JOURNAL OF RESEARCH IN SCIENCE TEACHING VOL. 45, NO. 2, PP.

197–218 (2008)

Transfer of Algebraic and Graphical Thinking between

Mathematics and Chemistry

Marietjie Potgieter,1 Ansie Harding,2 Johann Engelbrecht2

1
Department of Chemistry, University of Pretoria, 0002 Pretoria, South Africa
2
Department Mathematics and Applied Mathematics,
University of Pretoria, 0002 Pretoria, South Africa

Received 18 April 2006; Revised 15 February 2007; Accepted 25 February 2007

Abstract: Students in undergraduate chemistry courses find, as a rule, topics with a strong
mathematical basis difficult to master. In this study we investigate whether such mathematically related
problems are due to deficiencies in their mathematics foundation or due to the complexity introduced by
transfer of mathematics to a new scientific domain. In the investigation we exposed a group of students to a
chemistry instrument based on the Nernst equation in electrochemistry, and an equivalent group of students
to a similar mathematics instrument in which the questions were stripped of all chemistry context. Both
tests contained items requiring algebraic as well as graphical skills. Students experienced few problems
with the algebraic questions in both the chemistry and mathematics tests. Their graphical construction and
interpretation skills, on the other hand, are inadequate, as can be seen from the poor performance in both the
mathematics and the chemistry results of the graphical question. Our conclusion is that the
problem seems to lie at the mathematics side and is not due to the transfer of mathematics to an
application. ß 2007 Wiley Periodicals, Inc. J Res Sci Teach 45: 197–218, 2008
Keywords: chemistry; interdisciplinary science; mathematics; college/university; statistics/multivariate

Mathematics and chemistry are traditionally taught as separate disciplines, with little
attention paid to the interface between them. However, the role of mathematics and also of more
general quantitative methods in chemistry is becoming increasingly important (Atkins, 1998;
Witten, 2005). The need for closer interdisciplinary collaboration is underlined by a recent report
by the National Science Foundation (Wright & Chorin, 2000):

. . . it is obvious . . . that students of mathematics should be able to understand problems in

science, and that students of science should understand the power and roles of
mathematics. Each area of science has its own unique features, but the different areas
share common features that are often of a mathematical nature.

Correspondence to: J. Engelbrecht; E-mail: [email protected]

DOI 10.1002/tea.20208
Published online 12 December 2007 in Wiley InterScience (www.interscience.wiley.com).

ß 2007 Wiley Periodicals, Inc.

198 POTGIETER, HARDING, AND ENGELBRECHT

It has been our experience over a number of years that students in chemistry find topics with a
strong mathematical basis difficult to master. For example, the use of the Nernst equation (1) in
electrochemistry and the application of the Henderson-Hasselbach equation (2) (Silberberg,
2006) for buffer systems of weak acids (both equations given below) generally pose problems for
first- and second-year chemistry students. Both these equations consist of a logarithmic
relationship between variables, which can be represented by the more general mathematical
expression y ¼ b þ a ln x. To gain insight into the reasons for the difficulty that students seem to
experience in the application of mathematics in a different discipline, chemistry in our case,
we have selected to focus on the Nernst equation (1) for this study. We expect the results to be
generalizable to the Henderson-Hasselbach equation because of its similar mathematical structure
and also to other chemical equations containing logarithmic functions. The extent to which the
results of this study are relevant to nonlogarithmic functions in chemistry needs to be investigated
in future.
 2þ 
RT Zn
E ¼ E8
lnQ; where Q ¼ ð1Þ
nF Cu2þ
½base
pH ¼ pKa þ log ð2Þ
½acid

The Nernst equation is a mathematical expression used to determine the potential of an

electrochemical cell (or battery) under nonstandard conditions. The cell potential will change in
relation to the variation in electrolyte concentrations in the individual half-cells. The potential of
an electrochemical cell under standard conditions can be determined quite easily with the use of
reference tables that are freely available. However, because electrochemical cells typically
operate under nonstandard conditions it is essential to be able to calculate the expected potential
under these conditions.
If one considers that the Nernst equation describes the decrease in potential of a battery as a
function of the chemical reaction that occurs inside its cells it is evident that a graphical (visual)
representation of the decline in potential is perhaps more powerful than the equation itself for
illustrative purposes. In fact, the graph presents the student with an inevitable challenge to
reconcile the continuous decline implied by the logarithmic relationship with the commercial
requirement for batteries to deliver a sustained voltage output at a specific level. This realization is
important from an engineering/technological perspective. The algebra without its visual
implications is clearly inadequate to demonstrate the practical implications of the Nernst equation.
We compared the approach of a number of standard general chemistry textbooks for the
presentation of the influence of electrolyte concentration on cell potential, a discussion that
typically includes a presentation of the Nernst equation and examples of its use (Brown, LeMay, &
Bursten, 2006; Chang, 2005; Kotz, Treichel, & Weaver, 2006; Oxtoby, Freeman, & Block, 2003;
Silberberg, 2006). It is clear that the way in which the Nernst equation is introduced and
demonstrated is strongly procedural and algebraic in nature. Textbooks typically show that the
Nernst equation can be derived from the relationship between free energy and concentration
(a concept introduced earlier) and then proceed to demonstrate its use with worked examples.
Graphical representations of any kind are absent in all of the textbooks, except in one (Silberberg,
2006, p. 918), where the author included a graph to demonstrate the mathematical relationship
between variables (the linear relationship between E and ln Q). It is reasonable to expect that
chemistry textbooks should concentrate on developing conceptual understanding of the
overarching chemical concepts of electrochemistry rather than the mathematical concepts
associated with the Nernst equation, which fulfills only a supportive role in the bigger
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 199

picture. However, the question arises first, to what extent a lack of conceptual understanding of
mathematics (or a lack of mathematical skills) contributes to the problems students experience
with such topics in chemistry, and second what the impact is of transferring these skills to an
applied field such as chemistry where the mathematics is often ‘‘obscured.’’

Research Question
Against the background sketched above, we investigate the following research question: Are
the mathematically related difficulties that students experience in chemistry due to deficiencies in
their mathematics foundation or due to the complexity introduced by transfer of mathematics to a
new scientific domain?
The relevance of this question was highlighted by Mevarech and Kramarsky (1997), who
listed the issue of students’ ability to construct and interpret graphs representing different contexts
(among other issues) as open and relevant.

Theoretical Perspective

Transfer of Knowledge
Theories of transfer of knowledge are based upon the idea that knowledge can be transferred
from one situation to another, that it links with a new situation, and retrieves information from
memory. This study offers an investigation into the success with which knowledge is transferred
from mathematics to chemistry. Knowledge is seen as a set of tools stored in the memory to be used
when needed (Boaler, 1999). In an effort to improve skills of knowledge transfer, learning transfer
theorists (e.g., Anderson, Reder, & Simon, 1996) developed strategies that assist students to select
the right tool at the right time.
However, others (e.g., Lave, 1988) cast doubt on the emphasis that cognitive studies place on
general problem solving skills and argue that learners’ mental processes are structured by the
context, the activity, the tools, and their interactions with others. Lave, in a study conducted with
school-going children, found that they could solve problems using arithmetic skills with a
very high success rate in everyday life, compared to a poor performance when the equivalent
arithmetic skills are required in school-like tests (Lave, 1988; McCormick, 1997). In other words,
the children in Lave’s study were unable to abstract general principles from their practical
knowledge. Our focus is on university students, and this study aims to investigate whether
chemical engineering students experience the same lack of abstraction of mathematical principles.
Much research has been conducted on the transfer of knowledge, where some of the
knowledge acquired in a first domain can be used in the learning of a second domain (e.g., Cormier
& Hagman, 1987; Harvey & Anderson, 1996; Singley & Anderson, 1989), a strong literature base
to which this study aims to add to. Positive transfer has been reported in the case of procedural
tasks (Harvey, 1990; Singley & Anderson, 1988). However, the success of transfer between
problem solving tasks in more conceptual thinking is less clear, and failures to transfer have been
reported (Catrambone & Holyoak, 1989; Scholtz & Wiedenbeck, 1989).
In the case of procedural knowledge, researchers have proposed different abstraction
mechanisms (Bovair, Kieras, & Polson, 1990; Harvey & Anderson, 1996; Singley & Anderson,
1989) resulting in gains in the time learning in the second domain, once the rule has been acquired
in the initial domain (Singley & Anderson, 1989).
In the case of conceptual knowledge, Brooks and Dansereau (1987) identified ways in which
conceptual knowledge gained in one domain may facilitate transfer into a new domain. The first
Journal of Research in Science Teaching. DOI 10.1002/tea
200 POTGIETER, HARDING, AND ENGELBRECHT

domain knowledge may provide a familiar analogy assisting understanding of new procedural
information in the second domain. Furthermore, the old declarative knowledge can serve as a
general framework for embedding and elaborating more detailed new knowledge (Harvey &
Anderson, 1996). The problems associated with knowledge transfer within mathematics were
described by Ross (1989a, 1989b), who found that the abstract knowledge embedded in prior
examples is often not accessed. Students attempting to solve mathematics problems were more
often reminded of previously learned examples with superficial similarities, rather than of
examples with common structure (Forbus, Gentner, & Law, 1995; Gentner, Rattermann, &
Forbus, 1993; Loewenstein, Thompson, & Gentner, 1999).
From the literature review it is clear that successful knowledge transfer from one
domain to another depends on a number of factors. Cognizance of these general complexities is
imperative before next focusing on the relationship between mathematics and chemistry in
particular.

Mathematics and Chemistry

Scholarship of teaching and learning acknowledges differences in traditional pedagogies of
different disciplines, but growing forums for crossdisciplinary communication, including the
increased establishment of new interdisciplinary programs in, for example, computational
chemistry, bioinformatics, and geostatistics, create opportunities and the need to trade their
insights, ideas, and findings, and to learn from each other (Huber & Morreale, 2002).
We first give an overview of the interplay between mathematics and chemistry. The existence
of mathematically related difficulties experienced by students doing chemistry is well-
documented. According to Bangash and Mustafa (2002), although chemistry is difficult with
mathematics, it is impossible without it. A variety of reasons for these difficulties are offered,
which we discuss.
Bangash and Mustafa (2002) claim that students face problems in learning chemistry mostly
because of insufficient mathematics preparation. This is especially true for physical chemistry,
analytical chemistry, inorganic, and environmental chemistry.

The students face problems in learning chemistry mostly because they in their
undergraduate courses opt for chemistry as a major without mathematics in
combination (p. 3).

This finding is significant in the light of this paper and relates directly to the research question.
Other research that pertains directly to our research question includes, among others, that of
Meltzer (2002), who agrees with this claim about the importance of mathematics preparedness. He
found a consistent lack of correlation between learning gain of physics students and the results of a
pretest in physics, but on the other hand, found a significant positive correlation between the
learning gain of these physics students and the results of a pretest in mathematics. Witten (2005)
sees a mismatch between current content of mathematical courses taken by chemistry students and
the quantitative skills required in a typical senior level chemistry course due mainly to the lack of
exposure of students to mathematical tools applied directly to chemical problems. Another
phenomenon, also mentioned by Witten (2005), and supported by Schliemann (1995) and Boaler
(1998), is that students do not naturally generalize what they have learned in the mathematics
course to new, out-of-the-mathematics-class situations. He mentions the simple example of
students that cannot identify the difference between variables and constants in a chemistry
equation. Nonscience mathematics students dislike applications from physics or chemistry
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 201

(Hughes-Hallett, 2001) and mathematics has seldom been popular among chemists (Bangash &
Mustafa, 2002).
Literature offers possible remedies for mathematically related difficulties experienced by
chemistry students.
Applied science students, such as engineering students, are encouraged in the NSF
report cited earlier (Wright & Chorin, 2000) to include more quantitative courses into their
programs, also hereby reducing the perception that mathematics is only needed for mathematics
courses. Witten (2005) argues for adapting the current undergraduate supporting mathematics
courses to become more contextual, demonstrating the mathematical techniques in the applied
context.
Another solution is investigated by Niklasson, Christie, Larsson, Öhrström, and Bowden
(2003) by running an experiment focusing on the integration of mathematics courses with
chemistry courses and some chemical engineering courses. They believe that students need to
experience variation in the application of mathematics in the different contexts of problem solving
to enable them to identify what is common and what is different and understand why the
differences exist. Their approach is based on basic variation theory (Bowden & Marton, 1998;
Marton & Booth, 1997), which argues ‘‘the ability to discern the relevant aspects of a situation
cannot be developed unless variation is experienced’’ (Niklasson et al., 2003, p. 4). This approach
can lead to improved learning in chemistry by exposure to its mathematical foundation and
improved learning in mathematics by exposure to its application in chemistry.
Hughes-Hallett (2001) notes that mathematics focuses on climbing the ladder of abstraction
and wants students to rise above context. Yet students have difficulty applying the mathematics
they have learned in a different context. Although the fact that different subjects are taught
separate from one another is part of the reason, recognizing mathematics in another field requires
thorough understanding of the context.
We next pause to review graphs and other mathematical representations, which emerged as
both prominent and central to our investigation.

Graphs and Other Representations

In mathematics and science, many forms of representation are taught explicitly in secondary
and postsecondary courses. Constructing and interpreting tables, graphs, and equations are taught
to students as tools for communicating and reasoning within the relevant subject, for example,
chemistry and mathematics. Over the last few decades education initiatives have been advocating
more emphasis on students’ learning to apply what they learn about representations in
mathematics to their activities outside the subject (National Council of Teachers of Mathematics,
1989). Such initiatives recognize the problem of transferring mathematical skills to applied fields
as investigated in this study.
As a backdrop to this study we focus on the use of graphical representations as a means of
visualization and of cultivating conceptual understanding. The concept of a representation was
described (many years ago) by Peirce (1955), stating that for a notation to function as a
representation, someone has to interpret it and give it meaning, thereby including three notions:
something that is represented, the referring expression, and the interpretation linking the referring
expression to that which is represented. Hence, tables, equations, and graphs become
representations when someone gives them meaning by interpreting them (Greeno & Hall,
2004). Larkin (1981) claims, ‘‘the central difference between expert and novice solvers in a
scientific domain is that novice solvers have much less ability to construct or use scientific
representations’’ (p. 121).
Journal of Research in Science Teaching. DOI 10.1002/tea
202 POTGIETER, HARDING, AND ENGELBRECHT

Of the different forms of representation in mathematics, graphs are probably the most
important because they visually illustrate the relationships between different measured variables
(Bastide, 1990; Chambers, Cleveland, Kleiner, & Tukey, 1983; Lemke, 1998). Graphs have not
always been popular; legend has it that Lagrange was quite proud that in all his work not one figure
could be found. The contribution of graphic representations is well established in mathematics and
science education, and it is generally agreed upon that it is practically impossible to address many
of the topics without using graphic representations (Testa, Monroy, & Sassi, 2002). Although
graphs summarize large amounts of information, they still allow details to be resolved; scientists
can see trends and identify subtle differences (Beichner, 1994; Mokros & Tinker, 1987).
Two processes are involved in that graphs and other representations can be constructed or
interpreted. Students construct representations based on partial understanding and then use the
representations to improve their understanding. This enables them to improve on their
representation, leading to an iterative learning process (Greeno & Hall, 2004). Interpretation
refers to the ability to gain meaning from the graph (Leinhardt, Zaslavsky, & Stein, 1990;
Mevarech & Kramarsky, 1997). Although representations may represent different things for
different people, graphs, equations, and tables have common conventions of interpretation, and it
is important for students to learn these so that they can understand the different forms when they
encounter them and also be able to construct them to communicate ideas (Greeno & Hall, 2004).
Many studies focus on interpretation of graphs (e.g., Beichner, 1993; Dreyfus & Eisenberg, 1982;
Leinhardt et al., 1990; McDermott, Rosenquist, & Van Zee, 1987; Vinner, 1983; Wainer, 1992)
and some on the construction aspect (Moschkovich, Schoenfeld, & Arcavi, 1993).
The theoretical approach used in this study is related to a dimension of the framework
developed by Moschkovich et al. (1993) and also used by Mousoulides and Gagatsis (2004).
According to their approach there are two different perspectives from which functions can be
viewed: the process perspective and the object perspective. The process perspective views a
function as a link between x and y-values: students who view functions as a process can substitute
an x-value into an equation to find the resulting value for y. In the object perspective, a function is
thought of an entity; for example, as a graph in the plane that can be manipulated by, for example,
rotation or translation (Moschkovich et al., 1993; Mousoulides & Gagatsis, 2004). The process
approach leads to an algebraic perspective and the object approach to a graphical perspective
(Yerushalmy & Schwartz, 1993). Deep conceptual understanding of a function can only happen
when also seeing a function as an object (Sfard, 1992).
Interpreting graphs is not an elementary process. During interpretation, students have to
reconstruct environments for which graphs could be viable representations. The reconstruction of
such a phenomenon is not unique, and there are potentially many situations to which a graph can be
related. Students not having participated actively in constructing and interpreting graphs as a
regular practice are unlikely to arrive at an appropriate interpretation (Bowen, Roth, & McGinn,
1999).
There is ample evidence that students experience problems with understanding and
interpreting graphs. Students often struggle to extract information from graphs (Wainer, 1992) and
they find it difficult to shift between different representations (Brassel & Rowe, 1993;
Moschkovich et al., 1993). In their study with grade-12 physics students, Brassel and Rowe
(1993) found that

[students] do not understand the fundamental properties and functions of graphs

in representing relationships among variables . . . Their facility with graphs was
generally superficial, grounded on a few, simplistic algorithms such as plotting data
points (p. 69).

Journal of Research in Science Teaching. DOI 10.1002/tea

 MATHEMATICS IN CHEMISTRY 203

Moschkovich et al. (1993) experienced that even competent students treat algebraic and
graphical reprentations as though they are independent. This and other findings mentioned
above are important for the present study, as it provides background to a focal point of this study,
namely the particular difficulties that students experience in an applied field of study such
as chemistry.
Also relevant to this study is research by Beichner (1994) on the use of graphs as a language
for communication, for which students need to have a firm conceptual understanding of what is
represented and how. Prior conceptions of students, even before instruction, play an important role
in the student’s ability to interpret graphs correctly (Confrey, 1993; Mevarech & Kramarsky,
1997). Testa et al. (2002) agree:

When examining image misinterpretation, the influence of previous ideas (or reasoning
strategies or mental representations) that conflict with disciplinary knowledge has been
observed in many cases. This result, which perhaps was foreseeable, confirms that in the
reading of an image it is not easy to disentangle the contribution of such cognitive
‘‘lenses’’ from more specifically iconic difficulties (p. 252).

A number of specific problems experienced by students on graph representation are cited in

literature. Although students get extensive exposure to graphs outside the mathematics classroom,
this exposure, which is not always in accord with the actual meaning (as used in mathematics),
may contribute to students’ problems with correctly interpreting graphs in science and
mathematics (Mevarech & Kramarsky, 1997).
Some students cannot apply what they have learned about graphs in mathematics to other
subjects (McDermott et al., 1987). Bowen et al. (1999) conducted a study in which they
investigated the activities in which university students in an ecology course and professional
scientists engaged when faced with a graph interpretation task. They found that students did not
appear to display any general interpretive skills for graphs, but rather tended to apply the
professor’s interpretation.
Cheng (1996) claims that algebra is too powerful and unconstrained a representation to be
used for straightforward learning by students. Without additional representations such as graphs
and other visual material, algebra can hardly be understood properly.
In chemistry, empirical studies (Ben-Zvi, Eylon, & Silberstein, 1988; Wu, Krajcik, &
Soloway, 2001) have shown that students find microscopic and symbolic representations
especially difficult to understand. Students also experience problems when moving between
different representations without proper conceptual understanding and visual-spatial ability (Keig
& Rubba, 1993; Wu et al., 2001). Many attempts have been made by teachers and researchers to
create conceptual understanding of chemical concepts (e.g., Gabel, 1998; Kozma & Russell,
1997). Students need substantial conceptual knowledge and visual–spatial abilities to be able to
generate interpretations of representations, make translations, and mentally manipulate these
representations (Kozma & Russell, 1997; Wu, Krajcik, & Soloway, 2001). Wu et al. (2001)
propose a model of visualization to investigate possible interaction between the visual and
conceptual aspects of chemical representations. To assist students in making translations between
representations, they develop referential connections between visual and conceptual systems. In
the visualization process students provide multiple linked representations, the linking being as
important as constructing the representations (Wu et al., 2001).
As for the theoretical framework, Scaife and Rogers (1996) point out that little is known about
the cognitive value of graphical representations. Despite the vast amount of research investigating
different aspects of graphical representations, they experience a need to integrate the findings into

Journal of Research in Science Teaching. DOI 10.1002/tea

204 POTGIETER, HARDING, AND ENGELBRECHT

an analytic framework enabling researchers to address issues such as ‘‘. . . the nature of the rela-
tionship between graphical representations and internal representations and to consider how
graphical representations are used when learning, solving problems and making inferences’’ (p. 6).
We conclude the literature review and next describe the research experiment.

The Research Experiment

Two test instruments were developed for this study: one for chemistry, and the other for
mathematics. The test instrument for chemistry consisted of six free-response items followed by
an item requiring students to produce a graph of the Nernst equation. The six items were designed
to lead students through a logical analysis of the Nernst equation to produce a valid graphical
representation in the final question. Of the six items, one assesses chemical knowledge combined
with simple mathematical reasoning (question 3), one can be answered by recall (question 4), and
three require more sophisticated mathematical reasoning for a correct answer (questions 2, 5, and
6). Question 1 can be answered superficially be recall, but can also be answered in depth by
analyzing the Nernst equation.
The test instrument for chemistry was converted to its exact equivalent in mathematics,
stripped of any chemistry context. To do this, the Nernst equation was expressed in generic
mathematical symbols. Items 1, 2, 5, and 6 in both instruments relate to significant points in a
graphical presentation of the Nernst equation and were intended to prompt students to link the
algebraic analysis with the graphic representation required in the last item.

Chemistry Test

The Nernst equation below is used to calculate cell potential (E) at nonstandard conditions:
RT
E ¼ E8
lnQ with E ; R; T; n; and F positive constants:
nF
Q ¼ reaction quotient and E ¼ standard cell potential:

Answer the questions below for the Cu-Zn voltaic cell:

 2þ 
2þ 2þ Zn
Cu þ Zn ! Cu þ Zn ; where Q ¼   for dilute electrolyte solutions
Cu2þ

1. What should the electrolyte concentrations, [Zn2þ] and [Cu2þ], be in order that E ¼ E8?
2. Could E have a larger (more positive) value than E8? Explain (refer to Q).
3. How would the value of Q change as the voltaic cell discharges?
4. How would the value of E change as the voltaic cell discharges?
5. When equilibrium is reached there is no net reaction occurring in the voltaic cell and
E ¼ 0. What is the value of Q at equilibrium?
6. Could E have a negative value? Explain in words and mathematically.

Draw a graph of the change in cell potential E against the reaction quotient Q. Show the
equilibrium point on the graph as well as the point where E ¼ E8.

Each of the test items was followed by a question in which students were asked to indicate
their certainty about the correctness of their answer on a four-point scale (certain, fairly certain,
not certain, total guess). Numerical values of 4, 3, 2, and 1 were assigned to the four categories,
respectively, for the calculation of an average confidence value for each item.
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 205

Mathematics Test

Consider the following equation:

y ¼ b
a ln x
with a and b positive constants and x a fraction of the form x ¼ s/t, (s, t > 0).

1. What should the relation be between s and t in order that y ¼ b?

2. Could y have a larger (more positive) value than b? Explain (refer to s and t).
3. How would the value of x change if s increases and t decreases?
4. How would the value of y change if s increases and t decreases?
5. For what x-value will y ¼ 0?
6. Could y have a negative value? Explain in words and mathematically.

Draw a graph of the function y against x. Show b on the graph and possible intersections
with the axes. Give the point (coordinates) and show the point on the graph with y-
coordinate b.

The two test instruments were piloted with a group of students at the end of their first
academic year and subsequently refined. Content validity of the chemistry test was established by
means of a think-aloud protocol using an expert (a chemistry lecturer) to work through the test.
The two tests were then administered to a group of chemical engineering students in the second
semester of their second year of study. These students are enrolled in an internationally accredited
program for the training of chemical engineers at the University of Pretoria, a foremost university
in South Africa. At the time of the experiment, these students had been exposed in the mathematics
department to three semester courses in calculus, a course in linear algebra and a course in
differential equations. In most of these courses the approach was to establish a balance between the
algebraic, graphical, numerical and verbal representations. In the introductory part of the calculus
courses, in particular, some time is spent on sketching and interpreting graphs of a variety of
functions, including logarithmic functions.
These students had also completed three chemistry modules at the time of the experiment, that
is, two semester modules in general and organic chemistry at first-year level and one semester
module in organic chemistry at second-year level. They were enrolled for a second-year semester
module in analytical chemistry at the time of data collection. During their first year students were
briefly introduced to the Nernst equation as part of the discussion of the fundamentals of
electrochemistry. The topic was again addressed in the second year as the theoretical basis of
electroanalytical chemistry.
Both tests were written at the same time—the mathematics and chemistry versions given
randomly to the 82 students that were present. As a result, every student wrote either the chemistry
or the mathematics test. The chemistry test was written by 26 male and 15 female students, and the
mathematics tests by 31 male and 10 female students. To verify the equivalence of the two groups,
we used the performance of these students in the first calculus course as a measure. A t-test for
independent groups showed no significant difference in the average performance of the two groups
(p ¼ .81). Five chemistry tests were not answered in full due to an oversight and could not be used
for detailed analysis.
We discuss correct responses to the questions, both of the chemistry and mathematics tests.
All general chemistry textbooks state explicitly that E ¼ E8 under standard conditions (where
[Zn2þ] ¼ [Cu2þ] ¼ 1M). Question 1 (chemistry) can therefore be answered by simple recall,
which will result in a partially correct answer, but a mathematical approach can also be followed.
Journal of Research in Science Teaching. DOI 10.1002/tea
206 POTGIETER, HARDING, AND ENGELBRECHT

E y

E0 b

Q x
1 E 0 nF 1 b

e RT e a

Figure 1. Chemistry and mathematics graphs for the Nernst equation.

The reasoning that for E to be equal to E8, ln Q should have a zero value, in which case Q ¼ 1, that
is, at all concentrations where [Zn2þ] ¼ [Cu2þ], will lead to a correct answer of general
application. The same reasoning, namely that ln x ¼ 0 if x ¼ 1, is expected in the mathematical
version of the question.
In question 2, the reasoning, requiring mathematics, is that E will be larger than E8 if ln Q is
negative. This happens when Q has a value <1, that is, when [Zn2þ] < [Cu2þ]. In the mathematics
part of the test the same reasoning is expected, namely that y will be > b when the ln-function is
negative, that is when x lies between 0 and 1, when t > s.
For the chemistry version of question 3, chemistry knowledge combined with simple
mathematical reasoning reveals that Q will increase as the voltaic cell discharges because [Zn2þ]
increases and [Cu2þ] decreases. The mathematics equivalent follows the same simple reasoning
that x increases as s increases and t decreases.
Respondents are expected to know intuitively or by simple recall that the cell potential E will
decrease as the cell discharges (question 4 in the chemistry version). Mathematical reasoning can
be followed saying that if Q increases (prerequisite chemistry knowledge) then ln Q increases
(becomes more positive) and so the potential E decreases. In the mathematical version, ln x
increases as the value of x increases, and in turn, y decreases.
Question 5 in the chemistry version could be answered from a purely chemical perspective
(E ¼ 0 when equilibrium is reached; therefore, Q is equal to K, the equilibrium constant),
or mathematical skills can be used to solve for Q when setting E ¼ 0 in order to obtain
E0 nF b
Q ¼ e RT . Similar manipulations are necessary in the mathematics version to obtain x ¼ ea .
Question 6 of the chemistry version again requires mathematical skills and cannot be
answered by only considering the chemical process. E can have a negative value when the term
RT 2þ
nF ln Q is numerically larger than E8, that is, when Q has a very large value, because [Zn ] is
significantly larger than [Cu ]. The mathematical equivalent of the question renders x > b/ea.
2þ

The graphs required for the chemistry and the mathematics versions of the graph question are
given in Figure 1.
Although these engineering students had encountered the algebraic form of the Nernst
equation in both their first and second years of study by the time the data was collected, neither in
the teaching nor in any of the textbooks prescribed for their chemistry courses (Kotz et al., 2006,
for the first-year course, and Skoog, Holler, & Nieman, 1998, for the second-year course) was the
Nernst equation presented graphically.

Relationships Investigated
The different relationships that were investigated in this study are presented in Figure 2. In
each of the tests the translation of information between algebraic computations and a graphical
representation was determined (relationships a and b).
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 207

Mathematics Chemistry

Algebra c Algebra

a b

Graphic d Graphic
presentation presentation

Figure 2. Relationships investigated.

The comparison of the performance on algebra items in the mathematics test with that of
similar items in the chemistry test provides information about relationship c, that is, the
translation of algebraic principles to a new scientific domain. A comparison of graphic
competency in mathematics and in chemistry provides information about the strength of
relationship d. Relationships a and b are therefore determined from direct evidence, because
both elements were present in the same test instrument, that is, in both the chemistry
and mathematics tests. The parallel design of the two instruments allowed the study of
relationships c and d indirectly from evidence of mathematical reasoning in the chemistry
scripts.
The reliability of the coding process was addressed by the fact that the tests were
independently coded by two of the researchers—one for chemistry and the other for
mathematics—and then all scripts were checked by the third researcher. In some instances there
were some negligible differences in opinion that were settled after discussion.

Results
The results of the analysis of the chemistry test are presented in Table 1. The responses to the
first six items in the chemistry test were assessed to be correct, partially correct, or wrong.
Responses were also coded to reflect whether answers contained any evidence of mathematical
reasoning. For example, a correct answer for question 1 would be [Zn2þ] ¼ [Cu2þ]. The response,
[Zn2þ] ¼ [Cu2þ] ¼ 1, provided by eight students, is valid for a special case (standard conditions)
and was judged to be partially correct. Answers to questions 2 to 6 were judged to be partially
correct when either too vague to rule out guessing or if a correct answer was provided, but the
explanation demonstrated flawed chemical or mathematical reasoning.

Table 1
Results for the chemistry test
Number of Responses (N ¼ 36)
Q1 Q2 Q3 Q4 Q5 Q6
Correct, no indication of maths reasoning 11 8 13 22 9 5
Partially correct, no indication of maths reasoning 8 2 1 0 0 0
Wrong, no indication of maths reasoning 1 2 12 7 9 4
Correct answer with maths reasoning 12 14 10 5 14 19
Partially correct answer with maths reasoning 0 4 0 1 2 4
Wrong answer with maths reasoning 1 6 0 1 1 3
Not answered 3 0 0 0 1 1

Journal of Research in Science Teaching. DOI 10.1002/tea

208 POTGIETER, HARDING, AND ENGELBRECHT

Table 2
Results for the mathematics test
Number of Responses (N ¼ 41)
Q1 Q2 Q3 Q4 Q5 Q6
Correct 40 25 41 36 32 8
Partially correct 0 2 0 1 3 26
Wrong 1 14 0 4 6 7

The response pattern obtained reflects the nature of the test items. The majority of responses
to questions 3 and 4 did not contain any indication of mathematical reasoning, whereas the
majority of responses to questions 2 and 6 did include such evidence. The responses to
question 1 were divided between students who used mathematical reasoning (40%) and students
that did not (60%). This division may be due to the fact that although the question could be
analyzed with the aid of mathematical reasoning, it could also be answered by recall.
The results of the analysis of the mathematics test are presented in Table 2. Questions 2 and
6 posed the biggest challenge and resulted in poorer performance than for the other questions. In
the majority of answers to question 6 students pursued correct mathematical reasoning but did not
draw the argument to a final conclusion.
Of specific interest in the analysis of the data in Table 1 are the items with a relatively high
frequency of incorrect or partially correct answers, because the question arises whether
breakdown occurred as a result of inadequate chemical or mathematical understanding.
For the analysis of response frequencies for lack of understanding or the presence of
misconceptions, a frequency of up to 11% of partially correct or wrong responses (4 of
36 responses) was not considered as significant. Based on that criterion, the responses without
explicit mathematical reasoning to questions 1, 3, 4, and 5 require explanation. The eight partially
correct answers given to the question 1 in the chemistry test stated that E ¼ E8 when
[Cu2þ] ¼ [Zn2þ] ¼ 1 M, an answer that clearly highlights the inadequacy of memory recall.
The incorrect answers given for question 3 (12 responses), question 4 (7 responses), and for
question 5 (9 responses) could all be ascribed to a lack of chemical conceptual understanding.
Similarly, the higher frequency of incorrect or partially correct answers with indications of
explicit mathematical reasoning for questions 2 and 6 require explanation. Analysis of the
reasoning patterns for question 6 indicated that in all but one of the cases incorrect chemical
assumptions (e.g., E8 < 0 or Q cannot be larger than K) resulted in errors despite sound
mathematical reasoning. Question 2 is the only item that exposed a significant extent of flawed
mathematical reasoning in the chemistry context (10 of 36 responses, 28%).
For this analysis we constructed a flow diagram of logical thinking steps for each of the six
items, based on the responses obtained from the students. The majority of respondents provided
multistep expositions as explanation for their answers, which revealed their logical thinking
patterns and the causes of flawed reasoning, even those that may have been somewhat obscured by
a correct final answer. All of the flow diagrams were analyzed in depth and will be discussed below,
but the diagram for question 2 (Figure 3) was found to be most informative in terms of the research
question for this study.
Five respondents arrived at the wrong answer, because sound chemistry assumptions were
followed by flawed mathematical reasoning (Figure 3). These students did not realize that ln Q is a
negative quantity when Q is a fraction. Similar problems with the understanding and application of
the logarithmic function were observed for three other respondents who reasoned that ln Q would
be a negative quantity when Q < 0 or Q > 1, respectively (Figure 3).
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 209

Incorrect Q<0 2 [Cu2+] > [Zn2+]

assumption: 0<Q<1
If Q becomes
3 smaller
Flawed 1
1 1 1
maths: Q changes
1
constantly
13 1
0<Q<1 ln Q < 0 Factor Y < 0 E > E° YES 1
No reason
given
Flawed 1 1
maths: 1
Flawed
maths Q = 1/[Cu2+]
If Q > 1 1
0<Q<1 E > E°
Writing error?

Wrong answer due to incorrect mathematical reasoning:

4 ln Q > 0
Q>0 Factor Y > 0 E < E° NO
for all values of Q

Q cannot be 0

N = 34. (Two responses were too ambiguous to allow reliable analysis.)

Figure 3. Flow diagram of logical reasoning for question 2 of the chemistry test. (Response frequencies are
reported by numbers next to the arrows).

Table 3
Performance comparison of students in question 2 of the two tests
Partially
Correct Correct Wrong
Chemistry test—24 responses demonstrating mathematical reasoning 58% 17% 25%
Mathematics test—41 responses 61% 5% 34%

The analogous item in the mathematics test (question 2) was also analyzed to determine
whether knowledge and understanding of the logarithmic function was similarly inadequate in
the mathematics context. The results from this question were then compared with those of the
chemistry students that used mathematical reasoning in their approach of question 2. From Table 3
it seems as if the performance of the mathematics students corresponds with that of the chemistry
students that used mathematical reasoning in approaching this question.
We next focus on the three aspects of mathematical skills necessary to analyse the features of
a function as illustrated in Figure 2, namely algebraic skills, graphical skills, and skills of
connecting algebraic and graphical aspects of a function.

Algebraic Skills
In the mathematical sense, our use of the term algebraic refers to the use of symbolic
presentation and manipulation, whereas in the chemistry sense it also demands basic knowledge of
chemistry concepts. An overall score for the algebraic questions (the first six items) was calculated
for every individual and a classification of good ( > 70%), average (50–70%), or poor (<50%)
Journal of Research in Science Teaching. DOI 10.1002/tea
210 POTGIETER, HARDING, AND ENGELBRECHT

was then assigned to each individual. A student displayed good algebraic proficiency when both
the manipulations and algebraic reasoning were performed correctly. Minor sign errors or a single
isolated reasoning error such as that ln x cannot be negative were overlooked. A rating of average
algebraic proficiency was awarded when answers were mostly correct but a display of more than
the odd faulty reasoning and/or manipulation was evident.

Graphical Skills
The graphs in both tests were analyzed as an entity independent of the algebra preceding it.
The ability of the respondent to present the particular logarithmic relationship correctly was
evaluated on a three-point scale (good, average, or poor) depending on the shape of the curve, its
placement on the axes, the presence or not of an asymptote, and most importantly whether there
were indications of graphical skills such as translating and stretching of the original logarithmic
graph. Good graphic proficiency was displayed when the logarithmic shape of the graph was
evident, the graph was decreasing, the intercept with the x-axis was shown, and the annotation was
correct. Average graphical proficiency meant that one or at most two of the above features were
faulty; more than that pointed to poor graphical proficiency.

Connecting Skills
The extent to which answers provided for the preceding algebraic questions (items 2, 5, and 6)
were connected with the graphical representation in the last item was judged on a similar three-
point scale for both tests. Calculating for what x-value the y-value will be zero, and using the
obtained value as the x-intercept of the graph is considered as indication of a good connecting skill.
These items can be assumed to be directly associated with the features of the graph. Good
connections meant that the information obtained from the algebra questions 2, 5, and 6 were in line
with the features of the graph; for example when an x-value for which y ¼ 0 was calculated, and
this was displayed as an annotated x-intercept. A rating of average meant that some of the algebraic
information was evident on the graph, but not all or there was at most one discrepancy between the
algebra and the displayed graph. Poor connections meant that none or very little of the algebraic
information was displayed on the graph.
The results for both instruments of the analysis of skills associated with the translation of an
algebraic function to its graphic presentation are presented in Tables 4 and 5.
Three examples of graphs obtained from the mathematics test are presented as illustration of
the assessment procedure in Figure 4.
When investigating the various shapes that emerged when students attempted to sketch the
graph of the logarithmic function, one noteworthy aspect emerged. On the mathematics side only

Table 4
Results of skills analysis for the chemistry test
Number of Students (N ¼ 36)
Good Average Poor
Algebraic proficiency (Items 1–6) 20 (>70% correct) 8 (50–70% correct) 8 (<50% correct)
Graphic proficiency 6 6 24
Connections between graph and 3 9 24
answers to items 1, 2, 5, and 6

Journal of Research in Science Teaching. DOI 10.1002/tea

 MATHEMATICS IN CHEMISTRY 211

Table 5
Results of skills analysis for the mathematics test
Number of Students (N ¼ 41)
Good Average Poor
Algebraic proficiency (Items 1–6) 34 (>70% correct) 6 (50–70% correct) 1 (<50% correct)
Graphic proficiency 9 4 28
Connections between graph and 10 7 24
answers to items 1, 2, 5, and 6

5% of students sketched a simple decreasing straight line, whereas on the chemistry side a much
larger percentage, 20%, sketched a straight line (similar to Figure 4b). We surmize that the larger
percentage in chemistry might be because students revert to their chemistry knowledge when their
mathematics knowledge is not robust enough. Students know that as a battery discharges, the
voltage decreases and almost intuitively sketch a decreasing straight line.

Figure 4. (a) A response showing good graphic competency and a good connection with the correct algebra
preceding it. (b) A response that shows poor graphic competency, but an effort toward connecting with
the preceding correct algebra. (c) A response that shows poor graphic competency and poor connection
with the preceding correct algebra.

Journal of Research in Science Teaching. DOI 10.1002/tea

212 POTGIETER, HARDING, AND ENGELBRECHT

Table 6
Confidence values
Q1 Q2 Q3 Q4 Q5 Q6 Graph question
Chemistry test (N ¼ 36) 3.7 3.2 2.8 2.8 3.0 3.1 2.1
Mathematics test (N ¼ 41) 3.8 3.5 3.7 3.4 3.3 3.4 2.4

In Table 6 we give the results for the self-reported confidence values. The numbers represent
the average values per item for each cohort, where certain ¼ 4, fairly certain ¼ 3, not certain ¼ 2,
and total guess ¼ 1.
From Table 6 it is clear that for the six algebraic questions in the chemistry as well as the
mathematics test the confidence values (average of confidence values for participating students)
were higher than the confidence values for the graphical question. When comparing this result
with performance it is clear that students had justified confidence in their ability to answer algebra
questions correctly, and were realistically less confident in their graphic competence. In general,
comparable confidence values were obtained for the chemistry and mathematics versions of each
question, except for questions 3 and 4. This could be ascribed to possible uncertainty introduced
by the need for basic knowledge of electrochemistry in the chemistry questions, thereby possibly
increasing the cognitive demand of the questions, in comparison with their mathematics
counterparts.

Discussion
It is clear from the results presented above that the vast majority of students are comfortable
with an algebraic approach to the analysis of the Nernst equation both in chemistry and in
mathematics. The performance of the majority of students on the first part of the mathematics test
was good (more than 80% of students scored more than 70%). The poorer performance for the
chemistry version of the test (56% of students achieved more than 70%) was to be expected,
because of the increase in inherent complexity. A noticeable feature of the answers provided for
questions in the chemistry test, however, was the explicit and efficient use of algebra to arrive at
answers. This was true for items 2, 5, and 6, as well as for item 1 where mathematical reasoning
was not required (ca. 60% of correct responses for item 1 included mathematical statements as
justification for the answer provided). Where breakdown in algebraic proficiency occurred in the
chemistry context it could be ascribed to inadequate mathematical conceptual understanding and
not to transfer between contexts (question 2). This conclusion is confirmed by the fact that the
results for question 2 in the mathematics test show a similar pattern (Table 3).
The more procedural algebraic skills (process perspective), obtained in a mathematical
setting, were successfully transferred to chemistry. In fact, it was noticeable how many and with
what ease students applied these algebraic skills to the new, applied problem situation. These
results argue for a strong relationship c as shown in Figure 2. The results support the positive
transfer of knowledge and skills in the case of procedural tasks reported by Harvey (1990)
and Singley and Anderson (1988). This finding also supports the opinion of Brooks and
Danserau (1987) in saying that knowledge in a first domain may provide a familiar analogy
assisting understanding of new procedural knowledge in a second domain. The observed
algebraic proficiency in both disciplines could be a result of the emphasis on algebra
at secondary school level in South Africa and also on the extended years of exposition (South
African students first encounter algebra at age 13). Furthermore, in the university
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 213

mathematics courses that these students had been exposed to, the algebraic approach tends
to continue.
Only six students succeeded in drawing an acceptable graph of the Nernst equation as part of
the chemistry test (N ¼ 36). Similarly, only nine students produced an acceptable graph as part of
the mathematics test (N ¼ 41). The nature of the mathematical expression in chemistry is
inherently more demanding to grasp, because the equation is more complex. There are more pieces
of information to process simultaneously, resulting in a higher cognitive demand. It was expected
that fewer students will be able to produce an acceptable graphical representation within the
chemistry context, but the poor graphic proficiency of students within the mathematics domain
came as a surprise. In this regard, the results of the present study are in line with those of previous
studies indicating that students cannot effectively use the graphical (object) approach (Knuth,
2000; Mousoulides & Gagatsis, 2004). Students seemed to have been aware of their lack of
graphical skills as suggested by the fact that poor performance in question 6 was accompanied by
significantly lower confidence values.
It was also evident in the majority of cases in both mathematics and chemistry that
despite their ability to perform algebraic manipulations correctly students do not connect
the obtained information to the graphical representation of the equation (relationships a and b
in Figure 2 are weak). Students, in general, do not use information obtained algebraically
in the preceding questions to aid in sketching the graph (see Figure 4c). Of special
concern were contradictions between answers to the algebraic questions and the graph,
for example, saying that the function can be negative yet sketching a graph that does
not become negative. Graphical representations seem to be considered by students as
unconnected to the corresponding algebraic representations and students fail to
make the necessary connections between them (Mousoulides & Gagatsis, 2004; Yerushalmy
& Schwartz, 1993).
This result agrees with the findings of Moschkovich et al. (1993) and with those of Wainer
(1992), who found that even good students find it difficult to move flexibly across representations.
The results also support the findings Knuth (2000), who experienced that students have difficulties
in making connections between different representations of functions in interpreting graphs, and
those of Mousoulides and Gagatsis (2004), who found that:

Many students have not mastered even the fundamentals of the geometric approach in the
domain of functions. Students’ understanding is limited to the use of algebraic
representations and approach, while the use of graphical representations is fundamental
in solving geometric problems . . . The majority of students’ work with functions is
restricted to the domain of algebraic approach. (p. 390)

The observed graphical weakness could have more than one cause. In South Africa,
manipulation of graphs is only taught at the university level, and students do not have the benefit of
ample and repeated exposure as in the case of algebra. Students are not adequately proficient in the
skills of translation and stretching of a well-known graph (in this case the logarithmic function).
Figure 4a illustrates the type of thinking one would have liked to see. Another reason might be that
students have poor understanding of the logarithmic function itself (see Figure 4b). The
logarithmic function and its graph are again first encountered at university level and the instruction
might not be adequate to foster true understanding. Logarithms offer a challenging topic for most
students, and although they seem to be able to deal with it algebraically, the combination of
familiarity with the graph and performing shifts could be too challenging and requires more
attention.
Journal of Research in Science Teaching. DOI 10.1002/tea
214 POTGIETER, HARDING, AND ENGELBRECHT

It is more problematic to report on the transfer of conceptual knowledge and understanding

(the object perspective) from the results of this study. Because of the poor performance in the
graphical skills category in both chemistry and mathematics, we cannot justly determine the
strength of relationship d. In a few isolated cases students did indeed use their skills obtained in
mathematics to construct the chemistry graph, but because of the overwhelming poor
performance, there was certainly not enough evidence for any substantiated conclusion. One
could nevertheless infer that relationship d in Figure 2 is weakened not only by the challenge to
transfer to a new domain, but perhaps more severely by the lack of skills of translation within
mathematics itself. This is an alarming result if one assumes that the graphical representation
of mathematical relationships conveys and contributes toward a deeper conceptual
understanding of the relationship (Sfard, 1992) but it supports the results of Moschkovich
et al. (1993), and those of Brassel and Rowe (1993) that students struggle to interact between
an algebraic and a graphical approach.
It should be noted that students may have experienced the algebra to be easier than the
graphical question, because these questions allow systematic procedural thinking. By contrast the
graph requires the integration of many components, associated with higher cognitive demand. It
was not the intention to have the levels of difficulty on par, but rather to investigate the
mathematics related to the Nernst equation.

Conclusions
The research question for this study was whether mathematically related difficulties that
students experience in chemistry problems are due to deficiencies in their mathematics foundation
or due to the complexity introduced by transfer of mathematics to a new scientific domain. The
answer seems to be clear; the problem lies at the mathematics side and is not due to the transfer of
mathematics to an application. In mathematics, the algebraic foundation of the students is solid.
This is also reflected in the ease with which they deal algebraically with the Nernst equation in
chemistry. Their graphical skills, on the other hand, are inadequate with regard to logarithmic
functions, as can be seen from the poor performance in the graphical question of the mathematics
test. This lack of graphical proficiency is also reflected in the chemistry application, and hampers
performance there.
There also seems to be a lack of connectivity—students either use an algebra window or a
graph window, however poor the view through the latter. Students could have compensated for
poor graphical skills by applying information obtained algebraically regarding points of
interception, etc. This lack of connectivity impacts negatively on the conceptual understanding of
a chemical process. When moving to an application field such as chemistry the lack of graphical
skills and of making connections in mathematics becomes more pronounced. The added
understanding required by an application is impeded by the weak background in graphical
mathematics.
The study has shown that three issues need to be addressed, of which the first two relate to
mathematics and the third to chemistry. First, more emphasis should be put on cultivating
graphical skills when teaching mathematics; this includes construction as well as interpretation of
graphs. This should be done, not only for the sake of deeper conceptual understanding within
mathematics itself, but especially for understanding within applied fields such as chemistry.
Second, care should be taken not to teach algebraic and graphical skills as two disconnected
entities. Students should be taught to use these skills as complementary to obtain a better
understanding of the behavioral patterns of functions. Third, a graphical approach to processes
in chemistry should be encouraged and expanded upon. It is regrettable that few textbooks
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 215

illustrate the process under discussion using a graphical representation to enhance conceptual
understanding.

References
Anderson, J.R., Reder, L.M., & Simon, H.A. (1996). Situated learning and education.
Educational Researcher, 25, 5–11.
Atkins, P.W. (1998). Physical chemistry. Oxford: Oxford University Press.
Bangash, F.K., & Mustafa, S. (2002). Essentials of mathematics in teaching chemistry.
Proceedings of 10th IOSTE Symposium, Vol. 2, Foz do Iquaso, Parana, Brazil, July 2002.
Bastide, F. (1990). The iconography of scientific texts: Principles of analysis. In M. Lynch &
S. Woolgar (Eds.), Representation in scientific practice (pp. 187–229). Cambridge: MIT Press.
Beichner, R. (1994). Testing student interpretation of kinematics graphs. American Journal of
Physics, 62, 750–762.
Beichner, R.J. (1993). Testing student interpretation of kinematics graphs. Paper presented at
the 5th European Association for Research on Learning and Instruction, France.
Ben-Zvi, R., Eylon, B., & Silberstein, J. (1988). Theories, principles and laws. Education in
Chemistry, May, 89–92.
Boaler, J. (1998). Open and closed mathematics: Students’ experiences and understanding.
Journal for Research in Mathematics Education, 29, 41–62.
Boaler, J. (1999). Participation, knowledge and beliefs: A community perspective on
mathematics learning. Educational Studies in Mathematics, 40, 259–281.
Bovair, S., Kieras, D.E., & Polson, P.G. (1990). The acquisition and performance of
text-editing skill: A cognitive complexity analysis. Human Computer Interaction, 5, 1–48.
Bowden, J., & Marton, F. (1998). The university of learning: Beyond quality and competence.
London: Kogan Page.
Bowen, G.M., Roth, W.-M., & McGinn, M.K. (1999). Interpretations of graphs by
university biology students and practicing scientists: Toward a social practice view of
scientific representation practices. Journal of Research in Science Teaching, 36, 1020–
1043.
Brassel, H.M., & Rowe, M.B. (1993). Graphing skills among high school physics students.
School Science and Mathematics, 93, 63–71.
Brooks, L.W., & Dansereau, D.F. (1987). Transfer of information: An instructional
perspective. In S.M. Cormier & J.D. Hagman (Eds.), Transfer of learning: Contemporary
research and applications (pp. 121–150). New York: Academic Press.
Brown, T.L., LeMay, H.E., & Bursten, B.E. (2006). Chemistry: The central science (10th ed.).
Upper Saddle River, NJ: Pearson/Prentice-Hall.
Catrambone, R., & Holyoak, K.J. (1989). Overcoming contextual limitations on problem
solving transfer. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15,
1147–1156.
Chambers, J., Cleveland, W., Kleiner, B., & Tukey, P. (1983). Graphical methods for data
analysis. Belmont: Wadsworth International Group.
Chang, R. (2005). Chemistry (8th ed.). New York: McGraw-Hill.
Cheng, P.C.-H. (1996). Problem solving and learning with diagrammatic representations. In
D. Peterson (Ed.), Forms of representation (pp. 47–66). Exeter: Intellect Books.
Confrey, J. (1993). Learning to see children’s mathematics: Crucial challenges in
constructivist reform. In K. Tobin (Ed.), The practice of constructivism in science education
(pp. 299–321). Washington, DC: AAAS Press.
Journal of Research in Science Teaching. DOI 10.1002/tea
216 POTGIETER, HARDING, AND ENGELBRECHT

Cormier, S.M., & Hagman, J.D. (Eds.). (1987). Transfer of learning: Contemporary research
and applications. New York: Academic Press.
Dreyfus, T., & Eisenberg, T. (1982). The function concept in college students: Linearity,
smoothness and periodicity. Focus on Learning Problems in Mathematics, 5, 119–132.
Forbus, K.D., Gentner, D., & Law, K. (1995). MAC/FAC: A model of similarity-based
retrieval. Cognitive Science, 19, 41–205.
Gabel, D. (1998). The complexity of chemistry and implications for teaching. In B.J. Fraser &
K.G. Tobin (Eds.), International handbook of science education (pp. 233–248). Boston, MA:
Kluwer Academic Publishers.
Gentner, D., Rattermann, M.J., & Forbus, K.D. (1993). The roles of similarity in transfer:
Separating retrievability and inferential soundness. Cognitive Psychology, 25, 524–575.
Greeno, J.G., & Hall, R.P. (2004). Practicing representation learning with and about
representational forms. Online journal Phi Delta Kappan. Retrieved 3 December 2005 from
http:/ /www.pdkintl.org/kappan/k9701gre.htm
Hrvey, L. (1990). Systèmes de productions et interaction humain-ordinateur. Thèse de
doctorat, Université Laval, Québec.
Harvey, L., & Anderson, J.R. (1996). Transfer of declarative knowledge in complex
information processing domains. Human–Computer Interaction, 11, 69–96.
Huber, M.T., & Morreale, S.P. (2002). Situating the scholarship of teaching and learning: A
cross-disciplinary conversation. In M.T. Huber & S.P. Morreale (Eds.), Disciplinary styles in the
scholarship of teaching and learning: Exploring common ground (pp. 1–24). Washington, DC:
Carnegie Foundation for the Advancement of Teaching and the American Association for Higher
Education.
Hughes-Hallett, D. (2001). Achieving numeracy: The challenge of implementation. In L.A.
Steen (Ed.), Mathematics and democracy: The case of quantitative literacy (pp. 93–98).
Washington, DC: The National Council on Education and the Disciplines.
Keig, P.F., & Rubba, P.A. (1993). Translation of representations of the structure of matter and
its relationship to reasoning, gender, spatial reasoning, and specific prior knowledge. Journal of
Research in Science Teaching, 30, 883–903.
Knuth, J.E. (2000). Student understanding of the Cartesian Connection: An exploratory study.
Journal of Research in Mathematics Education, 31, 500–508.
Kotz, J.C., Treichel, P.M., & Weaver, G.C. (2006). Chemistry and chemical reactivity (6th
ed.). Belmont: Brooks/Cole-Thomson Learning.
Kozma, R.B., & Russell, J. (1997). Multimedia and understanding: Expert and novice
responses to different representations of chemical phenomena. Journal of Research in Science
Teaching, 34, 949–968.
Larkin, J. (1981). Understanding and problem solving in physics. In J. Robinson (Ed.),
Research in science education: New questions, new directions (pp. 115–130). Louisville: Center
for Educational Research and. Evaluation.
Lave, J. (1988). Cognition in practice: Mind, mathematics and culture in everyday life.
Cambridge: Cambridge University Press.
Leinhardt, G., Zaslavsky, O., & Stein, M.K. (1990). Functions, graphs, and graphing: Tasks,
learning, and teaching. Review of Educational Research, 60, 1–64.
Lemke, J.L. (1998). Multiplying meaning: Visual and verbal semiotics in scientific text. In
J.R. Martin & R. Veel (Eds.), Reading science (pp. 87–113). London: Routledge.
Loewenstein, J., Thompson, L., & Gentner, D. (1999). Analogical encoding facilitates
knowledge transfer in negotiation. Psychonomic Bulletin & Review, 6, 586–597.
Marton, F., & Booth, S. (1997). Learning and awareness. New York: Erlbaum.
Journal of Research in Science Teaching. DOI 10.1002/tea
 MATHEMATICS IN CHEMISTRY 217

McCormick, R. (1997). Conceptual and procedural knowledge. International Journal of

Technology and Design Education, 7, 141–159.
McDermott, L.C., Rosenquist, M.L., & Van Zee, E.H. (1987). Student difficulties in connecting
graphs and physics: Examples from kinematics. American Journal of Physics, 55, 503–513.
Meltzer, D.E. (2002). The relationship between mathematics preparation and conceptual
learning gains in physics: A possible ‘‘hidden variable’’ in diagnostic pretest scores. American
Journal of Physics, 70, 1259–1268.
Mevarech, Z.R., & Kramarsky, B. (1997). From verbal descriptions to graphic representa-
tions: Stability and change in students’ alternative conceptions. Educational Studies in
Mathematics, 32, 229–263.
Mokros, J.R., & Tinker, R.F. (1987). The impact of microcomputer-based labs on children’s
ability to interpret graphs. Journal of Research in Science Teaching, 24, 369–383.
Moschkovich, J., Schoenfeld, A.H., & Arcavi, A. (1993). Aspects of understanding: On
multiple perspectives and representations of linear relations and connections among them. In T.A.
Romberg, E. Fenemma, & T.P. Carpenter (Eds.), Integrating research on the graphical
representation of functions (pp. 69–100). New York: Erlbaum.
Mousoulides, N., & Gagatsis, A. (2004). Algebraic and geometric approach in function
problem solving. Proceedings of the 28th conference of the international group for the psychology
of mathematics education. (Vol. 3, pp. 385–392). Bergen, Norway: Program Committee of the
28th PME Conference.
National Council of Teachers of Mathematics. (1989). Curriculum and evaluation standards
for school mathematics. Reston: Author.
Niklasson, C., Christie, M., Larsson, S., Öhrströ, L., & Bowden, J. (2003). Integration of
mathematics/numeric analysis with chemistry/chemical engineering. Retrieved on 6 December
2005 from http:/ /www.math.chalmers.se/stig/papers/JEE%20paper%20031101.pdf
Oxtoby, D.W., Freeman, W.A., & Block, T.F. (2003). Chemistry: The science of change (4th
ed.). Belmont: Brooks/Cole-Thomson Learning.
Peirce, C.S. (1955). Logic as semiotic: The theory of signs. In J. Buchler (Ed.), Philosophical
writings of Peirce 1893–1910 (pp. 98–119 ) [Reprint]. New York: Dover.
Ross, B.H. (1989a). Distinguishing types of superficial similarities: Different effects on the
access and use of earlier problems. Journal of Experimental Psychology: Learning, Memory and
Cognition, 15, 456–468.
Ross, B.H. (1989b). Remindings in learning and instruction. In S. Vosniadou & A. Ortony
(Eds.), Similarity and analogical reasoning (pp. 438–469). New York: Cambridge University
Press.
Scaife, M., & Rogers, Y. (1996). External cognition: How do graphical representations work?
International Journal of Human–Computer Studies, 45, 185–213.
Schliemann, A. (1995). Some concerns about bringing everyday mathematics to mathematics
education. In L. Meira & D. Carraher (Eds.), Proceedings of the 19th international conference for
the psychology of mathematics education (Vol. III, pp. 230–237). Assisi, Italy: Program
Committee of the 15th PME Conference.
Scholtz, J., & Wiedenbeck, S. (1989). Learning second and subsequent programming
languages: A problem of transfer. Technical report #80 of Department of Computer Science and
Engineering, University of Nebraska.
Sfard, A. (1992). Operational origins of mathematical objects and the quandary of
reification—The case of function. In G. Harel & E. Dubinsky (Eds.), The concept of function:
Aspects of epistemology and pedagogy (MAA Notes, 25, pp. 59–84). Washington, DC:
Mathematical Association of America.
Journal of Research in Science Teaching. DOI 10.1002/tea
218 POTGIETER, HARDING, AND ENGELBRECHT

Silberberg, M.S. (2006). Chemistry: The molecular nature of matter and change (4th ed.).
New York: McGraw-Hill.
Singley, M.K., & Anderson, J.R. (1988). A keystroke analysis of learning and transfer in text-
editing. Human–Computer Interaction, 3, 223–274.
Singley, M.K., & Anderson, J.R. (1989). The transfer of cognitive skill. Cambridge, MA:
Harvard University Press.
Skoog, D.A., Holler, F.J., & Nieman, T.A. (1998). Principles of instrumental analysis (5th
ed.). Philadelphia, PA: Saunders College Publishing.
Testa, I., Monroy, G., & Sassi, E. (2002). Students’ reading images in kinematics: The case of
real-time graphs. International Journal of Science Education, 24, 235–256.
Vinner, S. (1983). Concept definition, concept image and the notion of function. International
Journal for Research in Mathematics Education, 20, 356–366.
Wainer, H. (1992). Understanding graphs and tables. Educational Researcher, 21, 14–23.
Witten, G.Q. (2005). Designing a mathematics course for chemistry and geology students.
Educational Studies in Mathematics, 58, 1–19.
Wright, M., & Chorin, A. (2000). Mathematics and Science. National Science Foundation
Report. Retrieved on 6 December 2005 from http:/ /www.nsf.gov/pubs/2000/mps0001/
mps0001.txt
Wu, H.-K., Krajcik, J.S., & Soloway, E. (2001). Promoting understanding of chemical
representations: Students’ use of a visualization tool in the classroom. Journal of Research in
Science Teaching, 38, 821–842.
Yerushalmy, M., & Schwartz, J.L. (1993). Seizing the opportunity to make algebra
mathematically and pedagogically interesting. In T.A. Romberg, E. Fennema, & T.P. Carpenter
(Eds.), Integrating research on the graphical representation of functions (pp. 41–68). Hillsdale:
Erlbaum.

Journal of Research in Science Teaching. DOI 10.1002/tea