Professional Documents
Culture Documents
Roberto Araya(*), Patricio Calfucura, Abelino Jiménez, Carlos Aguirre, María Angélica
Universidad de Chile
1
ANALOGIES IN EQUATIONS
Abstract
A total of 236 seventh grade students who had never been taught algebraic equations
divided into two groups at each school. The students in one group watched a 15-minute
video teaching them how to solve 5 different first degree linear equations using a
traditional symbolic strategy, while in the other group the students watched a 15-minute
video teaching them how to solve the same equations using four analogies for solving
an equation: a two-pan balance for the equals sign, a box for a variable, candies for
numbers, and guessing the number of candies inside a box The students were then tested
on 12 equation solving problems, all of them written using only symbolic notation. The
group that watched the analogies video performed significantly better. Students with a
below average mathematics GPA who watched the analogies video did as well as
students with an above average GPA who watched the symbolic strategy video.
Students who watched the analogies video also reached a better conceptual
2
ANALOGIES IN EQUATIONS
Understanding and solving first degree algebraic equations is one of the main goals
of middle school algebra. However, a large number of students do not understand them
and consistently make procedural and conceptual mistakes. For example, the Task
Education (DOE), 2008) concluded in their Algebra section that “Too many students in
high school algebra classes are woefully unprepared for learning even the basics of
algebra. The types of errors these students make when attempting to solve algebraic
equations reveal they do not have a firm understanding of many basic principles of
arithmetic (e.g., commutativity, distributivity), and many do not even understand the
concept of equality. Many students have difficulty grasping the syntax or structure of
(e.g., adding or subtracting the same value from both sides of the equation) or why
Some of the key misconceptions are: the notion of equality (McNeil et Al., 2006;
Knuth & Stephens, 2006; Booth & Koedinger, 2008); difficulty in applying the
commutative and distributive properties (Siegler, 1998); failure to correctly order the
operations, to correctly add and subtract numbers and do divisions on both sides of the
equation into another (DOE, 2008); misunderstanding the notion of solution when there
is no solution or when there are an infinite number of solutions (Vollrath, 1980); and
algebra (Koedinger, Alibali, & Natham, 2008; Hass, 2005; Austin & Vollrath, 1989;
3
ANALOGIES IN EQUATIONS
Holyoak & Stigler, 2004). A recent comparative study (Richland, Zur & Holyoak,
2007) between Japan, Hong Kong and the United States of teachers’ use of analogies in
10 eighth grade lessons sampled from the TIMSS 1999 video database found a
significantly higher frequency in Japan and Honk Kong than in the United States of
teachers’ use of analogies and strategies such as visual presentation of the source of the
analogy, keeping sources visible during comparisons, using spatial cues to highlight the
association between the corresponding elements of the source and target, using hand or
arm gestures to indicate an intended comparison, and using mental imagery. Several
other studies show the positive effect of concrete representations on students’ learning
of algebra and algebraic equations (Hass, 2005; DOE, 2008), and particularly the use of
the balance analogy for solving first degree algebraic equations (Skaggs, 2007).
However, there is still the need to measure the total impact on students’ performance
when using strategies that employ analogies for algebraic equation solving.
We pursued four main goals in this study. The first was to examine the effect of the
two-pan balance and box analogies on algebraic equation solving. Among students not
yet exposed to algebra we wanted to test whether students taught to solve equations
using a two-pan balance as an analogy for the equals sign, candies to represent a
number, boxes as an analogy for variables, and the process of guessing the number of
analogy for the process of finding the solution to an equation, would perform better than
The two-pan balance analogy for equation solving is well known; in most textbooks
the variable is represented as the unknown weight of an object; in some cases the
4
ANALOGIES IN EQUATIONS
balance is a stationary plastic balance scale and the unknown is represented by pawns
analogy, closer to the container analogy commonly used in computer science, and at the
same time as a physical object that can be grasped and manipulated. In this study, the
identical boxes, each containing the same number of candies. The container analogy
used here is similar to that proposed by Austin and Vollrath, but the weight of the empty
box is zero. This way it is easier to translate solutions from the two-pan balance to the
symbolic equation and vice-versa. We hypothesized (Araya, 2000) that this analogy
facilitates the manipulation of expressions where the addition of similar terms (3x + 2x)
is mapped to the addition of 3 boxes and 2 boxes, all of which have the same content.
This map is similar to that between numbers and a collection of objects, where the
hypothesized that this more complete representation involves a lower cognitive load
and the storage of partial results are mapped to simple physical manipulations and
positions of objects. Therefore, the physical world automatically computes what would
facilitate computations even in the case of mental calculations, since the mental imagery
of the physical movements of objects probably requires a lower cognitive load than the
mental calculation and manipulation of symbols. Thus, we wanted to test the impact of
analogy on the more cognitively demanding situation where the solution had to be
5
ANALOGIES IN EQUATIONS
Some of the challenges of measuring the effect of the proposed analogy were
controlling important factors such as the teacher, since his or her experience in using
manipulatives has an important impact on the students’ learning (Moyer, 2001). Other
critical factors are the students’ ability and knowledge of mathematics, as well as their
socioeconomic status.
In order to control these factors, in each class at the 10 schools the students were
randomly split into two groups. In each class, one group of students watched a 15-
minute video teaching them how to solve 5 different first degree linear equations using
a traditional symbolic strategy, whereas the other group watched a 15-minute video
teaching them how to solve the same 5 equations using the four abovementioned
analogies. The teachers were not present during the training stage nor the testing, since
all training was done through the videos and the testing through a computer system. A
main challenge, however, is to study the long-term effect of the use of analogies in
algebra. This would require tracking the students’ activities for several months and
maintaining strict control over the use of analogies for teaching one group and symbolic
training for the other, over several weeks and months and at various schools. This is
very difficult to achieve because in each class there are students that belong to both
groups, and teachers and school administrators could not guarantee that these students
would not share their acquired knowledge. Given the difficulties of such a long term
study, we designed a short term study restricted to a total of 3 hours, with all the
teaching done through a 15-minute video. The video was presented after 30 minutes’
training in the use of the computer software that was later employed in the pre- and
post-tests in order to test and record the students’ answers and response times.
The second goal was to study the improvement impact on the post-tests for students
with a below average mathematics GPA. Given that there is no standardized index of
6
ANALOGIES IN EQUATIONS
student performance on first degree algebraic equations after only an initial, short
performance of students with a below average mathematics GPA that watched the
analogies video with that of students with an above average mathematics GPA that
watched the traditional symbolic video. We presumed that the eventual effect could be
measured as a reduction in the gap between these two groups when compared to both
groups learning with symbolic instruction. This means that, if the analogies video
improves the performance of students with a below average mathematics GPA, then the
difference between the performance of these students and that of students with an above
average mathematics GPA that watch the symbolic video should be a fraction of the
GPA and students with a below average mathematics GPA that watch the symbolic
video.
The third major goal of the present study was to analyze the effect of the four
performed in a wide range of problem solving situations: (i) solving simple, previously
unseen algebraic equations with the variable on a different side of the equation to the
7
ANALOGIES IN EQUATIONS
examples used in training, (ii) with variables on both sides of the equal sign, (iii) using a
different letter to represent the variable, (iv) completing the steps followed to solve an
equation, (v) solving equations with no solution, (vi) solving equations with an infinite
number of solutions, and (vii) reasoning tasks where the students had to detect and
middle and high school, and are therefore rarely studied. Equations with an infinite
equations (Vollrath, 1980) and are instead considered algebraic laws. The error
detection problems where any number is the solution are particularly difficult because,
after a sequence of manipulations, the expression 0 = 0 is obtained and the student must
understand that from 0 = 0 it does not follow that the only solution is x = 0. It is a
solution, but not the only solution. We also wanted to examine whether the analogies
video improves the students’ ability to generalize for new kinds of equations. In
particular, we wanted to measure the students’ performance when they had to generalize
the procedures for an equation where the variables are located in previously unseen
positions. In order to do this we designed the training videos with two types of
equations: with variables only on the left-hand side of the equal sign, such as 2x + 1=5,
and with variables on both sides of the equal sign, such as 2x + 1 = 5 + x. We wanted to
measure the decline in performance when students were tested with equations similar to
the ones used on the training video but with variables only on the right-hand side of the
equal sign, such as 15 = 3 + 4x. Were the students who watched the analogies video less
8
ANALOGIES IN EQUATIONS
A fourth goal was to examine the emotional impact of learning algebraic equation
solving with the two-pan balance and box analogy. It is well known that mathematics,
particularly algebra, induces anxiety and stress in some students (DOE, 2008). We
wanted to test whether the proposed analogy helps to control the typical increase in a
Methods
Participants
The participants were 236 seventh grade students (45.8% female), ranging in age
mathematics GPAs ranging from 2.6 to 7 (the higher the better, 1 being the minimum
and 7 the maximum, with 4 as the passing grade) (M = 5.085, SD = 0.8551). The
participants were randomly selected from 10 schools in different regions of Chile. They
fee-paying private school of high socioeconomic status. At each school, the participants
in each class were randomly assigned to one of two groups: the analogies training group
and the symbolic training group. The analogies training group included 118 children
0.8570); the symbolic training group included 118 children (age: M = 13 years 2.27
age: 0.898, mathematics GPA: 0.661]. At all of the schools, the students had never been
taught algebraic equations. The experimenters comprised one female and two male
research assistants. One experimenter trained the students in the online computer system
used in the testing. Previous to the computer training and following each of the two
9
ANALOGIES IN EQUATIONS
post-tests, the other experimenters distributed sheets on which the students completed
[Figure 1]
In each school, the participants were first trained to answer basic mathematics
were related to equations or algebra. The purpose of this was to ensure that all students
were capable of using the system to answer different items. Following the training, the
children were randomly assigned to one of the two groups. For each group, the children
were immediately shown the corresponding video (Figure 1). The children then
answered the first set of 6 equation solving problems on the computer system, and, after
finishing, took a break of approximately 45 minutes. During this break, the other group
watched their corresponding video and immediately answered the same six equation
solving problems. After finishing, they took their break of approximately 45 minutes.
Answering the 6 problems took less than 15 minutes. After their first break, each group
of children then answered 6 further equation solving problems. All 12 problems were
the same for both conditions and were written using symbolic notation only. The
questions were presented in 4 different orders and randomly assigned to the participants.
All calculations were performed mentally; during the post-tests, paper and pencils were
not allowed in order to ensure that all calculations were done mentally, as well as
minimizing any possible cheating. All of the answers were entered into the system,
which measured the response time for each item. The students knew that their
performance would only be used for this research and would not affect their future
GPA. Before watching the first video, an affective assessment of the students was
10
ANALOGIES IN EQUATIONS
carried out using the Self-Assessment Manikin (SAM) (Bradley & Lang, 2007). After
solving the first 6 algebraic equation problems, another SAM affective assessment was
applied, with a third assessment following the final six algebraic equation problems
(Figure 2).
[Figure 2]
The symbolic training group. Students watched a 15 minute video, as shown in Table 1
and Figure 3.
[ Figure 3 ]
[Table 1]
The analogies training group. The students watched a 14 minute video as shown in
After watching the video, the children were shown 6 equation solving problems on a
with another set of 6 equation solving problems. The system displayed the items, one
screen for each problem, in one of 4 different sequences. For every problem, it
registered whether or not the answer was correct, the time at which the screen started to
display each problem, and the time at which the student sent the corresponding answer.
Once an answer had been sent, the system stopped displaying the problem and this
could no longer be seen by the student. The system then gave feedback as to whether or
not the answer was correct before the next problem appeared on the screen. The
[Table 2]
Results
11
ANALOGIES IN EQUATIONS
Analyses
and age made no difference [gender: F(1,234)=0.000, p-value: 0.994 bilateral, age:
students with a below average mathematics GPA their post-test performance was M=
22.58, SD=14.81, while for students with an above average mathematics GPA their
bilateral. There was a 7.36 post-test performance gain per mathematics GPA (beta =
2, and on the post-tests in general was statistically higher than in the symbolic training
[Table 3]
The response time was significantly higher in the analogies training group on post-
[ Figure 4 ]
[Table 4]
The post-test performance by children with below average mathematics GPAs and in
the analogies group was (M=25.84, SD=17.25) significantly better than children in the
bilateral). Also, the post-test performance by children with above average mathematics
GPAs and analogies training was significantly better (M=35.78, SD=19.84) than
value=0.007 bilateral).
12
ANALOGIES IN EQUATIONS
mathematics GPAs in the analogies training group with that by students with above
average mathematics GPAs in the symbolic training group. It turned out that there was
shown in Figure 5. This is a particularly impressive result, since it shows that with the
help of analogies, students with below average mathematics GPAs perform similarly to
students with above average mathematics GPAs who learn with traditional symbolic
training.
[ Figure 5 ]
We also divided the children into three categories according to their mathematics
GPA. The first third had the lowest GPAs, while the second and third thirds had the
higher GPAs. As shown in Figure 5, students in both the middle third and the higher
statistically significant difference was found for the third with the lowest mathematics
GPAs.
where the children had to detect and locate an eventual error in the steps to solving an
algebraic equation that were shown in several lines on the screen and where a “no error”
choice was available. The post-test performance by children in the analogies training
group was statistically higher than in the symbolic training group (in the analogies
bilateral).
13
ANALOGIES IN EQUATIONS
training group was statistically higher than in the symbolic training group (in the
was obtained by comparing the performance on two very similar questions. One had the
variable on the left-hand side of the equal sign (2x + 1= 9), similar to the questions on
the training videos, and the other had the variable on the right-hand side of the equal
sign (15 = 3 + 4x). As shown in Figure 6, the decline in performance was not
significant for students in the symbolic training group (M=39.91, SD=44.71, M=22.12,
[ Figure 6 ]
We compared the performance by the 40 students who chose not to fill out the SAM
manikin, or selected ‘one’ on the subjective stress arousal (one being the highest on the
scale) before starting the experiment. Only 25% of these students had an above average
performance on the post-tests, whereas among the rest of the students 48% had an above
p=0.008 bilateral).
the end of post-test 1 and post-test 2, compared with the initial measurement taken just
before the experiment. On post-test 1, 24% percent of the children increased their stress
arousal in the analogies training group whereas 42% percent of the children increased
their stress arousal in the symbolic training group. The difference was statistically
14
ANALOGIES IN EQUATIONS
children increased their stress arousal in the analogies training group whereas 33% of
the children increased their stress arousal in the symbolic training group. The difference
Multivariate Analyses
performance. Mathematics GPA had a greater effect, with standardized beta = 0.36 for
standardized beta = 0.216 for group (t(233) = 3.70 , p-value =0.000 bilateral). Also, as
shown in Figure 7, in the analogies training group, mathematics GPA had a greater
impact (9.57 performance gain per GPA, R² = 0.184) on post-test performance than in
the symbolic training group (4.88 performance gain per GPA, R² = 0.089). This means
that the higher the student’s mathematics GPA, the greater the effect of the analogies.
[ Figure 7 ]
Discussion
The four main goals pursued in this study were achieved. The main finding can be
summarized quite simply. The use of analogies in teaching algebraic equation solving
has an important impact on student learning. Moreover, we have been able to quantify
such an effect. In the realm of algebra, when the domain is completely new to the
students, after just 15 minutes of passive exposure students with below average
mathematics GPAs taught using analogies perform similarly to students with above
average mathematics GPAs taught using the traditional symbolic strategy. Thus, our
second goal was achieved with very impressive results. Given the enormous difficulty
15
ANALOGIES IN EQUATIONS
analogies are a powerful way to improve student performance. The third goal was to
examine the impact on conceptual understanding. Here, the results are very
achieved. Contrary to research on younger students which shows that the effective use
student knows the manipulatives well enough to use them automatically, in this study
the effects of analogies were achieved with students that had never used a two-pan
balance or boxes for learning mathematics; all they did was watch a short video with 5
worked examples. Finally, the fourth goal was to examine the affective impact. Even
though we did not see a positive impact on pleasure, there was a positive impact on
These findings raise the question of why analogies can produce such profound
equilibrium and the procedures of adding and subtracting the same amount of candies or
boxes on both sides of the two-pan balance are part of our biologically primary
cognition (Geary, 2007). This is probably folk physics knowledge. The use of analogies
establishes a mapping between such biologically primary knowledge and the abstract
mathematical concepts of algebraic equation solving. This mapping is not very difficult
to understand and use. Even though students in the analogies training group initially
took more time to answer the problems, by post-test 2 they had already achieved the
same response time as the students in the symbolic training group. We hypothesize that
the initially longer response time could be due to the need to translate the equation
The results of this study are very encouraging. However, it is still necessary to test
whether the analogy effect will persist and eventually increase with further exposure.
16
ANALOGIES IN EQUATIONS
This would require measuring the effect when teaching is extended from 15 minutes of
is necessary to measure the effect of analogies not only through passive video watching,
but when students are exposed to interactive teaching and active learning using
analogies. It would also be of great interest to measure the effect on other conceptual
saw no effect of the analogies on students with very poor mathematics grades. It is
important to further examine why we got this negative result. In order to detect a
Even though there was a greater effect on students with above average mathematics
performance students in other, more meaningful ways, for example, to determine how
much the teaching time can be reduced to achieve a given level of performance when
discover the following: different strategies that students use when learning through
analogies; the dynamic of the transition between strategies (Siegler, 1996); the inertia of
previously used strategies (Siegler & Araya, 2005); the impact of the type of problems
used by the teacher (direct solving, no solution, any solution, error detection, etc.); the
impact of the specific order in which the problems are presented; the impact of
immediate feedback to the student on the size of the analogies effect; the impact of the
signaling and gestures to explicitly show the mapping between source and object
domains of the analogy, as well as the mapping of operations and procedures (Alibali &
Natham, 2009); and how students translate back and forth between symbolic and
analogical representations.
17
ANALOGIES IN EQUATIONS
help children learn through analogies. They should teach abstract concepts using
appropriate analogies and ask students to actively use them. In algebraic equation
solving, the two-pan balance equilibrium as the equality between both sides of an
equation, candies as numbers, boxes as variables, and guessing a box’s contents as the
process for finding the solution to an algebraic equation, are four simple, yet powerful
analogies that the teacher can use to radically improve comprehension of the concepts
At present, use of the two-pan balance in textbooks is very limited. At most, a quick
analogy is included on one or two pages, but they do not make systematic use of it. A
them at the beginning before helping the students make the transition to symbolic
concepts and procedures. In this study, the effect of analogies was achieved simply by
having the students watch a video. There was no use of manipulatives or concrete
objects. There is therefore no need to have teachers with experience in using concrete
materials, nor to allocate a budget for physical items or spend class time preparing and
organizing physical objects. Hence, this teaching strategy is very simple and quick to
Our results indicate that improvements in subjective stress arousal can be fostered by
18
ANALOGIES IN EQUATIONS
References
Goldman, R. Pea, B. Barron, & S. J. Derry (Eds.). Video Research in the Learning
Austin J. D. & Vollrath H-J. (1989). Representing, Solving, and Using Algebraic
Equations. The Mathematics Teacher 82, pp. 608-612 . Retrieved, January 15,
2010. http://www.didaktik.mathematik.uni-wuerzburg.de/history/vollrath/papers
University Press.
30th Annual Cognitive Science Society (pp. 571-576). Austin, TX: Cognitive
Science Society.
(IAPS) in the Study of Emotion and Attention. In J.A. Coan & J.J.B. Allen (Eds.),
University Press.
Department of Education (2008). Foundations for success. The final Report of the
19
ANALOGIES IN EQUATIONS
24 - 46
Kieran, C. (1985). The Equation Solving Errors of Novice and Intermediate Algebra
Knuth, E. J., Stephens, A.C., McNeil, N. M., & Alibali, M. W. (2006). Does
Understanding the Equal Sign Matter? Evidence from Solving Equations. Journal
Koedinger, K., Alibali, M., & Natham, M. (2008). Trade-offs Between Grounded and
McNeil, N., Grandau, E., Knuth, E., Alibali, M., Stephens, A., Hattikudur, S., & Krill,
They Read Can´t Help. Cognition and Instruction, 24(3), 367 – 385.
Moyer, P. (2001). Are We Having Fun Yet? How Teachers Use Manipulatives to
Richland, L. E., Zur, O., & Holyoak, K. J. (2007). Cognitive supports for analogies in
20
ANALOGIES IN EQUATIONS
91 (1), 175-189.
Siegler, R. (1998). Children's thinking. Upper Saddle River, NJ: Prentice Hall.
33, 1-44.
http://www.borenson.com/Portals/25/MSKAGGSDISSERTATION
%20Complete.pdf
Vollrath H-J. (1980). A case study in the development of algebra teaching in the
21
ANALOGIES IN EQUATIONS
http://www.didaktik.mathematik.uni-wuerzburg.de/history/vollrath/papers
22
ANALOGIES IN EQUATIONS
Figure Captions
Figure 3. Screenshots of the symbolic training video (left) and the analogies training
video (right)
Figure 4. Response time in the symbolic training and analogies training conditions
23
ANALOGIES IN EQUATIONS
conditions
24
ANALOGIES IN EQUATIONS
25
ANALOGIES IN EQUATIONS
Procedure: Procedure,
1) grasp and lift one candy on
1) subtract 1 from both the left and simultaneously
sides 2x + 1 – 1 = 5 –1 grasp and lift one candy on
2) this is 2x = 4 the right-hand side.
3) divide by 2 on both 2) Now divide the content of the
sides 2x / 2 = 4/2 4 candies on the right-hand
4) this is x = 2 side into two groups.
3) Therefore each box has to
have 2 candies.
4) This is mathematically
written as x = 2.
26
ANALOGIES IN EQUATIONS
15 = 3 + 4x
x = ___
Solve equation with Find the value of x that
variables on both sides of solves the equation:
the equal sign 8x + 20 + 4x = 40 + 2x
(2 items on post-test 1) x = ___
27
ANALOGIES IN EQUATIONS
solutions in post-test 1 and expressions. Find out if Find out if there is an error
in post-test 2 there is an error and if you and if you find one choose
find one choose the line the line with the error. If
with the error. If there is no there is no error choose the
error choose the “No error” “No error” label.
label. Laura’s sequence
George’s sequence 1. subtracting 8
1. subtracting 2 from from both sides
both sides we get we get 10x + 3 =
10x + 10 = 5x + 10 2x + 2x + 27
+ 5x 2. subtracting 3
2. subtracting 5x from from both sides
both sides we get 5x we get 10x = 2x
+ 10 = 5x + 10 + 2x + 24
3. subtracting 10 from 3. subtracting 4x
both sides we get 5x from both sides
= 5x we get 6x = 4x+
4. subtracting 5x from 24
both sides we get 0 4. subtracting 4x
=0 from both sides
5. therefore x = 0 is the we get 2x = 24
only solution. 5. dividing by 2 on
both sides we get
x = 12
28
ANALOGIES IN EQUATIONS
Post-test 1 Post-test 2
Analogies M=85.40 SD=68.73 M = 51.32 SD =31.58
training
Symbolic M=67.51 SD=30.87 M =49.02 SD =29.93
training
F(1,233)=6.607 F(1,227)=0.320
p-value=0.011 bilateral p-value=0.572 bilateral
29