1. VR can improve training and learning for novices through experiential learning, but may not benefit experts or factual learning as much. High presence VR engages users but can also distract from content if overstimulating.
2. Creating a sense of embodiment and co-presence through matching user movements to their virtual avatar and natural responses from others increases engagement. Signs of sentience like motion and voice are more important than graphics quality.
3. Exposure to virtual environments and avatars can influence users' behaviors through the proteus effect. To avoid uncanny valley effects, avatars must react and move naturally while maintaining good visual quality.
3. Background
• I’ve been researching a book for a few months
• Have ‘collected’ a bunch of findings on VR &
Immersive tech – we’ll review some of the
most interesting.
4. Working Theses
VR will dramatically improve training and learning
for some things, but not necessarily others.
• VR instruction will be better for novices than
experts
• VR collaboration will work for experts and might
for novices
• VR will be better for experiential learning than
factual learning
6. What & Why of Presence
• “Presence” is the sense of being there:
– How much is the illusion of VR really working?
• To the degree this lowers distraction from the
new medium, and increases engagement, it
will help learning
7. Examples
• The “Pit” – heart rates go up, GSR increases
when presented with a high elevation
• Subjects in an experience with high levels of
perceived presence more likely to be
persuaded by health messages
• Children reported more false memories
In other words, high presence can mean you
experience the VR simulation as if it were real.
8. Inducing Embodiment in Presence
• “Rubber Hand” experiment is relatively old
and established.
• If what we see reacts like our bodies, moves
when we want it to
• If the body we see is doing the things we’re
making our body do, we’ll assume its ours
9. Sense of Embodiment
• Three factors:
– the sense of self-location – ego centric view is the
baseline.
– the sense of agency – am I able to act?
– the sense of body ownership – do those hands
move where I tell them to?
10. Involvement as Moderator
• How much do you care if it’s really your body?
• This is important because we don’t think about
embodiment very hard, and we accept that the
virtual body is our avatar (vs. actually believing it
is our body)
• The less we care, the less the immersive
environment has to convince us for us to just
accept it.
11. HIGH PRESENCE VR IS THE MOST HIGH-
BANDWIDTH COMMUNICATION EVER
DEVISED.
When you’re in a VR rig, and fully engaged, all of
your visual processing system is focused on the
content presented.
Nothing else comes close.
12. High Presence, Low Recall?
• Some studies have used highly vivid
experiences, and found that recall was low
• Why?
• Just like overly stimulating advertising, the
experience can distract from the message.
13. Immersion is not the same as presence
• Presence is a psychological thing
– Am I committed to the experience?
• Immersion is a technological thing
– Is the outside world shut off, and the experience
compelling?
14. Avatars vs. Agents
• When thinking of instructional
tech – are we using a person or
a machine-generated
instructor?
• Avatar is representing a human
• Agent is representing a
machine
Blascovich theorizes that users
will respond as if the virtual
human is real no matter the
technical richness, as long as
they think there’s a real person
behind the avatar.
Studies show that:
1. Virtual Humans exert influence
if:
a. User believes the Virtual
Human is computer or
human controlled
b. Accuracy of behaviors
c. Realistic low-level
reactions
d. Situational relevance – do
they care?
15. Mo-Cap is More Important Than
Massive Polygons
We care more about signs of sentience than
perfect rendering.
This is because of what I’m calling the Twilight
Effect
16. Presence in “Classroom”
• WE know that students who have the sole
attention of a teacher learn more.
• In VR, studies have shown that this works as
well – with the effect that everyone in the
class can have the sole attention of the
teacher.
17. Co-Presence
• In addition to being there, you are there with others.
• What determines the sense that others are there?
• Not richness of tech, not really movement – mostly
what they say, how naturally they respond. Essentially
social cues.
– Several sessions have shown that users barely move their
avatars, preferring instead to focus on conversation.
• This ‘how they respond’ means the sense of being
perceived – that is a critical factor.
18. Designing Space
• Three kinds of space around you:
– Peripersonal space – as far as your arms will go
– Extrapersonal space – as far as you’re likely to
walk/navigate
– Vista space – background
• Judging distances is hard. Several cues, like
shadows and brightness, don’t really work.
Perspective, and ‘motion parallax’ are most
effective for users.
19. Use of Space
• Several researchers have tested co-location as
a means to foster creativity.
• Space in VR doesn’t have to be ‘stateless’ we
can make notes in thin air, make personal
audio recordings. The whole environment can
be a toolkit.
21. Humans are Very Tolerant of Poor
Visual Input
• Visual processing is at least as much top-down
as bottom up: we see largely what we expect
to see
• We are also evolved to deal with poor visual
conditions – dark, fog, water, etc.
• We look for cues that survive bad visual
conditions
22. Motion & Reactions as Key Cues
• We’re looking for signs
of sentience:
– Verbal response
– Natural-looking motion
– Emotional interaction
Think of Pixar:
Famous Voices lend
personality
Motion-capture
lends natural
reactions
23. Twilight Effect
• Signs of sentience will dominate poor
execution, poor VR tech
• These signs must be relatively constant, as the
evaluation of an ‘other’ as sentient may be
provisional
25. Embodying a Different Body
What happens when you become:
• Taller = more confident
• Beautiful = more social
• Minority = more understanding, less biased
Proteus effect is robust – we take on, sometimes
for weeks, elements of the identities we inhabit
in VR
27. Hypothesis
As one makes artificial humans,
especially faces, more and more
real, there is a degree of ‘realness’
between completely real and
some midpoint, that users find
aversive.
Some research on the subject – VR
research I’ve found said that the
only uncanny valley effects found
were when features were
exaggerated or otherwise
disturbing on their own.
28. Avatars & Uncanny Valley
• Research on this will be
fraught, as a badly drawn
execution is just as aversive
as the ‘uncanniness’ of an
almost-human.
• Make avatars look as good
as we can, and make 100%
sure their reactions, and
movements, and voices all
sound human – as these
are not affected by the
‘twilight effect’.
29. Some Key Takeaways
• Presence is key to holding attention
• Creating presence is about embodiment – feeling you
“own” the thing in the virtual environment (VE)
• That comes from matching what your real body does
to what it sees in the virtual environment
• Richness of VE isn’t the key, richness of human
communication is
• Users will forgive all sorts of technical failings if the
voice at the other end sounds human, reacts
sentiently.
31. Cognitive Load
Intrinsic: Keep the tasks just beyond the
learner’s ability
Extraneous: Keep distraction to a minimum
Germane: Enable reflection on the material to
be learned
32. Extraneous Load
• Immersion, as described above, should enable
minimization of extraneous load
• Presence, as a psychological construct, should
further enable minimization of distractions &
extraneous load
33. Intrinsic Load
• More complex problems can be handled if
working memory tasks can be offloaded – by
storing thoughts in the environment.
• Notes, but multi-modal, sitting in thin air,
automatically stored or deleted.
• Relatedly, worked examples can be experienced.
• Faded examples can also be experienced vs. just
imagined.
34. Germane Load
• 100% environmental control allows for
pointers to appear, reminders to think about
the lesson to be learned vs. the task to be
completed.
• Exact duplicate of an experience can be
replicated, allowing learners to re-experience
a half-learned lesson.
• An immersive experience can be rewinded.
35. Expertise
• Multi-media learning studies have shown that
novices benefit from video & graphics,
whereas experts often find them distracting.
• VR, with it’s overwhelming experience of a
virtual world, can be expected to have an
enhanced version of this:
– Novices will learn faster when conceptual support
is helpful
– Experts will find it distracting
36. Multi-Media Learning Lessons Apply
• Don’t put text over graphics
• Use VO to add to graphic
• Use graphic instead of text whenever possible
• Avoid music
• Avoid accents
• Keep lessons short
37. Summary
• We know a fair bit about how to create good
VR Education.
• By this Sunday, I’ll have a 10-15 page
summary. Would love your input!
– I’ll post it on the meetup
• About 100+ papers went into this – they’ll all
be on a google drive I’ll share via meetup.