Slide 1
Slide 1 text
Instilling Behavior Change during Web Surfing through
Subliminal Priming
On the Design of Subly:
Ana Caraban, Vítor Teixeira, Pedro Campos, Madeira Interactive Technologies Institute, PT
Evangelos Karapanos, Cyprus University of Technology, CY
Sean A. Munson, University of Washington, USA
Slide 2
Slide 2 text
Modes of thinking
AUTOMATIC
REFLECTIVE
2
Dual Process Theory
Slide 3
Slide 3 text
AUTOMATIC REFLECTIVE
Adams A, et al. (2015). Mindless Computing.
6%
94%
Systems designed for...
Slide 4
Slide 4 text
Behavior change through knowledge
Slide 5
Slide 5 text
One third of owners of activity trackers discard the
tracker within the first
6 months
Ledger, D. (2014)
Slide 6
Slide 6 text
How can we design technologies that
trigger the Automatic Mind?
Slide 7
Slide 7 text
Subliminal Priming of behavior
concepts (e.g., to walk)
Slide 8
Slide 8 text
50% people spend over 6 hours
per day surfing on the web
Slide 9
Slide 9 text
Subly: Priming through changing the opacity of words
Slide 10
Slide 10 text
1. Can we determine the threshold of subliminal perception?
STUDIES
Slide 11
Slide 11 text
FirstWordSecondWord
Opacity was decreased at rate of 1% every 0.5 sec
STUDY 1 Thresholds of Subliminal Perception
Slide 12
Slide 12 text
FirstWordSecondWord
Participants were asked to look at the screen
and press the spacebar when noticing a change
STUDY 1 Thresholds of Subliminal Perception
Slide 13
Slide 13 text
FirstWordSecondWord
STUDY 1
Foveal vision
FirstWord SecondWord
Peripheral vision
Thresholds of Subliminal Perception
Slide 14
Slide 14 text
FirstWordSecondWord
STUDY 1
Black
FirstWordSecondWord
Blue
Thresholds of Subliminal Perception
Slide 15
Slide 15 text
FirstWordSecondWord
STUDY 1
Black
FirstWordSecondWord
Blue
5 repetitions
25 participants
Thresholds of Subliminal Perception
Slide 16
Slide 16 text
Median Opacity threshold (black font)
= 0.76
Median Opacity threshold (blue font)
= 0.73
This difference between blue and black text (3%) was not statistically significant
Findings Thresholds of Subliminal Perception
Slide 17
Slide 17 text
1. Can we determine the threshold of subliminal perception?
2. Can subliminal priming influence users’ decisions?
2 STUDIES
Slide 18
Slide 18 text
e.g “A bird, a cat and a bear were standing on a tree”
STUDY 2 Efficacy of subliminal priming
Slide 19
Slide 19 text
STUDY 2
e.g “A bird, a cat and a bear were standing on a tree”
Efficacy of subliminal priming
Slide 20
Slide 20 text
STUDY 2
e.g “A bird, a cat and a bear were standing on a tree”
Does priming a concept lead to higher likelihood of selecting the
corresponding image?
Efficacy of subliminal priming
Slide 21
Slide 21 text
Does priming a concept lead to higher likelihood of selecting the
corresponding image?
STUDY 2
e.g “A bird, a cat and a bear were standing on a tree”
30 participants 9 sentences
3 condition
Efficacy of subliminal priming
Slide 22
Slide 22 text
Supraliminal
STUDY 2
66 % Control
Subliminal
Are participants more likely to select a relevant image when a hint is primed
in the sentence?
Sentences, images and primes were randomized
76 %
e.g “A bird, a cat and a bear were standing on a tree”
Efficacy of subliminal priming
Slide 23
Slide 23 text
Findings
Slide 24
Slide 24 text
Were individuals more likely to select a relevant image when
we primed a hint?
0%
25%
50%
subliminal control supraliminal
Chance of selection = 33%
STUDY 2
33%
33%
33%
Slide 25
Slide 25 text
Were individuals more likely to select a relevant image when
we primed a hint?
0%
25%
50%
28%
STUDY 2
control
Slide 26
Slide 26 text
Were individuals more likely to select a relevant image when
we primed a hint?
0%
25%
50%
42%
28%
Subliminal primes leaded to
a higher chance of selection
p < 0.05
STUDY 2
subliminal control
Slide 27
Slide 27 text
0%
25%
50%
28%
24%
Aversion effect
p = 0.61
STUDY 2
Were individuals more likely to select a relevant image when
we primed a hint?
subliminal control supraliminal
Slide 28
Slide 28 text
STUDY 2 Did priming disrupt the reading process?
F(2)=0.04, p=0.96
F(16)=1.43, p=0.13
F(8)=9.14, p<0.001
Effect of condition on reading time:
Effect of the sentence on reading time:
Interaction effect between condition and the sentence:
Slide 29
Slide 29 text
STUDY 2 Did priming disrupt the reading process?
Pauses in reading
High interrater agreement (Cohen’s Kappa=0.85)
No significant difference among the three
conditions (Fisher’s exact test, p=0.99)
a pause in reading leading to silence, a hesitation or
a repetition in reading
Slide 30
Slide 30 text
Subliminal priming of behavioral
concepts can be a promising
Behavior change intervention
Slide 31
Slide 31 text
No content
Slide 32
Slide 32 text
Yet, bear in mind the
task..
Slide 33
Slide 33 text
No content
Slide 34
Slide 34 text
https://github.com/SublyM-ITI/Subly
subly.m-iti.org
www
Database connection
Events definition
Interaction Logs
http request
External applications or web
browsing behaviors
Interactions with primes and
webpages
Slide 35
Slide 35 text
Thank you
Ana Caraban
([email protected])
https://github.com/SublyM-ITI/Subly
subly.m-iti.org
www