Dimensions of Teacher Self-Efficacy and Relations With Strain Factors,
Perceived Collective Teacher Efficacy, and Teacher Burnout
Einar M. Skaalvik and Sidsel Skaalvik
Norwegian University of Science and Technology
In this study, the authors developed and factor analyzed the Norwegian Teacher Self-Efficacy Scale.
They also examined relations among teacher self-efficacy, perceived collective teacher efficacy, external
control (teachers’ general beliefs about limitations to what can be achieved through education), strain
factors, and teacher burnout. Participants were 244 elementary and middle school teachers. The analysis
supported the conceptualization of teacher self-efficacy as a multidimensional construct. They found
strong support for 6 separate but correlated dimensions of teacher self-efficacy, which were included in
the following subscales: Instruction, Adapting Education to Individual Students’ Needs, Motivating
Students, Keeping Discipline, Cooperating With Colleagues and Parents, and Coping With Changes and
Challenges. They also found support for a strong 2nd-order self-efficacy factor underlying the 6
dimensions. Teacher self-efficacy was conceptually distinguished from perceived collective teacher
efficacy and external control. Teacher self-efficacy was strongly related to collective teacher efficacy and
Keywords: teacher self-efficacy, teacher burnout, collective teacher efficacy
During the past 2 decades, we have witnessed a growing interest
in teacher self-efficacy. Despite using different instruments, sev-
eral researchers have found that teacher self-efficacy predicts both
teaching practices and student learning. Teacher self-efficacy has
been shown to predict student motivation and achievement (Ash-
ton & Webb, 1986; Midgley, Feldlaufer, & Eccles, 1989; Moore &
Esselman, 1992; Ross, 1992), students’ self-efficacy and attitudes
(Anderson, Greene, & Loewen, 1988; Cheung & Cheng, 1997),
teachers’ goals and aspirations (Muijs & Reynolds, 2002), teach-
ers’ attitudes toward innovation and change (Fuchs, Fuchs, &
Bishop, 1992; Guskey, 1988), teachers’ tendency to refer difficult
students to special education (Meijer & Foster, 1988; Soodak &
Podell, 1993), teachers’ use of teaching strategies (Allinder, 1994;
Woolfolk, Rosoff, & Hoy, 1990), and the likelihood that teachers
will stay in the teaching profession (Burley, Hall, Villeme, &
Brockmeier, 1991; Glickman & Tamashiro, 1982). There is also
some research evidence of a relation between teacher self-efficacy
and teacher burnout (Chwalisz, Altmaier, & Russell, 1992). For
instance, Friedman and Farber (1992) reported that teachers who
considered themselves less competent in classroom management
and discipline reported a higher level of burnout than teachers with
a higher level of self-efficacy.
Despite these promising results, a problem with research on
teacher self-efficacy is that the construct has been conceptualized
and measured differently by different researchers. One purpose of
this study was to develop and factor analyze a scale of teacher
self-efficacy built on an analysis of role expectations in Norwegian
schools. A second purpose was to test whether (individual) teacher
self-efficacy could be conceptually distinguished from perceived
collective teacher efficacy and external control (teachers’ general
beliefs about limitations to what can be achieved through educa-
tion). A third purpose was to examine relations between teachers’
perception of strain factors in school, external control, perceived
collective teacher efficacy, teacher self-efficacy, and level of
The Construct of Self-Efficacy in Social Cognitive
Bandura (1986) offered a formal definition of self-efficacy:
“Perceived self-efficacy is defined as people’s judgments of their
capabilities to organize and execute courses of action required to
attain designated types of performance” (p. 391). It is a belief
about what a person can do rather than judgments about one’s
attributes, which are characteristic of self-concept (Bong & Skaal-
vik, 2003; Zimmerman & Cleary, 2006). Furthermore, self-
efficacy is a multidimensional and context-specific construct
(Zimmerman & Cleary, 2006), and there is no all-purpose measure
of self-efficacy beliefs (Bandura, 2006b). Bong (2006) under-
scored that context specificity should not be confused with level of
generality and that self-efficacy beliefs may be skill specific, task
specific, or domain specific.
Self-efficacy is grounded in the theoretical framework of social
cognitive theory, emphasizing the evolution and exercise of human
agency—the idea that people can exercise some influence over
what they do (Bandura, 2006a). Bandura (2006a) maintained that
in this conception, people are self-organizing, proactive, self-
regulating, and self-reflecting. He emphasized that people form
Einar M. Skaalvik and Sidsel Skaalvik, Department of Education, Nor-
wegian University of Science and Technology, Trondheim, Norway.
This research was supported by grants from Utdanningsforbundet
(Union of Education Norway) and the Norwegian University of Science
Correspondence concerning this article should be addressed to Einar M.
Skaalvik, Department of Education, Norwegian University of Science
and Technology, Dragvoll, Trondheim, Norway 7491. E-mail:
Journal of Educational Psychology Copyright 2007 by the American Psychological Association
2007, Vol. 99, No. 3, 611–625 0022-0663/07/$12.00 DOI: 10.1037/0022-06126.96.36.1991
intentions, set goals, anticipate likely outcomes, monitor and reg-
ulate actions, and reflect on their personal efficacy. This theory
underscores the interactions among personal factors (e.g., cogni-
tions), behaviors, and environmental conditions. From this per-
spective, self-efficacy affects one’s goals and behaviors and is
influenced by conditions in the environment (Schunk & Meece,
2006). Efficacy beliefs determine how environmental opportuni-
ties and impediments are perceived (Bandura, 2006a) and affect
choice of activities, how much effort is expended on an activity,
and how long people will persevere when confronting obstacles
Bandura (1986, 1997) pointed out four major sources of self-
efficacy beliefs: enactive mastery experiences, vicarious experi-
ences, verbal persuasion, and physiological reactions. Mastery
experiences are regarded as the most influential source of self-
efficacy (Bong & Skaalvik, 2003; Pajares, 1997). Outcomes inter-
preted as successful raise self-efficacy, whereas those interpreted
as failures undermine it. Vicarious experiences are observations of
similar others’ performance on given tasks. This source of self-
efficacy is particularly influential when people are uncertain of
their own abilities or when they have little prior experience with
the relevant activity (Schunk, 1987). Persuasive communication
may also raise self-efficacy. It is most effective when those who
convey the efficacy information are viewed as competent and
reliable. Physiological responses, such as sweating, heartbeats, and
fatigue, may be associated with prior failure and may send signals
to people that affect their efficacy expectations in given situations.
Teacher Self-Efficacy: Construct and Measurement
A common conceptualization of teacher self-efficacy is that it
refers to teachers’ beliefs in their ability to influence valued
students outcomes (e.g., Soodak & Podell, 1996; Wheatley, 2005).
Still, teacher self-efficacy has been conceptualized and measured
differently by different researchers. Historically, research on
teacher self-efficacy has been approached from two different the-
oretical bases: Rotter’s (1966) concept of internal and external
control and Bandura’s (1997) concept of self-efficacy (see
Tschannen-Moran & Woolfolk Hoy, 2001). On the basis of Rot-
ter’s distinction between external and internal control, teacher
self-efficacy has been assumed to increase if teachers believe that
students’ achievement and behavior can be influenced by educa-
tion (Guskey & Passaro, 1994; Rose & Medway, 1981). Accord-
ingly, teacher self-efficacy has also been assumed to decrease if
teachers believe that factors external to teaching (e.g., students’
abilities and home environments) are more important to the stu-
dents’ learning than the influence that a teacher may have. In
contrast, as shown earlier, Bandura defined self-efficacy as beliefs
in one’s own capability to organize and execute the courses of
action required to produce given attainments (Bandura, 1997, p. 3).
On the basis of this construct, teacher self-efficacy may be con-
ceptualized as individual teachers’ beliefs in their own abilities to
plan, organize, and carry out activities required to attain given
In an early attempt to measure teacher self-efficacy, Armor et al.
(1976) asked participants to rate two statements. We cite these
statements because several later instruments are based on them.
The two statements are as follows: (a) “When it comes right down
to it, a teacher really can’t do much because most of a student’s
motivation and performance depend on his or her home environ-
ment” and (b) “If I really try hard, I can get through to even the
most difficult and unmotivated students.” The former of these
statements measures teachers’ general beliefs about limitations to
what can be achieved through education, which is often referred to
as teaching efficacy (e.g., Soodak & Podell, 1996). To emphasize
that this is a measure of the degree to which teachers believe that
factors external to their teaching put limitations on what they can
accomplish, we refer to this dimension as external control (see also
Ho & Hau, 2004). The latter question taps teachers’ beliefs about
their personal teaching ability, which we refer to as teacher self-
efficacy. Following the work of Armor et al. (1976), researchers
have developed a number of different instruments for measuring
teacher self-efficacy (for an extensive overview, see Tschannen-
Moran & Woolfolk Hoy, 2001). Analysis of a 30-item scale
developed by Gibson and Dembo (1984) clearly shows that the
two statements used by Armor et al. measure different constructs
(see also Soodak & Podell, 1996). A possible explanation of these
results is that external control and self-efficacy are different and
relatively independent constructs. External control may be concep-
tualized as a general and relatively stable belief about limitations
to what can be achieved through education. In contrast, teacher
self-efficacy is conceptualized as a context-specific and malleable
belief about what the individual teacher can accomplish given the
limitations caused by external factors (see also Bong & Skaalvik,
On the basis of Bandura’s (1997) definition of self-efficacy,
several instruments have been developed to measure (personal)
teacher self-efficacy. To measure teacher self-efficacy as a single
dimension, Schwarzer, Schmitz, and Daytner (1999) developed a
short instrument on which teachers responded to each of 10 state-
ments on a 4-point scale from not true at all to exactly true. One
example of an item is “I am convinced that I am able to success-
fully teach all relevant subject content to even the most difficult
students.” As can be seen from this item, Schwarzer et al. followed
Bandura’s (1997) recommendation in the item construction. First,
the object in each statement was Ibecause the aim was to assess
each teacher’s subjective belief about his or her own capability.
Second, the items contained verbs like can or be able to so that the
items clearly asked for mastery expectations because of personal
competence. Moreover, each item contained a barrier, which in the
example item was to successfully teach “the most difficult stu-
dents.” This point was underlined by Bandura’s statement that “If
there are no obstacles to surmount, the activity is easy to perform,
and everyone has uniformly high perceived self-efficacy for it”
(Bandura, 1997, p. 42).
A limitation in the scale by Schwarzer et al. (1999) is that it
measures teacher self-efficacy as a one-dimensional construct,
making it less useful both for research purposes and for assessing
the need for school development. As shown earlier, self-efficacy is
a multidimensional construct (E. M. Skaalvik & Bong, 2003), and
Bandura (1997) pointed out that multifaceted teacher self-efficacy
scales will enable researchers to select those dimensions that are
most germane to the domain of functioning that the research is
designed to study. Following this reasoning, Bandura (n.d.) pre-
sented a 30-item scale measuring seven dimensions of self-
efficacy. The dimensions are influence decision making, influence
school resources, instruction, discipline, enlist parental involve-
ment, enlist community involvement, and create a positive school
612 SKAALVIK AND SKAALVIK
climate. Even though Bandura’s (n.d.) scale is multidimensional,
the dimensions seem not to be equally central to teachers’ daily
work. Furthermore, important role expectations, at least in Nor-
wegian schools, are not represented in the scale. One example is
the expectation that teachers should differentiate instruction and
assignments to meet individual student needs, which is strongly
emphasized in the Norwegian national curriculum (Læreplanver-
ket for Kunnskapsløftet, 2006).
Recognizing the need for a multidimensional scale, Tschannen-
Moran and Woolfolk Hoy (2001) developed a 24-item Teachers’
Sense of Efficacy Scale consisting of three dimensions: instruc-
tional strategies, classroom management, and student engagement.
Each dimension has high reliability, and factor analysis confirmed
the existence of three separate dimensions. Although the scale
seems well designed, a limitation is that most of the items lack
clear obstacles, which is strongly recommended by Bandura
(1997). Another problem is that teacher self-efficacy is reduced to
three dimensions. The variety of tasks and demands put on a
teacher cannot, in our view, be reduced to three dimensions.
One of the aims of this study was to develop and test a multi-
dimensional teacher self-efficacy scale following Bandura’s
(1997) recommendations for item construction. After analyzing
central tasks in teachers’ daily work (e.g., as they were described
in the national curriculum as well as other political signals given in
Norwegian school reform), we developed a scale consisting of six
subscales: Instruction, Adapting Education to Individual Students’
Needs, Motivating Students, Keeping Discipline, Cooperating
With Colleagues and Parents, and Coping With Changes and
Challenges. Another aim was to test whether the dimensions of
self-efficacy could be separated from external control (see earlier
discussion) and perceived collective teacher efficacy.
Perceived Collective Teacher Efficacy
Bandura (1997, p. 45) emphasized that self-efficacy items must
represent beliefs about personal abilities to produce specified lev-
els of performance. However, teachers do not always work alone.
In most Norwegian schools, teachers now work in teams sharing
responsibility for a larger group of students. The actual instruction
is partly done by individual teachers in smaller groups and partly
by pairs of teachers in a larger group. Much of the organizing and
the planning are done in teacher teams. The individual teachers’
self-efficacy may therefore be dependent on the functioning of the
team. Moreover, the individual teacher may also have beliefs about
the ability both of the team and of the faculty of teachers at the
school to execute courses of action required to produce given
attainments. Such beliefs represent perceived collective teacher
efficacy (Bandura, 1997; Goddard & Goddard, 2001; Goddard,
Hoy, & Woolfolk Hoy, 2004). The four major sources of self-
efficacy beliefs are also important for perceived collective efficacy
beliefs (Bandura, 1997). In particular, past school successes build
teachers’ belief in the capability of the faculty, whereas failures
tend to undermine their belief (Goddard, 2001; Goddard et al.,
2004). Still, the extent to which past school successes are evidence
based may vary among schools. For instance, Parker, Hannah, and
Topping (2006) suggested that perceived collective teacher effi-
cacy is grounded in joint experiences in schools where teachers
interact frequently to plan, observe, and evaluate teaching. In
contrast, they speculate that perceived collective teacher efficacy
may be based more on guessing in schools where teachers do not
Little research has been done on perceived collective teacher
efficacy, and Tschannen-Moran, Woolfolk Hoy, and Hoy (1998)
underlined the need to examine the relation between perceived
collective teacher efficacy and individual teacher self-efficacy. It is
reasonable to predict that perceived collective efficacy affects
individual teacher self-efficacy. One reason for this expectation is
that perceived collective efficacy may serve as a normative expec-
tation for goal attainment (Goddard et al., 2004). Schools with a
high degree of perceived collective teacher efficacy set challeng-
ing goals and are persistent in their efforts to meet these goals.
Goddard et al. (2004) argued that these high expectations create a
normative pressure that encourages all teachers to do what it takes
to excel and discourages them from giving up when faced with
difficult situations. We propose that such a cultural context pro-
motes students’ achievements, which again enhances individual
teachers’ sense of self-efficacy. Furthermore, a high degree of
perceived collective efficacy constitutes a frame factor for indi-
vidual teachers’ efficacy beliefs. The better the instruction given
by other teachers at a school, the more able and motivated the
students will be. Hence, the goals set by all teachers in a school
may be more challenging. One may also speculate that high
perceived collective efficacy may have the opposite effect on some
teachers. By comparing themselves with highly efficacious col-
leagues, some teachers may feel that they are not able to reach the
same standard. However, social comparison is more important for
development of self-concept than self-efficacy (Marsh, Walker, &
Debus, 1991; E. M. Skaalvik & Bong, 2003).
Unfortunately, few studies have explored the relations among
perceived collective efficacy, student achievement, and individual
teacher self-efficacy. The few available studies suggest moderate
positive relations between perceived collective efficacy and stu-
dent achievement (Bandura, 1993; Goddard, Hoy, & Woolfolk
Hoy, 2000; Mawhinney, Haas, & Wood, 2005; Tschannen-Moran
& Barr, 2004) and between perceived collective efficacy and
individual teacher self-efficacy (Goddard & Goddard, 2001). We
hypothesize that perceived collective teacher efficacy is predictive
of individual teacher self-efficacy.
Maslach and Jackson (1981; see also Maslach, Jackson, &
Leiter, 1996) described burnout as a syndrome of emotional ex-
haustion, depersonalization, and reduced personal accomplish-
ment. An increased feeling of emotional exhaustion is described as
the key aspect of burnout (Maslach et al., 1996). Burnout is
conceptualized as resulting from long-term occupational stress,
particularly among human service workers, including teachers
(Jennett, Harris, & Mesibov, 2003). Although the reasons may
differ, all teachers experience stress in their work (Jennett et al.,
2003). The stressors may include students with behavioral prob-
lems, problems in the parent–teacher relationship, conflict with
colleagues, or having to organize teaching in new ways as a
consequence of working in teams or because of school reforms.
Most teachers cope successfully with such stress, for instance,
through active problem solving, social and emotional support from
colleagues, reorganizing the teaching situation, cooperating with
parents, or changing their teaching strategy. However, burnout
may be the endpoint of coping unsuccessfully with chronic stress
(Jennett et al., 2003).
Research evidence reveals a moderate but systematic relation
between teacher self-efficacy and teacher burnout (Chwalisz et al.,
1992; Evers, Brouwers, & Tomic, 2002; Friedman & Farber,
1992). There is less agreement about how to explain the relation
between teacher self-efficacy and teacher burnout. A possible
explanation is that low teacher self-efficacy may result in feelings
of burnout (Bandura, 1997; Evers et al., 2002). For instance,
Brouwers and Tomic (2000) speculated that teachers who doubt
their ability to manage disruptive students can blame students for
their doubts and therefore develop negative attitudes toward stu-
dents. Low expectations of classroom management also increase
occupational stress, which may increase emotional exhaustion.
Bandura (1997) noted that teachers with low self-efficacy view
many aspects of their environment as fraught with danger, dwell
on their coping deficiencies, and magnify the severity of possible
threats. This pattern of cognitive and emotional responses may be
expected to heighten emotional exhaustion and depersonalization.
Consequently, we hypothesized a negative relation between
teacher self-efficacy and teacher burnout. Furthermore, we ex-
pected a positive relation between perceived strain factors and
teacher burnout, partly mediated through teacher self-efficacy.
With the exception of the Maslach Burnout Inventory (Maslach
et al., 1996; see later discussion), all instruments were developed
for the current study and were administered in Norwegian. Sample
items as well as the scales presented in Appendixes A through D
represent translations from Norwegian into English.
Norwegian Teacher Self-Efficacy Scale (NTSES)
As stated earlier, one purpose of this study was to develop and
test a scale of teacher self-efficacy. Through an analysis of role
expectations in Norwegian schools, we decided to develop a scale
that consisted of six subscales: Instruction, Adapting Education to
Individual Students’ Needs, Motivating Students, Keeping Disci-
pline, Cooperating With Colleagues and Parents, and Coping With
Changes and Challenges. Our aim was to develop a 24-item scale
with 4 items measuring each of the six dimensions. We had two
reasons for deciding to develop a 24-item scale. First, the scale was
to be used together with other scales. To maintain the teachers’
motivation to participate, the scale could not be too large. Second,
we needed enough items in each subscale to ensure satisfactory
reliability. As a compromise, we decided to use 4 items in each
subscale. For that purpose, we started with 6 items for each
dimension, with the exception of the Coping With Changes and
Challenges subscale, which had 5 items. We followed Bandura’s
(1997) recommendations for item construction, including barriers
in the item formulations. Responses were given on a 7-point scale
from not certain at all (1) to absolutely certain (7). The six
subscales are described next.
Instruction. An important task for all teachers is to explain
subject matter so that students understand the basic principles. This
dimension focuses on the teacher’s expectation of being able to
instruct students, explain subject matter, advise students in their
work, and answer questions to improve students’ understanding.
An example of an item is as follows: “How certain are you that you
can provide good guidance and instruction to all students regard-
less of their level of ability?”
Adapting Education to Individual Students’ Needs. Since
1985, the national Norwegian curriculum has emphasized that
education should be adapted to individual students’ needs (Møn-
sterplan for grunnskolen, 1985, p. 24). This is also strongly em-
phasized in the current school reform (Læreplanverket for
kunnskapsløftet, 2006, pp. 33–34). Adapting education to the
needs of individual students is seen as a key element in the
movement toward inclusive education (Leithwood, Edge, & Jantzi,
1999, p. 99). However, research in Norwegian schools has shown
that teachers perceive this goal as extremely demanding and that
many teachers do not know how to address the diversity of
students’ needs and abilities (E. M. Skaalvik & Fossen, 1995). One
example of an item measuring this dimension of teacher self-
efficacy is as follows: “How certain are you that you can provide
realistic challenge for all students even in mixed ability classes?”
Motivating Students. Optimal learning is dependent on student
motivation. Motivating students is therefore an important task for
all teachers. This is also strongly emphasized in the current na-
tional curriculum (Læreplanverket for kunnskapsløftet, 2006, pp.
32–33). One of the dimensions included in the scale was therefore
teacher self-efficacy for motivating students. An example of an
item is as follows: “How certain are you that you can wake the
desire to learn even among the lowest-achieving students?”
Keeping Discipline. A national Norwegian survey of 7th-,
10th-, and 11th-grade students revealed that 23% of the students
often experienced so much noise at school that it was disturbing
(E. M. Skaalvik, Furre, Danielsen, & Jamt, 2006). The ability to
maintain order and discipline was therefore included as one di-
mension of teacher self-efficacy. An example of an item is as
follows: “How certain are you that you can get students with
behavioral problems to follow classroom rules?”
Cooperating With Colleagues and Parents. Teachers in Nor-
wegian schools are increasingly required to cooperate with col-
leagues and parents. In most schools, teachers now work in teams
sharing responsibility for a larger group of students. Additionally,
they are expected to cooperate extensively with parents, partly
informing parents about schoolwork and partly making decisions
together with parents. An example of an item from this dimension
is as follows: “How certain are you that you can collaborate
constructively with parents of students with behavioral problems?”
Coping With Changes and Challenges. During the past 10 to
15 years, Norwegian schools have undergone a number of reforms,
and the demands put on teachers have changed. One example is the
change from classroom teaching, where each teacher was primarily
responsible for a class of 20 to 30 students, into team teaching,
where teacher teams are responsible for all students at a given
grade level. A second example is the radical integration of disabled
students into regular educational settings followed by a strong
demand for differentiation of instruction. A third example is the
increasing power of school principals to decide instructional meth-
ods, which may vary both among schools and within schools over
time. Accordingly, the ability to cope with ongoing changes and
new challenges was included as a dimension of teacher self-
efficacy. An example of an item from this dimension is as follows:
614 SKAALVIK AND SKAALVIK
“How certain are you that you can teach well even if you are told
to use instructional methods that would not be your choice?”
Because the Teachers’ Sense of Efficacy Scale (Tschannen-
Moran & Woolfolk Hoy, 2001) is a much used measure of teacher
self-efficacy, it may be useful to compare the NTSES with this
scale. The Teachers’ Sense of Efficacy Scale includes three sub-
scales: Instructional Strategies, Classroom Management, and Stu-
dent Engagement. Although these subscales are not identical with
those in the NTSES, they are similar to Instruction, Keeping
Discipline, and Motivating Students, respectively. The remaining
NTSES subscales, Adapting Education to Individual Students’
Needs, Cooperating With Colleagues and Parents, and Coping
With Changes and Challenges, are not represented as separate
subscales in the Teachers’ Sense of Efficacy Scale.
Perceived Collective Teacher Efficacy
Perceived collective teacher efficacy was measured by a seven-
item scale. The scale was intended as a one-dimensional measure.
The items focused on instruction, motivation, controlling student
behavior, addressing students’ needs, and creating a safe environ-
ment. To mark the difference from items in the NTSES, all items
focused on what “we” or “teachers at this school” were able to do.
An example of an item is as follows: “As teachers of this school
we can get even the most difficult students engaged in their
schoolwork.” Responses were given on a 5-point scale from false
(1) to true (5).
External control was measured by a five-item scale. Each item
stated a limitation to what can be achieved through education
concerning students’ learning, achievement, motivation, or behav-
ior. The limitations were described as students’ abilities or home
environment. An example of an item is as follows: “How much
students can learn in school is primarily determined by their
abilities.” Responses were given on a 6-point scale from false (1)
to true (6).
Four strain factors in teachers’ daily work were identified
through conversations with 24 teachers. Analysis of the conversa-
tions revealed four strain factors, which were brought up by
approximately half of the teachers: students with behavior prob-
lems, conflicts with parents, conflicts among the teachers, and
having to organize teaching in ways one did not believe were the
best. An example of an item is as follows: “There are students with
severe behavioral problems in my class (or group of students).”
Responses were given on a 5-point scale from false (1) to true (5).
Teacher burnout was measured by means of the 22-item
Maslach Burnout Inventory—Educators Survey (Maslach et al.,
1996). The scale includes three subscales that measure dimensions
of teacher burnout: Emotional Exhaustion (9 items), Depersonal-
ization (5 items), and Reduced Personal Accomplishment (8
items). Participants rated statements indicating that their work
makes them feel emotionally drained (emotional exhaustion), that
they do not care about some students (depersonalization), and that
their jobs have allowed them to accomplish many things (personal
accomplishment). Responses were given on a 7-point scale from
never (0) to every day (6). Responses to items measuring personal
accomplishment were scored so that high scores indicated reduced
feeling of accomplishment.
Participants and Procedure
Participants in this study were 246 teachers from 12 elementary
schools and middle schools (1st–10th grade) in a large region in
Norway. The schools were drawn at random from two small cities
and a large rural area. In 10 of the schools, a particular time was
set aside for all teachers to respond to the questionnaire, and the
teachers were instructed not to discuss the items or to collaborate.
All teachers in these schools participated in the study. In 2 schools,
the principals did not allow teachers to participate during working
hours, and the teachers responded to the questionnaire by taking it
home. In these 2 schools, 60% of the teachers returned the ques-
tionnaire. Two of the returned questionnaires had missing values
and were excluded from the analysis. Thus, the analysis was based
on 244 teachers.
In the sample, 63% were women. The age of the teachers varied
from young teachers (the youngest was 27 years old) to teachers
close to retirement (the oldest was 65 years old). The mean age
was 45 years. The average number of years in the teaching pro-
fession was 14. The schools varied with respect to size from
schools with 5 teachers to schools with 44 teachers. About half of
the teachers in the sample (53%) worked in elementary schools
(Grades 1–7), whereas 27% worked in middle schools (Grades
8 –10), and 20% worked in combined elementary schools and
middle schools. Of the teachers, 41% worked in schools with
traditional classes of students, whereas 59% worked in schools
where a team of teachers shared responsibility for all students at a
given grade level.
The NTSES was first analyzed by means of Cronbach’s alpha.
In each dimension, the 4 items giving the highest internal consis-
tency (alpha) were selected. The resulting 24 items were then
analyzed by means of both exploratory and confirmatory factor
analyses. Confirmatory factor analyses were used to test three
models. The NTSES was constructed to measure a multidimen-
sional construct with six subdimensions, and Model 1 specified six
factors consistent with these dimensions. To further test the as-
sumption that teacher self-efficacy is multidimensional, we also
tested a model in which all 24 items loaded on a single factor
(Model 2). Because we expected the six factors in Model 1 to
correlate, we also specified a model with six primary factors and
one higher order factor (Model 3).
To assess model fit, we used well-established indices, such as
the comparative fit index (CFI), the incremental fit index (IFI), the
Tucker–Lewis index (TLI), and the root-mean-square error of
approximation (RMSEA), as well as the chi-square test statistics.
For the CFI, IFI, and TLI indices, values greater than .90 are
typically considered acceptable, and values greater than .95 indi-
cate good fit to the data (Bollen, 1989; Byrne, 2001; Hu & Bentler,
1999). For well-specified models, an RMSEA of .06 or less re-
flects a good fit (Hu & Bentler, 1999).
After testing Cronbach’s alpha for the Perceived Collective
Teacher Efficacy Scale and the External Control Scale, a confir-
matory factor analysis was conducted by means of the AMOS 5
program (Arbuckle, 1999). The model specified eight factors: the
six dimensions of teacher self-efficacy as well as collective teacher
efficacy and external control.
The next step in the data analysis was to conduct a series of
regression analyses predicting external control, collective teacher
efficacy, teacher self-efficacy, and teacher burnout. Finally, we
tested a path model for four latent traits (external control, per-
ceived collective teacher efficacy, teacher self-efficacy, and
teacher burnout) by means of structural equation modeling analy-
sis with the AMOS 5 program.
Analysis of Scales
Table 1 shows correlations among the study variables as well as
statistical means, standard deviations, and Cronbach’s alphas. The
six dimensions of teacher self-efficacy had satisfactory-to-high
reliability in terms of Cronbach’s alpha. Moreover, Cronbach’s
alpha changed only negligibly as a result of reducing the assess-
ment of each dimension to the four best items. This was done by
selecting the four items resulting in the highest internal consis-
tency. Even with the resulting four items, five of the dimensions
had alpha values between .80 and .91. The lowest alpha (.74),
which is still satisfactory, was found for Cooperation With Parents
and Colleagues. The correlations among the six subscales were
moderate, ranging from .33 to .54. All dimensions of teacher
self-efficacy were positively related to perceived collective effi-
cacy. The correlations ranged from .29 to .46. However, the six
dimensions of teacher self-efficacy correlated close to zero with
external control. The correlations ranged from .01 to ⫺.16.
The three subscales in the Maslach Burnout Inventory had
varying reliability. Cronbach’s alphas for Emotional Exhaustion,
Depersonalization, and Reduced Personal Accomplishment were
.89, .61, and .79, respectively. Whereas Emotional Exhaustion and
Reduced Personal Accomplishment had satisfactory reliability, the
third subscale, Depersonalization, had marginal internal consis-
tency. The correlation between the three subscales ranged between
.32 and .37.
As expected, the four strain factors (students with behavioral
problems, conflicts with parents, conflicts among the teachers, and
having to organize teaching in ways one did not believe were the
best) were weakly correlated (between .03 and .26). In accordance
with these expectations, Cronbach’s alpha, which is not shown in
Table 1, displayed low internal consistency (.49). The four items
measuring strain factors were therefore treated as separate vari-
ables in the following regression analyses and were not included in
the structural equation modeling analyses.
The 24 NTSES items were further analyzed by means of ex-
ploratory factor analysis with maximum likelihood extraction,
varimax rotation, and eigenvalues greater than 1 (Table 2). The
analysis extracted six factors consistent with the theoretical model
(see description of dimensions in the Method section). These
factors explained 61% of the variance in the equation. With two
exceptions, all expected factor loadings were greater than .5, and
none of the remaining factor loadings were greater than .4.
The next step in the analyses was to conduct confirmatory factor
analyses testing three different models. All three models were
based on the resulting 24 items in the NTSES. The fit indexes are
shown in Table 3. Model 1 defined six primary factors, and none
of the correlations between error terms were set free. All indicators
had high correlations with their respective factors, and the model
had acceptable fit to the data (Table 3). The correlations among the
six factors ranged between .40 and .67, with 11 of the 15 corre-
lations ranging between .50 and .60. To further test the multidi-
mensionality of teacher self-efficacy, all 24 items loaded on a
single primary factor in Model 2. This model did not fit the data
well (see Table 3). The third model defined six primary factors (as
in Model 1) and a single secondary (higher order) factor. This
model had good fit to the data and strong relations between all
primary factors and the secondary factor (.77, .76, .76, .73, .73, and
The model testing clearly indicated that teacher self-efficacy is
a multidimensional construct. However, the model testing also
showed that for research purposes, teacher self-efficacy may be
analyzed as a latent trait based on the six subscales.
Collective Teacher Efficacy and External Control
Both the Perceived Collective Teacher Efficacy Scale and the
External Control Scale had acceptable reliability. Cronbach’s al-
phas were .79 for both scales. To test that these constructs were
conceptually distinct from teacher self-efficacy, we conducted a
confirmatory factor analysis specifying eight factors: the six di-
mensions in the NTSES as well as collective teacher efficacy and
external control. The model had acceptable fit to the data,
N⫽244) ⫽920.02, CFI ⫽.92, IFI ⫽.92, TLI ⫽.91, RMSEA ⫽
.050, and with one exception, all indicators had high correlations
with their respective factors (Table 4). The result of the analysis
clearly indicated that teacher self-efficacy should be conceptually
distinguished from perceived collective teacher efficacy and ex-
Relations Among Variables
Analyses of relations among the variables were first explored by
means of a series of regression analyses (Table 5). Two separate
analyses were calculated with size of school, organization of
instruction, gender, number of years in the teaching profession,
and the four strain factors as predictors of external control and
perceived collective teacher efficacy. All predictors were entered
simultaneously into the equation. Gender of the teachers and two
of the strain factors (discipline and conflict with parents) were not
significantly related to external control or to perceived collective
teacher efficacy. Perceived collective teacher efficacy was nega-
tively related to number of years in the teaching profession (␤⫽
⫺.23). Moreover, perceived collective teacher efficacy was
slightly lower for teachers who experienced conflict among the
teachers (␤⫽⫺.18) and for teachers who felt that they had to
organize teaching in ways they did not believe were the best (␤⫽
616 SKAALVIK AND SKAALVIK
Zero-Order Correlations and Descriptive Statistics
Study variable 1 2 3 4 5678 9 10111213141516171819
1. Size of school — .10 .04 .05 .08 .09 .05 .01 .05 .06 .11 .05 .07 .12 .12 .13 ⫺.09 ⫺.04 ⫺.12
2. Organization — ⫺.01 ⫺.03 ⫺.05 ⫺.02 .06 .01 ⫺.08 .10 .13 .06 .08 .05 .04 .01 ⫺.13 ⫺.06 ⫺.01
3. Gender — .10 .07 ⫺.02 .11 .04 .08 ⫺.13 ⫺.10 .23 ⫺.02 ⫺.05 ⫺.11 .09 .02 .16 .11
4. Length — ⫺.13 ⫺.06 ⫺.01 ⫺.08 .06 ⫺.20 ⫺.17 ⫺.10 ⫺.20 ⫺.18 ⫺.35 ⫺.19 .11 ⫺.01 .11
5. Strain 1 (behavior) — .22 .08 .03 .09 ⫺.12 ⫺.01 .01 ⫺.07 .06 ⫺.01 .01 .18 .08 .07
6. Strain 2 (conflict with
— .11 .15 .09 ⫺.10 ⫺.15 ⫺.19 ⫺.19 ⫺.04 ⫺.15 ⫺.15 .26 .30 .12
7. Strain 3 (method) — .26 .19 ⫺.20 ⫺.19 ⫺.06 ⫺.21 ⫺.19 ⫺.21 ⫺.20 .23 .23 .21
8. Strain 4 (conflict among
—⫺.06 ⫺.21 ⫺.07 ⫺.01 ⫺.09 ⫺.11 ⫺.12 ⫺.14 .11 .13 .05
9. External control — ⫺.25 ⫺.15 ⫺.12 ⫺.16 ⫺.09 ⫺.09 .01 .19 .22 .16
10. Collective efficacy — .35 .28 .45 .29 .46 .35 ⫺.26 ⫺.25 ⫺.34
11. TSE 1 (adapting) — .48 .52 .45 .45 .46 ⫺.36 ⫺.36 ⫺.34
12. TSE 2 (discipline) — .54 .48 .33 .48 ⫺.31 ⫺.29 ⫺.31
13. TSE 3 (motivating) — .52 .44 .41 ⫺.29 ⫺.31 ⫺.32
14. TSE 4 (instruction) — .44 .49 ⫺.22 ⫺.34 ⫺.39
15. TSE 5 (coping) — .53 ⫺.30 ⫺.30 ⫺.31
16. TSE 6 (cooperating) — .34 ⫺.35 ⫺.35
17. Burnout 1 (exhaustion) — .37 .35
18. Burnout 2
19. Burnout 3 (RPA) —
M35.55 1.61 1.40 14.30 3.21 1.55 2.48 1.53 14.88 24.71 18.61 18.63 18.41 21.34 19.45 21.14 27.51 9.12 17.50
SD 17.83 0.49 0.49 10.85 1.51 1.03 1.15 0.80 4.90 3.46 3.63 3.85 3.16 2.61 3.58 2.64 9.49 3.42 5.71
Cronbach’s alpha — — — — ———— .79 .79 .87 .90 .91 .81 .80 .74 .89.61.79
Note. Correlations of .13 or higher are significant ( p⬍.05). Size of school ⫽number of teachers at the school; organization ⫽teaching organized as school classes (1) or teams of teachers with
responsibility for a larger group of students (2); length ⫽number of years as a teacher; method ⫽having to teach in ways one does not believe are the best; TSE ⫽teacher self-efficacy; adapting ⫽
adapting instruction to individual needs; discipline ⫽maintaining discipline; motivating ⫽motivating students; coping ⫽coping with changes and challenges; cooperating ⫽cooperating with parents
and colleagues; RPA ⫽reduced personal accomplishment.
⫺.14). External control was positively related to having to orga-
nize teaching in ways the individual teachers did not believe were
the best (.21).
Each of the six NTSES subscales was regressed on size of
school, organization of instruction, gender, number of years in the
teaching profession, the four strain factors, external control, and
perceived collective teacher efficacy. The organization of instruc-
tion in traditional classes with one teacher or in teams of teachers
sharing responsibility for larger groups of students was not sys-
tematically related to any of the self-efficacy measures. Efficacy
for Coping With Changes and Challenges and efficacy for Coop-
erating With Colleagues and Parents were positively but weakly
related to size of the school (.13 and .14, respectively). Compared
with female teachers, male teachers had significantly higher self-
efficacy for maintaining discipline (␤⫽.27) and for cooperating
with colleagues and parents (␤⫽.15). Length of service as a
teacher was negatively related to three of the NTSES subscales:
Motivating Students (␤⫽⫺.14), Coping With Changes and
Challenges (␤⫽⫺.28), and Cooperating With Colleagues and
Parents (␤⫽⫺.15). Two of the strain factors, students with
behavioral problems and conflicts among the teachers, were not
significantly related to any of the NTSES subscales. However,
conflict with parents was negatively related to five of the six
NTSES subscales (␤values between .14 and .19). Also, having to
organize teaching in ways one did not believe were the best was
negatively related to four of the subscales (␤values between .13
and .17). External control was not systematically related to the
NTSES subscales. Perceived collective teacher efficacy was mod-
erately related to all NTSES subscales (␤values between .27 and
The three subscales of the Maslach Burnout Inventory were
regressed on size of school, organization of instruction, gender,
number of years in the teaching profession, the four strain
factors, external control, perceived collective teacher efficacy,
and a total score of teacher self-efficacy. Burnout was most
strongly related to teacher self-efficacy (␤values varying be-
tween ⫺.32 and ⫺.40). Additionally, exhaustion was positively
related to two of the strain factors, students with behavioral
problems and conflict with parents (.15 and .14, respectively).
Depersonalization was positively related to conflict with par-
ents (.19) and to external control (.15). Moreover, depersonal-
ization decreased with length of service as a teacher (⫺.13) and
was higher among male than among female teachers (.17).
Reduced accomplishment was not significantly related to any of
the predictor variables except self-efficacy.
Exploratory Factor Analysis of the 24-Item Teacher Self-Efficacy Scale
Adapting 1 .77 .16 .15 .15 .15 .09
Adapting 2 .77 .19 .23 .09 .13 .26
Adapting 3 .76 .15 .20 .14 .14 .26
Adapting 4 .75 .20 .22 .20 .21 .06
Discipline 1 .11 .81 .27 .19 .16 .09
Discipline 2 .22 .78 .20 .21 .14 .12
Discipline 3 .21 .71 .09 .13 .10 .24
Discipline 4 .16 .67 .31 .13 ⫺.03 .23
Motivating 1 .20 .16 .76 .18 .11 .12
Motivating 2 .27 .14 .70 .21 .15 .15
Motivating 3 .18 .29 .68 .19 .15 .11
Motivating 4 .15 .23 .64 .09 .21 .10
Instruction 1 .20 .18 .17 .76 .17 .14
Instruction 2 .13 .11 .18 .70 .12 .15
Instruction 3 .13 .25 .14 .63 .04 .25
Instruction 4 .09 .14 .31 .40 .16 .27
Coping 1 .18 .14 .09 .08 .77 .22
Coping 2 .08 .11 .16 .05 .63 .08
Coping 3 .22 ⫺.07 .21 .21 .62 .30
Coping 4 .15 .13 .16 .27 .48 .34
Cooperating 1 .20 .26 .09 .22 .05 .67
Cooperating 2 .15 .08 .05 .21 .28 .57
Cooperating 3 .05 .10 .16 .15 .18 .52
Cooperating 4 .21 .23 .12 .05 .23 .50
Note. Numbers in bold represent factor loadings. Adapting ⫽adapting instruction to individual needs; discipline ⫽maintaining discipline; motivating ⫽
motivating students; coping ⫽coping with changes and challenges; cooperating ⫽cooperating with parents and colleagues.
df TLI CFI IFI RMSEA
Model 1 489.32 237 .91 .92 .92 .066
Model 2 1,572.71 252 .56 .60 .61 .147
Model 3 516.98 246 .91 .92 .92 .067
Note. Model 1 specified six primary factors, and none of the correlations
between error terms were set free. Model 2 specified one single primary
factor. Model 3 specified six primary factors and one second-order factor.
TLI ⫽Tucker–Lewis index; CFI ⫽comparative fit index; IFI ⫽incre-
mental fit index; RMSEA ⫽root-mean-square error of approximation.
618 SKAALVIK AND SKAALVIK
Relations among the variables were further analyzed by means
of structural equation modeling with the AMOS 5 program. We
tested a theoretical model including four latent variables: external
control, perceived collective teacher efficacy, teacher self-efficacy,
and teacher burnout (see Figure 1). External control was indicated
by the five items of the External Control Scale, and perceived
collective teacher efficacy was indicated by the seven items of the
Perceived Collective Teacher Efficacy Scale. Teacher self-efficacy
was indicated by the six NTSES subscales representing the six
dimensions of the construct, and burnout was indicated by the
three Maslach Burnout Inventory subscales representing emotional
exhaustion, depersonalization, and reduced personal accomplish-
External control was defined as teachers’ general beliefs about
limitations to what can be achieved through education. Although
Confirmatory Factor Analysis of Items in the Teacher Self-
Efficacy Scale, the Perceived Collective Teacher Efficacy Scale,
and the External Control Scale
Adapting 1 .88
Adapting 2 .86
Adapting 3 .84
Adapting 4 .81
Discipline 1 .89
Discipline 2 .88
Discipline 3 .79
Discipline 4 .77
Motivating 1 .82
Motivating 2 .81
Motivating 3 .81
Motivating 4 .73
Instruction 1 .82
Instruction 2 .75
Instruction 3 .75
Instruction 4 .59
Coping 1 .80
Coping 2 .73
Coping 3 .71
Coping 4 .61
Cooperating 1 .77
Cooperating 2 .67
Cooperating 3 .64
Cooperating 4 .57
Collective 1 .66
Collective 2 .65
Collective 3 .64
Collective 4 .61
Collective 5 .55
Collective 6 .52
Collective 7 .52
External 1 .86
External 2 .78
External 3 .54
External 4 .37
External 5 .31
Note. Adapting ⫽adapting instruction to individual needs; discipline ⫽
maintaining discipline; motivating ⫽motivating students; coping ⫽cop-
ing with changes and challenges; cooperating ⫽cooperating with parents
and colleagues; collective ⫽perceived collective teacher efficacy; exter-
nal ⫽external control.
Summary of Regression Analyses Predicting External Control, Perceived Collective Teacher Efficacy, Teacher Self-Efficacy, and Teacher Burnout
Teacher self-efficacy Teacher burnout
Adapting Discipline Motivating Instruction Coping Cooperating Exhaustion Depersonalization RPA
Size of school .03 .08 .11 .06 .05 .10 .13
⫺.08 ⫺.03 ⫺.10
Organization ⫺.09 .09 .09 .04 .02 .02 ⫺.01 ⫺.02 ⫺.09 ⫺.01 .05
Gender .05 ⫺.08 ⫺.05 .27
.05 .01 ⫺.03 .15
Length .06 ⫺.23
⫺.11 ⫺.07 ⫺.14
Strain 1 (behavior) .06 ⫺.12 .05 .06 .00 .10 .04 .04 .15
Strain 2 (conflict with parents) .07 ⫺.05 ⫺.14
Strain 3 (method) .21
⫺.13 ⫺.04 ⫺.14
.10 .06 .09
Strain 4 (conflict among teachers) ⫺.13 ⫺.18
.04 .07 .04 ⫺.01 .00 ⫺.03 .05 .05 ⫺.04
External control — — ⫺.03 ⫺.04 ⫺.01 .02 .05 .12 .11 .15
Collective efficacy — — .27
.00 .02 ⫺.13
Teacher self-efficacy — — — — — — — — ⫺.32
.08 .16 .19 .20 .26 .20 .33 .23 .27 .31 .25
Note. Teacher self-efficacy is a total score. PCTE ⫽perceived collective teacher efficacy; adapting ⫽adapting instruction to individual needs; discipline ⫽maintaining discipline; motivating ⫽
motivating students, coping ⫽coping with changes and challenges; cooperating ⫽cooperating with parents and colleagues; RPA ⫽reduced personal accomplishment; size of school ⫽number of
teachers at the school; organization ⫽teaching organized as school classes (1) or teams of teachers with responsibility for a larger group of students (2); length ⫽number of years as a teacher;
behavior ⫽students with behavioral problems; method ⫽having to teach in ways one does not believe are the best.
such beliefs are formed through experiences, we regard external
control as a relatively stable belief about limitations to what can be
achieved through education (see earlier discussion). We therefore
specified a model with external control as an exogenous variable.
Following our introductory discussion, we also let perceived col-
lective teacher efficacy predict teacher self-efficacy. Moreover, the
model was designed to let teacher self-efficacy predict teacher
burnout. None of the correlations between error terms were set
The model had an acceptable fit to the data,
(183, N⫽244) ⫽
307.83; TLI ⫽.91, CFI ⫽.92, IFI ⫽.92, and RMSEA ⫽.053.
-to-df ratio was 1.68, which is adequate on the basis of
Kline’s (1998) rule of values of less than 3 being considered
adequate. Teacher burnout was strongly related to teacher self-
efficacy (⫺.76). Teacher self-efficacy was also strongly related to
perceived collective teacher efficacy (.64). Perceived collective
teacher efficacy was not directly related to burnout; however, there
was a moderate indirect relation between perceived collective
teacher efficacy and burnout that was mediated through teacher
self-efficacy (⫺.49). External control was negatively but weakly
related to perceived collective teacher efficacy (⫺.27) and was
weakly but directly related to teacher burnout (.21), whereas
external control did not relate directly to teacher self-efficacy.
Because one of the subscales in the Maslach Burnout Inventory,
Reduced Personal Accomplishment, conceptually may overlap
measures of self-efficacy, we analyzed an alternative structural
model in which teacher burnout was indicated only by the Emo-
tional Exhaustion subscale and the Depersonalization subscale.
The alternative model had acceptable fit to the data,
244) ⫽283.05, TLI ⫽.91, CFI ⫽.92, IFI ⫽.92, and RMSEA ⫽
-to-df ratio was 1.73. The relations among perceived
collective teacher efficacy, teacher self-efficacy, and burnout did
not change compared with the results shown in Figure 1. The path
from external control to teacher burnout changed from .21 to .24,
and the path from external control to perceived collective teacher
efficacy changed from ⫺.27 to ⫺.28. Hence, the pattern of results
did not change, and changes in the coefficients were negligible.
One purpose of the present study was to develop and test the
NTSES on the basis of role expectations in Norwegian schools.
The analysis clearly supports the conceptualization of teacher
self-efficacy as a multidimensional construct. We found strong
support for six separate but correlated dimensions of teacher
self-efficacy, which were included in the following subscales:
Instruction, Adapting Education to Individual Students’ Needs,
Motivating Students, Keeping Discipline, Cooperating With Col-
leagues and Parents, and Coping With Changes and Challenges.
We also found support for a strong second-order self-efficacy
factor underlying the six dimensions. Each dimension had high
reliability in terms of Cronbach’s alpha.
The study revealed a particularly strong correlation between
teacher self-efficacy and teacher burnout, supporting the validity
of the NTSES. The relation between teacher self-efficacy and
teacher burnout is stronger than has been found in previous re-
Figure 1. Structural model of external control (EC), perceived collective teacher efficacy (CE), teacher
self-efficacy (SE), and teacher burnout (TB). Standardized solution (ns ⫽not significant).
620 SKAALVIK AND SKAALVIK
search. A possible explanation is that we examined the relation
between latent traits by means of structural equation modeling.
Furthermore, the validity of the NTSES was likely strengthened by
including as many as six dimensions of this construct as indicators
of the latent trait.
Both the regression analyses and the structural equation mod-
eling analysis were designed to let teacher self-efficacy predict
teacher burnout. We should therefore warn against causal inter-
pretation of the relation and emphasize that the study merely
shows relations between the constructs. The relation between these
constructs is likely reciprocal. Efficacy beliefs determine how
environmental opportunities and impediments are perceived (Ban-
dura, 2006a). People with low self-efficacy tend to dwell on their
coping deficiencies and magnify the severity of possible threats
(Bandura, 1997). Such construal of new situations as threatening
may lead to increased anxiety, which is energy consuming in itself.
Additionally, Bandura (1997) emphasized that individuals with
low self-efficacy resort to escapist modes of coping that create
even more strain and distress (see also S. Skaalvik, 2004). Low
mastery expectations may be particularly stressful for teachers
because they may be accompanied by expectations of disciplinary
problems and lower student performance, followed by possible
conflict with parents and school principals. Such expectations may
also represent a threat to an individual’s identity as a teacher and
may elicit defensive mechanisms that heighten emotional exhaus-
tion and depersonalization. Hence, the strong relation between
teacher self-efficacy and teacher burnout found in this study un-
derscores the importance of teacher self-efficacy. However,
teacher burnout may also affect teacher self-efficacy. Self-efficacy
beliefs are constructed largely on the basis of one’s prior mastery
experiences. Hence, emotional exhaustion may result in reduced
accomplishments, which again may affect self-efficacy negatively.
More research is needed to explore causal relations between these
This study clearly demonstrates that teacher self-efficacy should
be distinguished from perceived external control (often called
teaching efficacy), which we have defined as teachers’ general
beliefs about limitations to what can be achieved through educa-
tion. Confirmatory factor analysis showed that external control
should be distinguished from the six dimensions of teacher self-
efficacy. Moreover, perceived external control had no predictive
value for teacher self-efficacy, although it was weakly but directly
related to teacher burnout. We may therefore hypothesize that
perceived external control and teacher self-efficacy are separate
constructs with independent influence on teacher stress and burn-
out. Future research should therefore include both constructs,
although the present study indicates that teacher self-efficacy is
most strongly related to teacher burnout.
The conclusion that perceived external control and teacher self-
efficacy are practically unrelated constructs needs to be further
analyzed in future research. We have defined external control as
teachers’ general beliefs about limitations to what can be achieved
through education. Within these limits, teachers vary in their
expectations of what they personally can accomplish, which con-
stitutes measures of teacher self-efficacy. We may speculate that
most teachers have optimistic views about what can be accom-
plished through education and that they do not perceive narrow
limitations. If so, the perceived limitations to what can be accom-
plished through education may not affect teachers’ expectations
about what they personally can accomplish. However, these spec-
ulations call for qualitative studies in which teachers reflect on
reasons for their efficacy expectations.
Confirmatory factor analysis (Table 4) also indicates that
teacher self-efficacy and perceived collective teacher efficacy
should be treated as separate constructs. Still, an important finding
is that these constructs are positively and strongly related. Al-
though there may be reciprocal causal relations between these
constructs, the structural equation modeling analysis (Figure 1)
was designed to let perceived collective teacher efficacy predict
teacher self-efficacy. There are several reasons why we expect
perceived collective efficacy to influence individual teachers’ self-
efficacy. As outlined in the introductory discussion, Goddard et al.
(2004) argued that perceived collective efficacy serves as a nor-
mative expectation for goal attainment. High collective self-
efficacy leads to challenging goals and persistence in teachers’
efforts to meet those goals. We have argued that such a cultural
context promotes student engagement and achievement, which
again enhance individual teachers’ sense of self-efficacy. The
more able and motivated the students are, the more challenging the
goals that may be set by all teachers in school. A possible effect of
perceived collective teacher efficacy on individual teacher self-
efficacy may therefore be mediated through student motivation
and achievement. The positive relation between perceived collec-
tive teacher efficacy and teacher self-efficacy may also be inter-
preted as an effect of vicarious experiences. Observing colleagues
managing different aspects of teaching may increase individual
teachers’ self-efficacy, particularly when teachers work in teams
and have ample opportunities to observe each other. A practical
implication of this reasoning may be that one should attempt to
raise teachers’ competencies collectively through school develop-
ment and in-service training, rather than sending individual teach-
ers to courses and workshops outside of the school. However,
these are theoretical speculations that need to be tested in longi-
This study was designed to test how external control, teacher
self-efficacy, and teacher burnout related to four strain factors:
teaching students with behavior problems, conflicts with parents,
conflicts among the teachers, and having to organize teaching in
ways one did not believe were the best. Measures of the four strain
factors were entered separately as predictor variables in regression
analyses. The strongest and most consistent relations were found
for conflicts with parents and having to organize teaching in ways
one did not believe were the best. Conflict with parents related
negatively to five of the six dimensions of teacher self-efficacy as
well as to emotional exhaustion and depersonalization. This result
underscores the importance of collaborating with parents and of
informing parents about academic results as well as about goals
and measures. The teachers’ feelings of having to organize teach-
ing in ways they did not believe were the best were negatively
related to four of the six dimensions of self-efficacy as well as to
perceived collective self-efficacy. Furthermore, it was positively
related to external control. Norwegian schools have a long tradi-
tion of teacher autonomy over educational methods. Whereas the
national curriculum has defined the learning content, the teachers
have been free to choose teaching and learning methods. The
increased tendency for teachers to work in teams and share re-
sponsibility for a larger group of students may, for some teachers,
result in a feeling that they are forced to use teaching and learning
methods that they do not believe are the best or that they do not
feel comfortable with. This is a possible side effect of working in
teams that should be given serious attention by researchers as well
as by school leaders. Researchers have also pointed out that,
internationally, the use of prepackaged and predesigned materials
tends to diminish teachers’ autonomy (e.g., Ballet, Kelchtermans,
& Loughran, 2006). Given the importance of teacher autonomy
indicated in this study, more research should focus both on pro-
cesses by which teacher autonomy is reduced and on the impact of
such a development.
Several limitations in the present study should be pointed out.
The sample of teachers in this study was drawn from only one
region in Norway. Although this was a large region, it may not be
representative of the nation. To use the NTSES in school evalua-
tion and school development, self-efficacy norms should be based
on a national representative sample. Also, the analysis of strain
factors in this study was based on a few factors. Future research
should study a variety of contextual conditions and possible strain
factors. Qualitative interviews of teachers might be used in pilot
studies to explore strain factors that teachers feel are particularly
important. It is also important to emphasize that the path model in
Figure 1 builds on a theoretical model and that it shows only
relations between constructs and not causality. In future research,
longitudinal studies are needed.
The six dimensions of the NTSES were developed through an
analysis of role expectations in Norwegian schools, with an emphasis
on recent school reforms. We therefore expected these dimensions to
be psychologically central to teachers. Put differently, we expected
that these dimensions would be perceived as important dimensions of
teacher competence. This expectation was supported by the strong
correlation between self-efficacy and burnout but should be tested
explicitly in future research. Also, the validity of the dimensions
should be tested in educational contexts other than Norwegian
schools. Furthermore, even though the existence of the six dimensions
was supported empirically, other possible dimensions of teacher self-
efficacy should be explored in future research.
Because the NTSES consists of six correlated dimensions, this
instrument is particularly well suited both for research purposes and
school development. For research purposes, we have demonstrated
the validity of a latent self-efficacy variable based on the six domains
of teacher self-efficacy. However, future analyses should also exam-
ine whether the six dimensions of teacher self-efficacy are differently
related to other constructs and if they have different impact on
teaching strategies and teacher behavior. Research addressing such
issues is of vital importance. The NTSES may also prove to be a
valuable instrument in school development and in-service training,
particularly when one of the goals is to raise teacher competence. By
using a representative sample to establish a norm, it is possible to
explore at which dimensions of competence teachers at particular
schools feel most uncertain. These areas of competence may then be
given particular attention in school development and in upgrading
teachers’ competence and self-beliefs.
Allinder, R. M. (1994). The relationship between efficacy and the instruc-
tional practices of special education teachers and consultants. Teacher
Education and Special Education, 17, 86 –95.
Anderson, R., Greene, M., & Loewen, P. (1988). Relationships among
teachers’ and students’ thinking skills, sense of efficacy, and student
achievement. Alberta Journal of Educational Research, 34, 148 –165.
Arbuckle, J. L. (1999). AMOS 4.0 [Computer software]. Chicago: Small-
Armor, D., Conroy-Oseguera, P., Cox, M., King, N., McDonell, L., Pascal,
A., et al. (1976). Analysis of the school preferred reading program in
selected Los Angeles minority schools (Report No. R-2007-LAUSD).
Santa Monica, CA: Rand Corporation. (ERIC Document Reproduction
No. 130 243)
Ashton, P. T., & Webb, R. B. (1986). Making a difference: Teachers’ sense
of efficacy and student achievement. New York: Longman.
Ballet, K., Kelchtermans, G., & Loughran, J. (2006). Beyond intensifica-
tion towards a scholarship of practice: Analysing changes in teachers’
work lives. Teachers and Teaching: Theory and Practice, 12, 209 –229.
Bandura, A. (1986). Social foundations of thought and action. A social
cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A. (1993). Perceived self-efficacy in cognitive development and
functioning. Educational Psychologist, 28, 117–148.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
Bandura, A. (2006a). Adolescent development from an agentic perspective.
In F. Pajares & T. Urdan (Eds.), Self-efficacy beliefs of adolescents (pp.
1– 43). Greenwich, CT: Information Age.
Bandura, A. (2006b). Guide for constructing self-efficacy scales. In F.
Pajares & T. Urdan (Eds.), Self-efficacy beliefs of adolescents (pp.
307–337). Greenwich, CT: Information Age.
Bandura, A. (n.d.). Bandura’s Instrument Teacher Self-Efficacy Scale.
Retrieved June 6, 2007, from http://www.coe.ohio-state.edu/ahoy/
Bollen, K. A. (1989). A new incremental fit index for general structural
models. Sociological Methods & Research, 17, 303–316.
Bong, M. (2006). Asking the right question. How confident are you that
you could successfully perform these tasks? In F. Pajares & T. Urdan
(Eds.), Self-efficacy beliefs of adolescents (pp. 287–305). Greenwich,
CT: Information Age.
Bong, M., & Skaalvik, E. M. (2003). Academic self-concept and self-efficacy:
How different are they really? Educational Psychology Review, 15, 1– 40.
Brouwers, A., & Tomic, W. (2000). A longitudinal study of teacher
burnout and perceived self-efficacy in classroom management. Teaching
and Teacher Education, 16, 239 –253.
Burley, W. W., Hall, B. W., Villeme, M. G., & Brockmeier, L. L. (1991,
April). A path analysis of the mediating role of efficacy in first-year
teachers’ experiences, reactions, and plans. Paper presented at the annual
meeting of the American Educational Research Association, Chicago.
Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic
concepts, applications, and programming. Mahwah, NJ: Erlbaum.
Cheung, W. M., & Cheng, Y. C. (1997, March). A multi-level analysis of
teachers’ self-belief and behavior, and students’ educational outcomes.
Paper presented at the annual meeting of the American Educational
Research Association, Chicago.
Chwalisz, K., Altmaier, E. M., & Russell, D. W. (1992). Causal attribu-
tions, self-efficacy cognitions, and coping with stress. Journal of Social
and Clinical Psychology, 11, 377– 400.
Evers, W. J. G., Brouwers, A., & Tomic, W. (2002). Burnout and self-
efficacy: A study on teachers’ beliefs when implementing an innovative
educational system in the Netherlands. British Journal of Educational
Psychology, 72, 227–243.
Friedman, I. A., & Farber, B. A. (1992). Professional self-concept as a
predictor of teacher burnout. Journal of Educational Research, 86, 28 –35.
Fuchs, L. S., Fuchs, D., & Bishop, N. (1992). Instructional adaptation for
students at risk. Journal of Educational Research, 86, 70 – 84.
Gibson, S., & Dembo, M. (1984). Teacher efficacy: A construct validation.
Journal of Educational Psychology, 76, 569 –582.
Glickman, C. D., & Tamashiro, R. T. (1982). A comparison of first-year,
622 SKAALVIK AND SKAALVIK
fifth-year, and former teachers on efficacy, ego-development, and
problem-solving. Psychology in the Schools, 19, 558 –562.
Goddard, R. D. (2001). Collective efficacy: A neglected construct in the
study of schools and student achievement. Journal of Educational Psy-
chology, 93, 467– 476.
Goddard, R. D., & Goddard, Y. L. (2001). A multilevel analysis of the
relationship between teacher and collective efficacy in urban schools.
Teacher and Teacher Education, 17, 807– 818.
Goddard, R. D., Hoy, W. K., & Woolfolk Hoy, A. (2000). Collective
teacher efficacy: Its meaning, measure, and effect on student achieve-
ment. American Educational Research Journal, 37, 479 –507.
Goddard, R. D., Hoy, W. K., & Woolfolk Hoy, A. (2004). Collective
efficacy beliefs: Theoretical developments, empirical evidence, and fu-
ture directions. Educational Researcher, 33, 3–13.
Guskey, T. R. (1988). Teacher efficacy, self-concept, and attitudes toward
the implementation of instructional innovation. Teaching and Teacher
Education, 4, 63– 69.
Guskey, T. R., & Passaro, P. D. (1994). Teacher efficacy: A study of construct
dimensions. American Educational Research Journal, 31, 627– 643.
Ho, I. T., & Hau, K.-T. (2004). Australian and Chinese teacher efficacy:
Similarities and differences in personal instruction, discipline, guidance
efficacy and beliefs in external determinants. Teaching and Teacher
Education, 20, 313–323.
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
structure analysis: Conventional criteria versus new alternatives. Structural
Equation Modeling: A Multidisciplinary Journal, 6, 1–55.
Jennett, H. K., Harris, S. L., & Mesibov, G. B. (2003). Commitment to
philosophy, teacher efficacy, and burnout among teachers of children with
autism. Journal of Autism and Developmental Disorders, 33, 583–593.
Kline, R. B. (1998). Principles and practice of structural equation mod-
eling. New York: Guilford.
Læreplanverket for kunnskapsløftet. (2006). Oslo, Norway: Kunnskapsde-
Leithwood, K., Edge, K., & Jantzi, D. (1999). Educational accountability:
The state of the art. Gu¨ tersloh, Germany: Bertelsmann Foundation.
Marsh, H. W., Walker, R., & Debus, R. (1991). Subject specific compo-
nents of academic self concept and self efficacy. Contemporary Educa-
tional Psychology, 16, 331–345.
Maslach, C., & Jackson, S. E. (1981). Maslach Burnout Inventory manual.
Mountain View, CA: Consulting Psychologists Press.
Maslach, C., Jackson, S. E., & Leiter, M. P. (1996). Maslach Burnout
Inventory manual (3rd ed.). Mountain View, CA: Consulting Psychol-
Mawhinney, H. B., Haas, J., & Wood, C. (2005, November). Teachers’
perceptions of collective efficacy and school conditions for professional
learning. Paper presented at the annual meeting of the University Coun-
cil for Educational Administration, Nashville, Tennessee.
Meijer, C. J. W., & Foster, S. F. (1988). The effect of teacher self-efficacy
on referral change. Journal of Special Education, 22, 378 –385.
Midgley, C., Feldlaufer, H., & Eccles, J. (1989). Change in teacher efficacy
and student self- and task-related beliefs in mathematics during the transi-
tion to junior high school. Journal of Educational Psychology, 81, 247–258.
Mønsterplan for grunnskolen. (1985). Oslo, Norway: Aschehoug.
Moore, W., & Esselman, M. (1992, April). Teacher efficacy, power, school
climate and achievement: A desegregating district’s experience. Paper
presented at the annual meeting of the American Educational Research
Association, San Francisco.
Muijs, D., & Reynolds, D. (2002). Teacher beliefs and behaviors: What
matters. Journal of Classroom Interaction, 37, 3–15.
Pajares, F. (1997). Current directions in self-efficacy research. In H. W.
Marsh, R. G. Craven, & D. M. McInerney (Eds.), International advances
in self research (pp. 1– 49). Greenwich, CT: Information Age.
Parker, K., Hannah, E., & Topping, K. J. (2006). Collective teacher
efficacy, pupil attainment and socio-economic status in primary school.
Improving Schools, 9, 111–129.
Rose, J. S., & Medway, F. J. (1981). Measurement of teachers’ beliefs in their
control over student outcome. Journal of Educational Research, 74, 185–190.
Ross, J. A. (1992). Teacher efficacy and the effect of coaching on student
achievement. Canadian Journal of Education, 17, 51– 65.
Rotter, J. B. (1966). Generalized expectancies for internal versus external
control of reinforcement. Psychological Monographs, 80, 1–28.
Schunk, D. H. (1987). Peer models and children’s behavioral change.
Review of Educational Research, 57, 149 –174.
Schunk, D. H., & Meece, J. L. (2006). Self-efficacy development in
adolescence. In F. Pajares & T. Urdan (Eds.), Self-efficacy beliefs of
adolescents (pp. 71–96). Greenwich, CT: Information Age.
Schwarzer, R., Schmitz, G. S., & Daytner, G. T. (1999). Teacher self-
efficacy. Retrieved June 6, 2007, from http://userpage.fu-berlin.de/⬃
Skaalvik, E. M., & Bong, M. (2003). Self-concept and self-efficacy revis-
ited: A few notable differences and important similarities. In H. W.
Marsh, R. G. Craven, & D. M. McInerney (Eds.), International advances
in self research (pp. 67– 89). Greenwich, CT: Information Age.
Skaalvik, E. M., & Fossen, I. (1995). Tilpassing og differensiering. Idealer
og realiteter i norsk grunnskole [Adapting education to individual needs:
Ideals and realities in Norwegian Schools]. Trondheim, Norway: Tapir.
Skaalvik, E. M., Furre, H., Danielsen, I.-J., & Jamt, R. (2006). Som elevene
ser det. Analyse av den nasjonale undersøkelsen “Elevinspektørene” i
2005—revidert utgave [As the students see it: An analysis of the national
survey “Elevinspektørene” (Student Inspectors) 2005]. Kristiansand,
Norway: Oxford Research.
Skaalvik, S. (2004). Reading problems in school children and adults:
Experiences, self-perceptions, and strategies. Social Psychology of Ed-
ucation, 7, 105–125.
Soodak, L. C., & Podell, D. M. (1993). Teacher efficacy and student
problem as factors in special education referral. Journal of Special
Education, 27, 66 – 81.
Soodak, L. C., & Podell, D. M. (1996). Teacher efficacy: Toward the
understanding of a multi-faceted construct. Teaching and Teacher Ed-
ucation, 12, 401– 411.
Tschannen-Moran, M., & Barr, M. (2004). Fostering student achievement:
The relationship between collective teacher efficacy and student
achievement. Leadership and Policy in Schools, 3, 187–207.
Tschannen-Moran, M., & Woolfolk Hoy, A. (2001). Teacher efficacy: Capturing
an elusive construct. Teaching and Teacher Education, 17, 783– 805.
Tschannen-Moran, M., Woolfolk Hoy, A., & Hoy, W. K. (1998). Teacher effi-
cacy: Its meaning and measure. Review of Educational Research, 68, 202–248.
Wheatley, K. F. (2005). The case for reconceptualizing teacher efficacy
research. Teaching and Teacher Education, 21, 747–766.
Woolfolk, A. E., Rosoff, B., & Hoy, W. K. (1990). Teachers’ sense of
efficacy and their beliefs about managing students. Teacher and Teacher
Education, 6, 137–148.
Zimmerman, B. J., & Cleary, T. J. (2006). Adolescents’ development of
personal agency. The role of self-efficacy beliefs and self-regulatory
skills. In F. Pajares & T. Urdan (Eds.), Self-efficacy beliefs of adoles-
cents (pp. 45– 69). Greenwich, CT: Information Age.
Norwegian Teacher Self-Efficacy Scale
How certain are you that you can:
1. Explain central themes in your subjects so that
even the low-achieving students understand.
8. Provide good guidance and instruction to all stu-
dents regardless of their level of ability.
12. Answer students’ questions so that they under-
stand difficult problems.
16. Explain subject matter so that most students un-
derstand the basic principles.
Adapt Instruction to Individual Needs
5. Organize schoolwork to adapt instruction and as-
signments to individual needs.
11. Provide realistic challenge for all students even in
mixed ability classes.
18. Adapt instruction to the needs of low-ability stu-
dents while you also attend to the needs of other
students in class.
23. Organize classroom work so that both low- and
high-ability students work with tasks that are
adapted to their abilities.
2. Get all students in class to work hard with their
10. Wake the desire to learn even among the lowest
15. Get students to do their best even when working
with difficult problems.
21. Motivate students who show low interest in
6. Maintain discipline in any school class or group
9. Control even the most aggressive students.
14. Get students with behavioral problems to follow
19. Get all students to behave politely and respect the
Cooperate With Colleagues and Parents
3. Cooperate well with most parents.
7. Find adequate solutions to conflicts of interest
with other teachers.
13. Collaborate constructively with parents of stu-
dents with behavioral problems.
22. Cooperate effectively and constructively with
other teachers, for example, in teaching teams.
Cope With Change
4. Successfully use any instructional method that the
school decides to use.
17. Manage instruction regardless of how it is organized
(group composition, mixed age groups, etc.).
20. Manage instruction even if the curriculum is
24. Teach well even if you are told to use instruc-
tional methods that would not be your choice.
(1) not certain at all, (3) quite uncertain, (5) quite certain,
(7) absolutely certain.
624 SKAALVIK AND SKAALVIK
Perceived Collective Teacher Efficacy Scale
1. As teachers of this school, we can get even the
most difficult pupils engaged in their school-
2. Teachers in this school prevent mobbing effec-
3. As teachers of this school, we handle conflicts
constructively because we work as a team.
4. At this school, we have a common set of rules and
regulations that enables us to handle disciplinary
5. Teachers in this school successfully address indi-
vidual pupils’ needs.
6. At this school, we are able to create a safe and
inclusive atmosphere even in the most difficult
7. Teachers at this school succeed in teaching math
and language skills even to low-ability pupils.
(1) false, (2) mostly false, (3) sometimes false/sometimes
true, (4) mostly true, (5) true.
External Control Scale
1. How much pupils can learn in school is primarily
determined by their abilities.
2. If the pupils have not learned discipline at home,
there is not much the school can do.
3. A teacher cannot do much to improve students’
achievements if they have low abilities for school-
4. It is practically impossible for a teacher to motivate
a student for academic work if he or she lacks
support and stimulation at home.
5. Good teaching is more important to students’ en-
gagement in schoolwork than is their home envi-
(1) false, (2) mostly false, (3) more false than true, (4)
more true than false, (5) mostly true, (6) true.
1. There are students with severe behavioral problems
in my class (or group of students).
2. My relation to some of the parents involves conflict.
3. There are many conflicts among the teachers in my
4. I often feel that I have to organize teaching in ways
that I do not believe are the best or that would not
be my choice.
(1) false, (2) mostly false, (3) sometimes false/sometimes
true, (4) mostly true, (5) true.
Received July 11, 2006
Revision received March 20, 2007
Accepted March 22, 2007 䡲