Blending Digital and Face-to-face
Interaction using Co-located
Social Media in the Classroom
Adrian Holzer
Shared
Messages
Anonymity
Messages
Know
others
opinions
Shyness
Participate
Get
feedback
Issues are
invisible
Ratings
Address
students’
issues
SpeakUp
First mockup
First version
Second version
SUS
Good-Excellent usability
83
5’700
39K
Since November 2013
27K
66’000
10’000
RQ1 How can co-located social media
affect interaction in classroom audiences?


RQ2 How to design for a synergy between
digital and face-to-face interaction?
RQ1 How can co-located social media
affect interaction in classroom audiences?
Evaluations
2025
11Courses
Students
61 Lectures
Responses412
0
150
300
450
600
Number of Students
M=100
0
300
600
900
1200
Number of Messages
M=119
0
2500
5000
7500
10000
Number of Votes
M=1206
1 Prof
100 Students 200 Students
Monday Wednesday
223 responses
4 hours every week for 6 weeks
More interaction?
Week1
60
30
0
Week2 Week3 Week4 Week5 Week6
Numberofinteractions
More interaction with SpeakUp
No SpeakUp
SpeakUp
What kind of
interaction?
Generally good! Positive comments in course evaluations (N=141)
All comments on the ComA course evaluation related to SpeakUp verbatim
U"lisa"on	
  à	
  l'app	
  SpeakUp	
  
à	
  généraliser.	
  
Réguler	
  speak	
  up.
Les	
  professeurs	
  sont	
  vraiment	
  
impliqués	
  !	
  Le	
  cours	
  est	
  beau,	
  
bien	
  préparé	
  !	
  +	
  SpeakUp	
  
Les	
  profs	
  ont	
  été	
  toujours	
  
clairs.	
  L'u6lisa6on	
  de	
  SpeakUp	
  
Ajouter	
  des	
  clickers.	
  SpeakUp	
  
intéressant	
  mais	
  pas	
  performant.	
  
😀
😀
😀
😐
#
Content
I don’t understand differentiation by
product compared to conformity
+35
Organisation
Be careful with the background color
of your slides, the text is sometimes
hard to read
+5
Not relevant
We’re qualified-46
Interactions were coded in three categories
Week1
60
30
0
Week2 Week3 Week4 Week5 Week6
Numberofinteractions
No SpeakUp
SpeakUp
60
30
0
Numberofinteractions
Not relevant
F2F Content
Content
Organisation
Week1 Week2 Week3 Week4 Week5 Week6
Many of the interactions are not relevant to to teacher
What kind of non
relevant messages?
Lecture related
Look for the sickest marlboro ad ever+1
Fun
Chuck Norris is the biggest blood
donor, but it’s never his own
+3
Social
Pez, do you wanna marry me?+2
News
We’re qualified-43
We’ve been qualified for a month+29
We further coded the interactions not relevant to the teacher in four categories
60
30
0
Numberofinteractions
Not relevant
F2F Content
Content
Organisation
Week1 Week2 Week3 Week4 Week5 Week6
60
30
0
Numberofinteractions
Week1 Week2 Week3 Week4 Week5 Week6
Not relevant
60
30
0
Numberofinteractions
Week1 Week2 Week3 Week4 Week5 Week6
Misc
Fun
Social
News
Lecture
30
0
Numberofinteractions
60
Week1 Week2 Week3 Week4 Week5 Week6
Unrelated interactions are of many types
Social
News
Lecture
Misc
Fun
How can we filter
these
messages?
Lecture related
Fun
Social
News
Misc
Social voting
Mixed results, doesn’t always work, there are many instances of false positive
Theoretically we would like to easily identify the bad apples
Numberofgoodmessages
Number of bad messages
10
5
0
5 10
Ban bad apples
Ban bad apples
Theoretically we would like to easily identify the bad apples
Numberofgoodmessages
Number of bad messages
10
5
0
5 10
😇
😈 😈
😇
😇
😇
😇
😇
😇
😈
😈😈😇
Practically It looks a bit more messy
Numberofgoodmessages
Number of bad messages
10
5
0
5 10
😇
😇
😇
😈😈
😋
😋
😋
😋
😋
😋
😋😋 😋
😋
😋
Ban bad apples
Require identity
Might work, but can backfire and reduce interaction
Censorship
Our experience has shown that it can backfire, peer tagging is to be tested
IS12
F2F etiquette
Our experience has shown that it can work well
Com14PM13IS12
Backchannel
Front channel
Are interactions directed to lecturers or peers?
60
30
0
Numberofinteractions
Not relevant
F2F Content
Content
Organisation
Week1 Week2 Week3 Week4 Week5 Week6
Week6
60
30
0
Numberofinteractions
Not relevant
Relevant
Week1 Week2 Week3 Week4 Week5
30
0
Numberofinteractions
Not relevant
Relevant
15
Week1 Week2 Week3 Week4 Week5 Week6
Not relevant
Relevant
Front channel Back channel
Front channel interactions are almost all relevant, most backchannel is noise
Students want to keep a backchannel open
Tag relevant messages
Can potentially hide the messages from the lecturer
@prof
RQ1 How can co-located social media
affect interaction in classroom audiences?
co-located Social Media can improve
interactions in the classroom
Interactions are of a rich mix of kinds
on both the front and back channels
anonymity increases interaction but can
also lead to non-relevant messages
RQ1 How can co-located social media
affect interaction in classroom audiences?


RQ2 How to design for a synergy between
digital and face-to-face interaction?
RQ2 How to design for a synergy between
digital and face-to-face interaction?
Guideline 1 – Facilitate
access with little or no
registration
Guideline 2 – Facilitate
usage with simple
features
Design for first use
Guideline 3 – Provide
concealed identity
Design for more digital interaction
Guideline 4 – Design
for privacy
Guideline 5 – Make the
interaction temporary
Guideline 6 – Make use
of f2f etiquette to guide
digital interaction
Design for richer digital interaction
Guideline 7 – Embrace
the backchannel and
try to separate it from
the front channel
Guideline 8 – Design
for awareness
Design for digital f2f transitions
Guideline 9 – Design
for reflection
RQ2 How to design for a synergy between
digital and face-to-face interaction?
Facilitate first use, get
users onboard
Improve content, more and better
interaction
Improve action, more awareness
during and after the interaction
Questions?

Blended interactions