Vous êtes sur la page 1sur 16

High Educ (2012) 64:421436

DOI 10.1007/s10734-011-9502-3

Development and validation of a questionnaire


measuring teachers motivations for teaching
in higher education
Gerda J. Visser-Wijnveen Ann Stes Peter Van Petegem

Published online: 6 January 2012


Springer Science+Business Media B.V. 2012

Abstract A lot of research has been done into the motivations in teachers in primary/
secondary education and into student motivation. However, teachers motivations for
teaching in higher education are rarely studied. A growing interest exists though in the
professional development of teachers in higher education, of which motivation is an
important aspect. This article, therefore, focuses on the development and validation of a
Dutch questionnaire for teachers motivations for teaching in higher education. The
questionnaire is based on three earlier developed questionnaires, including the following
motivational aspects: efcacy, interest, and effort. Conrmatory factor analyses were
conducted with the data of 231 higher education teachers. The results of our explorative
study showed that after modications, including the removal of two out of three efcacyaspects (outcome efcacy and teaching efcacy), this instrument is sufciently reliable
and valid to use in educational practice and research. Future research into the use of the
questionnaire in different contexts is desirable.

Keywords Motivation Efcacy Teachers Professional development


Validation study

Introduction
In the various research studies conducted into the professional development of teachers in
higher education over the last few years a great deal of attention has been paid to changes
in teachers conceptions (McKenzie 2002; Postareff et al. 2007) and, to a lesser extent, into
changes in teaching behaviour (Gibbs and Coffey 2004; Stes et al. 2010). However, an
important issue has been overlooked: the motivation for teaching. In addition to behaviour
and conceptions, motivation is also a very important factor in how individuals function and
therefore, also, of how teachers function. The combination of competencies and motivation
G. J. Visser-Wijnveen (&) A. Stes P. Van Petegem
Instituut voor Onderwijsen Informatiewetenschappen, Expertisecentrum Hoger Onderwijs (ECHO),
Universiteit Antwerpen, Stadscampus, Venusstraat 35, 2000 Antwerpen, Belgium
e-mail: gjvisser@iclon.leidenuniv.nl

123

422

High Educ (2012) 64:421436

is a determining factor with regard to the performance achieved (Ambrose and Kulik
1999). This is why so many professional development programs designed for teachers
rightly place so much emphasis on improving teaching competencies, but the motivation of
teachers for teaching is a dimension which has been relatively little explored, despite the
fact that this is also a key component in achieving the desired result. There is therefore a
clear need to have instruments which can identify teachers motivations. Many of the
existing instruments are designed for use with teachers in secondary or primary education.
In the context of higher education Bailey (1999) and Trigwell et al. (2004) have used
questionnaires to look at certain aspects of motivation. However, a simple, easy-to-use,
questionnaire with which to research the various aspects that have been identied as
relevant is still lacking. This article describes the development and validation of a questionnaire which aims to provide a way of measuring the relevant aspects of teachers
motivations in higher education with regard to teaching.

Theoretical framework
Motivation
The long history of motivational research has given rise to a wide variety of different
theories which have been surveyed in several review articles (inter alia Ambrose and Kulik
1999; Eccles and Wigeld 2002). It is not our intention, therefore, to discuss all these
theories again here. Instead, we will restrict ourselves to explaining the various concepts
taken from these different theories that we have included in our questionnaire and how
these are related to some of the existing theories, specically: self-efcacy, interest, and
effort. All these concepts are part of intrinsic motivation. Cognitive evaluation theory
distinguishes between two types of motivation: intrinsic motivation and extrinsic motivation (Ambrose and Kulik 1999; Deci and Ryan n.d.). Intrinsic motivation means that the
activity in question is carried out for enjoyment or for the satisfaction derived from that
activity, while extrinsic motivation means that the activity in question is carried out as a
means to something else or from a sense of duty (Vallerand and Bissonnette 1992).
Intrinsic motivation consists of three different types of objectives: an intrinsic motivation
to know; an intrinsic motivation to achieve; or an intrinsic motivation to be stimulated and
is associated with improved performance (Vallerand and Bissonnette 1992).
There is also a long tradition of motivational research in the eld of education, with a
wide variety of studies looking at the motivation of pupils and students at different levels
so as to arrive at a better understanding of how performance can be improved or explained
(e.g., Hidi 2001; Moneta and Spada 2009). One of the inuencing factors involved is the
motivation of their teachers (Trigwell et al. 2004), in particular, how teachers rate their
own teaching abilities and the effort they are willing to put into teaching based on this
estimation (Tschannen-Moran and Hoy 2007). So far, however, research into teachers in
higher education has paid only limited attention to the issue of motivation (Bailey 1999).
Menges (1997) emphasizes the importance of motivation for the success of professional
development programs. The most important concern of the teachers taking part in schemes
of this sort is whether the program will help them to deal more effectively with the
problems they encounter in their day-to-day teaching practice and whether it will help them
to grow personally and professionally. This can only be achieved by specically focusing
on the teaching motivation of the teachers in question (Menges 1997). Postareff et al.
(2007) found that teachers belief in their own abilities (self-efcacy) can decline

123

High Educ (2012) 64:421436

423

temporarily while they are participating in professional development programs, because


teachers become aware of their shortcomings, and that, therefore, during the program,
instructors should work on trying to build up more self-condence in this area.
Additionally, motivation is a context-related concept (Woolfolk Hoy 2008; Vallerand
and Bissonnette 1992). From a relational perspective, motivation mediates between personal and contextual characteristics and actual performance (Trigwell et al. 2004). In their
review article on teacher efcacy Tschannen-Moran et al. (1998)emphasized this aspect
with respect to teacher efcacy, suggesting that teachers can only make a correct assessment of their own abilities and the expected outcome, if the context is clearly specied: It
is precisely this explicit emphasis on the teachers assessment of his or her personal
competency in the light of an analysis of the task and the situation that makes our model
better than previous models (Tschannen-Moran et al. 1998, p. 233). The idea that the
extent to which a teacher perceives him or herself as effective depends on the analysis of
the specic task is therefore a component which must be included when measuring teacher
efcacy (see also Tschannen-Moran and Hoy 2001).
With regard to motivational aspects, we concentrate on intrinsical aspects of motivation.
Starting from the perspective of efcacy, effort is a related and complementary concept
since both aspects include teacher characteristics. While a teacher can easily adapt his or
her effort into teaching, this is much harder with regard to self-efcacy. By including both
self-efcacy and effort we focus on core elements of motivation (Tschannen-Moran and
Hoy 2007). An additional aspect needs to be included when we look at motivation for
teaching of higher education teachers, which is interest. In higher education this aspect is
most likely even more important than in teachers in other contexts, although interest is an
important aspect in, for example, secondary education as well (Watt and Richardson 2008).
The higher education sector differs from other sectors, because there is competition
between research and teaching. Most teachers either have a strong interest in research or in
teaching (Macfarlane and Hughes 2009). In the following paragraphs we will discuss these
three concepts (efcacy, interest, and effort).

Efcacy
A major area of interest in the research into motivation is research into self efcacy.
Motivation is very closely related to how people see their abilities in particular situations
(Bailey 1999; Bandura 1994). Hence failure is interpreted differently depending on
whether people consider their own capacity to be high or low. Individuals who expect a
high personal efcacy see failure as a situation that they can learn from and this helps them
to do better the next time and such individuals generally continue to be strongly motivated,
whereas individuals with a low personal efcacy see failure as a conrmation that they are
indeed incapable of achieving the desired level and this may cause their motivation to
decline further. Bandura (1994), p. 71 denes self-efcacy as peoples beliefs about
their capabilities to produce designated levels of performance that exercise inuence over
events that affect their lives. He suggests that conceptions of self-efcacy are determinant
for how people feel, think, motivate themselves, and behave. Self-efcacy is also a relevant
concept from the perspective of goal theory. Perceived efcacy appears to inuence the
accepted or chosen level of difculty of an objective, commitment to the objective,
reactions to failure, and choice of strategy (Locke 1996). Furthermore, the expected value
theory is related to self-efcacy: the extent to which teachers are able to motivate themselves to teach, is linked to the expected results (Bandura 1994).

123

424

High Educ (2012) 64:421436

Perceived teacher efcacy is a specic variant of self-efcacy and is one of the factors
which appear to be very closely related to teaching behaviour and student outcomes
(Gibson and Dembo 1984; Woolfolk et al. 1990). Based on Banduras research (1977), the
literature on teacher efcacy makes a distinction between efcacy expectations, in the form
of teacher efcacy, and outcome expectations, in the form of outcome efcacy (Soodak
and Podell 1996). Teacher efcacy thus focuses on the question as to what extent teachers
assume that they possess the required abilities, while outcome efcacy focuses on the
question of how far the effort that they are expected to make also leads to the desired result
(Soodak and Podell 1996). It is assumed that there is a correlation between these two forms
of self-efcacy. Several researchers have tried to clarify both concepts (Guskey and
Passaro 1994; Soodak and Podell 1996; Woolfolk et al. 1990). An important nding was
that what has often been equated with outcome efcacy, based on the work of Gibson and
Dembo (1984), in fact relates to a third concept: teaching efcacy. The issue here is not
whether teachers expected that their own effort would actually lead to a result, but whether
the teacher concerned was (in general terms) able to overcome external factors, such as
genetics and pupils home environment. The dimension teaching efcacy was shown to
operate independently of outcome efcacy (Woolfolk et al. 1990). Soodak and Podell
(1996) developed a questionnaire for teachers in primary and secondary education that
included a new scale for outcome efcacy in addition to scales for personal efcacy and
teaching efcacy. However, Guskey and Passaro (1994) offered an alternative explanation
for the distinction between personal efcacy and teaching efcacy. Taking their lead from
locus of control research, they suggested that the underlying construct was actually an
internal versus an external focus. Items from the personal efcacy scales referred to the
individual teacher personally, whereas items from the teaching efcacy scales referred to
teachers in general. Their subsequent study, carried out with reformulated items, conrmed their hypothesis (Guskey and Passaro 1994).

Interest
Another relevant motivational aspect is interest: A recent study among mathematics
teachers revealed that those primarily interested in teaching showed higher quality
instructional behaviour than their colleagues whose main interest was the subject area
(Kunter et al. 2008). According to Deci and Ryan (n.d.) the interest and enjoyment that
someone experiences in an activity are important indicators of the intrinsic motivation for
that activity. In stating this, they join the self-dertermination theory, in which a distinction
is made between external and intrinsic motivation. The concept of interest has been
developed further within the eld of language teaching (Eccles and Wigeld 2002) and is
regarded as an important motivational factor which promotes performance (Hidi 2001). In
the context of research into interest a distinction is made between individual interest and
situational interest (Hidi 2001; Schiefele 1999). Individual interest is relatively stable and
relates to the intrinsic value assigned to a topic or activity and differs greatly in each
individual. Situational interest can be created by environmental factors and is thus related
to extrinsic factors (Schiefele 1999).

Effort
Effort plays a role in various motivational theories: for example, it is a relevant factor in
both expectancy theory and goal theory. Effort is affected by goal clarity and difculty and
the involvement with the task objective, and affects performance (Ambrose and Kulik

123

High Educ (2012) 64:421436

425

1999). Effort is also a closely related concept from the perspective of self-efcacy
(Tschannen-Moran and Hoy 2007). In attribution theory, which is an extension of expected
value theory, effort is regarded as one of the four aspects to which people attribute their
success or failure (Graham 1991). Of these four, only two are related to intrinsic aspects,
namely abilities (or, strictly speaking, perceived abilities, which are closely related to selfefcacy) and effort. The extrinsic aspects are the difculty of the task and luck (good or
bad). There are three more dimensions which exert an inuence, namely: locus of control,
stability and controllability (Vockell 2001). From the perspective of attribution theory,
effort is the prime element that can be inuenced, not only because effort is something
which can be controlled by the individual himself (controllability), but also because it is
changeable (stability) and can be ascribed to the individual (locus of control). Capacity is
likewise intrinsic (locus of control), but relatively unchangeable (stability) and less easy to
inuence (controllability) (Vockell 2001).

Research focus and context


In the present study we developed and validated an easy-to-use questionnaire for the study
of teachers motivation with respect to their teaching designed for use in higher education.
The following efcacy aspects were included: personal efcacy (the expectation that a
teacher has about his or her personal teaching abilities), outcome efcacy (the expectation
that someone has with respect to their own inuence on student learning) and teaching
efcacy (the expectation of the inuence which teaching in general has on student
learning). Other motivational aspects are interest in teaching and effort for teaching. See
Fig. 1 for a visual representation.
Our questionnaire is specically aimed at teachers in higher education. The higher
education setting is very different from other education settings in that there is competition
between research and teaching. Many teachers in higher education have a stronger individual interest in either research or teaching (Macfarlane and Hughes 2009). Bailey (1999)
suggests that this might be related, inter alia, to self-efcacy conceptions: if someone
thinks that he or she is good at teaching, he or she will not rate his or her research abilities
highly, which will denitely inuence that persons motivation for research (negatively) as

Fig. 1 Original model

123

426

High Educ (2012) 64:421436

well as teaching (positively). This reasoning also works in the opposite direction, of course.
Teachers who have achieved a certain status on the basis of their research, will often be
very strongly motivated for their research, while they have usually had no (or only very
limited) training as teachers and may therefore be less condent about their abilities in that
area and may be of the opinion that it is not worthwhile putting a lot of energy into their
teaching (Colbeck 1998).

Method
The design of the questionnaire
In compiling the questionnaire we used translated items taken from three questionnaires
which have already been used: the motivation questionnaire created by Trigwell et al.
(2004) designed for use in higher education (self-efcacy and task value); the Intrinsic
Motivation Inventory (Deci and Ryan n.d.); and Soodak and Podells Teacher Efcacy
Questionnaire (1996), which is a modied version of the questionnaire developed by
Gibson and Dembo (1984). The items we selected were translated and (where necessary)
modied for use in higher education. Table 1 shows the items in the various scales and
their origin. When referring to a specic context teachers were asked to think in terms of a
typical course module (in other words the kind of courses they usually teach, not for
example, one aimed at a special target group or involving an unusual approach). The items
were scored on a 4 point Likert scale (from strongly disagree to strongly agree). The
nal questionnaire consisted of 34 items, ten of which had to be recoded because of a
negative formulation. The 34 items were spread over three main scales and three subscales, as follows:

Efcacy

Personal efcacy (9 items)


Teaching efcacy (4 items)
Outcome efcacy (5 items)

Interest/Enjoyment (9 items)
Importance/Effort (7 items)
Respondents
The questionnaire was given to teachers at the University of Antwerp and at three university colleges in Antwerp, Belgium (N = 1367) on a voluntary basis. The questionnaire
was returned by 262 teachers, resulting in an initial response of 19%. The response rates at
the various institutes differed slightly: 16, 17, and 28% for the university colleges, and 18%
for the University of Antwerp. After correction for missing answers 231 questionnaires
were used for our conrmatory factor analyses. The nal usable response was 17%.

Analysis
The analytical phase consisted of an analysis of non-response and a conrmatory factor
analysis. We started by subjecting the items to an item-analysis. Items for which more than
5% of the respondents had not lled in an answer were considered insufciently

123

High Educ (2012) 64:421436

427

Table 1 Origin of questionnaire items


Item
no.

Item

Scale

Origin

I am absolutely condent that I have a sufciently good


knowledge of the content to teach this course module.

Efcacy
personal

Trigwell et al. (2004):


self-efcacy

How much a student learns is determined more by innate


talent than by teaching. (R)

Efcacy
teaching

Soodak and Podell


(1996): teaching
efcacy

When a student performs better, very often this is just


because I have made an extra effort.

Efcacy
outcome

Soodak and Podell


(1996): outcome
efcacy

I am very interested in the content of this course module.

Interest/
Enjoyment

Trigwell et al. (2004):


interest

I believe that it is important that students learn thanks to


my teaching of this course module.

Effort/
Importance

Trigwell et al. (2004):


importance

I am sure that I have the necessary skills to teach this


course module.

Efcacy
personal

Trigwell et al. (2004):


self-efcacy

How well a student performs is more strongly determined


by genetic factors than by teaching. (R)

Efcacy
teaching

Soodak and Podell


(1996): teaching
efcacy

When a student achieves a better score than he or she


usually gets, this is often because I have found a better
way to help that student.

Efcacy
outcome

Soodak and Podell


(1996): outcome
efcacy

I enjoy teaching the content of this course module.

Interest/
Enjoyment

Trigwell et al. (2004):


interest

10

I think it is important that students should learn the


content of this course module.

Effort/
Importance

Trigwell et al. (2004):


importance

11

I am absolutely condent that I have sufcient didactic


knowledge to teach.

Efcacy
personal

Trigwell et al. (2004):


self-efcacy

12

If a student performs well, this is usually due to his or her


innate abilities. (R)

Efcacy
teaching

Soodak and Podell


(1996): teaching
efcacy

13

If students scores improve, that is usually because I have


found a more effective teaching approach. (R)

Efcacy
outcome

Soodak and Podell


(1996): outcome
efcacy

14

I really enjoy teaching.

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

15

I put a lot of energy into my teaching.

Effort/
Importance

Deci and Ryan (n.d.):


effort

16

I think that I am quite a good teacher.

Efcacy
personal

Deci and Ryan (n.d.):


perceived competence

17

Even a good teacher doesnt reach all the students. (R)

Efcacy
teaching

Soodak and Podell


(1996): teaching
efcacy

18

If a student grasps a new concept quickly, that may be


because I knew the right way to explain that concept.

Efcacy
outcome

Soodak and Podell


(1996): outcome
efcacy

19

Teaching is fun.

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

20

I dont really make much of an effort to be a good


teacher. (R)

Effort/
Importance

Deci and Ryan (n.d.):


effort

123

428

High Educ (2012) 64:421436

Table 1 continued
Item
no.

Item

Scale

Origin

21

I think that I am quite a good teacher, compared to other


teachers.

Efcacy
personal

Deci and Ryan (n.d.):


perceived competence

22

I am absolutely condent that my students are learning


something from me in this course module.

Efcacy
outcome

Trigwell et al. (2004):


self-efcacy

23

Teaching is a boring activity. (R)

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

24

I try my absolute best in my teaching duties.

Effort/
Importance

Deci and Ryan (n.d.):


effort

25

Now that Ive been teaching for a while, I think that I am


a competent teacher.

Efcacy
personal

Deci and Ryan (n.d.):


perceived competence

26

I dont pay a lot of attention to teaching. (R)

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

27

Its important to me to be a good teacher.

Effort/
Importance

Deci and Ryan (n.d.):


effort

28

I am satised with my competencies as a teacher.

Efcacy
personal

Deci and Ryan (n.d.):


perceived competence

29

Teaching is a very interesting activity.

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

30

I dont put a lot of effort into my teaching. (R)

Effort/
Importance

Deci and Ryan (n.d.):


effort

31

I have a talent for teaching.

Efcacy
personal

Deci and Ryan (n.d.):


perceived competence

32

Teaching is fairly enjoyable.

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

33

I am not a very good teacher. (R)

Efcacy
personal

Deci and Ryan (n.d.):


perceived competence

34

When I am teaching, I notice that I enjoy it.

Interest/
Enjoyment

Deci and Ryan (n.d.):


interest/enjoyment

(R) indicates that the scoring of the item has to be reversed

interpretable and were eliminated from subsequent analyses. One of the items (21: I think
that I am quite a good teacher, compared to other teachers) was left blank by 6% of the
respondents and was therefore eliminated, while item 16 (I think that I am quite a good
teacher), which was similar in nature, had a 100% response. We then eliminated all the
respondents with one or more missing answers (apart from item 21) from the subsequent
analysis in order to meet the requirements of the conrmatory factor analysis program.
In order to check the t of the conrmatory factor structures we used a variety of
indices. The rst index we used was v2 and the corresponding p value. The chi2-test is only
regarded as a reliable measure to check the goodness-of-t, if the sample size is between
200 and 500 observations (Bollen 1989; Hoyle 1995), which meant that this test could be
used with the number of respondents we had (n = 231). The corresponding p value must
be greater than 0.05 in order to say that there is a good t of the data with the assumed
model (Bollen 1989; Hoyle 1995). Other indices we used to determine the t were: the
goodness-of-t index (GFI), the adjusted goodness-of-t index (AGFI), the Tucker-Lewis
Index (TLI), the comparative t index (CFI), and the Root-Mean-Square-Error-ofApproximation (RMSEA). In the case of GFI, AGFI, TLI, and CFI a value equal to or

123

High Educ (2012) 64:421436

429

greater than 0.90 is required in order to be able to claim a good t; a RMSEA value equal to
or less than 0.05 is used as an indication of a relatively good t of the data with the
assumed model (Hoyle 1995).
So as to achieve the best possible t and thus to arrive at a valid and useful questionnaire we conducted a variety of analyses. First of all, we carried out a factor analysis
forced to ve factors in order to check whether the original division into ve (sub) factors
was tenable. Beyond a limiting value of 0.32 in factor correlation, an oblique rotation is
more appropriate than an orthogonal rotation (Tabachnick and Fidell 2007). Given that two
factors had a higher correlation and one of the factors was also close to this limiting value
with respect to a third factor we opted for the Oblimin with Kaiser Normalisation rotation
method. When assigning items to a factor we used the relatively high limiting value of
0.400. Our next step, based on the estimates and modication indices, was to look at to
what extent the ve scales really did measure one concept together (in this case motivation for teaching). Thirdly, in order to arrive at an economical and more reliable and
valid questionnaire, we looked at the individual items, and those items which met the
following three criteria were eliminated: 1) the Cronbachs alpha of the scale concerned
rises if the item is removed; 2) the items loading on the scale is lower than 0.500; 3) the
modication-indices suggest putting covariates between the item and other scales. Finally,
we examined the modication indices again looking for signicant error-covariates from
the theory, in order to arrive at an improved t.

Results
Development of a valid instrument
The original version of the model with the three main scales efcacy, effort, and
interest, in which the scale efcacy was broken down into three sub-scales (personal
efcacy, outcome efcacy, and teaching efcacy) gave an insufcient t (see
Table 2model A). We subjected it to a factor analysis forced to ve factors with oblique
rotation (Direct Oblimin with Kaiser Normalisation) which resulted in 52% explained
variance. This solution agreed fairly well with the scales from the original model, but there
were also a few clear differences. Some items did not function as expected and failed to

Table 2 Results of the conrmatory factor analysis

Index

A. Original
model
(33 items)

B. 5 factors
model
(28 items)

C. 3 factors
model with
sub-factors (25 items)

v2

865.080

535.002

352.431

df

489

345

267

\0.001

\0.001

\0.001

v /df

1.769

1.551

1.320

GFI

0.824

0.863

0.892

AGFI

0.798

0.839

0.869

CFI

0.874

0.932

0.967

TLI

0.864

0.925

0.963

RMSEA

0.058

0.049

0.037

123

430

Fig. 2 Factor structure and standardised values for model C

123

High Educ (2012) 64:421436

High Educ (2012) 64:421436

431

load on any of the scales higher than 0.400. One item (22) seemed to be part of another
scale. A conrmatory analysis on the modied model (model B, see Table 2) had a
noticeably better t than that of the original model, after leaving out the low loading items
and moving item 22, although the GFI (0.863) and AGFI (0.839) still suggested an
insufcient t.
Studying the modication indices and Cronbachs alphas of the scales resulted in the
following adjustments. Due to the very low loadings (0.267 and 0.178, respectively) of the
scales outcome efcacy and teaching efcacy we decided to include these as covariates
in addition to the main scale, instead of including them as sub scales of the main scale
motivation, as originally intended. Moreover, three items which had low estimates (0.382,
0.383 and 0.453) and a negative inuence on the Cronbachs alphas of the scales concerned (difference average 0.023) were eliminated. The suggested error-covariates included, inter alia, the items 1, 6, and 11 from the scale personal efcacy. These items are
concerned with the self-rated personal efcacy in a specic context. In the nal model
(model C, see Table 2; Fig. 2) the scale personal efcacy was therefore broken down into
two sub-scales, in which the context-related items were grouped separately from the
general items. This model resulted in an improved t that could be regarded as good, based
on three indices (CFI, TLI, and RMSEA) and as reasonable, based on the GFI and AGFI.

Description of the nal model


Model C is shown, as the best tting model, in Fig. 2; Table 3 gives the averages and
standard deviations of this model for each scale. What is particularly noticeable here are
the high average scores on almost all the scales, except outcome efcacy and teaching
efcacy. The corresponding standard deviations are also relatively limited. Similarly,
Table 3 gives the Cronbachs alpha for each scale. Most of the (sub/main) scales exhibited
a good internal consistency: personal efcacy (a = 0.90), personal efcacy: general
(a = 0.87), interest (a = 0.83), teaching efcacy (a = 0.88), and motivation for teaching
(a = 0.89). Of the remaining three (sub) scales the internal consistency can be termed
sufcient: personal efcacy: context (a = 0.77), effort (a = 0.79), and outcome efcacy
(a = 0.74). The Cronbachs alphas of the main scale and the scale personal efcacy based
on the underlying scales are lower than on the basis of the individual items (which is a
logical consequence of the smaller number), but can be termed good (0.84 for personal
efcacy) or acceptable (0.66 for motivation for teaching).
Finally, we calculated the correlations between the various (sub/main) scales (see
Table 4). A noteworthy nding is that the sub-scales exhibit a high correlation with the
scale personal efcacy (0.91 and 0.95, respectively) and the scales personal efcacy and

Table 3 Characteristics of the


(sub)scales for model C

(Sub)scale

SD

Cronbachs a

Motivation for teaching

3.63

0.32

0.89a (0.66)

3.48

0.43

0.90a (0.84)

Personal efcacy: context

3.55

0.44

0.77

Personal efcacy: general

3.43

0.49

0.87

Interest

3.74

0.36

0.83

Effort

3.82

0.35

0.79

Outcome efcacy

2.66

0.58

0.74

Teaching efcacy

2.98

0.66

0.88

Personal efcacy

Calculated on the basis of the


underlying items. The value in
brackets is the value calculated
on the basis of the underlying
sub-scales

123

432

High Educ (2012) 64:421436

Table 4 Pearson correlation coefcient between the scales and sub-scales for model C
(Sub)scale

Pers.
eff.

Pers. eff.:
context

Pers. eff.:
general

Interest

Effort

Motivation
for teaching

Outcome
eff.

Teaching
eff.

Personal
efcacy

1.00

0.91**

0.95**

0.51**

0.26**

0.90**

0.13

0.10

1.00

0.73**

0.46**

0.24**

0.81**

0.13*

0.11

1.00

0.50**

0.24**

0.85**

0.11

0.09

1.00

0.43**

0.80**

0.21**

0.17*

1.00

0.56**

0.14*

0.06

1.00

0.19**

0.14*

1.00

0.15*

Personal
efcacy:
context
Personal
efcacy:
general
Interest
Effort
Motivation
for teaching
Outcome
efcacy
Teaching
efcacy

interest also exhibit a high correlation with the main scale motivation for teaching (0.90
and 0.80, respectively). The scale effort shows a reasonable correlation with the main
scale (0.56), while the scales outcome efcacy and teaching efcacy show a limited
correlation with the main scale motivation (0.19 and 0.14 respectively).

Conclusion and discussion


Methodological considerations
Nine items which had been found valid in another context, appeared in the present study to
be poor or only limited measures of what was intended to be measured. Thus a large
number of items which refer to (the content of) the module were removed. Only on the
scale personal efcacy was the number of items that were specically context-related
large enough to justify a separate sub-scale, in the other cases the items were removed due
to the limited number of context-related items. It is also noticeable that a number of
negatively formulated items did not meet the expectations. Finally, the high number of
missing answers to item 21 is an indication that some teachers nd it challenging to rate
their own abilities in comparison to those of their colleagues. One might question the
extent to which those teachers who did answer this question did so in an ecologically valid
manner, given that they are part of a culture in which teachers do not generally talk to their
colleagues about teaching. Other items with relatively high numbers of missing values
(2%) were item 13 (6 out of 262) and item 7 (5 out of 262). In general it appeared that the
items of the scales outcome efcacy and teaching efcacy held slightly higher numbers of
missing values than items in the other scales.
The high average values and limited standard deviations suggest that there was a ceiling
effect in the use of the questionnaire. This is also conrmed by the relatively high kurtosis
and negative skewness of the items in the scales interest (median: 3.08 and -1.75) and

123

1.00

High Educ (2012) 64:421436

433

effort (median: 6.10 and -2.49). In our study the questionnaire was given to teachers in
higher education on a voluntary basis. The fact that it was voluntary may have meant that it
was primarily teachers who were more than averagely motivated who lled out the
questionnaire. Especially since the original response rate was 19%, although only slightly
lower than normal for web surveys (Kaplowitz et al. 2004), the group might be quite
selective. This may have inuenced the average scale scores and the range within which
respondents scored. Following the example of the study by Trigwell et al. (2004) which
also looked at self-efcacy and motivation in a higher education context, we used a 4-point
Likert scale. However, a 6-point Likert scale would probably have given a better picture of
possible differences between respondents. In our case the overwhelming majority of the
respondents replied strongly agree to items relating to interest and effort, while giving
them the added option agree between the options partly agree and strongly agree
would probably have achieved the desired diversication. The fact that this problem
mainly arose with the scales effort and interest and not with the original self-efcacyscale is in contradiction to the idea that motivation is an easier concept to measure than
self-efcacy with regard to scale width (Bailey 1999). However, even with a 6-point Likert
scale it is likely that there would still be limited kurtosis and negative skewness. Nevertheless this statistical nding does not automatically mean that the items concerned would
not function well (cf. Lombaerts et al. 2009). In fact, it is quite possible that only a very
limited group of teachers is completely different with regard to motivation for teaching than
most of their colleagues, giving rise to signicant outliers. Explanations for the relatively
low scores, and higher number of missing values, on teaching efcacy and outcome efcacy
might be that modesty is highly regarded within the Flemish culture, so that to advocate that
students success depends mainly on the teacher might be considered as being pretentious.
Futhermore, Flemish higher education is characterized by no pre-selection of students; the
entrance into higher education is very open. The selection of students capable for a study
takes place during (the rst years of) higher education. This might inuence the way
teachers interprete questionnaire items concerning both teaching and outcome efcacy, for
not all students might actually succeed no matter how hard the teacher tries.
An important change with regard to the original model was made when deciding to no
longer include outcome efcacy and teaching efcacy as part of the motivation construct,
but rather to include them as covariates. This decision was primarily based on methodological grounds: a model with these scales as part of motivation did not t sufciently and
these specic scales did hardly contribute to the motivation, while a model with these
scales as covariates had an acceptable t. Other options were to fully exclude the scales or
to include them nonetheless. On the one hand, entirely excluding the two efcacy scales
would not do justice to the important position these constructs have in the literature on
efcacy and motivation. Furthermore, although the scales hardly contributed, signicant
relations were found between these scales and other scales, such as interest, and the main
scale of motivation. On the other hand, including the scales in the construct of motivation,
as suggested by the literature and originally planned, would not do justice to the ndings of
this study that showed that these aspects were, although partly signicant, only weakly
related to other motivational aspects for higher education teachers, and functioned apart
from the other aspects. Therefore, we decided to include them in the model as being related
to motivation (with correlations of 0.17 and 0.26, for teaching efcacy and outcome
efcacy, respectively), but not part of the construct of motivation itself. In future research,
we would recommend to continue including both outcome efcacy and teaching efcacy as
separate scales, and study their relation with motivational components, to see whether
these scales function comparable in other higher education contexts.

123

434

High Educ (2012) 64:421436

Theoretical considerations
In our research the context-relatedness of teacher efcacy emphasized by TschannenMoran et al. (1998, 2001) is recognizable to certain extent in the distinction that can be
made empirically between the general and context-specic items. The items which asked
teachers specically to evaluate the statement from a particular context (usually from this
course module) worked differently from the items which made a more general statement.
This resulted in the inclusion of a context-related sub-scale concerning personal efcacy.
With the scales interest and effort the number of context-related items was low (two out
of seven or nine) and, partly for that reason, these were left out in the validation study. In
subsequent research it would therefore be a good idea to include both sufciently general
items and context-related items for all the scales.
In this study we observed the fundamental distinction described by Guskey and Passaro
(1994) between an internal dimension of teacher efcacy, which relates to the teacher
himself and an external dimension, relating to the teacher/teaching in the general, but only
to a limited extent. While the items from the scale teaching efcacy are all externally
formulated, the items not included in the main scale from the scale outcome efcacy are
all internally formulated. Moreover, the scale interest contains both items in which
explicit reference is made to the teacher (in the rst person: I or me) as well as items in
which a general statement about enjoyability of teaching as an activity is given.

Conclusion
In our research we developed and validated a Dutch-language instrument to measure the
teaching motivation of teachers in higher education. Previous research into motivation has
been mostly developed on an independent basis (Ambrose and Kulik 1999) and motivation
research in education has generally been carried out in primary and secondary education
and with a focus on pupils (Bailey 1999). In the present study we set out to develop a valid
questionnaire which can be used in a Dutch language context in order to be able to arrive at
conclusions concerning relevant aspects with respect to motivation. We started by using
the existing literature on self-efcacy, which has long maintained a distinction between
personal efcacy (teacher efcacy) and outcome efcacy. At a later stage this was rened
by also considering teaching efcacy as an independent concept. We added the motivational aspects interest/enjoyment and importance/effort. The validation revealed that
the aspects outcome efcacy and teaching efcacy play only a limited role in the
motivation of teachers for teaching. With regard to motivation for teaching it can therefore
be said that this is an interaction of, in the rst place, the interest which a teacher has in
teaching (cf. Deci and Ryan n.d.), the effort which he or she is prepared to put in, and the
expected personal efcacy. In the case of teachers in higher education it appears that the
question as to how far the effort made also contributes to learning (outcome efcacy) and
the question as to how far teaching is even relevant (teaching efcacy) bears only a limited
relationship to their motivation, although both aspects are part of the concept self-efcacy. This may be connected to the fact that university teachers cannot choose not to
teach, nor have they chosen to become academics just because they want to teach; for
many of them this is something which is part of their duties as members of the academic
staff.
On the basis of our validation study conducted in the context of Flemish higher education it appears that the instrument of 25 items is a good tool for establishing teachers
motivation for teaching as part of educational research or educational practice. The

123

High Educ (2012) 64:421436


436

High Educ (2012) 64:421436 435

Menges, R. J.can
(1997).
faculty motivation
to teach: Approaches
faculty development.
instrument
be Fostering
used to analyze
the relationships
between to
teaching
motivation and other
In J. L.inBess
Teaching environment
well and liking at
it: Motivating
faculty
to teach effectively.
Baltimore:
aspects
the(Ed.),
educational
the level of
the teacher
or at the level
of the
The John Hopkins University.
student.
in M.
this
studyCoping
was voluntary,
likely ourbetween
respondents
were and
Moneta, G.Participation
B., & Spada, M.
(2009).
as a mediatorsoofmost
the relationships
trait intrinsic
overall
more
highly
motivated
for
teaching
than
university
teachers
in
general.
Therefore,
extrinsic motivation and approaches to studying during academic exam preparation. Personality and
Individual
Differences,
46, 664669.
further
research
is especially
needed in other contexts, inter alia in a non-voluntary setting,
Postareff,
L.,
Lindblom-Ylanne,
& Nevgi, and
A. (2007).
The effect
pedagogical training on teaching in
in order to further conrm its validity and S.,
reliability
to optimize
theofquestionnaire.
higher education. Teaching and Teacher Education, 23, 571577.
Schiefele, U. (1999). Interest and learning from text. Scientic Studies of Reading, 3, 257280.
Soodak, L. C., & Podell, D. M. (1996). Teacher efcacy: Toward the understanding of a multi-faceted
Acknowledgments
Theand
authors
would
like to thank
Prof. Dr. F. Lardon, Prof. Dr. S. De Maeyer and
construct. Teaching
Teacher
Education,
12, 401411.
L.
Coertjens
for
their
help
in
creating,
distributing,
and
analyzing
this questionnaire.
Stes, A., Coertjens, L., & Van Petegem, P. (2010). Instructional
development
for teachers in higher
education: Impact on teaching approach. Higher Education, 60, 187204.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Upper Saddle River, NJ:
Pearson Allyn & Bacon.
Trigwell, K., Ashwin, P., Lindblom-Ylanne, S., & Nevgi, A. (2004, Juni). Variation in approaches to
university teaching: the role of regulation and motivation. Paper presented at the European Association
Bibliography
for Research on Learning and Instruction (EARLI) Higher Education Special Interest Group conference, Stockholm, Sweden.
Tschannen-Moran,
& Hoy,
W. (2001).
Teachernew
efcacy:
elusive construct.
Teaching
Ambrose, M. L., & M.,
Kulik,
C. T.A.
(1999).
Old friends,
faces:Capturing
Motivationanresearch
in the 1990s.
Journal of
Management,
and
Teacher Education,
25, 231292.
17, 783805.
Tschannen-Moran,
& Hoy, A.motivation
W. (2007).and
Theself-efcacy
differential for
antecedents
beliefsEducation
of novice
Bailey, J. G. (1999).M.,
Academics
teaching of
andself-efcacy
research. Higher
Research
and
experienced
& Development,
teachers. Teaching
18, 343359.
and Teacher Education, 23, 944956.
Tschannen-Moran,
Hoy, A. W.,Toward
& Hoy,aW.
K. (1998).
Teacher
Efcacy:
Its meaning
and measure.
Bandura, A. (1977).M.,
Self-efcacy:
unifying
theory
of behavior
change.
Psychological
Review, 84,
191215.of Educational Research, 68, 202248.
Review
Vallerand,
J., & Bissonnette,
R.In
(1992).
Intrinsic, extrinsic,
and
amotivational
predictors
of 4,
Bandura, A.R.(1994).
Self-efcacy.
V. S. Ramachaudran
(Ed.),
Encyclopedia
of styles
humanasbehavior
(Vol.
pp. 7181).
behavior:
A New
prospective
York, NY:
study.
Academic
Journal Press.
of Personality, 60, 599620.
Vockell,
E.A.
L. (1989).
(2001). Structural
Educational
psychology.
practical
approach
Ed.),
http://education.calumet.
Bollen, K.
equations
with A
latent
variables.
New(Online
York, NY:
Wiley.
Colbeck,
purdue.edu/vockell/EdPsyBook/.
C. (1998). Merging in a seamless
Accessed
blendhow
9 Feb 2011.
faculty integrate teaching and research. Journal of
Watt,
H. M.Education,
G., & Richardson,
P. W. (2008). Motivations, perceptions, and aspirations concerning teaching
Higher
69(6), 647671.
Deci,
as aE.career
L., & for
Ryan,
different
R. M. types
(n.d.)of
Intrinsic
beginning
motivation
teachers.inventory
Learning(IMI).
and Instruction,
http://www.psych.rochester.edu/
18, 408428.
Woolfolk
Hoy, A. (2008). What motivates
teachers?
Important
SDT/measures/IMI_description.php.
Accessed
9 Feb
2011. work on a complex question. Learning and
Eccles,
Instruction,
J. S., &18,
Wigeld,
492498.
A. (2002). Motivational beliefs, values, and goals. Annual Review of Psychology,
Woolfolk,
A. E., Rosoff, B., & Hoy, W. K. (1990). Teachers sense of efcacy and their beliefs about
53, 109132.
Gibbs,
managing
G., & Coffey,
students.M.
Teaching
(2004). A
and
proposal
Teacherfor
Education,
an international
6, 137148.
collaborative research programme to
identify the impact of initial training on university teachers. Active Learning in Higher Education, 5,
87100.
Gibson, S., & Dembo, M. H. (1984). Teacher efcacy: A construct validation. Journal of Educational
Psychology, 76, 569582.
Graham, S. (1991). A review of attribution theory in achievement contexts. Educational Psychology Review,
3, 539.
Guskey, T. R., & Passaro, P. D. (1994). Teacher efcacy: A study of construct dimensions. American
Education Research Journal, 31, 627643.
Hidi, S. (2001). Interest, reading, and learning: theoretical and practical considerations. Educational Psychology Review, 13, 191209.
Hoyle, R. H. (Ed.). (1995). Structural equation modeling: Concepts, issues, and applications. Thousand
Oakes, CA: Sage.
Kaplowitz, M. D., Hadlock, T. D., & Levine, R. (2004). A comparison of web and mail survey response
rates. Public Opinion Quarterly, 68(1), 94101.
Kunter, M., Tsai, Y.-M., Klusmann, U., Brunner, M., Krauss, S., & Baumert, J. (2008). Students and
mathematics teachers perceptions of teacher enthusiasm and instruction. Learning and Instruction,
18(5), 468482.
Locke, E. A. (1996). Motivation through concious goal setting. Applied and Preventive Psychology, 5,
117124.
Lombaerts, K., De Backer, F., Engels, N., Van Braak, J., & Athanasou, J. (2009). Development of the selfregulated learning teacher belief scale. European Journal of Psychology of Education, 24, 7996.
Macfarlane, B., & Hughes, G. (2009). Turning teachers into academics? The role of educational development in fostering synergy between teaching and research. Innovations in Education and Teaching
International, 46, 514.
McKenzie, J. (2002). Variation and relevance structures for university teachers learning: Bringing about
change is ways of experiencing teaching. In A. Goody, J. Herrington, & M. Northcote (Eds.), Research
and development in higher education (pp. 434441). Perth: Higher Edcuation Research and Development Society of Australasia.

123

123

Vous aimerez peut-être aussi