The Construct of Attitude in Mathematics

Education

Pietro Di Martino and Rosetta Zan

Abstract This chapter addresses a number of crucial theoretical issues about

research on attitude towards mathematics, a field that has a very long tradition in
mathematics education, with early studies on attitude being published more than
60 years ago. Over time, research on attitude in mathematics education has
developed a range of perspectives and methodologies, dealing with a variety of
questions concerning the construct of attitude: discussion and development of
tools for measuring/assessing/observing it; analysis of the relationship with other
affective constructs and with cognition; investigation of the relationship with
achievement; critique of the lack of a suitable theoretical framework. The chap-
ter traces the ‘story’ of the construct of attitude, providing a theoretical discus-
sion of the issues mentioned above that are crucial to understanding the mosaic
of relationships and interactions within the affect field. Through the theoretical
debate, the aim of the chapter is to highlight new directions for research on atti-
tude in mathematics education.

Keywords Attitude towards mathematics • Affect in mathematics education •

Emotions • Beliefs

Introduction

Research on attitudes towards mathematics can be viewed as paradigmatic of

research in mathematics education. This research field lies at the intellectual cross-
roads of many different domains (e.g. mathematics, psychology, cognitive science,
epistemology, semiotics, anthropology), and often deals with constructs that have
been developed in those domains to face (new) emerging issues in mathematics
education (Sierpinska et al. 1993). The construct of attitude was introduced in the
first decades of the nineteenth century in the context of social psychology in order

P. Di Martino (*) • R. Zan

Dipartimento di Matematica, Università di Pisa, Largo Pontecorvo 5, 56127, Pisa, Italy
e-mail: [email protected]; [email protected]

© Springer International Publishing Switzerland 2015 51

B. Pepin, B. Roesken-Winter (eds.), From beliefs to dynamic affect
systems in mathematics education, Advances in Mathematics Education,
DOI 10.1007/978-3-319-06808-4_3
52 P. Di Martino and R. Zan

to foresee individuals’ choices in contexts such as voting or buying goods. Attitude

is seen as a trait of an individual that has a direct influence upon his/her behaviour:
An attitude is a mental and neural state of readiness, organized through experience, exerting
a directive and dynamic influence upon the individual’s response to all objects and situa-
tions with which it is related. (Allport 1935, p. 810)

In mathematics education, early studies about attitude appear in the middle of the
twentieth century. These pioneering studies were deeply affected by the field (social
psychology) in which the construct was born, both regarding the characterization of
attitude, seen as an individual’s trait capable of influencing his/her own behaviour
(Aiken 1970), and the methods used to assess and measure it.
In this context, the main goal was the search for a measurement of attitude:
Dutton (1951), in one of the first studies concerning attitude and mathematics,
stated his aim to measure pupils’ and teachers’ attitude towards arithmetic using
Thurstone scales. As a matter of fact, following the trend in social psychology, the
measurement of attitude was mainly carried out by the means of unidimensional ad
hoc scaling methods, such as Thurstone and Likert scales.
Many things have changed in the field of research on attitude from those early
studies up to now; some of those changes have been deeply influenced by a change
of perspectives in mathematics education. At present, attitude is considered (together
with beliefs, emotions and values) as one of the constructs that characterize a new
field of research: that of affect.
Research on attitude, as often happens, has not followed a linear path. Over the
years, the researchers’ position on basic issues such as the definition itself of atti-
tude and the instruments used to assess the construct has dramatically changed and
new issues and goals have been identified.
This feature of research on attitude has increased the need for a clear theoretical
systematization of research results, which has also emerged as a priority in the
whole mathematics education field in the last two decades. As a matter of fact, this
need has now become a necessity in mathematics education, due to the considerable
development of the research field in the last few years and, in particular, to the iden-
tification of its cumulative and universal characters (Boero and Szendrei 1998).
This view of the field is strictly linked with the characterization of the nature of
research findings:
Researchers in education have an intellectual obligation to push for greater clarity and
specificity (…) [in mathematics education] findings are rarely definitive; they are usually
suggestive. Evidence is not on the order of proof, but is cumulative. (Schoenfeld 2000,
pp. 647–648)

Therefore, coherently with the cumulative characterization of research in

mathematics education, we believe that tracing, with critical eyes, the history of
research on attitude may bring forward an understanding – through a theoretical
lens – of the mosaic of the relationships and interactions between definitions of
attitude and instruments to measure it, and of the influence the shift from a normative
to an interpretive paradigm had on both these issues.
The Construct of Attitude in Mathematics Education 53

Moreover, this systematization is necessary to map out the future of research on

attitude, including the identification of new issues, the development of suitable
methods, and a warning against repeating the same old mistakes.

Early Studies of Attitude in Mathematics Education:

The Problematic Relationship Between Attitude
and Achievement

During its early period (ranging from the first half of the twentieth century to the
end of the 1980s), research on attitude within mathematics education followed the
trend of research in social psychology. The definition of attitude was rarely made
explicit, and implicitly it seemed to refer to the tendency to behave in a certain way.
A central research topic was the development or refinement of measuring instru-
ments and sampling methods:
The search for more adequate questionnaire and sampling techniques and factors underlying
attitudes toward these subjects [arithmetic and mathematics] continues to be an important
area for research. (Dutton 1951, p. 418)

In this period, the predominant methodology was quantitative and statistical: as

a matter of fact the quantitative and statistical approach seems to have been consid-
ered a sort of warrant for the scientific nature of the discipline.
Research on attitude at this stage reflects the evolution of the field of mathemat-
ics education: an in-depth discussion about the very nature of this emerging field
had not yet been developed. According to Kilpatrick (1992, p. 15), in that period
“the measurement movement begins”. The quantitative primacy in the methods
used had its roots in the search for scientific acceptance of a young discipline that
was just beginning to take its first steps:
From the beginnings of the century through its three-quarter point, such inquiry [inquiry in
math education] becomes increasingly “scientific”, that is, ostensibly objective and rigor-
ously quantified. (Schoenfeld 1994, p. 698)

On the other hand, the attention paid to measurement instruments was also linked
to the main goal of early studies on attitude, which was the identification of causal
correlations between attitude and other significant factors.
In the first review of the construct of attitude within mathematics education,
Feierabend (1960) highlighted two main reasons for the increasing academic inter-
est in this construct. Drawing on the development of such construct in social psy-
chology, the first reason was related to the view of attitude as a selective factor
because of its correlation with the choice of enrolling/not enrolling in advanced
mathematical courses:
Mathematics, geometry, and algebra are the courses which, when disliked in high school,
have the highest percentage of students who never take a course in this area again. This implies
54 P. Di Martino and R. Zan

the operation of such a strong selective factor that by the time students reach college,
only the students with a strong positive attitude will still be taking mathematics; the rest
have negative attitudes which may increase in strength with the operation of time and the
lack of counteracting influences. (Feierabend 1960, p. 19)

The second reason concerned the relationship between attitude and mathematical
achievement:
A series of recent investigations have attempted to explain differences in school
performance among students of equal abilities on the basis of their attitudes. (Feierabend
1960, p. 11)

This point also implied taking into account gender differences in mathematics
achievement and in problem-solving ability:
There are sex differences in problem-solving ability unrelated to general mental ability,
special abilities, or specific knowledge (…) [he] attempted to show that the differential
performance of the two sexes was due to a difference in attitude toward problem-solving.
(Feierabend 1960, p. 17)

In his review Feierabend advanced some criticism towards research on attitude,

but his criticism was limited to some aspects related to the development of instru-
ments and to the statistical analysis. There was no reference to the lack of theoreti-
cal clarity and in particular no explicit definition of attitude was provided: a naïve
view of the construct emerges. The term ‘attitude’ was used to address different
constructs, such as preference, interest, motivation.
Ten years later Aiken (1970) summarized early research on attitude as follows:
The major topics covered were: methods of measuring attitudes towards arithmetic and
mathematics; the distribution and stability of mathematics attitudes; the effects of attitudes
on achievement in mathematics; the relationship of mathematics attitudes to ability and
personal factors. (Aiken 1970, p. 592)

It is interesting to notice that Aiken’s list also does not include reference to the
topic ‘nature of the construct of attitude’ (that would become a major topic in
research on attitude in the early 1990s).
What emerges from the reviews carried out by Feierabend and Aiken and from
the analysis of other literature of that period (Reyes 1984) is that most studies
were focused on the search for evidence of a causal relationship between “some-
thing called attitude” (Neale 1969, p. 631) and other variables, in particular math-
ematical achievement. This causal relationship is even seen as a hypothesis of the
aetiology of attitude towards mathematics (Aiken and Drager 1961). The search
for a causal relationship reveals a normative approach, that seems to drive research
on attitude and provide a justification, and in some way a reinforcement, for the
great attention paid to measurement instruments, rather than to the theoretical
clarification of the construct.
Despite its theoretical limitations, this first phase of research on attitude was
fruitful and produced meaningful results that, coherently with a cumulative view of
research, contributed significantly to the new research era that would follow. The
most significant contribution was what became the initial assumption of this kind of
research, that is, that non-cognitive factors strictly interact with cognitive factors
and have a crucial role in the learning of mathematics. This assumption is a sort of
The Construct of Attitude in Mathematics Education 55

break in the wall of the purely cognitive approach to mathematics education, and
was to be decisive in the development of the specific field of affect in mathematics
education: not purely cognitive factors – and in particular attitude – would become
a relevant topic in the study of mathematical learning:
The attitudes of students toward mathematics play a vital part in their learning (…)
Important for the study of attitudes toward mathematics is the idea that an attitude involves
both cognitive and non-cognitive aspects. (Corcoran and Gibb 1961, p. 105)

In addition to this, the great emphasis placed on methods caused a refinement of

many observational instruments. This brings to light important issues related to the
observation of attitude (and more in general of affective constructs), as for example
the tendency of individuals to reply to questionnaires according to what is socially
accepted and valued, rather than expressing their own thoughts – the so-called social
desiderability phenomenon (Kloosterman and Stage 1992).
Moreover, in this early period, research on attitude consolidated two significant
findings. A first result – confirmed by many studies – was the relationship between
attitude towards mathematics and the choice of mathematics courses. For instance,
in his literature review of research on attitude, Aiken (1970) stated that there is a
good body of evidence showing that the choice of enrolling in advanced mathemat-
ics courses is significantly affected by attitude towards mathematics. A second
important finding refers to gender differences in mathematical achievements. In
particular, the valuable work of Elizabeth Fennema and Julia Sherman highlighted
the differences in attitude towards mathematics between males and females, offer-
ing a new and important key for the interpretation of gender differences in mathe-
matics achievements:
Since the study of mathematics appears not to be sex-neutral, attitudes toward mathematics
may reflect cultural proscriptions and prescriptions (…) These data certainly indicate that
many females have as much mathematical potential as do many males. The generalized
belief that females cannot do well in mathematics is not supported. (Fennema and Sherman
1977, p. 69)

This result, that may now appear unquestionable, was not so obvious before the
work of Fennema and Sherman.
Even if the first period of research on attitude provides several important findings
and suggests a number of research hypotheses, the above discussion has also shown
its strong limitations from the very beginning. The identification and analysis of
these limitations has been crucial for the development of research on attitude in the
following years.
According to Bishop (1992), carrying out a research study in mathematics edu-
cation requires taking into account three components: enquiry (which concerns the
reason for the research activity), evidence and theory. The initial studies on attitude
are motivated by the assumption of the existence of a causal relationship between
attitude and achievement in mathematics, and seem to be focused on searching for
evidence of this relationship rather than developing a theoretical framework or clari-
fying the nature of the construct. But in spite of the efforts devoted to developing
measuring instruments, research fails to show a causal relationship in the direction
attitude → achievement, or a clear correlation between them.
56 P. Di Martino and R. Zan

Aiken (1970) reported the results of several studies in which this correlation is
far from being clear, highlighting the need for clarifying its very nature. Almost
30 years later, Ma and Kishor (1997), analysing the results of 113 different studies,
conclude that this correlation is not statistically significant. Assuming that this cor-
relation does exist, Ma and Kishor identify the cause of the failure to prove it in
the inappropriateness of the observing instruments used in the research on attitude
towards mathematics up to that point. At that stage, the instruments used to mea-
sure attitude towards mathematics have been criticized by many researchers,
because their nature is considered “exceptionally primitive” (Leder 1985).
However, starting from the 1980s, researchers increasingly acknowledge that the
major weakness of this kind of research lies in the lack of clarity at the theoretical
level and in the definition of the construct itself. Kulm (1980) suggests the existence
of a trend that tends to avoid an explicit definition of attitude towards mathematics
and instead adopts operational definitions determined by the types of instruments
used to measure attitude. This lack of interest in characterizing the construct pro-
duces a gap between the definition of attitude and its measurement (Leder 1985),
and results in the lack of reliability of the observational instruments.
Germann’s words below summarize the criticism towards the first phase of
research about attitude:
First, the construct of attitude has been vague, inconsistent, and ambiguous. Second,
research has often been conducted without a theoretical model of the relationship of attitude
with other variables. Third, the attitude instruments themselves are judged to be immature
and inadequate. (Germann 1988, p. 689)

In other words, the naïve theoretical approach that characterizes early studies on
attitude appears to be inadequate within the normative-positivistic paradigm in
which those studies were conducted. As a matter of fact, this paradigm demands
isolating and clearly identifying variables in order to interpret statistical results and
to be able to compare them across studies:
Sometimes no description or definition of what is meant by a particular variable is even
included in the research report. This makes interpretation of results difficult and detracts
from efforts to compare results across studies. (Hart 1984, p. 573)

For this reason, the process of re-thinking research on attitude began at the end
of the 1980s, addressing many aspects such as: the paradigm in which it is framed,
the goals that it pursues, the construct definition, the relationship between the con-
struct and other (affective and cognitive) factors, the development of observational
tools and the discussion about methods for analysing data.

The Theoretical Debate About Attitude in Mathematics

Education

In 1992, in the well-known Handbook of research on mathematics teaching and

learning, McLeod traces the way for a reconceptualization of research on affect in
mathematics education. He identifies three different constructs – beliefs, attitudes
The Construct of Attitude in Mathematics Education 57

and emotions1 – that, in his view, vary in stability and differ in the degree of the role
played by cognition. McLeod’s work starts with a crucial premise:
Affective issues play a central role in mathematics learning and instruction (…)
If research on learning and instruction is to maximize its impact on students and teachers,
affective issues need to occupy a more central position in the minds of researchers.
(McLeod 1992, p. 575)

He emphasizes the relationship between the newly acknowledged role assigned

to affective factors and the constructivist view of mathematics learning:
If we believe that the learner is someone who only receives knowledge rather than someone
who is actively involved in constructing knowledge, our research program could be entirely
different in terms of both the affective and the cognitive domain. (McLeod 1992, p. 576)

The need for reconceptualization is strictly connected with the criticism of the
previous research on attitude:
Research on affect has been voluminous, but not particularly powerful in influencing the
field of mathematics education. It seems that research on instruction in most cases goes
on without any particular attention to the affective issues (…) A major difficulty is that
research on affect has not usually been grounded in a strong theoretical foundation.
(McLeod 1992, p. 590)

Therefore, McLeod highlights that research on affect has to pay particular atten-
tion to three strictly intertwined aspects: the discussion of theoretical issues, the
development of a wider variety of methods, and the analysis of the relationships
among affective constructs and between affect and cognition.
Once again, the development of research on attitude is deeply influenced by the
simultaneous development of the field of mathematics education at the end of the
1980s. In this period, many scholars debate on the nature of mathematics education
and on the criteria for establishing quality of research in this field. In particular,
consistently with the goal of universalization of research results, the request for a
theoretical clarification of the constructs used in research is emphasized:
A community of scholars engaged in the research of common areas with common themes,
however, has responsibility to communicate ideas and results as clearly as possible using
common terms. For these reasons, it is important to use the terms consistently, accurately,
and appropriately once their definitions have been agreed on. (Pajares 1992, p. 315)

What Is Attitude Towards Mathematics?

The discussion about the theoretical aspects of research on attitude starts with a
‘definition problem’: what is attitude towards mathematics?

1
Later, De Bellis and Goldin (1999) propose ‘values’ as the fourth construct of the affective
domain.
58 P. Di Martino and R. Zan

A first critical issue relates to the object that attitude is oriented towards – that is,
mathematics. Some researchers refer to a ‘unique’ attitude toward mathematics
(Haladyna et al. 1983), while others claim that many different attitudes exist accord-
ing to the different topics and activities that are considered (Tirosh 1993); still other
scholars propose to distinguish between attitude towards mathematics seen as a
branch of scientific knowledge and towards mathematics as school subject
(Schoenfeld 1989), or even that attitude can refer to different objects and situations,
such as mathematical content, characteristics of mathematics, kind of teaching,
mathematical activities in the classroom and mathematics teacher (Kulm 1980).
Moreover, this complexity grows when, in addition to the variety of objects atti-
tude is oriented towards, we also consider the variety of subjects: whose attitude?
Research on attitude has dealt with a large variety of individuals: students, prospec-
tive and in-service teachers, students’ parents, and, more in general, adults.
But the most significant aspect of the complexity regarding the ‘definition prob-
lem’ is that it involves not only the characterization of the construct ‘attitude’, but
also that of positive/negative attitude, a dichotomy that pervades research, both
implicitly and explicitly. Classic studies regarding the relationship between attitude
and achievement in practice investigate the correlation between positive attitude and
success. In the same way, studies aiming to change attitude actually end up in set-
ting the objective of transforming a negative attitude into a positive one.
As already mentioned, a large portion of studies show the lack of a clear defini-
tion of the construct: attitude tends to be defined implicitly and a posteriori through
the instruments used to measure it (Kulm 1980; Leder 1985; Daskalogianni and
Simpson 2000).
In social psychology, the most recent theories agree on the multidimensionality
of the construct, and make reference to a tripartite model, according to which atti-
tude has a cognitive, an affective, and a behavioural component (Eagly and Chaiken
1998). Within the field of mathematics education many explicit definitions of atti-
tude refer to this tripartite model, describing attitude by means of three components:
the emotional disposition towards mathematics, the set of beliefs regarding mathe-
matics, and the behaviour related to mathematics (Hart 1989; Leder 1992; Ruffel
et al. 1998). However, some studies – generally in the earliest period of research –
adopt a ‘simple’ characterization, seeing attitude as a general emotional disposition
(Haladyna et al. 1983).
Both definitions show their theoretical but also operational and didactical limita-
tions (Di Martino and Zan 2001). The simple definition does not make explicit refer-
ence to cognitive aspects, although many researchers who subscribe to this definition
use models (see Mandler 1984; Ortony et al. 1988) that emphasize the relationship
between emotion and cognition, describing emotional experience as the result of a
combination of cognitive analyses and physiological responses. In this framework,
it is the interpretation given by an individual to an experience that elicits the emo-
tion, and not the experience itself:
First, the meaning comes out of the cognitive interpretation of the arousal. This meaning
will be dependent on what the individual knows or assumes to be true. In other words, the
The Construct of Attitude in Mathematics Education 59

individual’s knowledge and beliefs play a significant role in the interpretation of the
interruption. (McLeod 1992, p. 578)

According to the simple definition, the characterization of positive/negative

attitude is clear: a positive (negative) attitude is a positive (negative) emotional
disposition towards the subject.
This characterization can be useful when dealing with issues such as the choice
of mathematics courses or the comparison between different groups of individuals,
but it seems inadequate to deal with complex issues such as success in mathematics.
In this context, the idea of positive attitude that emerges from the simple definition
is not considered very significant by many mathematics education researchers, who
underline the importance of linking a positive emotional disposition with an episte-
mologically correct view of the discipline (Ernest 1988). In the same vein, the cru-
cial issue of promoting a positive attitude risks losing its significance if the goal of
developing a positive emotional disposition toward mathematics is not associated to
the goal of promoting a positive view of the discipline. Considering only the emo-
tional aspects poses an even greater didactical threat, since teachers may choose to
avoid complex tasks in order to prevent producing negative emotions.
Kulm (1980) discusses similar issues about the attitude definition in the early
research period and concludes:
It is probably not possible to offer a definition of attitude towards mathematics that would
be suitable for all situations, and even if one were agreed on, it would probably be too gen-
eral to be useful. (Kulm 1980, p. 358)

The awareness that the appropriateness of the construct depends on the studied
issues will lead to the idea of a ‘working definition’ (Daskalogianni and Simpson
2000).
As regards the tripartite model, the main critical aspect is that the implicit
assumption of a link between attitude and behaviour becomes part of the construct
definition itself. This theoretical choice exposes research to the risk of circular rea-
soning, as eloquently described by Lester (2002) in relation to the belief-construct:
A central difficulty is that the fundamental assumption undergirding much of this research
rests on a shaky logical foundation. Specifically, a basic assumption is that beliefs influence
peoples’ thinking and actions. However, it is also often assumed that beliefs lie hidden and
so can be studied only by inferring them from how people think and act. For researchers to
claim that students behave in a particular manner because of their beliefs and then infer the
students’ beliefs from how they behave involves circular reasoning. (Lester 2002, p. 346)

In the light of these critical aspects, a third definition of attitude emerges in

which behaviours are not explicitly mentioned: attitude towards mathematics is
described as the pattern of beliefs and emotions associated with mathematics
(Daskalogianni and Simpson 2000).
This choice overcomes the risk of circularity, but the theoretical problem of iden-
tifying a positive/negative attitude according to a multidimensional definition still
remains (Di Martino and Zan 2003). As a matter of fact, there is not only a need for
characterization of the positive/negative dichotomy for each dimension (emotions,
beliefs, possibly behaviour), but it is also essential to identify if and how the dichotomies
60 P. Di Martino and R. Zan

related to the single components can result in a unique characterization of positive/

negative attitude. This issue is strictly connected with the choice of the instruments
used to measure attitude.

Instruments Used to Measure Attitude

As Leder (1985) claims, the lack of interest in characterizing the construct produces
a gap between the definition of attitude and its measurement: as a matter of fact the
instruments traditionally used to assess and measure attitudes are not consistent
with the different definitions and with whether an explicit definition of attitude is
given or not.
The instruments used are almost exclusively self-report scales (Kulm 1980;
Leder 1985; McLeod 1987) such as Thurstone or Likert scales. These instruments
propose items that take into consideration beliefs and behaviours as well as emo-
tions: for example ‘Mathematics is useful’, ‘I think about arithmetic problems out-
side school’, ‘I like problem solving’. Therefore, they make implicit reference to the
tripartite model, regardless of whether this definition is explicitly selected as a start-
ing point or not. Even if the instruments used appear to be increasingly sophisti-
cated, the measurement generally results in a reduction to the positive/negative
bipolarity, which is obtained by summing up the scores related to each of the three
dimensions: cognitive, affective and behavioural.
While some scholars underplay this operation by observing that ‘the correlation
among measures of the three components, although leaving room for some unique
variance, are typically of considerable magnitude’ (Ajzen 1988, p. 22), others con-
sider this reduction as contradicting the recognized complexity of the tripartite
model (Eagly and Chaiken 1998). Reducing the description of attitude to a single
score is also in contrast with the original idea of Thurstone and Chave (1929) who
claim that attitude is a complex construct that cannot be measured by a single score,
but requires several indices. Thurstone and Chave underline that the choice of the
characteristics (indices) to be measured depends on the context – in the same way
as when measuring a physical object like a table one can decide whether to measure
length, width or height.
But the theoretical debate about research on attitude highlights other critical
issues in the measurement process. First, the separate measurement of each compo-
nent presents significant problems, due to the limitations of questionnaires. As far
as beliefs are concerned, the mismatch between exposed beliefs and beliefs-in-
action is well known (Schoenfeld 1989), just like the already mentioned social
desiderability phenomenon (Kloosterman and Stage 1992). Regarding emotions,
researchers have discussed the difference between an opinion about an emotion and
the emotion itself (Ruffel et al. 1998) and the limitations of instruments such as
questionnaires and interviews in capturing emotional reactions that are not con-
scious (Schlöglmann 2002).
The Construct of Attitude in Mathematics Education 61

A second critical point concerns the choice of items that, in the case of question-
naires, is fully determined by the researchers, while respondents are only asked to
express their agreement/disagreement with these items: how can we be sure that the
topic of the items is relevant to the respondent? In other words, using the terminol-
ogy introduced by Green (1971), how can we be sure that the corresponding beliefs
are psychologically central in the respondent’s belief system?
A third critical aspect concerns the choice of the score to be attributed to each of
the items, requiring identification of what a positive emotion/belief/behaviour is
(this shows the strong relationship that exists between the definition and the mea-
surement problem). Essentially:
– When positive refers to an emotion, it normally means ‘perceived as pleasurable’.
So anxiety when confronting a problem is seen as negative, while pleasure in
doing mathematics is evaluated as positive.
– When positive refers to beliefs, it is generally used with the meaning ‘shared by
experts’. The first limitation of this approach is brought to light by a number of
studies which highlight that there is no single pattern of beliefs shared by experts
in mathematics (Mura 1993, 1995; Grigutsch and Törner 1998). In light of this,
identifying several different typical patterns of beliefs towards mathematics
shared by experts becomes necessary. At present, this still remains an issue for
discussion that could lead to the definition of a number of different patterns to act
as models of ‘successful views of mathematics’.
– When it refers to a specific behaviour, positive generally means ‘successful’.
In the school context, a successful behaviour is generally identified with high
achievement. This characterization leads to the problem of how to assess achieve-
ment (Middleton and Spanias 1999).
A further problem is that the differences between the various meanings of posi-
tive attitude are rarely made explicit. If the researcher does not declare his/her initial
choices, interpreting the results of a study and comparing different studies becomes
problematic.
Even if this ambiguity is overcome by making explicit the initial choices and
assumptions, in our opinion other problems remain. In some studies the three mean-
ings for ‘positive’ (related to emotion, belief and behaviour) overlap thanks to
implicit assumptions: for example, that a ‘positive’ belief (i.e. shared by experts) is
associated with a successful behaviour and elicits a pleasurable emotion; or that a
pleasurable emotion is necessarily associated with a positive behaviour in mathe-
matics, and vice versa for negative emotion.
Evaluating a belief (or an emotion) as ‘positive’ or ‘negative’ according to the
emotion and behaviour related to it implies using a cause/effect model, according to
which the same belief causes the same emotion or the same behaviour in all indi-
viduals. Moreover, this evaluation not only assumes that a certain belief has an
emotional component, but also looks at the significance of that emotional compo-
nent, that is, not just that it is linked to a behaviour, but also which type of
behaviour.
62 P. Di Martino and R. Zan

In this case, the cure seems worse than the disease, since this approach does not
take into account the very complex nature of the relationship among beliefs,
emotions and behaviour.
As a matter of fact, a number of studies about emotions (Evans 2000) suggest the
possibility that for certain subjects, an optimal level of anxiety exists, above which,
but also below which, performance is reduced.
The relationship between beliefs and emotions was investigated in a study with
211 high school students aged between 14 and 18 (Di Martino and Zan 2002) in the
case of the belief ‘In mathematics there is a reason for everything’, which is an item
that is also used in many scales for measuring attitude towards mathematics.
Students are asked to fill in a questionnaire including the following item:

Choose the option you most agree with:

◻ In mathematics there is always a reason for everything (B)
◻ It is not true that in mathematics there is always a reason for everything (not B)
And: ◻ I like ◻ I don’t like ◻ I am indifferent to this characteristic of
mathematics

The findings show that only 51.7 % of the sample fell in the two expected groups
(i.e. ‘B – I like’ and ‘not B – I don’t like’). But overall there was no difference in the
percentage of belief B-holders between the groups of high achievers and low achiev-
ers. The distinction between these two groups is related to the emotion associated
with this belief: 76 % of the high-achievers who are belief B-holders liked this char-
acteristic of mathematics, while this percentage dramatically decreased to 28 %
within the low-achievers group.
About the combination ‘epistemological correct belief – negative emotion’, we
suggest two possible interpretations. The first interpretation is that the negative
emotion is directly related to the belief. On the other hand, we also need to consider
the possibility that the emotional disposition is not directly linked to that single
belief, but to its interaction with other beliefs. This remark questions the possibility
of characterizing a single belief as positive or negative, without considering its con-
nection with other beliefs an individual may have (belief system):
Because they [single beliefs] offer a limited glimpse into a much broader system and
because understanding their connections and centrality is essential to understanding the
nature of their effect, researchers must study the context-specific effects of beliefs in terms
of these connections. (Pajares 1992, p. 326)

More specifically, when describing belief systems Rokeach (1968) recognizes

the dimension of centrality for a particular belief, highlighting that not all beliefs
have the same importance for an individual. Central beliefs play a prominent role
in people’s belief systems, and consequently in influencing their behaviour. As
Eagly (1967) observes, beliefs about self are generally considered more central
than other ones.
The Construct of Attitude in Mathematics Education 63

Consider for example the relationship between belief B used in our study
(‘In mathematics there is a reason for everything’) and the (likely) central self-
belief ‘I am not able to understand these reasons’: the presence of such relationship
may result in unproductive behaviours such as avoiding answering a question or
giving random answers (Di Martino 2004).
The discussion above highlights that the assumption of the existence of a
cause-effect relationship between a specific belief and emotion or behaviour is inad-
equate. The interaction is more complex, since it involves an individual’s belief
system (and not only the single belief) and is strongly dependent on the individual.
Following the results of this debate, a movement towards the overcoming of the
normative approach and the use of an interpretive approach for research on attitude
emerges with the aim of attending to the complexity of the issues at stake.

The Attitude Construct in the Reconceptualization

of the Affective Domain in Mathematics Education

Once again, the history of research on attitude reflects the evolution of the mathe-
matics education field: the theoretical debate about attitude develops in parallel with
a new interpretive perspective that begins to emerge within the field of mathematics
education. This perspective, in contrast with the normative-positivistic one, signifi-
cantly affects the discussion about the theoretical characterization of constructs.
The gradual affirmation of the interpretive paradigm in the social sciences,
including a greater attention paid to the complexity of human behaviour, leads
researchers in mathematics education to abandon the attempt of explaining behav-
iour through measurements or general rules based on a cause-effect model, and to
search instead for new interpretive tools (once again drawing on other domains):
The purpose of doing interpretivist research (…) is to provide information that will allow
the investigator to “make sense” of the world from the perspective of participants. (Eisenhart
1988, p. 103)

This implies a significant shift in focus: an emerging attention to the understanding

of a phenomenon (‘making sense of the world’) replaces the description of the
phenomenon itself, which could be seen as a shift from product to process
(Schoenfeld 1994).
The interpretive approach has a direct influence on the process of re-thinking
research methods since the limitations of the statistical methods become evident:
Through the 1980s and into the 1990s (…) with a shift in focus there was a concomitant
shift in methods (including the reporting of clinical interviews, process and simulation
models, field observations and participant observations), because a new class of
phenomena required a new set of explanations a new set of tools to uncover them.
(Schoenfeld 1994, p. 703)

This shift of perspectives gives new strength to research on attitude that until this
point had remained stuck in the causal-relationship paradigm. In particular, attitude
64 P. Di Martino and R. Zan

gains renewed popularity in the studies about problem-solving activities aimed at

interpreting the failure of students who seem to have the required cognitive resources.
The book Affect and mathematical problem solving (McLeod and Adams 1989),
collecting contributions by several authors, represents the turning point for research
on affective constructs, and in particular on attitude. For the first time, affective
constructs are used not only to prove the existence of a numerical correlation with
an outcome (mathematical achievement), but also to interpret a process (the interac-
tions between affective and cognitive aspects in problem-solving activities).
Therefore, the need for a theoretical clarification in mathematics education (that is
also related to the possibility and the intention for a cumulative development in the
field) appears to become a fundamental issue also specifically for research on affect:
There was a lack of definition, lack of clarity, and lack of connections to mathematics.
It is possible to avoid making the same mistakes again as new ideas and research method-
ologies are employed. It is hoped that new researchers on affect will be clear about what is
being studied, precise in definition, and respectful of what has been learned previously.
(Fennema 1989, p. 209)

The double occurrence of the adjective ‘new’ in Fennema’s words is not casual:
it shows awareness of the fact that new perspectives and new more complex issues
force a rethinking of the affective constructs. In particular, the shift from a norma-
tive paradigm to an interpretive one provokes a discussion (re-definition) of goals,
definitions and methods.
The belief that research on attitude towards mathematics may offer interpretive
instruments to understand the reasons for an individual’s intentional actions in the
mathematical context grows (Zan et al. 2006). This belief is supported by the initial
evidence coming from related research in the field of neuroscience:
There is apparently some neurological basis for asserting a link between affective and
cognitive aspects of human functioning. (Silver 1985, p. 253)

More recently, Damasio (1996) highlighted the close relationship between affect
and decision-making processes.
The theoretical construct of ‘attitude towards mathematics’ is no longer a
construct aimed at explaining causes of behaviour, thus enabling researchers to pre-
dict it, but instead it becomes a flexible and multidimensional interpretive tool,
aimed at describing the interactions between affective and cognitive aspects in
mathematical activity. In particular, attitude becomes a tool to interpret people’s
decisions in mathematical activities, and, if necessary, suggest strategies to modify
them. In this context, particularly significant is Ruffel, Mason and Allen’s position
about the definition of the construct of attitude itself:
Reflecting on them [some previous studies about attitude] led us to challenge the very
construct of attitude. We are also led to challenge the cause-and-effect model underlying
much attitudinal research. We now see attitude as at best a complex notion, and we conjec-
ture that perhaps it is not a quality of an individual but rather a construct of an observer’s
desire to formulate a story to account for observation. (Ruffel et al. 1998, p. 1)

It could be argued that the same thing can be said about every theoretical
construct, not only in mathematics education. In fact, in our opinion, this position
shows awareness of the fact that any phenomenon can only be observed from a
The Construct of Attitude in Mathematics Education 65

particular point of view and thus highlights the role of the researcher/observer, who
cannot be a mere measurer. This position represents the overcoming of a naïve
approach, in which attitude is seen as an objectively measurable quality of an indi-
vidual, and the transition to a theoretical approach.
In line with this perspective, Daskalogianni and Simpson (2000) assume that the
definition of attitude becomes a working definition, which is functional to the research
questions that researchers pose in each study. Therefore, having different definitions
of the construct appears natural, and a definition is no longer evaluated in terms of
correctness (is it the right definition?) but in terms of suitability to address a specific
research problem in mathematics education (Di Martino and Zan 2010). According to
the classification of research proposed by Bishop (1992), this kind of approach char-
acterizes the new trend of research on attitude as problem-led.
The theoretical re-thinking of research on attitude leads to the exploration of new
methods of inquiry in the field. Coherently with their position, Ruffel et al. (1998)
emphasize the inadequacy of the measurement approach by replacing the verb
‘measuring’ with the verb ‘probing’.
At the beginning of the new millennium, a strong criticism of the use of quantitative
methods in the research on attitude emerged, and a movement towards the use of
qualitative approaches has begun. It is understood that qualitative methods, and in
particular the use of narratives, enable researchers to take into account those beliefs
and emotions which are psychologically central for the respondents. A number of
studies using essays, diaries, interviews and also the observation of behaviour in
natural settings or in structured situations appear (Karsenty and Vinner 2000;
Hannula 2002; Zan and Di Martino 2007; Kaasila 2007).
Differently from what happens with the traditional attitude scales, the respon-
dents are not requested to express agreement/disagreement with respect to items
chosen by others, but are asked to tell their mathematical ‘stories’, through which
they can recount all the aspects that they consider relevant in their relationship with
mathematics. As a matter of fact, the pivotal motivation for using narratives in edu-
cational research is the following:
Humans are storytelling organisms who, individually and socially, lead storied lives.
The study of narrative, therefore, is the study of the ways humans experience the world.
(Connelly and Clandinin 1990, p. 2)

As a consequence, almost 10 years after McLeod’s manifesto, the shift in focus

in research on attitude provokes a shift in perspectives and methods: a real
revolution.

The TMA Model: A Definition of Attitude Grounded

in Students’ Narratives

Within the described framework, we have investigated how mathematics teachers

use the diagnosis ‘this student has a negative attitude’ to interpret students’ mathe-
matical difficulties in the context of an Italian National Project. The results of the
66 P. Di Martino and R. Zan

study (Polo and Zan 2006) show that this diagnosis is frequently used (at all school
levels) by mathematics teachers to explain students’ difficulties, and, above all, that
in the majority of cases it represents a case of surrender instead of being used as an
interpretive step capable of steering future action.
These findings persuade us that, in order to turn the ‘negative attitude diagnosis’
into a useful instrument for both practitioners (teachers) and researchers, it is neces-
sary to link the theoretical construct of attitude to practice. This fits with the strong
incentive put forward by Phillip “to develop constructs that might be applied to help
make sense of teaching and learning environments” (Philipp 2007, p. 264).
Therefore we have designed a study based on the collection and analysis of
students’ autobiographical narratives and aimed at constructing a characterization
of students’ attitude towards mathematics in relation to their experience (Di Martino
and Zan 2010).
Our reason for choosing to use autobiographical essays is that through this
format pupils tend to explicitly evoke events about their past that they deem impor-
tant and also to paste fragments by introducing causal links, not in a logical perspec-
tive but rather in a social, ethical and psychological one (Bruner 1990). We believe
that in order to describe the kind of relationship an individual has with mathematics,
and consequently to suggest a characterization of attitude towards mathematics
strictly linked to experience, this pasting process is more important than an objective
report of one’s experience with the discipline at school. As Bruner claims:
It does not matter whether the account conforms to what others might say who were
witnesses, nor are we in pursuit of such ontologically obscure issues as whether the account
is ‘self-deceptive’ or ‘true’. Our interest, rather, is only in what the person thought he did,
what he thought he was in, and so on. (Bruner 1990, pp. 119–120)

In doing so, a theoretical model for attitude emerges from the data collected
through a cyclical analytical process, that is, through what Glaser and Strauss (1967)
call grounded theory. In this kind of process, the autobiographical texts are analysed
in order to systematically make meaning out of the individuals’ narrations: the final
outcome is the identification of a set of categories and relationships aimed at under-
standing and interpreting different behaviours (Demazière and Dubar 1997).
We have collected and analysed 1662 anonymous essays entitled “Maths and me:
my relationship with maths up to now”, written by students whose school levels
ranged from grade 1 to grade 13.2 The results of our study show that when students
describe their relationship with mathematics, almost all of them refer to one (or
more) of the following three dimensions:
• emotional disposition towards mathematics,
• vision of mathematics,
• perceived competence in mathematics.
This result suggests the Three-dimensional Model for Attitude (TMA)
represented in Fig. 1.

2
The sample of the study was not chosen on a statistical basis, but we relied on the collaboration
of teachers who voluntarily agreed to participate in our research.
The Construct of Attitude in Mathematics Education 67

Fig. 1 The three-

dimensional model for
attitude

TMA takes explicitly into account the close relationship amongst the three
dimensions. The research study also highlights the subjectivity of these interactions,
confirming the need for designing suitable observational tools to track it:
The proposed model of attitude acts as a bridge between beliefs and emotions, in that it
explicitly takes into account beliefs (about self and mathematics) and emotions, and also
the interplay between them. However, in order for it to become effective theoretical and
didactical instruments, the construction and use of consistent instruments for observation,
capable of taking into account its complexity, is needed. (Di Martino and Zan 2011, p. 479)

The analysis of the students’ autobiographical essays also suggests the need
for the development of a new approach to the positive/negative characterization
of attitude, confirming that the reduction of the dichotomy positive/negative
attitude to the emotional dimension is questionable. As a matter of fact, we find
that negative emotional dispositions towards mathematics may be associated
with different patterns of attitude, depending on the student’s perceived compe-
tence and vision of mathematics as well as on the relationships amongst the
three dimensions. Coherently with this observation, and with the multidimen-
sional characterization of the construct in TMA, we have developed a definition
of ‘negative attitude’ that explicitly makes reference to the negativity of at least
one of the three dimensions:
The multidimensionality of the model underlines the inadequacy of the positive/negative
dichotomy for attitude referred only to the emotional dimension (like/dislike), and rather
suggests considering an attitude as negative, when at least one of the dimensions is nega-
tive. In this way, we can outline profiles of negative attitude, depending on the dimension
that appears to be negative. (Di Martino and Zan 2010, p. 44)

We identify two polarities for each dimension, and define as negative an

emotional disposition resulting in a dislike for mathematics, a low perceived com-
petence, and – according to the characterization of Skemp (1976) – an instrumental
vision of mathematics. This definition of negative profiles of attitude within TMA
suggests two new interrelated research avenues. One the one hand, the development
of observational tools aimed at identifying a student’s profile of attitude towards
mathematics, in particular at recognizing a possible negative component in this pro-
file. On the other hand, the theoretical construction and implementation of didacti-
cal interventions, aimed at preventing or overcoming a negative attitude towards
mathematics and differentiated according to the different profiles of negative atti-
tude identified in TMA.
68 P. Di Martino and R. Zan

The TMA model, originally created as a model for students’ attitudes towards
mathematics, also appears suitable for characterizing attitudes towards specific
mathematics topics (geometry, algebra, etc.) and for investigating the attitudes
towards mathematics held by different groups of people (teachers, adults, etc.). For
this reason, the TMA model has recently been used to study and analyse in-service
and pre-service primary teachers’ attitude towards mathematics and its teaching
(Coppola et al. 2012).

Summing Up and Looking Ahead

In mathematics education, research on attitude has a very long tradition, based on

the interest, shared by mathematicians, teachers and mathematics educators, in
identifying a causal relationship between something called ‘positive attitude’ and
achievement. In the first period of the research most studies aimed at refining or
developing measuring instruments, rather than at clarifying theoretical aspects.
With the evolution of mathematics education as a research field, and with the
development of a specific research field on affect, research on attitude towards
mathematics has evolved by identifying critical points in the previous phase and
setting the need for a theoretical framework as a crucial item in the research agenda.
This change has also provoked a shift from a normative paradigm to an interpretive
one. Attitude is no longer seen as an individual’s trait, useful for predicting his/her
behaviour, but as an observer’s construct, capable of suggesting an understanding of
the individual’s intentional actions in a complex context, as is the learning of math-
ematics: a multidimensional construct that involves beliefs and emotions and acts as
a bridge between them (Di Martino and Zan 2011).
The development of research on attitude also suggests new issues to be explored,
such as: constructing new observation tools that are consistent with the interpretive
approach and the multidimensional characterization of attitude; investigating atti-
tude toward mathematics of different groups of individuals; identifying possible
motives underlying a change of attitude; designing and testing didactical paths to
prevent or modify attitude.
But the theoretical debate about the quality of research about attitude persists. As
a matter of fact, the need for comparing results from different studies and different
theoretical frameworks is still a crucial issue, even when studies using question-
naires and statistical analysis have been replaced by qualitative case studies. New
paradigms and new methods require the identification of new criteria for research
quality: this is an important topic for future research in the affect field.
Despite the fact that many studies on attitude ‘look ahead’, drawing on the most
important findings produced so far, in our opinion some critical issues still remain.
The gap between the definition of the construct of attitude and the methods used to
assess it is far from being bridged: many studies still use the term ‘attitude’ without
defining it, or propose questionnaires that are not consistent with the chosen charac-
terization of attitude, and, in particular, without clarifying the theoretical choices
The Construct of Attitude in Mathematics Education 69

underlying the studies themselves. Moreover, although the normative approach in

the research on attitude has showed all its theoretical limitations, many recent stud-
ies place themselves in a normative paradigm, even if, perhaps, this is not a con-
scious choice made by the researcher.
This lack of a cumulative character in research on attitude is, in our opinion, one
of its main weaknesses, a historical weakness that has not yet been overcome. In
1976, in his update on research on attitude, Aiken wrote:
Regardless of the efforts of this writer and others to bring to the educational research
community periodic reviews of studies concerned with attitudes and anxiety toward
mathematics, many investigators in this area continue to be unaware or unappreciative of
previous research on the topic (…) This oversight is almost certainly due to a failure to
search the relevant literature, the first step in any scientific inquiry (Aiken 1976, p. 293)

More than 30 years later, we notice exactly the same phenomenon, in a sort of
theoretical and meta-theoretical déjà vu that, we are sure, has limited the develop-
ment of stronger results in the field. For that reason, we believe that tracing the
‘story’ of the construct of attitude and discussing the results obtained so far is a very
significant step in the development of research in this field.

References

Aiken, L. (1970). Attitudes toward mathematics. Review of Educational Research, 40, 551–596.
Aiken, L. (1976). Update on attitudes and other affective variables in learning mathematics. Review
of Educational Research, 46(2), 293–311.
Aiken, L., & Drager, R. (1961). The effect of attitudes on performance in mathematics. Journal of
Educational Psychology, 52, 19–24.
Ajzen, I. (1988). Attitudes, personality, and behavior. Chicago: Dorsey Press.
Allport, W. (1935). Attitudes. In C. A. Murchison (Ed.), A handbook of social psychology.
Worcester: Clark University Press.
Bishop, A. (1992). International perspectives on research in mathematics education. In D. A.
Grouws (Ed.), Handbook of research on mathematics teaching and learning (pp. 710–723).
New York: Macmillan.
Boero, P., & Szendrei, J. (1998). Research and results in mathematics education: Some contradic-
tory aspects. In J. Kilpatrick & A. Sierpinska (Eds.), Mathematics education as a research
domain: A search for identity (pp. 197–212). Dordrecht: Kluwer.
Bruner, J. (1990). Acts of meaning. Cambridge, MA: Harvard University Press.
Connelly, F. M., & Clandinin, D. J. (1990). Stories of experience and narrative inquiry. Educational
Researcher, 19(5), 2–14.
Coppola, C., Di Martino, P., Pacelli, T., & Sabena, C. (2012). Primary teachers’ affect: A crucial
variable in the teaching of mathematics. Nordic Studies in Mathematics Education, 17(3–4),
101–118.
Corcoran, M., & Gibb, G. (1961). Appraising attitudes in the learning of mathematics. In
Evaluation in mathematics: Twenty-sixth yearbook of the NCTM. Reston: NCTM.
Damasio, A. (1996). Descartes’ error: Emotion, reason and the human brain. Basingstoke:
Papermac.
Daskalogianni, K., & Simpson, A. (2000). Towards a definition of attitude: The relationship
between the affective and the cognitive in pre-university students. In T. Nakahara & M. Koyama
(Eds.), Proceedings of the 24th conference of the IGPME (Vol. 2, pp. 217–224). Hiroshima,
Japan.
70 P. Di Martino and R. Zan

De Bellis, V., & Goldin, G. (1999). Aspect of affect: Mathematical intimacy, mathemati-
cal integrity. In O. Zaslavsky (Ed.), Proceedings of the 23rd conference of the IGPME
(Vol. 2, pp. 249–256). Haifa, Israel.
Demazière, D., & Dubar, C. (1997). Analyser les entretiens biographiques. Paris: Édition Nathan.
Di Martino, P. (2004). From single beliefs to belief systems: A new observational tool. In M. J.
Hoines & A. B. Fuglestad (Eds.), Proceedings of the 28th conference of the IGPME
(Vol. 2, pp. 271–278). Bergen, Norway.
Di Martino, P., & Zan, R. (2001). Attitude toward mathematics: Some theoretical issues. In M. van
den Heuvel-Panhuizen (Ed.), Proceedings of the 25th conference of the IGPME
(Vol. 3, pp. 351–358). Utrecht, The Netherlands.
Di Martino, P., & Zan, R. (2002). An attempt to describe a ‘negative’ attitude toward mathematics.
In P. Di Martino (Ed.), Proceedings of the MAVI XI European workshop (pp. 22–29). Pisa:
Università di Pisa Press.
Di Martino, P., & Zan, R. (2003). What does ‘positive’ attitude really mean? In N. Pateman,
B. Doherty, & J. Zilliox (Eds.), Proceedings of the 27th conference of the IGPME
(Vol. 4, pp. 451–458). Honolulu, Hawaii.
Di Martino, P., & Zan, R. (2010). ‘Me and maths’: Towards a definition of attitude grounded on
students’ narratives. Journal of Mathematics Teacher Education, 13(1), 27–48.
Di Martino, P., & Zan, R. (2011). Attitude towards mathematics: A bridge between beliefs and
emotions. ZDM – The International Journal on Mathematics Education, 43(4), 471–482.
Dutton, W. (1951). Attitudes of prospective teachers toward arithmetic. The Elementary School
Journal, 42, 84–90.
Eagly, A. (1967). Involvement as a determinant of response to favorable and unfavourable informa-
tion. Journal of Personality and Social Psychology, 7, 1–15.
Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T. Fiske,
& G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 1, pp. 269–322).
New York: McGraw-Hill.
Eisenhart, M. (1988). The ethnographic research tradition and mathematics education research.
Journal for Research in Mathematics Education, 19(2), 99–114.
Ernest, P. (1988). The attitudes and practices of student teachers of primary school mathematics.
In A. Barbos (Ed.), Proceedings of the 12th conference of the IGPME (Vol. 1, pp. 288–295).
Veszprém, Hungary.
Evans, J. (2000). Adults’ mathematical thinking and emotions: A study of numerate practices.
London: Routledge Falmer.
Feierabend, R. (1960). Review of research on psychological problems in mathematics education.
Research Problems in Mathematics Education, 3, 3–46.
Fennema, E. (1989). The study of affect and mathematics: A proposed generic model for research.
In D. McLeod & V. Adams (Eds.), Affect and mathematical problem solving: A new perspec-
tive (pp. 205–219). New York: Springer.
Fennema, E., & Sherman, J. (1977). Sex-related differences in mathematics achievement, spatial
visualization and affective factors. American Educational Research Journal, 14(1), 51–71.
Germann, P. (1988). Development of the attitude toward science in school assessment and its use
to investigate the relationship between science achievement and attitude toward science in
school. Journal of Research in Science Teaching, 25(8), 689–703.
Glaser, B., & Strauss, A. (1967). The discovery of grounded theory: Strategies for qualitative
research. Chicago: Aldine.
Green, T. (1971). The activities of teaching. New York: McGraw-Hill.
Grigutsch, S., & Törner, G. (1998). World views of mathematics held by university teachers of
mathematical science. Duisburg: Gerhard Mercator University, Preprint 420.
Haladyna, T., Shaughnessy, J., & Shaughnessy, M. (1983). A causal analysis of attitude toward
mathematics. Journal for Research in Mathematics Education, 14, 19–29.
Hannula, M. (2002). Attitude toward mathematics: Emotions, expectations and values. Educational
Studies in Mathematics, 49, 25–46.
The Construct of Attitude in Mathematics Education 71

Hart, L. (1984). Affective variables and mathematics education. The Elementary School Journal,
84(5), 558–581.
Hart, L. (1989). Describing the affective domain: Saying what we mean. In D. McLeod & V. Adams
(Eds.), Affect and mathematical problem solving (pp. 37–45). New York: Springer.
Kaasila, R. (2007). Using narrative inquiry for investigating the becoming of a mathematics
teacher. ZDM – The International Journal of Mathematics Education, 39(3), 205–213.
Karsenty, R., & Vinner, S. (2000). What do we remember when it’s over? Adults’ recollections of
their mathematical experience. In T. Nakahara & M. Koyama (Eds.), Proceedings of the 24th
conference of the IGPME (Vol. 4, pp. 451–458). Hiroshima, Japan.
Kilpatrick, J. (1992). A history of research in mathematics education. In D. A. Grouws (Ed.),
Handbook of research on mathematics teaching and learning (pp. 3–38). New York: Macmillan.
Kloosterman, P., & Stage, F. (1992). Measuring beliefs about mathematical problem solving.
School Science and Mathematics, 92(3), 109–115.
Kulm, G. (1980). Research on mathematics attitude. In R. J. Shumway (Ed.), Research in mathe-
matics education. Reston: NCTM.
Leder, G. (1985). Measurement of attitude to mathematics. For the Learning of Mathematics, 5(3),
18–22.
Leder, G. (1992). Measuring attitudes to mathematics. In W. Geeslin & K. Graham (Eds.),
Proceedings of the 16th conference of the IGPME (Vol. 2, pp. 33–39). Durham, USA.
Lester, F. (2002). Implications of research on students’ beliefs for classroom practice. In G. Leder,
E. Pehkonen, & G. Törner (Eds.), Beliefs: A hidden variable in mathematics education?
Dordrecht: Kluwer.
Ma, X., & Kishor, N. (1997). Assessing the relationship between attitude toward mathematics and
achievement in mathematics: A meta-analysis. Journal for Research in Mathematics Education,
30, 65–88.
Mandler, G. (1984). Mind and body: Psychology of emotion and stress. New York: Norton.
McLeod, D. (1987). A constructivist approach to research on attitude toward mathematics. In
L. Streefland (Ed.), Proceedings of the 9th conference of the IGPME (Vol. 1, pp. 133–140).
Montreal, Canada.
McLeod, D. (1992). Research on affect in mathematics education: A reconceptualization. In D. A.
Grouws (Ed.), Handbook of research on mathematics teaching and learning (pp. 575–596).
New York: Macmillan.
McLeod, D., & Adams, V. (Eds.). (1989). Affect and mathematical problem solving: A new per-
spective. New York: Springer.
Middleton, J. A., & Spanias, P. A. (1999). Motivation for achievement in mathematics: Findings,
generalizations, and criticism of the research. Journal for Research in Mathematics Education,
30, 65–88.
Mura, R. (1993). Images of mathematics held by university teachers of mathematical sciences.
Educational Studies in Mathematics, 25(4), 375–385.
Mura, R. (1995). Images of mathematics held by university teachers of mathematics education.
Educational Studies in Mathematics, 28(4), 385–399.
Neale, D. (1969). The role of attitudes in learning mathematics. The Arithmetic Teacher, 16,
631–641.
Ortony, A., Clore, G., & Collins, A. (1988). The cognitive structure of emotions. Cambridge:
Cambridge University Press.
Pajares, F. (1992). Teachers’ beliefs and educational research: Cleaning up a messy construct.
Review of Educational Research, 62(3), 307–332.
Philipp, R. A. (2007). Mathematics teachers’ beliefs and affect. In F. K. Lester (Ed.), Second
handbook of research on mathematics teaching and learning (pp. 257–315). Reston: NCTM.
Polo, M., & Zan, R. (2006). Teachers’ use of the construct ‘attitude’. Preliminary research findings.
In M. Bosch (Ed.), Proceedings of the 4th CERME. Barcelona: FundEmi.
Reyes, L. (1984). Affective variables and mathematics education. The Elementary School Journal,
84(5), 558–581.
Rokeach, M. (1968). Beliefs, attitudes and values. San Francisco: Jossey-Bass.
72 P. Di Martino and R. Zan

Ruffel, M., Mason, J., & Allen, B. (1998). Studying attitude to mathematics. Educational Studies
in Mathematics, 35, 1–18.
Schlöglmann, W. (2002). Affect and mathematics learning. In A. D. Cockburn & E. Nardi (Eds.),
Proceedings of the 26th conference of the IGPME (Vol. 4, pp. 185–192). Norwich, UK.
Schoenfeld, A. (1989). Exploration of students’ mathematical beliefs and behavior. Journal for
Research in Mathematics Education, 20, 338–355.
Schoenfeld, A. (1994). A discourse on methods. Journal for Research in Mathematics Education,
25(6), 697–710.
Schoenfeld, A. (2000). Purposes and methods of research in mathematics education. Notices of the
American Mathematical Society, 47(6), 641–649.
Sierpinska, A., Kilpatrick, J., Balacheff, N., Howson, G., Sfard, A., & Steinbring, H. (1993). What
is research in mathematics education, and what are its results? Journal for Research in
Mathematics Education, 24(2), 274–278.
Silver, E. (1985). Research on teaching mathematical problem solving: Some underrepresented
themes and needed directions. In E. Silver (Ed.), Teaching and learning mathematical problem
solving: Multiple research perspectives (pp. 247–266). Hillsdale: Lawrence Erlbaum.
Skemp, R. (1976). Relational understanding and instrumental understanding. Mathematics
Teaching, 77, 20–26.
Thurstone, L., & Chave, E. (1929). The measurement of attitude: A psychophysical method and
some experiments with a scale for measuring attitude toward the Church. Chicago: University
of Chicago Press.
Tirosh, D. (1993). Affective and cognitive variables: Students’ perspectives. In I. Hirabayashi,
N. Nohda, K. Shigematsu, & F. L. Lin (Eds.), Proceedings of the 17th conference of the IGPME
(Vol. 1, pp. 11–13). Tsukuba, Ibaraki, Japan.
Zan, R., & Di Martino, P. (2007). Attitude toward mathematics: Overcoming the positive/negative
dichotomy. The Montana Mathematics Enthusiast, Monograph 3, 157–168.
Zan, R., Brown, L., Evans, J., & Hannula, M. (2006). Affect in mathematics education: An
introduction. Educational Studies in Mathematics, 63(2), 113–121.