Jl.

of Computers in Mathematics and Science Teaching (2012) 31(2), 93-113

Making Sense of Integer Arithmetic: The Effect of Using

Virtual Manipulatives on Students’ Representational
Fluency

Johnna Bolyard
West Virginia University, USA
[email protected]

Patricia S. Moyer-Packenham
Utah State University, USA
[email protected]

This study investigated how the use of virtual manipulatives

in integer instruction impacts student achievement for integer
addition and subtraction. Of particular interest was the in-
fluence of using virtual manipulatives on students’ ability to
create and translate among representations for integer compu-
tation. The research employed a quasi-experimental pretest-
posttest design. Ninety-nine sixth-grade students participated
over a four-week period. Six classes were randomly assigned
to one of three treatment groups. Students increased in inte-
ger computation achievement and demonstrated facility with
pictures and written representations. Students had more dif-
ficulty creating symbolic representations and making connec-
tions among this and other representational forms.

Many mathematics educators agree that the gateway to higher mathe-

matics is the first year algebra course. One topic important to the study of
algebra is integer arithmetic. Students must have the opportunity to engage
in experiences that enable them to develop a deep understanding of these
fundamental concepts. Although integers and integer arithmetic concepts
are generally considered important topics in mathematics, there has been lit-
tle research conducted in this area (Kilpatrick, Swafford, & Findell, 2001).
94 Bolyard and Moyer-Packenham

Some studies have examined students’ pre-instructional knowledge (Muk-

hopadhyay, Resnick, & Schauble, 1990; Peled, Mukhopadhyay, & Resn-
ick, 1989). These results indicate that students have some knowledge prior
to instruction and often use their knowledge of whole number arithmetic to
reason through problems involving negative integers. Other research has fo-
cused on appropriate instructional models for integer arithmetic. Even fewer
studies have examined how models for integers can be used for instruction
using technology such as virtual manipulatives.
The present study was designed to address the need for understanding
how virtual manipulatives might support students’ representational fluency
in the addition and subtraction of integers. Specifically, we investigated how
the use of virtual manipulatives, representing two different models for inte-
gers, impacts student achievement for integer addition and subtraction. Also
of interest in this study was the influence of virtual manipulatives on stu-
dents’ ability to create and translate among representations for integer addi-
tion and subtraction.

Literature Review

Integers

Representational metaphors used to facilitate the learning of integer

arithmetic include positively and negatively charged particles, debts and as-
sets, and movements along a number line. Different models emphasize dif-
ferent attributes of number. The number line model focuses on the measure-
ment aspect of number, or the notion of how much; debts and assets models
focus on the quantity aspect of number, or how many. While these models
may help students conceptualize integer values and, in some cases, integer
addition, a common criticism is that they do not adequately represent other
operations with integers, particularly subtraction (Kilpatrick, Swafford, &
Findell, 2001). Other proposed models, including Janvier’s (1985) hybrid
model and a model based on algebraic geometry (Moses, Kamii, Swap, &
Howard, 1989; Carson & Day, 1995) focus on quantity and directional as-
pects of number used in algebra.
While proponents of the various models advocate for one over anoth-
er, few attempts have been made to compare the effectiveness of different
models. Two studies comparing quantity and number line models found
that students using quantity models outperformed those using number line
models (Liebeck, 1990; Sherzer, 1973). One study comparing Janvier’s hy-
Making Sense of Integer Arithmetic 95

brid model and a quantity model found that students using the hybrid model
outperformed those using a quantity model (Janvier, 1985). The limited re-
search does not provide adequate evidence to determine conclusively any
difference in student achievement among the models.
Regardless of the model used, it is clear that integer arithmetic remains
an area of difficulty for many students (Bruno et al., 1997; Peled, 1991).
Many of these difficulties are presented when students work with symbolic
representational forms for integer addition and subtraction. Lesh, Post, and
Behr (1987) discuss five different representational forms that support con-
cept development: manipulative models, pictures, written symbols, real-
world situations, and oral language. Each form has the potential to highlight
different features of a mathematical idea or relationship (NCTM, 2000);
Translation among representational forms builds understanding (Hiebert,
1990). Research indicates that after instruction, students can work with
story problems to solve integer addition and subtraction situations (Peled
& Carraher, 2007). However, students are not as successful with symbol-
ic-only forms. Some attribute this difficulty, in part, to the fact that stu-
dents’ whole number interpretations of addition and subtraction do not eas-
ily translate to these new situations (Kilpatrick, Swafford, & Findell, 2001;
Moses, Kamii, Swap, & Howard, 1989). To make sense of integer compu-
tation, the learner needs to develop a different perception of number com-
pared to that used in whole number arithmetic. In whole number arithme-
tic, the focus on number is mainly one of magnitude or quantity. However,
when students expand their work with number to include negative numbers,
they must also consider direction (Moses, Kamii, Swap, & Howard, 1989;
Peled & Carraher, 2007).
Students must also expand their understanding of the meaning of opera-
tions. Vershcaffel, Greer, and De Corte (2007) categorize additive situations
in whole number arithmetic as involving changes in quantities, combina-
tions of two discrete sets, and comparison of two discrete sets. When the
largest quantity is unknown, adding the two given quantities will produce
the solution; when one of the smaller quantities is unknown, subtracting the
smaller set from the larger set will produce the solution. However, this in-
terpretation does not translate to operations with integers in which students
can no longer think of addition as making bigger and subtraction as making
smaller (Peled & Carraher, 2007). Subtraction presents particular difficulty.
Vlassis (2008) points out that integer computation involves multiple and
varied uses of the minus sign, including: a unary structural signifier (e.g., to
indicate a value, -2, as opposed to 2); a binary operational signifier (e.g., to
indicate subtracting in both arithmetic [take away or difference] and alge-
96 Bolyard and Moyer-Packenham

braic [adding the opposite] contexts; and a symmetric operational signifier

(e.g., taking the opposite of a number). The need to expand one’s interpreta-
tion of number, operations, and symbols adds to the complexity of integer
study.

Technology in Integer Instruction

Technology can support students’ developing understanding of integer

concepts. Cognitive technology tools (Pea, 1985), like virtual manipula-
tives, can connect verbal (e.g., written directions and feedback) and nonver-
bal (e.g., visual/pictorial representations and dynamic objects) representa-
tions. The interconnections between verbal and nonverbal systems proposed
in Dual Coding Theory (DCT, Paivio, 1991) increase the efficiency of the
working memory. Logogens (i.e., language units, such as words, phrases or
sentences), that process information sequentially, and imagens (i.e., repre-
sentation units that activate different types of imagery), that process con-
cepts in synchronous hierarchies, support the formation of representations
of mathematical concepts by allowing the learner to manipulate objects, ob-
serve changes, and make connections.
A thorough review of the research on virtual manipulatives (see Moy-
er-Packenham, Westenskow, & Salkind, manuscript under review) pro-
duced only four studies specifically focusing on using virtual manipulatives
to teach integer concepts (i.e., positive and negative integers, addition and
subtraction of integers). Three of these studies were dissertations (Bolyard,
2006; Smith, 1995; Smith, 2006) and one was a journal publication (More-
no & Mayer, 1999). Overall, these studies found that students using virtual
manipulatives made gains in achievement of integer concepts.
The limited research base on the teaching and learning of integer arith-
metic using virtual manipulatives leaves room for further exploration. This
study was designed to contribute to that knowledge. The following research
questions guided the study:
1. How does the use of three different web-based virtual manipulatives for
integer addition and subtraction impact students’ achievement in com-
putation? And, are there differences in achievement among the three
treatment groups using each web-based virtual manipulative?
2. How does the use of web-based virtual manipulatives for integer ad-
dition and subtraction influence students’ creation of and translation
among representations during task solutions?
Making Sense of Integer Arithmetic 97

Methodology

This study employed a quasi-experimental pretest-posttest design.

The research was conducted over a four-week period in two public middle
schools in a single school district. Six classes were randomly assigned to
one of three treatment groups: Virtual Integer Chips (VIC), Virtual Integer
Chips with Context (VICC), and Virtual Number Line (VNL).

Participants and Site Description

Ninety-nine sixth-grade students (46% female; 54% male) participated

in the study during regular mathematics class instruction. The student de-
mographics included 73% White, 10% Asian/Pacific Islander, 7% Black,
5% Hispanic, 4% Other, and 1% American Indian/Alaskan Native. Each
middle school in the district was equipped with three computer labs and
four computers per classroom. However, the participating teachers reported
little use of computers during mathematics instruction.
The classes were taught by four regular classroom teachers using an in-
structional unit designed by the researchers. The regular classroom teachers
participated in a two-week summer institute taught by one of the researchers
designed to provide professional development in the use of virtual and phys-
ical manipulatives for middle school mathematics. Choosing these teachers
controlled for variability due to teacher expertise with the tools.

Instructional Materials and Procedures

The student participants used one of three different web-based virtual

manipulatives: Virtual Integer Chips (VIC), Virtual Integer Chip with Con-
text VICC), and Virtual Number Line (VNL). The VIC and VICC manipu-
latives were chosen as representative of a quantity model for integer addi-
tion and subtraction. The VNL manipulative was chosen as representative of
a number line model. Each of these tools is presented electronically in the
form of an “applet” or small, stand-alone application program.
The VNL manipulative was the Adding and Subtracting Integers applet,
available from www.explorelearning.com and required a subscription. This
applet presents the user with a number line, one red and one blue slider for
representing integer values on the number line, and the symbolic sentence 0
+ 0 = 0. A button at the top of the screen allows the user to choose either ad-
98 Bolyard and Moyer-Packenham

dition or subtraction. Users move the sliders left (to indicate a negative val-
ue) or right (to indicate a positive value) resulting in an arrow of appropriate
length and direction to appear on the number line. Simultaneously, the val-
ues in the displayed symbolic sentence adjust according to the user’s actions
on the slider. Users also have the option of typing in a symbolic value in a
box at the end of each slider. (See Figure 1).

Figure 1. Virtual number line applet. Copyright © 1999-2003 ExploreLearn-

ing (http://explorelearning.com). All rights reserved. Used by permission.

The VIC manipulatives were the Color Chips Addition and Color Chips
Subtraction applets, available at http://matti.usu.edu/nlvm/nav/ and free on
the Internet. These applets present the user with a problem statement in
symbolic form (i.e., -2 + 3), a color chip bank showing positive (black) and
negative (red) chips, and a workspace. The applet instructs the user to drag
chips into the workspace to model the problem presented and then click
“Continue.” In the Color Chips Addition applet, the user drags pairs of posi-
tive and negative chips together to create zero pairs. As the zero pairs are
created, the chips disappear from the screen. The remaining chips represent
the simplified value of the original problem statement. The user then types
in an answer (See figure 2).
Making Sense of Integer Arithmetic 99

Figure 2. Virtual integer chips applet. The National Library of Virtual Ma-
nipulatives (http://matti.usu.edu). © 2003 Utah State University. All rights
reserved. Used with permission.

In the Color Chips Subtraction applet, the user models the minuend of
the problem using chips, then clicks “Continue.” Next, the user drags zero
pairs (combinations of one positive and one negative chip) to enable him or
her to remove the designated value of positive or negative chips. The user
then enters the answer. In both applets, the program controls actions and
provides feedback. For example, the program prompts the user to re-exam-
ine his or her response with suggestions such as “Your chips don’t represent
the problem shown” or “Check the color of the chips and try again.” The
VICC manipulative is a modified version of the VIC manipulative. Howev-
er, the VICC applet presents the initial problem using the context of debts
and assets. After the initial introduction of the problem, the program contin-
ues in the same manner as the VIC manipulative.
The researchers created the following instructional materials for teach-
ers: a detailed instructional plan, integer task cards, technology guides, stu-
dent recording sheets, and student practice sheets. Instructional activities us-
ing the VNL model used the metaphor of walking along a line a specified
number of spaces in the direction indicated. Instructional activities using the
VIC and VICC models employed the metaphor of debts and assets. These
metaphors were used during all classroom discourse as well as on all in-
structional handouts.
100 Bolyard and Moyer-Packenham

During instruction, introductory activities focused on establishing the

respective metaphor using related physical and pictorial models. Students
also participated in extended computer lab sessions (30 to 40 minutes) in
which they used the virtual manipulatives to explore integer addition and
subtraction. Following each computer session, students engaged in follow-
up discussions focused on articulating their observations. Instruction took
place during four, 90 minute alternating day periods over two weeks to ac-
commodate the middle school block scheduling format used in the school.

Data Sources and Analysis

Data were collected and analyzed from three sources: integer addi-
tion and subtraction pretests and posttests and student interviews. Prior to
instruction, students took a 24-item paper and pencil pretest to assess stu-
dents’ prior knowledge of integer addition and subtraction. The test was
researcher created and contained 12 addition and 12 subtraction items, all
presented in symbolic form (ex. 2 − ¯3). The problems were similar to those
explored during the instructional sessions and were drawn from district and
state assessment standards. Following treatment, students were given a 24-
item paper and pencil posttest. The posttest was of the same format as the
pretest but with different items in order to reduce the threat of test reactivity.
For each assessment, five individuals identified as expert reviewers conduct-
ed an analysis of the test items, examining 1) how appropriately each mea-
sured student achievement in the concept and 2) level of difficulty. During
analysis, the addition and subtraction sections of the pre- and posttest were
examined separately. Each item was worth one point for a total of 12 pos-
sible points for each section.
Following the treatment, one researcher conducted task-based inter-
views (Goldin, 1997) with three randomly selected students from each treat-
ment group (n=9). Each interview was audiotaped and transcribed to main-
tain a written record of students’ responses. During the interviews, students
solved six problems (three addition and three subtraction) presented in sym-
bols, pictures, or written/metaphorical (story) form. For each item, partici-
pants were asked to 1) simplify given problems, 2) represent the problems
in two additional forms (other than the one initially used), and 3) explain
their thinking and solution process. Student work on the tasks was collected
for analysis. During analysis, a paired samples t test was performed using
pre- and posttest scores for the three virtual treatment groups to determine
if there was an overall difference in achievement after using the virtual ma-
Making Sense of Integer Arithmetic 101

nipulatives during instruction. To determine if there were differences in stu-

dent achievement, an Analysis of Variance (ANOVA) was performed on the
posttest measures of the three treatment groups.
Qualitative analysis procedures examined students’ abilities to flexibly
apply knowledge of integer addition and subtraction to various representa-
tions of the concept and make connections among representations. Inter-
view transcripts were coded using conventional content analysis (Hsieh &
Shannon, 2005). Codes were then categorized to compare the data, identify
patterns, and determine emerging themes and categories with respect to stu-
dents’ level of conceptual understanding (Glesne, 1999; Maxwell, 1996).
Two major themes used in this analysis were flexibility in using different
representational forms and the ability to make connections among represen-
tational forms.

Results

The results below are presented based on our two major research ques-
tions. To answer our first research question (How does the use of three dif-
ferent web-based virtual manipulatives for integer addition and subtraction
impact students’ achievement in computation? And, are there differences in
achievement among the three treatment groups using each web-based vir-
tual manipulative?), we used paired t tests to compare pretest and posttest
scores on the integer addition and integer subtraction tests for each treat-
ment group. The size of each significant effect was assessed using Cohen’s
d, with .20, .50, and .80 representing the lower limit of small, medium, and
large effect sizes, respectively. A summary of the means and standard devia-
tions for the pre- and posttests for each treatment group are presented in Ta-
ble 1. There were significant differences for all three groups for the addition
(VIC, t[34] = 4.46, p < .01, d = .75, VICC, t[36] = 4.44, p < .01, d = .73,
and VNL, t[26] = 3.26, p < .01, d = .63) and subtraction (VIC, t[34] = 8.05,
p < .01, d = 1.36, VICC, t[36] = 10.23, p < .01, d = 1.68, and VNL, t[26] =
9.53, p < .01, d = 1.83) portions of the pretest and posttest scores. Thus, all
treatment groups showed a significant increase in student achievement.
102 Bolyard and Moyer-Packenham

Table 1
Pre and Posttest Means and Standard Deviations by Treatment Group
Integer Addition
Treatment Group Pretest Posttest
Virtual Integer Chips (VIC) M 71.90 M 93.57
n = 35 SD 30.79 SD 11.63
Virtual Integer Chips with Context (VICC) M 70.86 M 95.05
n = 37 SD 30.21 SD 9.71
Virtual Number Line (VNL) M 69.14 M 89.51
n = 27 SD 28.76 SD 19.69
Integer Subtraction
Treatment Group Pretest Posttest
Virtual Integer Chips (VIC) M 45.95 M 81.90
n = 35 SD 22.90 SD 19.23
Virtual Integer Chips with Context VICC) M 47.75 M 86.04
n = 37 SD 18.39 SD 19.94
Virtual Number Line (VNL) M 40.43 M 86.73
n = 27 SD 24.31 SD 15.20

The next part of the analysis for our first research question examined
differences for each web-based virtual manipulative. An analysis of variance
(ANOVA) was performed on the integer addition and subtraction pretest
scores to determine if there were pre-existing differences among the three
treatment groups. There were no significant differences among the groups
for addition, F(2, 96) = .07, ns, or subtraction, F(2, 96) = .92, ns pretests.
The ANOVA indicated no significant differences in posttest addition scores
among the three treatment groups: VIC (M = 93.57, SD=11.63), VICC (M
= 95.05, SD = 9.71) or VNL (M = 89.51, SD = 19.69), F(2, 96) = 1.32, ns.
Similarly, there were no significant differences among the groups for sub-
traction: VIC (M = 81.90, SD=19.23), VICC (M = 86.04, SD = 19.94), and
VNL (M = 86.73, SD = 15.20), F(2, 96) = .66, ns.

Analysis of Students’ Work with Representational Forms for Integer

Computation

To answer our second research question (How does the use of web-
based virtual manipulatives for integer addition and subtraction influence
Making Sense of Integer Arithmetic 103

students’ creation of and translation among representations during task solu-

tions?), we analyzed student work on the interview tasks. These interviews
revealed that students were generally able to work flexibly within the writ-
ten/metaphorical and picture representational forms for integer addition and
subtraction. Students had more difficulty working with the symbolic repre-
sentation for integer addition and subtraction. Table 2 summarizes the num-
ber of correct responses for each type of interview task.

Table 2
Number of Correct Responses on Interview Tasks by Representational Form (N = 9)
Presentation Mode Correct Responses by Student Response
Form Types
Addition Items
Written/Metaphorical Picture Symbols
Response Response Response
Item 1: ¯7 + 10 9 9 9
(presented in written/metaphorical)
Item 2: ¯3 + 2 9 9 7
(presented in pictures)
Item 3: ¯4 + ¯5 9 9 9
(presented in symbols)
Subtraction Items
Written/Metaphorical Picture Symbols
Response Response Response
Item 4: 3 − 5 8 8 6
(presented in pictures)
Item 5: ¯9 − ¯3 9 9 6
(presented in written/metaphorical)
Item 6: ¯4 − 5 4 4 4
(presented in symbols)

Facility with Representational Forms: Pictures and Written/Metaphorical

Students in this study were successful in evaluating integer addition and

subtraction items presented in written/metaphorical and picture representa-
tional forms. Further, the students were generally successful in translating
items to written/metaphorical and picture forms.
104 Bolyard and Moyer-Packenham

Items given in picture form. All students were able to correctly evalu-
ate the addition item (representing -3 + 2) using the picture model; all stu-
dents but one correctly evaluated the subtraction problem (representing 3 –
5) presented in pictures. On the addition item, most students described using
the picture to arrive at their answer giving responses such as, “there were
three negatives and two positives then we would cancel that out and there
would be one negative left,” (VIC student) or “the ending position was at
negative one” (VNL student). Although most students were able to correctly
complete the picture on the subtraction item, they showed evidence of strug-
gling with this item and relied on other representations to help them cre-
ate and verify their responses. When asked to create a story representing the
picture, students personalized the stories by inserting themselves into the
narrative. For example, a VICs student said, “I owed my sister three dollars
and then my mom gave me two dollars.” A VNL student related his story to
football: “You gain three yards then lose five yards.”

Facility with Representational Forms: Symbolic

Students demonstrated less facility working with symbolic representa-

tional forms. When students were given an addition item, some could ex-
plain that they knew to use a negative seven and a positive ten to write the
statement, ¯7 + 10, by stating: “You move seven steps left on a number line.
When you go left those are the negatives” (VNL student); and “Um, debt,
because a debt is something that you take away” (VICC student). However,
after writing ¯7 + 10, one VICC student explained, “sometimes when I do
these I do, like, ten minus seven.” Asked why she thought of it as subtrac-
tion even though she had written it as an addition statement, the student re-
sponded uncertainly, “Because this (referring to the ¯7) is a negative num-
ber?” The student used whole number reasoning to simplify the problem.
Another VICC student explained his decision to use addition in his math-
ematical sentence by stating, “I thought it must be adding because most
of the time when it talks about this and it doesn’t say to subtract it usually
means add.” These examples illustrate, that while students were able to con-
nect the words of the story problem to the values of the addends used in
their mathematical sentences, they were not as confident in describing the
connection between the problem statement and their use of the addition op-
eration signifier (+).
Work on the addition item given in pictures (¯3 + 2) further demonstrat-
ed students’ incomplete understanding of the symbolic form of represent-
Making Sense of Integer Arithmetic 105

ing integer addition. While most students were able to produce an accurate
symbolic statement, this task presented difficulties for some students. One
student in the VIC group wrote the statement as subtraction (¯3 − +2) rather
than addition and explained: “Um, there was negative three chips and posi-
tive two chips so I figured it would be a subtraction problem because there
were more of one. So, I, my final answer was -1 because negative three mi-
nus positive 2 was negative 1.” A VNL student, who evaluated the picture
representation correctly, wrote a symbolic statement of ¯3 + ¯2. The student
indicated that she looked at the arrow indicating three steps left to determine
¯3 and the arrow indicating two steps right to determine ¯2. These students
did not appear to notice any inconsistencies between the result of their ini-
tial evaluation of the given problem statement and the symbolic statement
they created in the second task.
The symbolic form presented even greater challenges for students in
subtraction contexts. For example, on the first item (3 – 5), presented in pic-
tures, six students correctly produced a symbolic statement of 3 − 5. How-
ever, three students had difficulties correctly translating their work from pic-
tures to symbols. Two VIC students drew a picture in which they added five
negatives rather than subtracting or removing five positives (by adding zero
pairs). One student used a story about owing her friends $5 to help her com-
plete her statement. These students recognized that removing five positive
tiles from the picture was equivalent to adding five negative tiles. However,
this connection failed to translate to symbols. Both students created problem
statements of 3 − (¯5) rather than 3 + (¯5) as they had represented in their
pictures. In this scenario, the minus sign serves a binary function (Vlassis,
2008). However, in the symbolic form, these students included two minus
signs: the first serving as a binary sign to indicate subtraction and the sec-
ond serving as a unary sign attached to the numeral 5.
On the second subtraction item (¯9 - ¯3), presented in written/meta-
phorical form, students correctly represented their interpretation of the story
problem using a picture. Similar to their work in addition, students’ com-
ments illustrated that they were making connections between the words of
the story and the integer values they represent, i.e., “if you owe your brother
$9 that means it’s a debt, so that would be negative nine” (VIC student).
Students used both negative three and positive three in their mathemati-
cal sentences. Students had more difficulty choosing a correct operation to
complete their symbolic statement for (¯9 - ¯3). Six students created correct
statements. These students were able to use the words of the story problem
to articulate their reasoning for the chosen operation. For example, one VIC
student explained her use of subtraction: “it says ‘erases’ and, so, that’s like
106 Bolyard and Moyer-Packenham

taking away.” The student made connections between the idea of erasing a
$3 debt and the symbolic notation − ¯3. Students creating ¯9 − ¯3 repre-
sented their interpretation of the story with the use of two minus signs in
their symbolic statement: the first as a binary sign indicating “take away”
and the second as a unary sign attached to the numeral 3. Students who pro-
duced the addition statement ¯9 + 3 to represent the problem also used the
context of the words to make connections to their choice of operation. One
VIC student describing his reasoning for using addition explained that if
someone takes away a debt, “it’s positive.” A VICC student who wrote ¯9 +
3 was able to see that either an addition or a subtraction statement was accu-
rate after rereading the problem concluding, “I guess I could also have it be
¯9 − ¯3.” The symbolic statement produced by these six students related to
their picture interpretation. In other words, students who wrote an addition
statement drew a picture indicating addition; those who wrote a subtraction
statement drew a picture that illustrated taking away or subtracting values.
The remaining three students (one in each group) created an incorrect state-
ment (¯9 − 3) and did not appear to make connections between their work in
pictures and their work in symbols. Each of these students created a picture
representing the addition problem, ¯9 + 3, interpreting the situation using a
binary function of adding the opposite.
The final subtraction item presented the students with the statement ¯4
− 5. By far, this item gave students the greatest difficulty with fewer than
half of the students able to respond correctly using written/metaphorical,
pictorial, or symbolic forms. Four students (one VICC, one VIC, and two
VNL) correctly interpreted the symbolic statement as ¯9 and completed the
remaining two tasks correctly. The other five students were not able to ac-
curately interpret the given problem statement. One VIC student initially
crossed out the negative on the first addend and changed subtraction sign
to an addition sign, resulting in 4 + 5. As the student created his picture and
story representations, he expressed uncertainty in his initial answer: “It’s
one or nine. I know it’s like one or nine, but I’m thinking that it’s probably
nine . . . .” The remaining four students misinterpreted the statement given
(¯4 − 5) as ¯4 + 5, reported an answer of one, and then continued to create
pictures and stories based on this interpretation of the problem. While some
students did not question their initial interpretation, it appeared that others
did continue to reason through the problem and attempt to make connec-
tions as they worked with the other representations. For example, one VIC
student who had created a pictorial representations for ¯4 + 5, began to cre-
ate a correct story for the original problem, ¯4 − 5. However, upon realiz-
ing that her scenario would not result in an answer of positive one, she then
Making Sense of Integer Arithmetic 107

changed her story to match the sentence ¯4 + 5. When in doubt, students

forced the other representations to support the symbolic statement.

Discussion

This research study investigated the impact of using virtual manipula-

tives to teach integer addition and subtraction on students’ computation
skills and use of representations. The four major findings of this study pre-
sented below will be discussed in detail in the sections that follow.
Finding 1: Students demonstrated significant gains in computation
achievement after using three different virtual manipulative applets dur-
ing instruction of integer addition and integer subtraction concepts.
Finding 2: There were no statistically significant differences in achieve-
ment among students using three different virtual manipulative applets
designed for integer instruction.
Finding 3: Students successfully translated between pictorial and
written/metaphorical representational forms for integer addition and
subtraction.
Finding 4: Students demonstrated less facility in creating an accurate
symbolic representation and connecting this with other representations
(i.e., pictorial and written/metaphorical), particularly for subtraction.

Discussion of Student Achievement

Our first two findings showed significant gains for all treatment groups
after using three different virtual manipulatives, and no significant differ-
ences among the three virtual manipulatives applets. Cohen’s d values cal-
culated on the integer pretest-to-posttest mean differences for all three treat-
ment groups met the standard of a medium effect size for addition items
(smallest d = .63) and the standard of a large effect size for subtraction
items (smallest d = 1.36). The fact that students made larger gains in sub-
traction than addition is not surprising. Analysis of the pretest scores indi-
cated that students had some prior knowledge of integer addition but less
prior knowledge of integer subtraction.
Analysis of achievement by treatment group for the three virtual manip-
ulatives applets indicated no significant differences. While these three vir-
tual manipulatives shared several key features (dynamic linked representa-
tions, interactivity, multiple representations, and immediate feedback), there
108 Bolyard and Moyer-Packenham

were some features (type of user input required, degree of guidance provid-
ed by the applet, and problem presentation) that were unique to each virtual
manipulative. One conclusion that might be drawn from this is that the fea-
tures shared by these three applets had a larger impact on students’ learning
of integer computation than those that were unique to any one specific tool.

Discussion of Students’ Facility with Different Representational Forms

Although the results of the pre- and post-testing in each of the three dif-
ferent virtual manipulatives integer groups demonstrate that the different ap-
plets had a positive influence on students’ abilities to perform integer com-
putation, we believe that a more important result of this teaching experiment
is the differences in the ways students translated among different represen-
tations for integer addition and subtraction problems. Specifically, students
had much more difficulty creating representations and making connections
when problems were presented in symbols or when a symbolic representa-
tion was requested, particularly for subtraction.
Students’ work on the interview tasks showed they were making con-
nections between written/metaphorical and pictorial representations of in-
teger values. For example, students expressed thinking that demonstrated
connections between phrases such as “debt” and “walking left” to negative
integer values and “assets” and “walking right” to positive integer values.
Students also used appropriate images (positive and negative tiles or left
and right arrows) to represent words indicating positive and negative integer
values in their pictorial models. There was also evidence that students were
able to make personal connections to the metaphors based on the frequency
with which students created personalized stories. English (1997) describes
such analogies and metaphors as “illuminating devices” which help learners
take concrete experiences and build them into mental models for abstract
ideas.
In contrast to their facility with written and pictorial representational
forms, students in this study had noticeably more difficulty working with
the symbolic form. On addition tasks, students showed evidence of making
connections among the words in the story and the images in the pictures to
the values of the integers used in their mathematical sentences. However,
they were not confident in describing the connection between the problem
situation and their use of the addition operation in their symbolic sentences.
Students’ difficulties with the symbolic form were evident across all
three subtraction interview items. Several students successfully translated
Making Sense of Integer Arithmetic 109

the subtraction items presented in pictures and words into related addition
situations as they worked in these forms (i.e., they equated removing five
positive tiles from the picture with adding five negative tiles to the picture).
However, students made errors on their symbolic statements for these items
because they failed to translate their understanding of this relationship (sub-
tracting a number is equivalent to adding the opposite) to the appropriate
symbols of the mathematical sentence. They appeared to treat the subtrac-
tion sign as a placeholder and did not fully understand its role and purpose
as binary sign, indicating algebraic subtraction (Vlassis, 2008). Interest-
ingly, once students misinterpreted the original symbolic problem, they
proceeded to create pictures and stories to match this interpretation. Rarely
were students able to overcome their initial error.
Many mathematics educators argue that students over-learn “take
away” as an interpretation of subtraction in whole number arithmetic (Mo-
ses, Kamii, Swap, & Howard, 1989). This interpretation does not always
translate to the modeling and counting up procedures many children natu-
rally use to solve these problems (Baroody, 1984; Fuson, 1984). When stu-
dents expand into arithmetic situations with negative integers, the complex-
ity increases and “take away” does not adequately model subtraction with
positive and negative integers. Thus, a more flexible interpretation of sub-
traction (i.e., one that includes comparison, difference, and other contexts)
that allows for both quantity and direction features of integers to be made
explicit is needed (Moses, Kamii, Swap, & Howard, 1989).
Confusion over the subtraction operator in integer arithmetic is com-
mon. Mathematics educators note the multiple and varied uses of the mi-
nus sign as one of the main obstacles students face in working with inte-
gers (Hativa & Cohen, 1995; Vlassis, 2008). Peled et al. (1989) found that
students often disregard or (inappropriately) relocate the minus signs in
problems in order to better accommodate their understanding. Students in
this study demonstrated these errors. Ashlock (1994) points out that instruc-
tional experiences must “relate the model to a number sentence, numbers
to their numerals, and operations to the sign for the operation.” (p. 246).
It appears that while the students’ work with the virtual manipulatives was
successful in relating models to number values in number sentences, it was
not completely successful in relating the model to the operation used in the
number sentences.
The structure of the virtual manipulative applets used in this study be-
gan with the symbolic statement and then created pictorial models from that
statement. Therefore, students were only required to attend to the values of
the integers in the statement (to produce the correct number of tiles or to
110 Bolyard and Moyer-Packenham

move the slider in the correct direction). The applet provided the necessary
structure to ensure that the pictorial model complied with the specific op-
eration symbol in the statement. As a result, when creating a corresponding
symbolic sentence, some students appeared to disregard the role of the “−”
sign as an operation indicator and ignored its effects on the value of the sub-
trahend. These students would have benefited from tasks targeted on making
that distinction explicit.

Limitations

Several limitations to the present study should be noted. First, the par-
ticipants in this study were enrolled in what was labeled as an above-grade
level course in the districts’ curriculum. Therefore, the results may not be
representative and generalizeable to all sixth-grade students. Second, the
time of instruction could have been longer to allow students adequate op-
portunities to explore and practice the concepts of integer addition and sub-
traction. Third, because the researchers were not the instructors, there may
have been inconsistencies in the students’ instructional experiences due to
individual teacher biases and practices. Finally, most of the students had
never used any virtual manipulative tool prior to this study. Therefore, the
novelty of these tools might have contributed to the positive results in com-
putation.

Conclusions and Implications for Further Research

Although students in this study made positive gains in achievement on

integer addition and subtraction computation, the students had not yet devel-
oped a complete understanding of integer addition and subtraction concepts.
Similar to other research (e.g., Peled & Carraher, 2007), after working with
the virtual manipulatives, students demonstrated a degree of flexibility in
working with and among some representations for integers. However, they
still showed some difficulties in interpreting and representing situations in
symbolic form, particularly for subtraction items. Students need targeted ex-
periences to help them build connections among symbolic and other repre-
sentational forms of integer subtraction (Kilpartick, et al. 2001). Designers
of future virtual manipulatives for integer instruction should consider add-
ing features to the applet that would focus the student’s attention on the pur-
pose of the operation sign allowing them to make distinctions between, for
Making Sense of Integer Arithmetic 111

example, the dual roles of the minus sign through their work with the virtual
manipulatives.
Building on these results, further study should compare virtual manipu-
latives representing other models for integers to determine their effects on
learning. For example, studies could examine the effectiveness of virtual
manipulatives created to represent Janvier’s hybrid model (Janvier, 1983) or
an integer model based on algebraic geometry (Carson & Day, 1995). Fi-
nally, further research should examine the effect of virtual manipulatives de-
signed specifically to help students make connections among interpretations
of integer addition and subtraction situations and how these are represented
in symbols. The process of working in different representational forms and
reconciling inconsistencies among them can facilitate the development of
a more complete and flexible understanding of integers and integer opera-
tions. The world of virtual manipulatives offers an excellent forum in which
to design and study effective approaches to providing such experiences.

References

Ashlock, R. B.(1994). Error patterns in computation: Using error patterns to

improve instruction (8th ed.). Upper Saddle River, NJ: Merrill Prentice
Hall.
Baroody, A. J. (1984). Children’s difficulties in subtraction: Some causes and
questions. Journal for Research in Mathematics Education, 15, 203-213.
Bolyard, J. J. (2006). A comparison of the impact of two virtual manipulatives
on student achievement and conceptual understanding of integer addition
and subtraction. (Doctoral dissertation, George Mason University, 2006).
Dissertation Abstracts International, 66 (11), 3960A.
Bruno, A., Espinel, M. C., & Martinon, A. (1997). Prospective teachers solve
additive problems with negative numbers. Focus on Learning Problems in
Mathematics, 19(4), 36-55.
Carson, C. L., & Day, J. (1995). Annual report on promising practices: How the
algebra project eliminates the “game of signs” with negative numbers. San
Francisco: Far West Lab for Educational Research and Development. (Eric
Document Reproduction Service No. ED 394 828)
English, L. D. (Ed.). (1997). Analgoies, metaphors, and images: Vehicles for
mathematical reasoning. In L. D. English (Ed.), Mathematical reasoning:
Analogies, metaphors, and images. Mahwah, NJ: Lawrence Erlbaum Asso-
ciates.
Fuson, K. (1984). More complexities in subtraction. Journal for Research in
Mathematics Education, 15, 214-225.
Glesne, C. (1999). Becoming qualitative researchers: An introduction (2nd ed).
New York: Addison Wesley Longman.
112 Bolyard and Moyer-Packenham

Goldin, G. A. (1997). Observing mathematical problem solving through task-

based interview. Journal for Research in Mathematics Education Mono-
graphs, 9, 40-62.
Hativa, N. & Cohen, D. (1995). Self learning of negative number concepts by
lower division elementary students through solving computer-provided nu-
merical problems. Educational Studies in Mathematics, 28, 401-431.
Hsieh, H., & Shannon, S.E. (2005). Three approaches to qualitative con-
tent analysis. Qualitative Health Research, 15(9),1277-1288. doi:
10.1177/1049732305276687
Janvier, C. (1985). Comparison of models aimed at teaching signed integers. In
L. Streefland (Ed.), Proceedings of the 9h International Conference for the
Psychology of Mathematics Education (vol. 1, pp. 135-140). International
Group for the Psychology of Mathematics Education. (ERIC Document Re-
production Service No. ED 411 130)
Kilpatrick, J., Swafford, J., & Findell, B. (2001). Adding it up. Helping children
learn mathematics. Washington, DC: National Academy Press.
Lesh, R., Post, T., & Behr, M. (1987). Representations and translations among
representations in mathematics learning and problem solving. In C. Janvier
(Ed.), Problems of representation in the teaching and learning of mathemat-
ics (33-40). Hillsdale, NJ: Erlbaum.
Liebeck, P. (1990). Scores and forfeits: An intuitive model for integer arithmetic.
Educational Studies in Mathematics, 21, 221-239.
Maxwell, J. A. (1996). Qualitative research design: An interactive approach.
Thousand Oaks, CA: Sage.
Moreno, R. & Mayer, R. E. (1999). Multimedia-supported metaphors for mean-
ing making in mathematics. Cognition and Instruction, 17, 215-248.
Moses, R. P., Kamii, M., Swap, S. M., & Howard, J. (1989). The Algebra Proj-
ect: Organizing in the spirit of Ella. Harvard Educational Review, 59, 423-
443.
Moyer-Packenham, P. S., Westenskow, A., & Salkind, G. (April, 2012). A review
of the research on virtual manipulatives in mathematics instruction. Paper
accepted for presentation at the annual meeting of the America Educational
Research Association, Vancouver.
Mukhopadhyay, S., Resnick, L. D., & Schauble, L. (1990). Social sense-mak-
ing in mathematics; Children’s ideas of negative numbers. In G. Booker, P.
Cobb, T. N. deMendicuti (Eds.), Proceedings of the 14th International Con-
ference for the Psychology of Mathematics Education (vol. 3, pp. 281-288).
International Group for the Psychology of Mathematics Education. (ERIC
Document Reproduction Service No. ED 411 139)
National Council of Teachers of Mathematics. (2000). Principles and standards
for school mathematics. Reston, VA: Author.
Peled, I. (1991). Levels of knowledge about signed numbers: Effect of age and
ability. In F. Furinghetti (Ed.), Proceedings of the 15th International Confer-
ence for the Psychology of Mathematics Education (vol. 3, pp. 145-152).
International Group for the Psychology of Mathematics Education. (ERIC
Document Reproduction Service No. ED 413164)
Making Sense of Integer Arithmetic 113

Peled, I. & Carraher, D.W. (2007). Signed numbers and algebraic thinking. In
J. Kaput. D. Carraher, & M. Blanton (Eds.), Algebra in the Early Grades.
Mahwah, NJ, Erlbaum, pp. 303-327 (now Taylor & Francis).
Peled, I., Mukhopadhyay, S., & Resnick, L. B. (1989). Formal and informal
sources of mental models for negative numbers. In G. Vergnaud, J. Rogal-
ski, & M. Artique (Eds.), Proceedings of the 13h International Conference
for the Psychology of Mathematics Education (vol. 3, pp. 106-110). Interna-
tional Group for the Psychology of Mathematics Education. (ERIC Docu-
ment Reproduction Service No. 411 142)
Sherzer, L. (1973). Effects of different methods of integer addition instruction
on elementary school students of different grade and aptitude levels. (Doc-
toral dissertation, University of Miami, 1973). Dissertation Abstracts Inter-
national, 34(05), 2465A. (UMI No. 7325919)
Smith, J. P. (1995). The effects of a computer microworld on middle school stu-
dents’ use and understanding of integers. Dissertation Abstracts Interna-
tional, 56(09), 3492A.
Smith, L. (2006). The impact of virtual and concrete manipulatives on algebraic
understanding. Unpublished doctoral dissertation, George Mason Univer-
sity, Fairfax, Virginia.
Vershcaffel, L., Greer, B., & De Corte, E. (2007) Whole number concepts and
operations. In F. K. Lester, Jr. (Ed.), Second Handbook of Research on
Mathematics Teaching and Learning. Charlotte, NC: Information Age.
Vlassis, J. (2008). The role of mathematical symbols in the development of
number conceptualization: The case of the minus sign. Philosophical Psy-
chology, 21, 555-570.