Reflective, Systemic and Analytic Thinking in Real Numbers

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

Educ Stud Math (2013) 82:5–22

DOI 10.1007/s10649-012-9413-y

Reflective, systemic and analytic thinking in real numbers

Theodossios Zachariades & Constantinos Christou & Demetra Pitta-Pantazi

Published online: 25 May 2012


# Springer Science+Business Media B.V. 2012

Abstract The aim of this paper is to propose a theoretical model to analyze prospective
teachers’ reasoning and knowledge of real numbers, and to provide an empirical verification
of it. The model is based on Sierpinska’s theory of theoretical thinking. Data were collected
from 59 prospective teachers through a written test and interviews. The data indicated that
mathematical tasks on real numbers, based on Sierpinska’s theory, could be categorized
according to whether they require reflective, systemic or analytic thinking. Analysis of the
data identified three different groups of prospective teachers reflecting different types of
theoretical thinking about real numbers. The interviews confirmed the empirical data from
the written test, and provided a better insight into the thinking and characteristic features of
the prospective teachers in each group. The analysis also indicated that the participants were
more successful in tasks requiring systemic and analytic thinking, and only when this was
achieved were they able to solve problems which required reflective thinking. Implications
for teaching related to the findings of the study are discussed.

Keywords Real number . Theoretical thinking . Reflective . Systemic . Analytic

1 Introduction

As is pointed out in a number of curricula, students’ understanding of number is significant


for the development of understanding of other strands of mathematics (NCTM, 2006).
However, much of the research about students’ understanding of the number system shows
that the concepts of rational, irrational and real numbers are difficult to grasp (Fischbein,
Jehiam, & Cohen, 1995; Vamvakoussi & Vosniadou, 2007). The main reason, according to
Fischbein et al. (1995), is that school mathematics does not emphasize the idea of

T. Zachariades
Department of Mathematics, University of Athens, Panepistimiopolis, 15784 Athens, Greece

C. Christou : D. Pitta-Pantazi (*)


Department of Education, University of Cyprus, P. O. Box 20537, 1678 Nicosia, Cyprus
e-mail: dpitta@ucy.ac.cy
6 T. Zachariades et al.

mathematics as a coherent, structurally organized body of knowledge. As a consequence,


less emphasis is put on theoretical thinking, although this is necessary for development of
the concept of real numbers. In this paper, we propose a theoretical model based on the
features described in Sierpinska’s model of theoretical thinking (2000, 2005). The model
constitutes our attempt to encompass the whole spectrum of students’ understanding of
rational, irrational and real numbers. Furthermore, the study provides empirical verification
of the proposed model, and traces the different types of thinking projected by prospective
teachers in the context of theoretical thinking about real numbers.
To this end, Section Two offers an overview of previous research and the theoretical
perspectives relating to number systems. Sierpinska’s (2000, 2005) model of theoretical
thinking, used for the analysis of the data in this study, is presented in detail, and its
relevance to the number system is discussed. In Section Three, we present the purpose
and aims of the study, together with a description of the participants, tasks, and statistical
analysis involved. The results are presented in Section Four, whilst findings and implications
for educational planning and teaching are discussed in Section Five.

2 Theoretical background

2.1 Real numbers

Recent research studies have shown that students and teachers experience difficulties with
numbers, and especially with rational and irrational numbers (Fischbein et al., 1995;
Giannakoulias, Souyoul, & Zachariades, 2007; Sirotic & Zazkis, 2007; Tirosh, Fischbein,
Graeber, & Wilson, 1998; Vamvakoussi & Vosniadou, 2004, 2007; Zazkis & Sirotic, 2004).
Despite the importance of real numbers, only a small number of research papers
have focused on this topic. Most of these emphasized the deficiencies in students’ and
teachers’ understanding of irrational numbers. Fischbein et al. (1995) hypothesized
that students’ difficulties with understanding number systems were the result of two
intuitive obstacles: the incommensurability of irrational magnitudes and the non-
denumerability of the set of real numbers. Other research studies have shown that
students’ and prospective teachers’ difficulties may also arise from the discreteness of
natural numbers, which is a barrier to understanding the dense structure of the rational
and real numbers (Malara, 2001; Merenluoto & Lehtinen, 2002, 2004; Neumann,
1998; Tirosh et al., 1998; Vamvakoussi & Vosniadou, 2004, 2007).
Several research papers by Zazkis and Sirotic (Sirotic & Zazkis, 2007; Zazkis & Sirotic,
2004) report a systematic study of irrational numbers from many perspectives: For instance,
Sirotic and Zazkis (2007) investigated prospective secondary mathematics teachers’ under-
standing of irrational numbers, examining their intuitions and beliefs about the relationship
between rational and irrational numbers. The authors found that there were inconsistencies
between participants’ intuitions and their formal, algorithmic knowledge regarding the
density of numbers and the operations among the elements of rational and irrational
numbers. Previously, these authors had examined the influence of different representations
of irrational numbers on prospective teachers’ understanding of irrationality (Sirotic &
Zazkis, 2004; Zazkis, 2005). They suggested that prospective teachers mainly relied on
calculators, showing a preference for the decimal rather than the fractional representation in
determining whether a number is rational or irrational. This is mainly due to the fact that
students fail to grasp the fact that rational numbers can be represented in multiple ways, with
different symbolic representations referring to the same mathematical entity; the latter, of
Reflective, systemic and analytic thinking in real numbers 7

course, requires the ability to move flexibly among different symbolic representations of
rational numbers. It also reinforces Vamvakoussi and Vosniadou’s (2004) results, providing
evidence that students, in their attempts to describe the structure of rational numbers, treat
different symbolic representations as if they were different numbers. More recently, Sirotic
and Zazkis (2007) focused on representations of irrational numbers on the number line, and
found that prospective teachers had difficulty in dealing with the concept of density,
especially with the idea that every number corresponds to a single point on the number line.
The main focus of research about real numbers, then, seems to be on concepts of density,
representations of rational and irrational numbers, and recognition of rational and irrational
numbers. However, all the above studies investigate these concepts in isolation. In the
present study we attempt to link the above concepts in one theoretical model.
Another important aspect of the research on number systems is the variety of theoretical
perspectives that were used. A number of authors (Merenluoto & Lehtinen, 2002, 2004;
Vamvakoussi & Vosniadou, 2004, 2007) used the conceptual change framework extensively,
and pointed out that the shift in students’ thinking about numbers is a gradual and time
consuming process. They argued that the shift from natural to rational and real numbers is
not an add-on process, and that wider explanatory frameworks for numbers cannot simply be
accomplished by enrichment of the initial frameworks. On the contrary, it requires a
reconstruction of students’ knowledge about the structure of number concepts.
Other authors used the theory of procedural and conceptual knowledge as a theoretical
framework, as well as the idea that mathematical knowledge is embedded in a set of
connections among algorithmic, intuitive and formal dimensions (Fischbein, 1987). For
example, Tirosh et al. (1998) argued that students’ and prospective elementary school
teachers’ conceptions of numbers were almost entirely based on natural numbers. This
allowed prospective teachers to operate partially on the algorithmic level, but blocked their
ability to cope with tasks that demanded formal or intuitive knowledge of rational numbers.
Sirotic and Zazkis (2007) emphasized the connections between algorithms, intuitions and
formal knowledge. They indicated that underdeveloped intuitions are related to weaknesses
in formal knowledge, and to a lack of algorithmic experience.
Finally, the theory of representations was used extensively as a theoretical perspec-
tive to explain the difficulties and inconsistencies in students’ and prospective teach-
ers’ knowledge about irrationality. Sirotic and Zazkis (2004) used the distinction
between transparent and opaque representations, and suggested that the connections
between decimal and other representations of a number should be emphasized during
teaching. Lack of understanding about the equivalence of different representations of
numbers can lead to inconsistencies in knowledge, and many students and prospective
teachers are unable to identify correctly various examples of numbers as being whole,
rational, irrational, or real numbers (Arcavi, Bruckheimer, & Ben-Zvi, 1987; Fischbein
et al., 1995). The belief that different symbolic representations stand for different
numbers may have a further implication for students’ thinking about the structure of
the rational numbers: for example, they may consider fractions and decimals to be
different, unrelated “sets” of numbers (Vamvakoussi & Vosniadou, 2007).
What is new in the present study is the use of a broader theoretical perspective, more
inclusive and more comprehensive than those employed in earlier studies. We try to
incorporate several aspects of mathematical thinking, previously only investigated in an
isolated fashion, which may allow us to investigate possible interrelationships. We used the
distinctions between theoretical thinking components introduced by Sierpinska (2005).
Sierpinska postulated features of theoretical thinking, naming the main categories “reflec-
tive,” “systemic” and “analytic” thinking. In the following section, we present some
8 T. Zachariades et al.

elements of these categories of thinking; and then explain how we interpret them in the
context of real numbers. Our theoretical perspective is articulated and tested with secondary
school graduates studying for a Bachelor Degree in Education, in order to validate some
aspects of it.

2.2 Categories of theoretical thinking

Sierpinska et al. (Sierpinska, Nnadozie, & Oktaç, 2002; Sierpinska, 2005) studied theoretical
thinking and its relevance to linear algebra. We modified these ideas in such a way as to
reflect the perspective we used in explaining students’ theoretical thinking while they were
working on tasks involving the number system.
According to Sierpinska (2005), theoretical thinking is thinking where thought and its
objects belong to distinct planes of action, and whose purpose is the production of internally
coherent conceptual systems based on specially created systems of signs. Theoretical
thinking therefore, is reflective, systemic and analytic (Sierpinska, 2005). Reflective think-
ing is “thinking for the sake of thinking” (Sierpinska et al., 2002, p. 17). In particular,
“theoretical thinking is reflective in that it does not take such techniques or procedures for
granted but considers them always open to questioning and change” (Sierpinska, 2005, p.
122). She notes that “theories do not grow by simple addition of new concepts, but that new
developments may cause a restructuring of the whole system. The system is always reflected
upon as a whole. This feature of theoretical thinking is sometimes called ‘reflexivity’”
(Sierpinska, 2005, p. 122). Therefore, an individual expressing reflective thinking may
reflect back on his/her own solution, seek different approaches, and notice relationships
with previously solved problems (Sierpinska, Bobos, & Pruncut, 2011).
In the case of numbers, we hypothesized that students who exhibit reflective thinking
participate in the construction of the meaning of real numbers, and require constant
justification and verification. The study of number systems is the result of reflection on
existing properties, aiming at generalization within the frame of a unified theory of number
systems. The construction of such a theory is not the result of improving students’ existing
knowledge structures, but principally comes about through the reorganization of their
knowledge. Some of the studies mentioned previously (Merenluoto & Lehtinen, 2004;
Sirotic & Zazkis, 2007; Tirosh et al., 1998; Vamvakoussi & Vosniadou, 2007) investigated
aspects of reflective thinking in the case of number. In particular, these studies explored their
participants’ understanding of the dense structure of real numbers where they needed to
restructure their previous knowledge. For example, in Sirotic and Zazkis’ study (2007),
prospective secondary mathematics teachers were asked to determine the correctness of the
statement that it was always possible to find a rational number between any two rational
numbers, and explain their thinking. In another study by Vamvakoussi and Vosniadou
(2007), ninth and eleventh grade students were asked to specify if there were any numbers
greater than one number they were given (e.g., 0.005) while at the same time less than
another (e.g., 0.006), and to provide examples of such numbers.
From an epistemological perspective, the conditions of a definition are sometimes the
result of a process of categorization in a domain of existing concepts. We assume that, in
theoretical thinking, this categorization is “systemic,” that is it has a “key” or well-defined
feature, which serves as a basis for separating concepts into disjoint classes (Sierpinska et al.,
2002). Therefore, “systemic thinking is thinking about systems of concepts, where the
meaning of a concept is established based on its relations with other concepts and not with
things or events” (Sierpinska et al., 2002, p. 35). Sierpinska et al. (2011) suggest that
systemic thinking is based on definitional and hypothetical thinking. Definitional thinking
Reflective, systemic and analytic thinking in real numbers 9

refers to concepts that are defined by other concepts within a system, and a statement is true
by means of a proof. These proofs rely on accepted definitions, conceptual and logical
relations within a system. Hypothetical thinking refers to being aware of assumptions and of
conditional mathematical statements.
The definitions of the subsets of real numbers (natural, integers, rational and irrational)
are the outcomes of a process of categorization in a domain of existing concepts. This
process, in secondary education, does not come after the construction of natural numbers
because it is defined within advanced mathematics. Students define integers as the additive
inverses of natural numbers, and rational numbers as fractions of integers. Irrational numbers
are then defined to be numbers which are not rational, and real numbers by reference to
rational and irrational numbers. In other words, these number systems are determined by
explicitly defined properties; the lack of this knowledge explains students’ difficulties with
real number tasks (Fischbein et al., 1995; Vamvakoussi & Vosniadou, 2007). When students
are advised to think about the meanings of these number systems, they rarely think about the
definitions that these concepts satisfy. They usually think about particular examples, and
this, of course, cannot lead them to producing generalizations. Students who exhibit
systemic thinking provide coherent definitions for natural, rational, irrational and real
numbers, and can justify their claims since they rely on acceptable definitions. However,
this categorization in itself does not ensure a conceptual understanding of real numbers.
Some students simply cannot grasp the real numbers as a whole systems and the relation-
ships and differences between the various subsets of this number system. Furthermore, they
do not seem to reflect on their solutions in order to improve them. In the latter case, reflective
theoretical thinking is necessary in order that students change the whole system in a
consistent and unified way. Sirotic and Zazkis (2007) studied aspects of systemic thinking
in the case of numbers. They asked prospective mathematics teachers to recognize which set,
rational or irrational numbers, was “richer,” and to explain their answer. The participants
responded to this question based on the definitions of these two number systems.
Analytic thinking entails the development of specialized representational systems such as
signs, symbolic notations and specialized terminology. If the meaning of signs is derived
from their relations with other signs, then this meaning is based on certain explicit con-
ventions and not on, say, some “natural” resemblance or contiguity between signs and their
intended objects. However, one convention can always be replaced by another. This
creates a distance between signs and objects in theoretical thinking: the relation is
indirect, mediated by language. This is what we mean by an “analytic approach to
signs,” which we shall sometimes refer to as “analytic thinking,” for the sake of brevity
(Sierpinska et al., 2002, pp. 28–29).
Sierpinska states that “an analytic approach to signs is extremely relevant for the learning
of linear algebra, because linear algebra could be seen, in fact, as all about languages and
expressing the same thing in many different ways” (Sierpinska et al., 2002, pp. 30–31), and
we believe that the same holds in the case of numbers. Number systems appear to develop
concurrently with methods for representing and manipulating them symbolically (Zazkis &
Sirotic, 2004). Analytic thinking is relevant to the learning of number systems, because in
identifying rational numbers, for example, students need to appreciate the different repre-
sentations of rational numbers. Different notation or terminology often implies the same
concept. For example, a rational number is a number that can be expressed as the ratio of two
integers, like 2/5, -2/5, etc. Alternatively, a rational number can be expressed either as a
simple decimal or as a repeating decimal, for example 0.23, 0.4, 0.3757575…, 0.333…, etc.
A number of researchers (Zazkis, 2005; Zazkis & Sirotic, 2004) studied various aspects of
analytic thinking in the case of numbers. For instance, Zazkis and Sirotic (2004) examined
10 T. Zachariades et al.

the impact of different representations of numbers, and prospective teachers’ ability to


recognize whether they were rational or irrational numbers.

3 The present study

3.1 The purpose of the study

The aim of this study was to explore whether, by using Sierpinska’s (Sierpinska et al., 2002;
Sierpinska, 2005) model of theoretical thinking, we could analyze prospective teachers’
reasoning and knowledge of real numbers. We hypothesized that the concept of real numbers
has distinct aspects that represent reflective, systemic and analytic theoretical thinking. We
investigated the relationship between these three components (reflective, systemic and
analytic) as it unfolded through prospective elementary school teachers’ responses to tasks
involving rational, irrational and real numbers. For this reason:

a. we investigated whether different groups of tasks in the context of number systems


could be used to explore prospective teachers’ reflective, systemic and analytic thinking;
and
b. we examined whether distinct groups of prospective teachers could be identified to
provide evidence of reflective, systemic and analytic thinking, and to investigate the
structure of, and relationships between, these three components of theoretical thinking.

3.2 Participants and tasks

The participants in this study were 59 first year students who had taken mathematics as a
major subject at secondary school, and who were studying at the University of Cyprus to
become elementary school teachers. These prospective teachers had been taught about real
numbers during their studies in secondary school in the following order. Instruction begun
with natural numbers, then proceeded to fractions with positive terms, then positive decimals
(including repeating decimals), followed by negative numbers. Finally, the students were
introduced to rational numbers as the set including all the aforementioned numbers. Once
this was accomplished, students were presented with the idea that there are numbers which
could not be included in the set of rational numbers, and that these are called irrational
numbers.
Data reported in this paper were collected through a written test and interviews. The test
was administered during the participants’ mathematics course towards the beginning of their
first semester. The duration of the test was 1 h. Interviews were conducted after quantitative
analysis of the data was completed.
The test included 25 tasks (see Table 1). Tasks T1 and T2 investigated the
participants’ understanding of the density of rational and real numbers with specific
examples. Tasks T3 to T7 addressed conceptual understanding of density, or concep-
tual understanding of different representations of real numbers. These tasks required a
restructuring of the participants’ pre-existing perceptions of the number system. Thus,
we assumed that, for a participant to be able to respond to this group of tasks,
reflective thinking was required (Sierpinska, 2005). Tasks T8 to T15 were intended to
explore participants’ ability to recognize and define rational and irrational numbers,
identifying the particular features that serve as a basis for separating these numbers
into disjoint classes. In particular, Tasks T8 to T15 referred to the recognition using
Reflective, systemic and analytic thinking in real numbers 11

the definition of rational numbers as repeating decimals (T8-T9), as decimals (T10-T11), as


fractions (T14-T15) and to the recognition of square roots of non-square numbers as irrational
numbers (T12-T13). We assumed that the ability to solve a large proportion of this group of
questions (T8-T15) required systemic thinking (Sierpinska et al., 2002). Finally, Tasks T16 to
T25 focused on different representations, such as representation of repeating decimals as
fractions (T16-T17), representation of decimals as fractions (T18-T19), decimal representation
of surds (square roots of natural numbers that are not perfect squares) (T20-T22), and repre-
sentation of fractions as decimals (T23-T25). We assumed that analytic thinking was required to
respond to these tasks.
Some of the tasks included in the test intentionally referred to similar things (for example,
whether a certain number is rational or irrational). This was in order to check the persever-
ance of participants’ knowledge. Previous studies showed that some students believe that
rational and irrational numbers are not distinct sets, or that there are numbers with no
decimal representation (Giannakoulias et al., 2007).
In the present study, we hypothesized that the aforementioned groups of tasks
constituted an a-priori model based on Sierpinska’s theory (see Fig. 1). The a-priori
model consists of three second order factors (reflective thinking, systemic thinking
and analytic thinking). Reflective thinking is composed of two first order factors that
refer to participants’ understanding of density (Density) and their conceptual under-
standing (Conceptual). Systemic thinking is composed of four first order factors that
reflect participants’ ability to recognize and define number systems depicted as
repeating decimals (Def_rep), in decimal form (Def_dec), square root form (Def_root),
and simple fractional form (Def_fract). Finally, analytic thinking is composed of four
first order factors, which purport to identify participants’ ability to represent repeating
decimals as simple fractions (Re_dec_rep), decimals as simple fractions (Re_dec),
square roots as decimals (Re_root), and simple fractions as decimals and repeating
decimals (Re_fract). The questions for each of the components of theoretical thinking
are presented in Table 1.

Table 1 Items representing reflective, systemic and analytic thinking

Reflective thinking
Density
T1: Write a number, if there is one, which is between 0.999 and 0.999…Explain your answer.
3 4
T2: Is there a rational number between 5 and 5 ? If yes, write it. Explain your answer.
Conceptual
T3: Can you find two real numbers so that between them there is no other real number? If yes, write them.
Explain your answer.
T4: One of the properties of natural numbers is the following: “If n is a natural number, the next
natural number is the number n+1. In other words, n<n+1 and between n and n+1 there is no
other natural number”. Does this property hold for rational numbers? Explain your answer.
T5: Is the following statement correct or not? Explain your answer. “Every real number can be written as a
repeating decimal”.
T6: Is the following statement correct or not? Explain your answer. “Each number which can be written in
decimal form is a real number”.
T7: Is the following statement correct or not? Explain your answer. “Every real number can be written as a
fraction”.
Systemic Thinking
Recognition and definition of rational numbers involving repeating decimals (Def_rep)
T8: Is 0.7777… a rational number? Explain your answer
12 T. Zachariades et al.

Table 1 (continued)
T9: Is 0.7777… a real number? Explain your answer
Recognition and definition of rational, real numbers involving decimals (Def_dec)
T10: Is the following statement correct or not? Explain your answer. “The number 3.14 is a rational number”.
T11: Is the following statement correct or not? Explain your answer. “The number 3.14 is an irrational number”.
Recognition and definition of irrational numbers involving square roots (Def_root)
T12: Is the following statement correct or not? Explain your answer. “The number √2 is a rational number”.
T13: Is the following statement correct or not? Explain your answer. “The number √2 is an irrational
number”.
Recognition and definition of rational numbers involving fractions (Def_fract)
T14: Is the following statement correct or not? Explain your answer. “The number 7
9 is a rational
number”.
T15: Is the following statement correct or not? Explain your answer. “The number 7
9 is an irrational
number”.
Analytic Thinking
Representation of repeating decimals as fractions (Re_dec_rep)
T16: Is the following statement correct or not? Explain your answer. “The number 0.7777… can be written as a
fraction”.
T17: Is the following statement correct or not? Explain your answer. “The number 0.7777… cannot be written
as a fraction”.
Representation of decimals as fractions (Re_dec)
T18: Is the following statement correct or not? Explain your answer. “The number 3.14 can be written as a fraction”.
T19: Is the following statement correct or not? Explain your answer. “The number 3.14 cannot be written as a
fraction”.
Decimal representation of square roots (Re_root)
T20: Is the following statement correct or not? Explain your answer. “The number √2 can be written as a
decimal with finite decimal digits”.
T21: Is the following statement correct or not? Explain your answer. “The number √2 can be written as a
decimal with an infinite number of decimal digits”.
T22: Is the following statement correct or not? Explain your answer. “The number √2 cannot be written as a
decimal”.
Decimal representation of fractions (Re_fract)
T23: Is the following statement correct or not? Explain your answer. “The number 7
9 can be written as a
decimal with finite decimal digits”.
T24: Is the following statement correct or not? Explain your answer. “The number 7
9 can be written as a
decimal with an infinite number of decimal digits”.
T25: Is the following statement correct or not? Explain your answer. “The number 7
9 cannot be written as a
decimal”.

3.3 Scoring and analysis

Correct responses were marked with 1, and incorrect ones with 0. If a participant gave a
partly correct response, for example if he/she gave a correct answer but the justification was
wrong, this again was marked with 0.
To fulfill the aims of the study, both quantitative and qualitative techniques were
used to analyze the data. It was our assumption that a combination of both types of
methods would give a clearer picture of the data (Richardson, 2001). First, to
investigate whether the model of theoretical thinking which was defined a-priori fitted
our data, we used confirmatory factor analysis (CFA). CFA is part of a more general
class of approaches called structural equation modelling (SEM). Generally, SEM
Reflective, systemic and analytic thinking in real numbers 13

represents “the translation of a hypothesized series of cause–effect relationships


between variables into a composite statistical hypothesis concerning patterns of sta-
tistical dependencies” (Shirpley, 2000, p. 206). So, “hypothesized relationships are
translated into mathematical models and a researcher can use SEM methodology to
quantify and test a wide variety of proposed models” (Marcoulides & Kyriakides,
2010, p. 278). In particular, CFA is appropriate in situations where, due to previous
research, the factors of a set of variables for a given population are known. It then
follows that the proposed theory can be tested with empirical data. Furthermore, CFA
can be used to examine construct validity and “evaluate the extent to which particular
instruments actually measure one or more latent variables they are supposed to assess”
(Marcoulides & Kyriakides, 2010, p. 279). In the present study, CFA was applied to test
hypotheses corresponding to Sierpinska’s theoretical conceptualization of what constitutes
prospective teachers’ mathematical reasoning and knowledge of real numbers. Our task was
not to determine the factors of a set of variables or to find the pattern of the factor loadings.
Instead, our purpose in using CFA was to investigate whether the model which was defined a-
priori (see Fig. 1) could be confirmed by our empirical data. To be precise, CFA was used to test
whether the three groups of questions (T1-T7, T8-T15 and T16-T25) were discrete factors, as
we expected, corresponding to reflective, systemic and analytic thinking respectively. An

Fig. 1 The proposed model. Def_rep: Recognition and definition of rational numbers involving repeating
decimals; Def_dec: Recognition and definition of rational numbers involving decimals; Def_root: Recognition
and definition of irrational numbers involving square roots; Def_fract: Recognition and definition of rational
numbers involving fractions; Re_dec_rep: Representation of repeating decimals as fractions; Re_dec: Repre-
sentation of decimals as fractions; Re_root: Decimal representation of square roots; Re_fract: Decimal
representation of fractions
14 T. Zachariades et al.

electronic structural equation modeling program, MPLUS (Muthén & Muthén, 1998), was used
to test for model fitting in this study. The model was estimated by using normal theory
maximum likelihood methods (ML). To evaluate model fit, more than one fit index was used:
the chi-square to its degree of freedom ratio (x²/df), the comparative fit index (CFI), and the root
mean-square error of approximation (RMSEA) (Marcoulides & Schumacker, 1996). The
observed values of x²/df should be less than 2, the values for CFI should be higher than 0.9,
and the RMSEA values should be close to zero.
To fulfill the aims of the study, clinical interviews and latent class analysis (LCA) were used
to trace whether there were groups of prospective teachers who exhibited reflective, systemic
and analytic thinking. LCA is one of the Mixture Modelling techniques, which aim to find
groups of people who have similar responses in specific variables from an unobserved
heterogeneity population (Muthén, 2004). In particular, LCA is a statistical method for finding
subtypes of related cases (latent classes) from multivariate data. The results of LCA can also be
used to classify cases according to their most likely latent class.
For this study, semi-structured clinical interviews were conducted. Participants were
presented with questions that they had addressed in the written test, and were asked to
clarify and justify their responses. In other words, they were given the opportunity in the
interview to expand on their solutions, or change their initial responses. These interviews
gave us the opportunity to further investigate which type (or types) of thinking led the
participants to respond in the way they did. They also allowed us to confirm our initial
premise that the three groups of tasks require reflective, systemic and analytic thinking.

4 Results

Our results are presented in relation to the aims of the study. First of all, we examined the
validity of the a-priori model (see Fig. 1), and then we categorized multiple groups of
prospective teachers according to the way in which they had responded to the real number tasks.

4.1 The distinctive nature of the components of theoretical thinking

Three factors, the reflective, the systemic and the analytic, should be sufficient to model the
participants’ performance in the tasks addressed. Figure 2 represents the model which best
describes Sierpinska’s theoretical model (2002). Confirmatory factor analysis (CFA) was used
to evaluate the construct validity of the model. CFA showed that each of the tasks employed in
the present study loaded adequately (i.e., they were statistically significant, since z values were
greater than 1.96) on each factor, as shown in Fig. 2. CFA also confirmed that the a-priori model
matched the data set of the present study and determined the “goodness of fit” of the factor
model (CFI00.921, x2 0415.470, df0267, x2/df01.55, RMSEA00.07), indicating that the
reflective, the systemic and the analytic components do represent three distinct functions of
prospective teachers’ thinking.

4.2 Groups of prospective teachers

The second aim of the study concerns the extent to which prospective teachers in the sample differ
according to the answers they provided in the test. We examined whether there were different
types of prospective teachers in our sample who could reflect the components of theoretical
thinking. Latent class analysis (Muthén & Muthén, 1998), was used to answer this question, since
it enables specification of models in which one model applies to one subset of the data, and
Reflective, systemic and analytic thinking in real numbers 15

Fig. 2 The model of theoretical thinking in real numbers. Def_rep: Recognition and definition of rational
numbers involving repeating decimals; Def_dec: Recognition and definition of rational numbers involving
decimals; Def_root: Recognition and definition of irrational numbers involving square roots; Def_fract:
Recognition and definition of rational numbers involving fractions; Re_dec_rep: Representation of repeating
decimals as fractions; Re_dec: Representation of decimals as fractions; Re_root: Decimal representation of
square roots; Re_fract: Decimal representation of fractions

another applies to another set. The modeling here used a stepwise method - that is, the model was
tested under the assumption that there are two, three, and four groups of subjects. The best fitting
model (see Muthén & Muthén, 1998) was the one involving three groups. In what follows, we
present the arguments that participants provided in the written questionnaire, and we exemplify
these arguments with excerpts from the interviews. The main purpose of this section is to describe
the participants’ responses in each of the three groups with respect to their understanding of
irrational numbers.

4.2.1 Group 3 prospective teachers

Group 3 prospective teachers’ ability to think reflectively became apparent from the number
of correct responses they provided in the tasks and the way in which they responded in the
16 T. Zachariades et al.

interviews. The characteristic feature of thinking in this group was the ability to think
reflectively while solving real number problems. These prospective teachers concep-
tually understood and explained the structure of the real number system, as well as
how rational and irrational numbers fit together, related to the density of both sets.
The purpose of Tasks 1 and 2 was to test whether a prospective teacher had achieved
what we considered to be deep understanding of the system of rational numbers. The
following excerpts from participants’ interviews clarify their thinking in solving these
two tasks:

Participant A: (answering Task 1) There is an infinite number of numbers between


0.999 and 0.999…

Interviewer: How do you know that?

Participant A: We know that 0.999… equals 1. Thus, between 0.999 and 1, there is an
infinite number of numbers.

Interviewer: Can you prove that 0.999…01?

Participant A: Yes (and she provides a proof).

Participant B: (answering Task 2) There is an infinite number of rational numbers


between the two fractions (3/5 and 4/5).

Participant C: (answering Task 2) Fractions are rational numbers and between two
rational numbers there is an infinite number of numbers.
Prospective teachers in this group did not simply answer by giving a number between the
two given numbers, but in fact voluntarily gave an answer which went beyond the requested
task. They answered consistently that between any two non-equal rational numbers there are
infinitely many rational numbers, regardless of the way they are represented (as fractions or
decimals). The participants in this group also responded to Task 3 in a sophisticated way.
Their explanations prompted the interviewer to discuss the construction of rational or
irrational numbers with them. For example, Participant A provided a detailed account of
all possible cases in justification of her answer:

Participant A: There are no such numbers because there is always a rational number
between any two irrationals; there is an irrational number between any two irrational
numbers. It is also possible to find an irrational number between any two rational
numbers, and finally it is possible to find a rational number between any two rational
numbers, as I explained in Task 1.

Interviewer: How did you explain that?

Participant A: I showed that if we have two rational numbers, then there is a midpoint
between them. We can find this midpoint by adding the numbers and dividing by 2 . . .
which yields a rational number.

Interviewer: How are you sure that there exists a rational number between two
irrational numbers?
Reflective, systemic and analytic thinking in real numbers 17

Participant A: You can add the two numbers again and divide by 2 and round the
quotient to some nearby rational number. And this number falls between the two
irrational numbers.
In a similar way, participants justified the existence of an irrational number
between two rational numbers. These responses revealed a deep conceptual under-
standing of the structure of real numbers. The latter was reaffirmed when they
exemplified their thinking during the discussion of similar tasks (e.g., Task 4). Such
responses were indications of participants’ ability to think reflectively.

4.2.2 Group 2 prospective teachers

From their responses to the tasks and during the interviews, it appears that these participants
recognized and understood the definitions and some of the characteristics of irrational
numbers, but failed in tasks that required reflective thinking. These prospective teachers
recognized rational and irrational numbers represented in decimal and fractional form using
definitions they had learned during their own education. Participant D, for example, based
her answer on the formal definition of rational numbers without being able to provide a
reasonable explanation of the structure of rational numbers:

Interviewer: You said that 0.7777… is a rational, real number. Can you explain why?

Participant D: It is, because a periodic decimal with an infinite number of digits is a


rational number. It is also a real number since the set of rational numbers is a subset
of the real numbers.

Interviewer: This is correct. However, can you show that 0.7777… is a rational
number?

Participant D: No. I cannot remember, but I am sure that 0.7777… is a rational


number.

Interviewer: What is a rational number?

Participant D: A rational number is a number that can be written as the quotient of two
integers.

Interviewer: Yes. Do you know an algorithm by which 0.7777…, a periodic infinite


decimal, may be transformed into a quotient of two integers?

Participant D: No.
Although knowledge of the definition of number sets helped participants to answer most of
the items in the questionnaire about the identification of rational and irrational numbers
correctly, the interviews revealed that this knowledge was not adequate for enabling participants
to think conceptually and flexibly about the structure and logic of rational and irrational
numbers. The latter is what distinguishes prospective teachers in Group 2 from those belonging
in Group 3. Those in Group 2, for example, had difficulty explaining why there is an infinite
number of numbers between 3/5 and 4/5…, although they knew and used the appropriate
definitions in recognizing rational and irrational numbers:
18 T. Zachariades et al.

Participant D: Both of these numbers are rational. So, there is a number between them,
such as the 3.5/5 which is equivalent to 7/10.

Interviewer: Yes, 7/10 is a number between 3/5 and 4/5. Is there any other number?

Participant D: There is an infinite number of numbers. For example, all the fractions,
which are equivalent to 7/10, such as 14/20, 21/30.

Interviewer: You said there is an infinite number of numbers . . .

Participant D: (interrupting the interviewer) Yes, there are infinite fraction numbers
that are equivalent to 7/10.

Interviewer: You mean only the equivalents to 7/10? Are there any other numbers
which are between the fractions and are not equivalent to 7/10?

Participant D: I don’t think so.


The concept of density is a very demanding one for the prospective teachers in this group.
As can be deduced from the above vignette, Participant D generated a limited response. In
fact she gave as an answer one number and its equivalents. However, this participant
changed her mind as the interview unfolded. This was probably due to the interviewer’s
probing questions, which gave her a chance to rethink the concept of density. When she was
asked to provide justification for her response to the above example, she realized that there
was an infinite number of numbers between 3/5 and 4/5:

Interviewer: Can you provide a justification for this answer?

Participant D: You can turn both fractions into decimals. So 3/500.6 and 4/500.8.
Thus, between them is 0.7. Please wait a minute . . . It could also be 0.71, 0.711 etc. I
think I was wrong. There are many numbers which are not equivalent to 7/10.
We see that in the interview, when prompted by the interviewer, Participant D changed
her representation of rational numbers. This allowed her to reflect on her answer using the
new representation she had created, and to recognize her mistake. She then proceeded to
provide a correct solution. Participant D has some characteristics of systemic thinking, but
shows no evidence of reflective thinking. She did not seem to have the capacity to reflect on
her own answer spontaneously, only doing so when prompted by the interviewer.

4.2.3 Group 1 prospective teachers

From participants’ responses, two subcategories of prospective teachers can be identified as


belonging in this group. Those in the first subcategory were not able to give a correct
definition of the concepts of rational, irrational and real numbers. For a significant number of
prospective teachers in this category, these definitions were not part of their active repertoire
of knowledge. Many prospective teachers in this group could not even correctly classify
various numbers as being rational, irrational or real. However, the most characteristic feature
of this group was that their conceptions of number were based almost entirely on their
experiences with natural numbers, as shown in the example that follows:
Reflective, systemic and analytic thinking in real numbers 19

Interviewer: Why did you say that the number 0.7777… is not a rational number?
(see Task 8).

Participant E: Because 0.7777… continues with an infinite number of 7 s. We cannot be


sure where it ends. If it continues in the same way, it is impossible to turn it into a fraction.

Participant F: (answering Task 9) This is not a real number because real numbers
should be integer numbers.

Participant G: (justifying answer in Task 21) √2 is a real number and it can be


represented as a fraction as all real numbers. . . . It may have an infinite number of digits.
Although participants in the first subcategory presented various inconsistencies in relation to
the definitions of rational, irrational and real numbers, those in the second subcategory were
able to recognize a number represented as a common fraction as being rational, and provided
correct justification for their claim. They also recognized non- repeating decimals as represen-
tations of irrational numbers, understood the relationship between fractions and their represen-
tations, and could distinguish rational and irrational numbers when represented as square roots.
Finally, the majority of those in this subcategory knew most of the representations of rational
and irrational numbers. Therefore, we can argue that the thinking of the prospective teachers in
the second subcategory has features of analytic thinking.

4.2.4 The structure of thinking of the groups of prospective teachers

The mean and standard deviation of each of the three components of theoretical thinking
(reflective, systemic and analytic) for each of the three groups are shown in Table 2. This
shows that prospective teachers in Group 3 ðMean ¼ 23:36Þ outperformed prospective
teachers in Groups 2 and 1 ( Mean ¼ 20:70, Mean ¼ 14:57, respectively) in all tasks,
while prospective teachers in Group 2 outperformed their counterparts in Group 1.
However, the percentage of success of prospective teachers in Group 1 in the
reflective and systemic tasks was below 50 %, showing that these prospective teachers
found the tasks difficult. Group 2 prospective teachers had difficulties in understand-
ing the reflective tasks, since their success percentage was lower than 50 %
ðMean ¼ 0:49Þ. However, they were successful in most of the systemic tasks (72 %), and

Table 2 Means and standard deviations of students’ performance in the test and in the reflective, systemic
and analytic tasks

Groups (N) Test Performance Reflective tasks Systemic tasks Analytic tasks

1 (N012) Mean 14.57 0.38 0.44 0.52


Std. Deviation 3.41 0.16 0.18 0.22
2 (N040) Mean 20.70 0.49 0.72 0.65
Std. Deviation 2.10 0.17 0.16 0.18
3 (N07) Mean 23.36 0.73 0.73 0.77
Std. Deviation 4.55 0.17 0.19 0.16
Total (N059) Mean 19.35 0.54 0.63 0.62
Std. Deviation 4.47 0.20 0.21 0.20
20 T. Zachariades et al.

answered more than 50 % ðMean ¼ 0:65Þ of the analytic tasks correctly. Finally, Group 3
prospective teachers seemed not only to understand the systemic (73 %) and analytic (77 %)
tasks, but also to have the ability to think reflectively (73 %).
From Table 2, it can be deduced that there is a variability in prospective teachers’
abilities to complete the assigned tasks, since success with any problem by more than
50 % of those in one group was associated with similar success by more than 50 %
of all those in subsequent groups. Thus, Group 2, which was the largest (N040), can
be considered as the systemic and analytic group, because this group included the
prospective teachers who demonstrated in both their written responses and interviews
that they could resolve the tasks that required systemic and analytic thinking. Group 1
prospective teachers (N012) seemed to have difficulties in most of the tasks, so we
could consider Group 1 as a transitional group, moving towards understanding
analytic, and to a lesser degree systemic, tasks. In summary, prospective teachers
belonging to Group 3 (N 07) can solve real number problems using reflective,
systemic and analytic thinking. What distinguishes prospective teachers in Group 2
from those in Group 3 is their inability to think reflectively. Prospective teachers in
Group 1 seem to think more analytically than systemically. However, their ability to
think in these two modes is inferior to that of the prospective teachers in the other
two groups.
The presence of a consistent trend in the difficulty level across systemic, analytic,
and reflective thinking supports the hypothesis for the existence of a specific trend.
The data imply that prospective teachers find it easiest to understand numbers in
various representations (analytic). Once this is achieved, they can then proceed to
provide definitions (systemic) of rational numbers. Reflective thinking is grasped only
after conceptualization of the systemic and analytic representations.

5 Conclusions

The primary aim of this study was to examine whether different groups of tasks in the context of
number systems could be used to explore students’ reflective, systemic and analytic thinking. It
can be deduced from the quantitative and qualitative data presented here that the three groups of
mathematical tasks may be categorized according to Sierpinska’s theory, in the sense that they
require reflective, systemic and analytic thinking.
The secondary aim of the study concerned the extent to which prospective teachers in the
sample showed evidence of reflective, systemic and analytic thinking. The mixture growth
modeling demonstrated that three different groups of prospective teachers could be identi-
fied. The interviews provided some insightful notions about the thinking and characteristic
features of participants belonging to each group. They also confirmed the hypothesis that
different components of theoretical thinking were necessary for the three groups of tasks to
be solved. Group 1 prospective teachers were able to make a partial response to the tasks that
required analytic thinking. However, most of them presented difficulties and misconcep-
tions, mainly because they were unable to implement the definitions and representations of
numbers in a proper way.
Group 2 prospective teachers were able to respond to the tasks which required systemic
and analytic thinking, while those in Group 3 were able to respond to all three types of tasks.
It was also found that if 50 % of the prospective teachers in one group succeeded in any
particular problem, then more than 50 % succeeded in all subsequent (“higher”) groups. The
presence of a consistent trend in the difficulty level across analytic, systemic and reflective
Reflective, systemic and analytic thinking in real numbers 21

tasks suggests the existence of a specific developmental trend. Analysis indicated that the
prospective teachers were more successful in carrying out systemic and analytic tasks first
and leaving reflective ones until later. Participants were only able to deal with tasks that
required reflective thinking once they had been successful with tasks only requiring analytic
and systemic thinking. Thus, it can be argued that for first year university students who have
studied mathematics as a major subject in upper grades, understanding of real numbers
progresses from analytic to systemic and then to reflective thinking. This sequence of
reasoning may be the result of the teaching of real numbers in the mathematics classroom,
or due to the nature of the particular mathematical topic. A research question arising from the
above is: Which abilities are required for the transition from analytic and systemic to
reflective thinking? The ability to use different representations of numbers in a flexible
way and to select appropriate representations to answer a question would seem to be helpful
for this transition.
The conceptual framework of theoretical thinking used in this study was applied by
Sierpinska et al. (2002) to linear algebra, and it is clear that it is applicable to other
mathematical domains as well. This, of course, requires further research. The idea was not
just to see if particular conceptions or misconceptions appeared in prospective teachers’
thinking, but to see if Sierpinska’s conceptual framework could explain the variation in
prospective teachers’ abilities to solve real number problems, and evaluate whether it could
be used as a model for future instructional practices.
A number of teaching implications arise from our findings. Teacher knowledge is a
prerequisite for developing an understanding of the challenges that the concepts of rational
and irrational numbers present to students at all levels. The development of the concept of
density appears to be a gradual process, and it seems that it can be understood by students
who can think reflectively, that is reach a level where they can use different ways to think
about fractions, decimals and representations of different number sets. They can give a larger
number of solutions to a given problem, and also more general answers. Reflective thinking
also requires students to reorganize their prior learning about natural numbers (Vamvakoussi
& Vosniadou, 2007).
While some errors made by participants in Group 1 may be due to the inherent difficulty
of the topic of irrational numbers (Sierpinska, 1994), other errors observed during interviews
with participants in Group 2 may be a direct outcome of instructional practices. The extent to
which the results of this study could be linked to educational choices and teaching practices,
both at secondary school and at university level, could be a matter for future research.
Teacher education programs may be an appropriate place for reconsidering these particular
choices and practices, as part of an overarching attempt to strengthen prospective teachers’
overall mathematical competence.

References

Arcavi, A., Bruckheimer, M., & Ben-Zvi, R. (1987). History of mathematics for teachers: The case of
irrational numbers. For the Learning of Mathematics, 7(2), 18–23.
Fischbein, E. (1987). Intuition in science and mathematics: An educational approach. Dordrecht: Reidel.
Fischbein, E., Jehiam, R., & Cohen, D. (1995). The concept of irrational number in high-school student and
prospective teachers. Educational Studies in Mathematics, 29(1), 29–44.
Giannakoulias, E., Souyoul, A., & Zachariades, T. (2007). Students’ thinking about fundamental real numbers
properties. In D. Pitta-Pantazi & G. Philippou (Eds.), Proceedings of the Fifth Congress of the European
Society for Research in Mathematics Education (pp. 416–425). Larnaka, Cyprus: Department of Educa-
tion, University of Cyprus.
22 T. Zachariades et al.

Malara, N. (2001). From fractions to rational numbers in their structure: Outlines for an innovative didactical
strategy and the question of density. In J. Novotná (Ed.), Proceedings of the 2nd Conference of the
European Society for Research Mathematics Education (pp. 35–46). Praga: Univerzita Karlova v Praze,
Pedagogická Faculta.
Marcoulides, G. A., & Kyriakides, L. (2010). Structural equation modelling techniques. In B. Creemers, L.
Kyriakides, & P. Sammons (Eds.), Methodological advances in educational effectiveness research
(Quantitative methodology series) (pp. 277–302). London, UK and New York, USA: Routledge.
Marcoulides, G. A., & Schumacker, R. E. (1996). Advanced structural equation modeling. New Jersey:
Lawrence Erlbaum Associates.
Merenluoto, K., & Lehtinen, E. (2002). Conceptual change in mathematics: Understanding the real numbers.
In M. Limón & L. Mason (Eds.), Reconsidering conceptual change: Issues in theory and practice (pp.
233–257). Dordrecht: Kluwer.
Merenluoto, K., & Lehtinen, E. (2004). Number concept and conceptual change: Towards a systemic model of
the processes of change. Learning and Instruction, 14(5), 519–534.
Muthén, B. (2004). Latent variable analysis: Growth mixture modeling and related techniques for longitudinal
data. In D. Kaplan (Ed.), Handbook of quantitative methodology for the social sciences (pp. 345–368).
Newbury Park, CA: Sage.
Muthén, L. K., & Muthén, B. O. (1998). Mplus user’s guide. Los Angeles, CA: Muthén & Muthén.
National Council of Teachers of Mathematics (NCTM). (2006). Navigating through number and operations in
grades 9–12. Reston, VA: NCTM.
Neumann, R. (1998). Students’ ideas on the density of fractions. In H. G. Weigand, A. Peter-Koop, N. Neil, K.
Reiss, G. Törner, & B. Wollring (Eds.), Proceedings of the Annual Meeting of the Gesellschaft fur
Didaktik der Mathematik on Didactics of Mathematics (pp. 97–104). Munich: Gesellschaft fur Didaktik
der Mathematik.
Richardson, V. (Ed.). (2001). Handbook of research teaching (4th ed.). Washington, DC: American Educa-
tional Research Association.
Shirpley, B. (2000). A new inferential test for path models based on directed acyclic graphs. Structural
Equation Modeling, 7(2), 206–218.
Sierpinska, A. (1994). Understanding in mathematics. London: The Falmer Press.
Sierpinska, A. (2000). On some aspects of students’ thinking in linear algebra. In J. L. Dorier (Ed.), On the
teaching of linear algebra (pp. 209–246). Dordrecht: Kluwer.
Sierpinska, A. (2005). On practical and theoretical thinking and other false dichotomies in mathematics
education. In M. Hoffmann, J. Lenhard, & F. Seeger (Eds.), Activity and sign - grounding mathematics
education (pp. 117–135). New York: Springer.
Sierpinska, A., Bobos, A., & Pruncut, A. (2011). Teaching absolute value inequalities to mature students.
Educational Studies in Mathematics, 78(3), 275–305.
Sierpinska, A., Nnadozie, A., & Oktaç, A. (2002). A study of relationships between theoretical thinking and
high achievement in linear algebra. Retrieved May 14, 2012, from Anna Sierpinska http://annasierpin
ska.wkrib.com/pdf/Sierpinska-TT-Report.pdf
Sirotic, N., & Zazkis, R. (2004). Irrational numbers: Dimensions of knowledge. In D. E. McDougall & J. A.
Ross (Eds.), Proceedings of the 26th Annual Meeting of the North American Chapter of the International
Group for the Psychology of Mathematics Education, Vol. 1 (pp. 171–178). Ontario: Ontario Institute for
Studies in Education, University of Toronto.
Sirotic, N., & Zazkis, R. (2007). Irrational number: The gap between formal and intuitive knowledge.
Educational Studies in Mathematics, 65(1), 49–76.
Tirosh, D., Fischbein, E., Graeber, A., & Wilson, J. W. (1998). Prospective elementary teachers’ conceptions
of rational numbers. Retrieved January 31, 2012, from http://jwilson.coe.uga.edu/Texts.Folder/Tirosh/
Pros.El.Tchrs.html
Vamvakoussi, X., & Vosniadou, S. (2004). Understanding the structure of the set of rational numbers: A
conceptual change approach. Learning and Instruction, 14(5), 453–467.
Vamvakoussi, X., & Vosniadou, S. (2007). How many numbers are there in a rational number interval?
Constraints, synthetic models and the effect of the number line. In S. Vosniadou, A. Baltas, & X.
Vamvakoussi (Eds.), Reframing the conceptual change approach in learning and instruction (pp. 265–
282). The Netherlands: Elsevier.
Zazkis, R. (2005). Representing numbers: Prime and irrational. International Journal of Mathematical
Education in Science and Technology, 36(2–3), 207–217.
Zazkis, R., & Sirotic, N. (2004). Making sense of irrational numbers: Focusing on representation. In M. J.
Høines & A. B. Fuglestad (Eds.), Proceedings of the 28th Conference of the International Group for the
Psychology of Mathematics Education, Vol. 4 (pp. 497–504). Norway: Bergen University College.

You might also like