the
Portal 2 level editor. For example, the impact of playing the main multiplayer campaign could be compared and contrasted with the impact of designing levels for a peer to play, and playing the peer’s levels over time. The main multiplayer campaign can be thought of as collaborative. In contrast, designing levels
for another could be made competitive
in the same manner as chess:

both cases involve each player trying
to understand their opponent’s thinking in order to outwit them. This would allow us to explore, for example,
whether competitive perspective-taking yields the same benefits
as collaborative perspective-taking. Additionally, pairs of players could work together to design a level
for other pairs to complete, allowing
for exploration of the effect of the level editor when used collaboratively. research with People
with Autism Exciting Opportunities For People on the Autistic Spectrum instead of 1 picture and 4 words to choose from, participants have 1 word and 4 pictures from which to choose. Autism manifests in very different ways in different people, but one relatively
common component is that people with autism
can have trouble working out what other people are thinking or feeling Often this may not always be
as intuitive as it can be
for non-autistic people There is some evidence that people with autism find it harder to read other people’s faces
or interpret body language The Mind in the Eyes test was designed in-part to explore this, although it would have been be truer to life if videos rather than static images had been used There is also some evidence to suggest
that one aspect of face-to-face social conversations that some people with autism
find difficult is that there is no particular reason or purpose for the interaction The Portal 2 multiplayer helps with these issues by: Removing face-to-face contact Providing a limited and unambiguous
set of points/gestures Providing both a purpose for interaction
and a framework to support that interaction The initial experiment provided evidence that collaboration in Portal 2
can help with later perspective-taking between players If this effect can be replicated
when people with autism play it,
it could become a powerful tool
for supporting communication
and social interaction
within the classroom. The ChaT Lab has extensive experience of working with children with autism,
and I personally have experience of working with children both with autism and other physical and learning difficulties. We are therefore very well placed to set-up and run a long-term intervention bringing Portal 2 into schools,
working with autistic children,
and measuring the results. Now We're Talking Through Portals? It would be exciting to try out the Portal 2 co-op game
in a number of small-ish experimental settings,
in order to explore its educational value,
e.g. by picking apart further its precisely how it
promotes collaboration and perspective-taking. This section provides some examples of some fairly minor modifications of the current experimental set-up,
partly because the current set-up proved to be
even more effective than I had hoped. However, the possibilities for exploration are
as broad as we want them to be... They suggest that collaboration may have several important dimensions
that may influence its impact upon perspective-taking. Portal 2 may therefore be an ideal medium to help
people with autism to work/play with others Collaboration has now been recognized as one of the most effective ways
to support both social and academic learning in the classroom The ease with which levels can be created and modified by both players and researchers makes Portal 2 an ideal tool for exploring different aspects of collaboration Use of collaborative scenarios within Teach with Portals lesson plans
may help the game to become an even more effective
educational tool - this is an area that I
would like to explore further. The director's models look like this: Whilst the builder starts off with a tray of bricks like this: The experience of undertaking this research should help clarify
the direction that longer-term research should take. When I first contacted the Teach with Portals team, it looked as if I would have to generate my own sources of funding in order to be able to undertake a PhD in the ChaT lab at the University of Sussex. One potential route was to get Economic Social Research Council (ESRC) collaborative studentship funding, and I approached you to investigate whether you might be interested in half-funding a PhD.

I received no response, so I assume this is beyond the remit for what you would like to do with the Teach with Portals project. However, since then, new sources of funding have become available that require
no financial input from Valve: The ESRC is providing the University of Sussex with 11 funded PhD places next year, so I will be applying for one of these. The applications favour interdisciplinary research.

My undergraduate degree was in Philosophy, and my masters in Psychology,
so my primary supervisor will be Dr Nicola Yuill in the ChaT lab,
supporting 70% of my research, and my secondary supervisor will be
Dr Andrew Chitty from the Philosophy department, supporting 30% of my research. The ESRC has also announced that it is looking to fund multidisciplinary collaborations between different subject areas. Either source of funding will cover the complete financial requirements of my research, so I am not seeking any money from Valve However, it would be nice to be able to report in my application that I have shown the Teach with Portals team my original experimental research I would also like to know whether my proposed research
could be of assistance to the Teach with Portals project,
and whether future developments in your project could potentially guide the direction of my research. It would be great to hear what you think about my research, whether any similar research has been done with Portal 2,
and whether Teach with Portals intends to do further work with the Portal 2 multiplayer. The overlap between philosophy and psychology research into collaboration is called 'Collective Intentionality', which covers how people interact and the processes that occur when they decide to work together. I am currently supporting my supervisors in writing a proposal for this multidisciplinary funding.
Full transcript