Beyond One-Size-Fits-All: How Interactive Tabletops
Support Collaborative Learning
Dept. of Ed. Tech.
University of Warwick
Coventry, CV4 7AL, UK
School of Psychology
University of Sussex
Brighton, BN1 9QH, UK
Previous research has demonstrated the capacity of interactive table-
tops to support co-located collaborative learning; however, these
analyses have been at a coarse scale—focusing on general trends
across conditions. In this paper, we offer a complimentary per-
spective by focusing on speciﬁc group dynamics. We detail three
cases of dyads using the DigiTile application to work on fraction
challenges. While all pairs perform well, their group dynamics are
distinctive; as a consequence, the beneﬁts of working together and
the beneﬁts of using an interactive tabletop are different for each
pair. Thus, we demonstrate that one size does not ﬁtall when char-
acterizing how interactive tabletops support collaborative learning.
Categories and Subject Descriptors
H.5.2 [Information Interfaces and Presentation]: User
Design, Human Factors
Interactive tabletops, collaborative learning, case study
1. INTERACTIVE TABLETOPS
With the advent of commercially available systems such as the Mi-
crosoft Surface and the SMART Table, user studies of multi-touch
tabletop interaction are progressing apace. Previous research ﬁnds
that interactive tabletops are enjoyable to use , promote play-
fulness [19, 20, 26], support awareness , encourage equity of
participation , and can promote learning [10, 17, 25, 27, 30,
34]. These studies have tended to be controlled comparative stud-
ies. Tabletop interfaces are compared to other interaction mecha-
nisms, such as with desktop computers , vertical displays ,
non-digital materials , or a different tabletop interface [9, 28,
30]. The goal here is to hold other factors constant between con-
ditions to test potential statistical regularities in behaviour related
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for proﬁt or commercial advantage and that copies
bear this notice and the full citation on the ﬁrst page. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior speciﬁc
permission and/or a fee.
IDC 2011, June 20–23, 2011, Ann Arbor, USA.
Copyright 2011 ACM 978-1-4503-0751-2...$10.00.
to using the tabletop system. Thus, the factors to be compared are
decided upon a priori.
A second strategy has been to apply qualitative analytic approaches
such as video interaction analysis  and ethnography to study,
in detail, interactions between people and with the material envi-
ronment when using tabletop systems. The goal here has been to
draw regularities from a corpus of speciﬁc instances . Work
has focused on how people approach, share, and come to under-
stand how to use tabletop systems in real-world contexts and some
of the problems that can result [3, 11, 12, 23, 24], on the physical
mechanisms used to mediate conﬂicts at the table , and on the
general mechanisms for learning at the tabletop [8, 27].
Individual differences between groups working in the same situa-
tion are often underemphasized. For controlled experiments, this
is to highlight differences between conditions. In video-analytic
or ethnographic studies, speciﬁc episodes are selected from a large
sample and regularities are drawn out in how people use this tech-
nology. To summarize, previous research has been carried out us-
ing a largely nomothetic approach—constructing a generalized un-
derstanding based on a large sample set (although not necessar-
ily generalized beyond a speciﬁc context of use). In this paper,
while we rely on detailed video analysis of interactions between
users of an interactive tabletop, we take an idiographic approach—
constructing speciﬁc understanding of compelling cases. Our inter-
est is in children’s collaborative learning with tabletop interfaces.
We purposefully select three dissimilar cases to emphasize how dif-
ferent group dynamics inﬂuence how children collaborate using in-
teractive tabletops. Through this analysis we demonstrate diver-
sity in the ways that interactive tabletops can support collaborative
DigiTile is an adaptation of DigiQuilt  to the DiamondTouch
 interactive tabletop . Like DigiQuilt, it is a construction
kit for learning about math and art by designing colorful mosaic
tiles. In addition to being aesthetically pleasing, these tiles lend
themselves to mathematical analysis. The designs embody fraction
concepts and can be symmetrical. For instance, the design in Figure
1 is half red and half yellow; it also has horizontal and vertical
lines of symmetry. While DigiQuilt was designed for a single user,
DigiTile was designed for two concurrent users positioned side-by-
side in front of the interactive tabletop.
DigiTile’s interface is split into four main areas (Figure 1). The
central tile (a) is a square grid (2-by-2, 3-by-3, 4-by-4, or 5-by-5)
Figure 1: Screenshot of DigiTile, with main areas marked
of snaps. Pieces can be dragged into these snaps to create a col-
orful tile. Both users are provided with their own palette area (b)
on the left and right side to choose pieces of different colors and
shapes. Pressing on a color button changes the color of the pieces
on that palette. For this study, each palette contained three different
shapes: a snap-sized square, a half triangle, and a quarter triangle.
To provide feedback on fraction tasks, each color button displays
the portion of the central tile corresponding to that color in three
representations separated by equal signs: 1) as a fraction with a
lowest common denominator, 2) as a reduced fraction, and 3) as a
decimal. In the case where the ﬁrst two are identical, only one frac-
tion is shown (e.g., in Figure 1, the red and yellow buttons display
2= 0.5). Users can drag pieces to one of the ﬁve work snaps (c).
There, multiple pieces can be assembled and rotated. When pieces
are dragged out of a work snap, a copy is automatically created;
thus, the assembly and rotation work does not have to be repeated.
To clear a work snap, the user presses the eraser button below that
snap. A scrollable bar contains a graphical history (d) of how the
tile was created. By clicking on a thumbnail, users can revert to an
older version of the tile.
3. PARTICIPANTS AND PROCEDURE
Participants, age 8 to 9, were selected from the same classroom.
DigiTile sessions were conducted in the back of that classroom
during normal class time. The sessions occured during a “Maths
Week,” where more emphasis was given to Mathematics; as such,
the learning goals of fraction understanding ﬁt in well with the stan-
dard curriculum. All sessions were video taped from two angles:
one on the learners and one on the display. These were edited to-
gether for analysis. The three sessions analyzed in this paper were
purposefully chosen using the following criteria: good audio and
video quality, success on completing the tasks, and to showcase
various styles of working together. Two researchers iteratively an-
alyzed the video. The cases are a synthesis of their interpretations.
At the beginning of each session, the researcher demonstrated how
to move and rotate pieces, change their color, and use the graphical
history. Participants brieﬂy tried out these features. After this intro-
duction, each pair attempted the same three tasks (i.e., challenges
that they needed to complete by working together). For each task,
the children were given a printout stating the task; the researcher
also read the task description aloud. Once students completed a
task, the researcher set them to work on the next task. After 30
minutes, the researcher ended the session, even if it meant that the
third task went unﬁnished. While they observed sessions, both the
classroom teacher and researcher took a hands-off approach, only
intervening a few times.
4. THE THREE TASKS
The ﬁrst task was to replicate the pattern in Figure 1. The children
were provided with a printout of the pattern, including grid lines, to
guide their work. The purpose of the task was mainly to serve as an
introductory exercise to familiarize students with the interface and
working together on the DiamondTouch. While this task was com-
pleted by all, matching the pattern was not trivial; all pairs made
errors placing pieces. Most participants found it difﬁcult to orient
a triangle to ﬁll a gap. The interaction pattern of orienting a piece,
dragging it to the intended location, realizing that it did not ﬁt, and
dragging it back for further rotation was common.
At the end of the ﬁrst task, the researcher asked, “What fraction
of the square is red and what fraction is yellow?” Most pairs were
able to come up with the correct answer on the ﬁrst try; if not,
the researcher helped them to realize it was one half. As a follow
up, the researcher asked, “What percentage of the square is red
and what percentage is yellow?” These questions were asked to
draw the participants’ attention to the different representation of
the fractions on the color buttons. This was important as the next
two tasks focused on fraction understanding.
For the second task, the researcher asked the participants to “make
the square three eighths orange and three eighths brown. Some of
the square can be left blank. That means you don’t have to ﬁll it all
in if you don’t want to.” This was a relatively easy fraction chal-
lenge for several reasons: 1) it only required ﬁguring out one frac-
8); 2) as it makes use of empty space, the likelihood of chil-
dren interfering with each other’s work was minimized; 3) the task
could be completed using only square pieces (6 out of 16 snaps).
For the third task, the researcher asked the participants to “make
the square one tenth red, four tenths green, two tenths blue, and
three tenths yellow. Now, this one is a bit different to the last one.
In this one, none of the squares should be left blank.” The third
task is signiﬁcantly more challenging than the second for several
reasons: 1) it uses a larger grid (5-by-5 versus 4-by-4 for the pre-
vious tasks); 2) it requires four different fractions; 3) participants
had to use triangular pieces (e.g., 1
10 is most easily composed of
two squares and one half triangle); 4) it requires using unreduced
fractions (e.g., 4
10 for green); 5) as the whole tile must be ﬁlled in,
the chances for clashes between users or accidental destruction of
previous progress is increased, particularly towards completion of
the task when the tile ﬁlls up.
Common fraction misconceptions were observed. First, students
often fail to realize that a larger denominator indicates a smaller
fraction (e.g., 3
8). In such cases, students mistakenly try to
take away pieces to achieve the larger fraction. Second, students
sometimes fail to realize that the position of pieces does not affect
the fraction. Third, students do not realize that they can multi-
plying a pattern of pieces to multiply the numerator of the frac-
tion. For instance, on the second task, a student might ﬁnd that
two square pieces equals 1
8but not realize that therefore six square
pieces equals 3
8. Fourth, students do not realize that two fractions
can be equivalent. This is particularly problematic in the third task
where teams observe 1
5blue and fail to realize that they have ac-
Table 1: Descriptive Statistics of the Pairs
Pair Amy & Ben Chris & Dave Emily & Ford
% of time with matching colors
Task 1 0% 69% 27%
Task 2 0% 97% 12%
Task 3 0% 93% 6%
verbal declarations / exchanges per minute
Task 1 0.8 / 0.5 2.4 / 3.1 0.8 / 0.8
Task 2 3.3 / 1.4 2.8 / 4.4 0.3 / 0.9
Task 3 4.7 / 1.7 2.2 / 2.4 1.0 / 1.2
complished the goal of 2
10 blue. As these cases demonstrate, stu-
dents engage these misconceptions by working with DigiTile on
5. THREE DIFFERENT CASES
A previous article reported the quantitative results of this ﬁeld study:
It detailed the learning gains in fraction understanding made by Di-
giTile users over a comparison group, using standardized pre- and
post-tests . Here, we report on the qualitative ﬁndings based
on video analysis. We present three cases as a case study  of
how individual group dynamics affect children’s collaboration on
interactive tabletops. The children are given corresponding-gender
pseudonyms in alphabetical order based on case order and position
at the tabletop (left to right in the ﬁgures). To further protect par-
ticipants and aid comprehensibility, ﬁgures based on the video are
presented as sketches.
Before detailing the cases, we preview the major differences be-
tween pairs. While these differences are qualitative, they are cor-
roborated by simple quantitative analysis of the interaction. Ta-
ble 1 shows what percentage of time the colors of the two palettes
matched and the rate at which pairs engaged in verbal declara-
tions (the partner does not respond) and exchanges (the partner re-
Amy & Ben All three tasks were designed to be split up by color.
Amy and Ben choose to strictly divide the task, never using
the same color. While they work on different parts of the
task, they use narration (a running commentary on their ac-
tions) to communicate; this manifests as a high rate of verbal
Chris & Dave They choose to work together on the same part of
the task. Whenever they move on to a new part, Dave switches
both palettes to that color. Because of their joint focus, they
are able to discuss strategies before executing them; this man-
ifests as a high rate of verbal exchanges.
Emily & Ford Like Amy and Ben, they choose to split up the task
by color; however, their verbal communication is minimal.
It is so noticeably lacking that the teacher prompts them to
communicate to solve the challenge in the ﬁrst and third task.
While this prompting sparked bursts of verbal communica-
tion, their overall rate was still relatively low.
While the collaboration styles diverged, all three pairs were suc-
cessful, demonstrating increasing competency throughout the ses-
sion and succeeding on the tasks. All three accomplished the ﬁrst
two tasks and made signiﬁcant progress on the third. Chris and
Dave were the only pair to ﬁnish the third task. Amy and Ben
brieﬂy completed the challenge, but failed to realize it as they had
problems understanding the equivalence of unreduced fractions.
While Emily understood how to complete the third task, Ford’s
contributions accidentally stopped her from ﬁnishing the task in
the allotted time frame; at several points, at least three out of the
four parts were solved.
Each case begins with an introduction to the pair’s dynamic. Then,
one episode of collaborative learning is detailed for each of the
fraction tasks. These episodes were chosen to exemplify the differ-
ences between the pairs.
5.1 Amy & Ben: Dividing the Task
Amy and Ben strictly divide each task by color. While all three
tasks were explicitly designed to allow for parallel work, the re-
searcher never gives any indication that they need to divide the
task. Amy and Ben chose that strategy for themselves from the be-
ginning. Although they work on different parts of the task, they
communicate well, speaking often and regularly checking in on
what the partner is doing. They do not get frustrated with each
other, even when the task seems difﬁcult or a technology malfunc-
tion makes their actions interfere with each other. Occasionally,
they will even share a laugh about it.
One remarkable feature is that they tend to mirror each other: mak-
ing eye contact, replicating the same tile pattern, or expressing the
same body language. When the researcher asks them what per-
centage of the tile is red, they have to think about it. Ben guesses
“twenty percent.” Amy doubts it, “Is it? No. Uh. I don’t know.”
They make eye contact and simultaneously answer “ﬁfty percent.”
That mirroring is most evident in the second task, where they fre-
quently place the same tile pattern (when reﬂected across a verti-
cal line of symmetry). In all, such physical mirroring occurs ten
times; in contrast, there is not one instance of mirroring for the
other pairs. At one such occasion, Ben states, “One eighth. .. I’ve
got an eighth.” Amy giggles and replies, “so do I.” Neither knows
where to go from there. They simultaneously scratch their heads.
After some more experimenting and narration, Amy discovers three
eighths with orange. She announces, “Oh [surprised], I’ve done
mine.” Five seconds later, Ben has mirrored hers to accomplish
three eighths brown (Figure 2).
The third task proves to be more difﬁcult with initial progress be-
ing slow. Finally, Amy stumbles onto 1
10 working in green; she
uses two squares and a half triangle. She announces, “I’ve got one
tenth.” Ben replies, “What? How did you get a tenth?” He exam-
ines her side of the board. She admits, “I don’t know. I think that’s
one tenth [points]. Isn’t it?” Ben points to her green button, which
10 , and replies, “It says so.” Amy proceeds to replace the
green pattern with a red one. Meanwhile, Ben narrates, “We need
three of those.” He replicates her pattern three times in yellow (Fig-
ure 3). Successful, he announces, “Three tenths.” “Is that what you
needed,” Amy enquires. “Yeah.” Here Ben demonstrates compe-
tence in the multiplication principle, which he was unable to use
during the second task.
While they choose to work on independent parts of the task, they
communicate well through both physical (as exempliﬁed by the
mirroring) and verbal (as exempliﬁed in the exchanges from the
third task) means. So, while they do not share a joint task focus
Figure 2: Ben mirrors Amy to complete Task 2
Figure 3: Amy checks the fraction as Ben completes 3
(something often considered essential to collaboration ), they
manage to communicate andlearn together. Furthermore, the phys-
ical mirroring they exhibit is an indicator of group ﬂow, a state
where the group is performing at thepeak of its abilities .
5.2 Chris & Dave: Sharing the Task
Chris and Dave work well together, but both contribute different
properties to their group dynamic. Chris is intense in both his ver-
bal and body language. His emotions, ranging from excitement to
doubt and frustration, are readily apparent. Dave is more relaxed.
While Chris talks signiﬁcantly more, Dave usually takes the leader-
ship in physically placing pieces. Unlike Amy and Ben, Chris and
Dave choose to share the task. They work together on one color at
As evident in the second task, Chris and Dave often take turns ac-
tively moving pieces and observing / commenting. Dave starts by
adding four triangular pieces to the second row. Chris comments,
“That’s one eighth.” Dave narrates, “[I’ll] do it again underneath
Figure 4: Chris explains Task 2 solution to Dave
here,” as he places triangular pieces in the row below it. Chris
wants to reserve the lower rows for orange pieces and starts mov-
ing the pieces up one row. Dave agrees and they end up with the
top two rows containing half triangles. Chris announces, “That’s
one fourth. It’s got to be three eighths.. .I think the squares [in-
dicating the central tile] need to be bigger.” Chris wants to keep
the brown pieces conﬁned tothetop half, but wants another row to
place more pieces. He comes up with a solution, rotating a brown
piece around to ﬁll in the gaps left by the existing triangles: “No,
but we could still do this.” Dave agrees. Both place brown pieces
into the gaps rapidly. At one point, Chris notices that the interface
8. Dave does not notice and places another piece, thereby
changing it. Chris exclaims, “Three eighths. Take that [pointing to
a triangle] off.” He removes the piece himself and they have the
brown part of the challenge. Dave changes both palettes to orange.
Dave claims, “Now, we just need to do the same. .. ” as he is about
to place orange triangles. Chris counters, “Hold on. Wait, wait,
wait, wait. Hold on, hold on, ‘cause you can do squares [moving
an orange square onto the board].” Dave agrees, “Oh yeah. Let’s
do squares then. It’s easier.” After they have placed six orange
squares, Chris notices that they are done. He is surprised that the
orange pattern is not the same as the brown pattern. He steps back
and shrugs while looking around. After twelve seconds, he real-
izes why both patterns represent the same fraction. He explains
it to Dave, “Oh yeah. Look! ‘Cause when you switch that round
[Figure 4], it’s three eighths.”
The third task proves more difﬁcult. After several failed attempts
to accomplish 1
10 red, Chris claims, “Oh! I know how to do it.”
He starts over and ﬁlls in one column with red half triangles. He
explains that he’ll ﬁll another column, but Dave notices that the
current arrangement is one tenth. Proudly Dave proclaims, “We’ve
done it.” Surprised, Chris accepts it: “That is one tenth.” Both are
noticeably excited. Encouraged verbally by Chris, Dave quickly
adds two columns of blue triangles. They are thrilled to discover
that it is two tenths as expected. With broad smiles, both exclaim,
“Yes!” Dave switches both palettes to yellow. As in the previous
task, Chris is worried that they will run out of space for yellow,
given that there are only two empty columns. Dave understands
his concern and comes up with a solution. He rotates a yellow half
triangle to ﬁll in the gap made by the red and blue triangles. Once
he realizes what Dave is doing, Chris gives his approval: “Yeah.
Figure 5: Chris and Dave accomplish 3
10 on Task 3
You’ve got it.” Dave ﬁlls in the gaps in the center column with yel-
low triangles. Dave is planning to continue in the empty columns,
but Chris motions him to the remaining gaps in the blue and red
columns: “That means, just do it over there [gesturing to the trian-
gular gaps].” They ﬁll in the empty spaces together. Reading the
display, Chris announces, “Boom! Three tenths” (Figure 5).
Chris and Dave’s interaction is a textbook example of collabora-
tion: they share a joint task focus, are able to smoothly alternate
between engaging the task and reﬂecting on it, and are invested in a
successful outcome. Because of the joint focus, each is able to halt
the physical activity to verbally negotiate the next step (something
that Amy and Ben do not do). They often also take different roles:
actor and observer. This role division is particularly well suited
to DigiTile as one can manipulate the tile representation while an-
other observes the fraction representation. As the examples above
demonstrate, the observer is able to notice unexpected progress,
leading to quicker task progress and gains in domain understand-
5.3 Emily & Ford: Working in the Same Space
Emily and Ford are generally cordial, but not particularly friendly
with one another. Unlike the other pairs, they seldom communicate,
either verbally or physically. Another marked difference is that
Emily is noticeably ahead of Ford in fraction understanding. By a
strict deﬁnition, their interaction is not collaboration. Instead, we
characterize it as working in thesame space on the same task.
When they start working on the second task. Both work indepen-
dently with a different color on their respective halves. Emily picks
brown; Ford picks orange. They work independently. He gets stuck
using quarter triangles. After some experimenting, Emily ﬁgures
out that six squares solve 3
8brown. She stops working. Soon after
she stops working, Ford notices. He inquires, “[Are] you done?”
Emily nods. Ford continues to work on orange with the small trian-
gles as Emily looks on. Ford fails to understand that he could just
replicate her pattern. At one point, she squints with concentration.
A few seconds later, she approaches him. “Hey Ford. You see how
I used six pieces to do my part (Figure 6). You could use the same
thing.” He accepts, “Oh, yeah.” As he quickly places blocks to
mirror her, she continues, “but you don’t need to put it in the same
Figure 6: Emily explains Task 2 solution to Ford
pattern, so you could put it over there [motions].” Ford accepts and
creates a different pattern to complete the task. Thus, she coaches
him to understand the shape and location independence principle.
Before task three, the researcher encourages them to talk. Again,
both of them work independently and mostly in silence. They seem
to quite systematically try out different combinations of shapes.
Emily works on yellow; Ford works on red. Then, Ford discov-
ers that 1
10 can be made with two wholes and a half triangle. He
pauses. Emily glances at Ford’s side of the board, noticing that
Ford has completed red (top of Figure 7). She proceeds to imme-
diately duplicate his pattern three times to achieve 3
10 (bottom of
Figure 7), thereby demonstrating her mastery of the multiplication
principle. While this example is similar to that of Amy and Ben’s
(i.e., Amy and Ford discover 1
10 ; Ben and Emily use it to achieve
10 ), there is, remarkably, no verbal communication. Ford has no
awareness that Emily has appropriated his ﬁnding or that she un-
derstands a principle that will be useful for solving the rest of the
Given the limited communication, it is even more difﬁcult to char-
acterize this pair’s work as collaboration than for Amy and Ben;
however, their interaction is still a successful example of collabo-
rative learning: Both learners beneﬁt from working together. As
evidenced by the former example, Ford beneﬁts by working with
a more knowledgeable peer who can, at times, act as a tutor. As
evidenced by the latter example, it is not a one way street: Emily is
able to appropriate a discovery made by Ford.
6. BEYOND ONE-SIZE-FITS-ALL
As these three cases demonstrate, collaborative learning using an
interactive tabletop takes different forms, even for children from
the same classroom working on the same task with the same ap-
plication. When trying to characterize how interactive tabletops
support collaborative learning, one size does not ﬁt all. Table 2
summarizes the differences between the pairs
6.1 Beneﬁts of Working Together
One oft-cited model of how two people engage in learning is Vy-
gotsky’s  zone of proximal development (ZPD). The range be-
tween tasks that a person can accomplish by themselves and those
Figure 7: Emily uses Ford’s solution to 1
10 red for 3
that they can accomplish with the other’s help is the ZPD. Vygotsky
based this model of learning on observations of how mothers play
with their young children. Inherent to this model is an inequity
between participants as Vygotsky is concerned with the learning
and ZPD of the child, rather than that of the mother. Of the three
cases, this model only applies to Ford and Emily. Emily is notice-
ably more proﬁcient in all three tasks. For the ﬁrst two tasks, her
actions provide Ford with a model of how to accomplish his part.
Occasionally, she provides explicit verbal guidance. In these ex-
changes, Emily acts as the tutor to Ford’s pupil. Ford beneﬁts by
working with a more capable peer. Emily beneﬁts by articulating
her understanding, a valuable meta-cognitive skill .
While Emily is ahead of Ford, it is not by that much; hence, she
cannot always articulate her understanding. Her mental effort in
assisting Ford is particularly visible towards the end of the second
task. In the third task, the tutor / pupil relationship breaks down.
Emily ignores Ford’s confusion and seems annoyed when his work
on blue interferes with her progress on green. When prompted by
the teacher to assist Ford, Emily does try to explain fraction equiv-
alence to Ford, but her explanation is not well formed and ineffec-
tive. Even though he fails to understand, she makes no effort to
follow up, perhaps because her own understanding is still tenuous.
For such a challenging task, the ZPD created by Emily working
with Ford is insufﬁcient to allow Ford to accomplish the task. We
should also remember that Emily is not a trained tutor; she mainly
chooses to help Ford when it directly beneﬁts her (e.g., avoiding
inactivity at the end of the second task).
A different model of collaboration is Roschelle’s  convergent
conceptual change: as two learners work with a tool that embodies
the domain concepts, their conceptual understanding tends to con-
verge with each other and the domain concepts. Unlike the zone
of proximal development, this theory allows for the learners to be
comparable in skills and content understanding. The criterion for
useful collaboration is that both learners are able to generate ideas
and evaluate them. As the tool allows them to test their theories,
correct concepts are more likely to endure than misconceptions. In
his characterization, Roschelle places emphasis on learners having
to work together (i.e., a joint task focus) to choose which ideas they
want to evaluate. This is a ﬁtting description of how Chris and Dave
work. They discuss their conﬂicts and negotiate which ideas they
will test. They ﬂuidly transition between the roles of suggesting
ideas and evaluating them.
Roschelle based his theory on observations of pairs using desktop
software with only one entry point—a shared mouse. Furthermore,
his software ran one simulation at a time. In contrast, interactive
tabletops allow users to interact concurrently and DigiTile allows
users to work on different parts (i.e., colors) of the task concur-
rently. The pairs from the ﬁrst and last case chose to divide the task
and work in parallel. For Ford and Emily, this allowed them to work
independently. Do-Lenh et al.  report on a study comparing the
same jigsaw task being performed on an interactive tabletop and a
desktop machine. As groups in the tabletop condition worked inde-
pendently, they had little need to share their individual understand-
ing. As a result, groups using a desktop machine learned more,
even though the tabletop interface was preferred. While interactive
tabletops can fail to support collaborative learning because there
is less impetus to interact and compromise, Amy and Ben demon-
strate that this is not inherently the case. While they worked in
parallel, they consistently shared ideas and converged on correct
6.2 Beneﬁts of Using an Interactive Tabletop
Two oft-cited beneﬁts of interactive tabletops are their ability to
support the awareness of other’s actions and the ability to support
concurrent input. While other studies have demonstrated these ben-
eﬁts in general, we further show how individual group dynamics
impact how these beneﬁts support collaborative learning.
Interactive tabletops are large horizontal displays that users engage
through direct manipulation. As the surface is large, actions tend
to be large enough that fellow users are able to be aware of them
even if they occur in their peripheral vision. In contrast, staying
aware of another’s mouse cursor in single display groupware 
requires more direct attention. As the objects are directly manipu-
lated, it is easy for a user to transition between interacting with the
application and gesturing to communicate with others at the table.
Thus, tabletop interfaces improve indirect and direct communica-
tion respectively. In a study that compared mouse and touch input
for collaborating on a planning task, awareness of others’ activity
was increased in the touch condition . The touch condition also
led to an increase in awareness work, deﬁned by Schmidt  as
Table 2: Summary of Learning and Interaction by Pair
Pair Amy & Ben Chris & Dave Emily & Ford
beneﬁts of working together
Learning Theory [no established theory] Convergent Conceptual
Change Zone of Proximal Development
Advantages They can work independently,
but still share their ﬁndings
with each other.
By running experiments and
reﬂecting verbally on the
results, they are able to
converge on a better domain
The more advanced student
acts as a model to help scaffold
her partner through tasks that
would have been too difﬁcult
beneﬁts of using an interactive tabletop
Awareness While they divide the task,
remain open through narration
and being visually able to
check the other’s progress.
Learners switch smoothly
between acting and observing,
thereby enabling verbal
exchanges centered on the
shared task goal.
The less advanced student is
aware of the progress of his
partner. Thus, these actions can
model the domain skills.
Concurrent Input They are able to smoothly split
up the task to explore their
solutions to the problem
They smoothly transition
between actor and observer. At
times, they simultaneously
The more advanced student is
unable to dominate the
interaction, allowing the
partner to make contributions.
reciprocal practices of monitoring others and designing actions so
as to render visible certain aspects of activity.
For Ford and Emily, there was little explicit verbal or physical com-
munication; however, both were aware of what the other was doing,
which facilitated collaborative learning. For Ford, awareness of the
activity of a more able peer provided a model that enabled him to
perform at a level above that at which he would have been able to
on his own. During the second task, Ford soon noticed that Emily
had ceased her efforts, which caused him to inquire as to whether
she had ﬁnished. It was perhaps this question that prompted Emily
to monitor Ford’s progress and ultimately provide him assistance
in completing the task. During the third task, Ford stopped adding
pieces when he completed 1
10 red. Likely triggered by his pause,
Emily glances over and then appropriates his discovery. Chris and
Dave’s activity was much more focused and integrated. The table-
top interface provided them a shared workspace enabling them to
form a “common ground” of understanding and supporting a close
coupling of actions, deictic dialogue, and gestures. As DigiTile
provided multiple linked representations  of the fraction con-
tent, one could manipulate the central-tile representation while the
other monitored the displayed fraction. As is common in tabletop
tasks , Amy and Ben continuously provided narration. Between
that verbal narration and the ability to see visual actions, Amy and
Ben were aware enough of each other to both execute their own
actions while still engaging their partner.
Unlike most educational technology (e.g., desktop computers, hand-
held devices, and even electronic whiteboards), interactive table-
tops allow multiple users to interact simultaneously. This has sev-
eral beneﬁts. First, it empowers every participant. If Emily and
Ford had worked on a desktop machine, it seems likely that Emily
would have dominated the team; as is, Ford, though a less skilled
partner, was actively able to contribute. At the beginning of the
third task, his discovery of 1
10 even directly helped Emily make
progress. Second, it enables working in parallel. Both the pairs
of Emily and Ford and Amy and Ben demonstrate how working
in parallel can be effective for collaborative learning. Third, unin-
tentional interference of actions can provide an impetus for groups
to engage with each other [8, 27]. As Chris and Dave work on the
same part, they must negotiate their actions to prevent those actions
from interfering with each other. For the other pairs, unintentional
interference of actions only became a problem towards the end of
the third task. When it arrives, neither Amy and Ben nor Emily
and Ford deal with it as well as Chris and Dave have been doing.
That is perhaps one reason why Chris and Dave are the only team
to successfully complete the third task.
6.3 Implications of Diverse Group Dynamics
Whereas the conventional approach to the evaluation of tabletop
systems has been to look for generalities in interactions with and
around the table, we have deliberately focused on what was id-
iosyncratic and related to individual group dynamics. What this has
highlighted is the diversity of ways that interacting with the Digi-
Tile system can facilitate (and hinder) collaborative learning. Ac-
knowledging the importance of group dynamics has implications
for both the design and evaluation of such systems.
From a developer’s perspective, the question is whether to embrace
or spurn such diversity. All three of our cases focus on successful
pairs. DigiTile works for these pairs, at least partially, because it
supports a variety of collaboration styles. On the one hand, design-
ers might thus beneﬁt learning by supporting such diversity. On
the other hand, other projects have demonstrated the value of en-
couraging (or even enforcing) particular collaborative behaviours
[2, 16, 22]. For instance, if the system only allowed one person
at a time to interact , this could focus the attention of the other
person on the relevant activity. That would create an incentive for
pairs to share a task focus, something considered beneﬁcial for col-
laborative learning . Such an interface might support Emily and
Ford to coordinate more closely and for Emily to take a more active
role in mentoring Ford; alternatively, it could have led to adverse
While all pairs were successful, Chris and Dave were the best ex-
ample of collaborative learning—fully completing the third task
and displaying a joint task focus. The system could be changed
to more subtly encourage their collaboration style. In this study,
special effort was given to allow pairs to split up the tasks. All
three challenges were designed to be split across different colors
and half of the pairs used a software version where each palette
only contained three colors (red, orange, green vs. brown, yellow,
blue). It was hoped that such a condition would enforce collabora-
tion through splitting the task; as it turned out, pairs did not need
that incentive to divide the tasks. Considering that all the pairs
chosen for this qualitative analysis based on their task performance
and diversity of collaboration styles were in the other condition, it
may have been a pedagogically poor choice to even encourage such
splitting; however, both conditions showed similar improvements
in learning gains . A useful middle ground between supporting
diversity and encouraging ideal behavior is to allow the teacher (or
even the learners) to select the software features that best support
the relative competence and social dynamics of the group.
In addition to having implications for the design of systems, such
diversity has implications for how we analyze these interactions.
For instance, previous work has shown that tabletops encourage
a higher equity of participation than other interfaces [9, 22, 28].
While equity of participation is a good indicator of successful group
work, it is not an absolute good. Chris and Dave demonstrate how
a group can work well together even when one person dominates
the action (i.e., low equity of physical participation). Enforcing eq-
uitable physical contributions (e.g., by enforced turn taking) would
have only disrupted theirsuccessful group dynamics.
While it is interesting to consider how design features can explicitly
address group dynamics, it is worth noting that other design choices
can have signiﬁcant implicit effects on those dynamics and the ef-
fectiveness of the system. Based on this study, DigiTile was altered
to include a pie chart representation of the fractions. That repre-
sentation is particularly useful for realizing that larger denomina-
tors imply smaller fractions, while larger numerators imply larger
fractions. This could have helped Amy and Ben realize that they
needed to add pieces to move from 1
8. It could have helped
Emily to better explain fraction equivalence to Ford. Also, the soft-
ware could be aware of the speciﬁc challenge and provide support
based on that. For instance, when Amy and Ben completed the
third task, DigiTile could have announced their success. While they
might have been surprised by their success, such a scenario could
have led to them to reﬂect on fraction equivalence.
This work was done as part of the ShareIT project funded by the
EPSRC, grant number EP/F017324/1. We thank the schools we
worked with and the children who participated in this study. We
thank MERL for loaning us the DiamondTouch. We especially
thank Kim Bryant for sketching the ﬁgures.
 A. C. Armstrong. The fragility of group ﬂow: The
experiences of two small groups in a middle school
mathematics classroom. Journal of Mathematical Behavior,
 S. Benford, B. B. Bederson, K.-P. Akesson, V. Bayon,
A. Druin, P. Hansson, J. P. Hourcade, R. Ingram, H. Neale,
C. O’Malley, K. T. Simsarian, D. Stanton, Y. Sundblad, and
G. Taxén. Designing storytelling technologies to encouraging
collaboration between young children. In Proceedings of
CHI ’00, pages 556–563, New York, 2000. ACM Press.
 X. Cao, S. E. Lindley, J. Helmes, and A. Sellen. Telling the
whole story: Anticipation, inspiration and reputation in a
ﬁeld deployment of TellTable. In Proceedings of CSCW ’10,
pages 251–260, New York, 2010. ACM Press.
 A. Collins, J. S. Brown, and S. E. Newman. Cognitive
apprenticeship: Teaching the crafts of reading, writing, and
mathematics. In L. B. Resnick, editor, Knowing, Learning,
and Instruction: Essays in honor of Robert Glaser, pages
453–494. Lawrence Erlbaum Associates, Hillsdale, NJ, 1989.
 P. Dietz and D. Leigh. DiamondTouch: A multi-user touch
technology. In Proceedings of UIST ’01, pages 219–226,
New York, 2001. ACM Press.
 P. Dillenbourg. What do you mean by collaborative learning?
In P. Dillenbourg, editor, Collaborative-learning: Cognitive
and computational approaches, pages 1–19. Elsevier,
 S. Do-Lenh, F. Kaplan, and P. Dillenbourg. Paper-based
concept map: The effects of tabletop on an expressive
collaborative learning task. In Proceedings of HCI 2009,
pages 149–158, New York, 2009. ACM Press.
 R. Fleck, Y. Rogers, N. Yuill, P. Marshall, A. Carr, J. Rick,
and V. Bonnett. Actions speak loudly with words:
Unpacking collaboration around the table. In Proceedings of
ITS ’09, pages 189–196, New York, 2009. ACM Press.
 A. Harris, J. Rick, V. Bonnett, N. Yuill, R. Fleck, P. Marshall,
and Y. Rogers. Around the table: Are multiple-touch
surfaces better than single-touch for children’s collaborative
interactions? In Proceedings of CSCL ’09, pages 335–344.
 A. Hatch, S. Higgins, A. Joyce-Gibbons, and E. Mercier.
ClassCalc: Using multi-touch technology to support within
and between-group mathematics learning. In Proceedings of
CSCL 2011. ISLS, 2011. to appear.
 U. Hinrichs and S. Carpendale. Gestures in the wild:
Studying multi-touch gesture sequences on interactive
tabletop exhibits. In Proceedings of CHI 2011, New York,
2011. ACM Press. to appear.
 E. Hornecker. "I don’t understand it either, but it is cool":
Visitor interactions with a multi-touch table in a museum. In
Proceedings of TABLETOP ’08, pages 121–128, Los
Alamitos, CA, 2008. IEEE.
 E. Hornecker, P. Marshall, N. Dalton, and Y. Rogers.
Collaboration and interference: Awareness with mice or
touch input. In Proceedings of CSCW ’08, New York, 2008.
 B. Jordan and A. Henderson. Interaction analysis:
Foundations and practice. The Journal of the Learning
Sciences, 4(1):39–103, 1995.
 J. Kaput. Linking representations in the symbolic systems of
algebra. In S. Wagner and C. Kieran, editors, Research
agenda for mathematics education: Research issues in the
learning and teaching of algebra, pages 167–194. National
Council of Teachers of Mathematics, Reston, VA, 1989.
 L. Kerawalla, D. Pearce, N. Yuill, R. Luckin, and A. Harris.
"I’m keeping those there, are you?" The role of a new user
interface paradigm—separate control of shared space
(SCOSS)—in the collaborative decision-making process.
Computers & Education, 50(1):193–206, 2008.
 A. Kharrufa, D. Leat, and P. Olivier. Digital mysteries:
designing for learning at the tabletop. In Proceedings of ITS
’10, pages 197–206, New York, 2010. ACM Press.
 K. K. Lamberty and J. L. Kolodner. Exploring digital quilt
design using manipulatives as a math learning tool. In
P. Bell, R. Stevens, and T. Satwicz, editors, Proceedings of
ICLS 2002, pages 552–553, Mahwah, NJ, 2002. Lawrence
 E. I. Mansor, A. De Angeli, and O. De Bruijn. The fantasy
table. In Proceedings of IDC ’09, pages 70–79, New York,
2009. ACM Press.
 J. Marco, E. Cerezo, S. Baldassarri, E. Mazzone, and J. C.
Read. Bringing tabletop technologies to kindergarten
children. In Proceedings of HCI ’09, pages 103–111,
Swinton, UK, 2009. British Computer Society.
 P. Marshall, R. Fleck, A. Harris, J. Rick, E. Hornecker,
Y. Rogers, N. Yuill, and N. S. Dalton. Fighting for control:
Children’s embodied interactions when using physical and
digital representations. In Proceedings of CHI ’09, pages
2149–2152, New York, 2009. ACM Press.
 P. Marshall, E. Hornecker, R. Morris, S. Dalton, and
Y. Rogers. When the ﬁngers do the talking: A study of group
participation for different kinds of shareable surfaces. In
Proceedings of TABLETOP ’08, Washington, DC, 2008.
IEEE Computer Society.
 P. Marshall, R. Morris, Y. Rogers, S. Kreitmayer, and
M. Davies. Rethinking ’multi-user’: An in-the-wild study of
how groups approach a walk-up-and-use tabletop interface.
In Proceedings of CHI 2011, New York, 2011. ACM Press.
 K. O’Hara. Interactivity and non-interactivity on tabletops.
In Proceedings of CHI 2010, pages 2611–2614, New York,
2010. ACM Press.
 A. M. Piper and J. D. Hollan. Tabletop displays for small
group study: Affordances of paper and digital materials. In
Proceedings of CHI ’09, pages 1227–1236, New York, 2009.
 A. M. Piper, E. O’Brien, M. R. Morris, and T. Winograd.
SIDES: A cooperative tabletop computer game for social
skills development. In Proceedings of CSCW ’06, pages
1–10, New York, 2006. ACM Press.
 T. Pontual Falcão and S. Price. What have you done! The
role of ’interference’ in tangible environments for supporting
collaborative learning. In Proceedings of CSCL ’09, pages
325–334. ISLS, 2009.
 J. Rick, A. Harris, P. Marshall, R. Fleck, N. Yuill, and
Y. Rogers. Children designing together on a multi-touch
tabletop: An analysis of spatial orientation and user
interactions. In Proceedings of IDC ’09, pages 106–114,
New York, 2009. ACM Press.
 J. Rick and Y. Rogers. From DigiQuilt to DigiTile: Adapting
educational technology to a multi-touch table. In
Proceedings of TABLETOP ’08, pages 79–86, Los Alamitos,
CA, 2008. IEEE.
 J. Rick, Y. Rogers, C. Haig, and N. Yuill. Learning by doing
with shareable interfaces. Children, Youth & Environments,
 Y. Rogers, Y.-K. Lim, W. R. Hazlewood, and P. Marshall.
Equal opportunities: Do shareable interfaces promote more
group participation than single users displays?
Human-Computer Interaction, 24(2):79–116, 2009.
 J. Roschelle. Learning by collaborating: Convergent
conceptual change. In T. Koschman, editor, CSCL: Theory
and Practice of an Emerging Paradigm, pages 209–248.
Lawrence Erlbaum Associates, Mahwah, NJ, 1996.
 K. Schmidt. The problem with "awareness". JCSCW,
 R. J. W. Sluis, I. Weevers, C. H. G. J. van Schijndel,
L. Kolos-Mazuryk, S. Fitrianie, and J. B. O. S. Martens.
Read-It: Five-to-seven-year-old children learn to read in a
tabletop environment. In Proceedings of IDC ’04, pages
73–80, New York, 2004. ACM Press.
 J. Stewart, B. B. Bederson, and A. Druin. Single display
groupware: A model for co-present collaboration. In
Proceedings of CHI ’99, pages 286–293, New York, 1999.
 L. Suchman and R. Trigg. Understanding practice: Video as
a medium for reﬂection and design. In J. Greenbaum and
M. Kyng, editors, Design at Work, pages 65–89. Lawrence
Erlbaum Associates, London, 1991.
 L. S. Vygotsky. Mind in Society: The Development of Higher
Psychological Processes. Harvard University Press,
Cambridge, MA, 1978.
 R. K. Yin. Case Study Research: Design and Methods. Sage,
Thousand Oaks, CA, third edition, 2003.