-
fee
, Lentzeallee 94, 14195 Berlin, Germany6, 10099 Berlin,
Germanyge 15, 60325 Frankfurt am Main, GermanyEducation,
Olshausenstrae 62, 24118 Kiel, Germany45, 1419
eitherstudy wcher efemotiothe pro
. All rights reserved.
cially stressful period in the socialization of beginning
teachers(Fantilli & McDougall, 2009; Gold, 1996; Veenman,
1984). Relativeto their more experienced colleagues, beginning
teachers tend toleave the profession at a higher rate (Ingersoll,
2001; Kukla-Acevedo, 2009), report lower teacher efcacy (Wolters
&
ational stress and& Baumert, 2012;se difculties and
the high degree of teacher attrition among beginning teachers
inthe United States, school districts and states have introduced
formalteacher induction programs that include orientation sessions,
in-service training, classroom observation, formative
assessments,and the support of colleagues (Barnett,
Hopkins-Thompson, &Hoke, 2002; Gold, 1996; Huling-Austin,1990).
Although the speciccontent of induction programs varies, many of
them pair upbeginning teachers with experienced teachers, otherwise
known asmentor teachers (Ingersoll & Kralik, 2004).
* Corresponding author. Humboldt University Berlin, Unter den
Linden 6, 10099Berlin, Germany. Tel.: 49 30 2093 46522.
Contents lists availab
Teaching and Tea
.e
Teaching and Teacher Education 36 (2013) 166e177E-mail address:
[email protected] (D. Richter).Cooperating
teacherMentor supportBeginning teacher inductionInduction
supportProfessional development
2013 Elsevier Ltd
1. Introduction
The rst years of teaching are frequently described as an
espe-
Daugherty, 2007), and perceive higher occupemotional exhaustion
(Klusmann, Kunter, Voss,Tynjl & Heikkinen, 2011). In response
to theAccepted 17 July 2013
Keywords:Mentor
that explains a successful career start. In particular,
mentoring that follows constructivist rather thantransmissive
principles of learning fosters the growth of teacher efcacy,
teaching enthusiasm, and jobsatisfaction and reduces emotional
exhaustion.a r t i c l e i n f o
Article history:Received 12 April 2012Received in revised form26
April 2013
a b s t r a c t
This study examines the extent to which the quality of mentoring
and its frequency during the rst yearsof teaching inuence teachers
professional competence and well-being. Analyses are based on a
sampleof more than 700 German beginning mathematics teachers who
participated in a pre-test/post-test studyover the course of one
year. Findings indicate that it is the quality of mentoring rather
than its frequencyaMax Planck Institute for Human
DevelopmentbHumboldt University Berlin, Unter den LindencGoethe
University Frankfurt, Senckenberganlad Leibniz Institute for
Science and Mathematicse Free University Berlin, Habelschwerdter
Allee
h i g h l i g h t s
Mentoring quality can be described as Constructs are reliably
measured in a Constructivist mentoring improves tea Constructivist
mentoring also reduces Transmissive mentoring barely
affects0742-051X/$ e see front matter 2013 Elsevier
Ltd.http://dx.doi.org/10.1016/j.tate.2013.07.0125 Berlin,
Germany
constructivist or transmissive.ith more than 700 beginning
teachers.cacy, enthusiasm and job satisfaction.nal
exhaustion.fessional development of teachers in these
respects.Jrgen Baumert a
Dirk Richter a,b,*, Mareike Kunter a,c, Oliver Ldtke a,b, Uta
Klusmann a,d, Yvonne Anders a,e,How different mentoring approaches
afdevelopment in the rst years of practic
journal homepage: wwwAll rights reserved.ct beginning
teachers
le at ScienceDirect
cher Education
lsevier .com/locate/ tate
-
eachThe role of mentor teachers and their effects in the
inductionprocess has received much attention in educational
research (seeoverviews in Hawkey, 1997; Hennissen, Crasborn,
Brouwer,Korthagen, & Bergen, 2008; Hobson, Ashby, Malderez,
&Tomlinson, 2009; Ingersoll & Strong, 2011; Wang &
Odell, 2002;Wang, Odell, & Schwille, 2008). Studies have
consistently shownthat novices perceive their mentor teachers to be
one of their mostimportant sources of support during the rst year
of teaching(Carter & Francis, 2001; Lindgren, 2005; Luft &
Cox, 2001; Marable& Raimondi, 2007). Lindgren (2005) reported
that mentors providecrucial practical advice about teaching and
teachers roles (e.g.,cooperating with parents). Furthermore,
together with their men-tee, they reect on positive and negative
occurrences in lessons. Inaddition to the qualitative literature on
teacher mentoring, quan-titative studies have shown that the
support of a mentor teacher ispositively associated with teacher
efcacy (LoCasale-Crouch, Davis,Wiens, & Pianta, 2012), teaching
commitment (Rots, Aelterman,Vlerick, & Vermeulen, 2007),
well-being (Kessels, Beijaard, Veen,& Verloop, 2008), and
improved instructional practice (Rozelle &Wilson, 2012;
Stanulis & Floden, 2009).
These empirical ndings indicate that mentoring can
positivelyaffect the transition to the teaching profession.
However, there isvery little longitudinal research investigating
howmentor teachersimpact the development of beginning teachers
professionalcompetence and well-being. In addition, few studies
examine dif-ferences in the quality of mentoring provided (Kessels
et al., 2008;Rots et al., 2007). Consequently, little is known
about which men-toring approaches best support teachers development
in the rstyears of practice. Wang et al. (2008, p. 138) note this
gap in theliterature and recommend that researchers explore the
connec-tions between these mentoring practices and what
beginningteachers think and do.
The goal of this study is therefore to examine whether
qualityand frequency of mentoring predict beginning teachers
develop-ment of professional competence and well-being in the rst
twoyears of their career. The analyses are based on a framework
ofteachers professional competence that differentiates
betweenprofessional knowledge, beliefs, motivational orientation,
and self-regulation (Baumert & Kunter, 2006). In particular, we
investigatethe effects of mentoring on teacher efcacy, teacher
enthusiasm,beliefs about learning, emotional exhaustion and job
satisfaction.
The literature review begins with a brief discussion of the
goalsof mentoring and its intended outcomes. Drawing on the
discussionon the quality of teacher mentoring, we outline two
theoreticalapproaches proposed by Cochran-Smith and Paris (1995)
andFeiman-Nemser (1998, 2001). These approaches provide
thefoundations for the concepts of constructivist- and
transmission-oriented mentoring that constitute the theoretical
framework ofthis study.
1.1. Goals of mentoring
Our conceptualization of the goals of mentoring is in
alignmentwith Golds (1996) approach to induction support, which
distin-guishes between instructional and psychological support. In
addi-tion, we draw on literature from organizational psychology,
whichdescribes role modeling as an additional function performed
bymentors (Scandura & Ragins, 1993).
Instructional support fosters the development of the
knowledgeand skills needed to succeed in the classroom (Gold,
1996). It in-cludes assistance with lesson planning, advice on
classroom man-agement, instruction-related advice and feedback,
help withassessing student work, and other activities that promote
begin-ning teachers instructional knowledge and skills. It can be
hy-
D. Richter et al. / Teaching and Tpothesized that this type of
support impacts not only beginningteachers competence but also the
quality of instruction and stu-dent learning. Indeed, beginning
teachers report that mentoringimproves their instructional skills
(Borko &Mayeld,1995; Feiman-Nemser & Buchmann, 1987;
Fletcher & Barrett, 2004). However,studies using observer
ratings of instructional quality have yieldedmixed results.
Stanulis and Floden (2009) compared the instruc-tional quality of
beginning teachers who received intensive men-toring with that of a
comparison group who did not. Their ndingsshowed that mentored
teachers outperformed non-mentoredteachers in the areas of
classroom atmosphere, instruction, andstudent engagement. In
contrast, Glazerman et al. (2008) comparedthe effectiveness of
comprehensive induction programs (developedby the Educational
Testing Service and the New Teacher Center atthe University of
California, Santa Cruz) with that of regular districtinduction.
This study showed that beginning teachers in thecomprehensive
programs received more mentoring and partici-pated more frequently
in professional development activities, butthat this was not
reected in either improved instruction orincreased student-test
scores. In conclusion, the subjectiveperception of beginning
teachers is that mentoring develops theirskills. However, there is
no clear evidence that mentoring neces-sarily leads to observable
improvements in the quality ofinstruction.
Psychological support includes building condence,
encouragingself-esteem, listening, and enhancing self-reliance
(Gold, 1996).This type of support is especially relevant at the
start of the rstyear of teaching, as beginning teachers adjust to
their new workenvironment. Psychological support is thought to
foster individualwell-being in terms of reduced stress levels and
enhanced jobsatisfaction. Strong psychological supportmay also
reduce attrition.A number of studies describing the interactions
between mentorand beginning teachers has shown that beginning
teachers receiveand value psychological support (Ballantyne,
Hansford, & Packer,1995; Hall, Draper, Smith, & Bullough,
2008; Odell & Ferraro,1992), but there has been little
empirical investigation of its ef-fects on their development. One
study investigated the antecedentsof beginning teachers well-being
and identied mentor support asan important predictor (Kessels et
al., 2008). Other studies haveinvestigated the relevance of mentor
support for teacher attrition,but ndings are inconclusive. Results
presented by Odell andFerraro (1992) and Smith and Ingersoll (2004)
indicate that men-toring has the potential to reduce beginning
teacher attrition. Incontrast, Glazerman et al. (2008) found no
difference in the attri-tion rates of beginning teachers in
comprehensive induction pro-grams and those receiving regular
district induction. Organizationalpsychologists have examined the
effects of mentoring in otherprofessional contexts (Noe,
Greenberger, & Wang, 2002; Wanberg,Welsh, & Hezlett, 2003).
Although the goals of mentoring in thecompany context differ,
employees who participated in a mentor-ing program showed greater
job satisfaction (Seibert, 1999). Thissuggests that mentoring may
facilitate individuals well-being,regardless of the organizational
context. However, there is noclear evidence that mentoring affects
beginning teachers devel-opment longitudinally.
Role modeling is provided when beginning teachers observetheir
mentors teaching. Although beginning teachers wereexposed to
thousands of hours of teaching during their own time atschool, they
can now use their professional knowledge to reect ontheir
observations. This gives them the opportunity to analyzeteaching
from an external perspective, which can provide newinsightsdfor
example, into how to organize instruction andinteract with
students. Classroom observation can also serve as abasis for
further discussion with the mentor. By providing a rolemodel for
beginning teachers, the latter can be socialized into the
er Education 36 (2013) 166e177 167teaching community and learn
how to act as professional.
-
achTherefore, role modeling differs from the two previous
goalsbecause it does not require active support on the part of
thementor.The empirical literature has shown that beginning
teachers drawconclusions for their own teaching from watching their
mentorsteach (Feiman-Nemser & Buchmann, 1987). However, beyond
thending that some beginning teachers perceive their mentors as
rolemodels, research has not investigated the distinct importance
ofrole modeling for the development of beginning teachers. It
seemsreasonable to hypothesize that role modeling contributes to
thedevelopment of instructional knowledge and skills,
althoughempirical evidence is scarce.
This overview of the three goals of mentoring provides a
briefsummary of mentor teachers responsibilities and it identies
po-tential outcomes of mentoring. Our review of the
literaturerevealed that many studies have investigated the support
begin-ning teachers receive without linking these ndings to
beginningteachers development. Furthermore, no common
theoreticalframework of mentoring has yet been developed, and
differentmeasures have been used to assess mentoring experiences.
Somestudies have investigated whether mentored and
non-mentoredteachers differed from each other (Stanulis &
Floden, 2009);others have examined the activities of mentor
teachers (Kesselset al., 2008; Rots et al., 2007) and related them
to externalcriteria (e.g., teacher efcacy, well-being). Due to this
heterogene-ity, caution must be exercised in comparing the results
of differentstudies. A more insightful understanding of the effects
of teachermentoring requires more careful differentiation of the
quality ofmentoring provided. It is not sufcient to examine whether
men-toring is available to beginning teachers. Rather, we need
toinvestigate the quality and quantity of interactions
betweenmentorand beginning teachers and to examine which factors
predictprofessional growth in the rst years of teaching. To this
end, werst drew on existing theoretical approaches to teacher
mentoringand linked them to the literature on learning theory.
1.2. Models of mentoring
Models of mentoring provide a theoretical distinction
betweendifferent (prototypical) styles of arranging mentorenovice
re-lationships (see overview in Wang & Odell, 2002). They
norma-tively describe the roles of mentor and novice, their
interactions,and the way learning is organized in this
relationship. A diversity ofsuch models has been developed in the
past two decades; we focuson two that still receive attention in
the literature (Bradbury &Koballa, 2008; Schwille, 2008;
Stanulis & Floden, 2009; Wang,2002). First, the model proposed
by Cochran-Smith and Paris(1995) characterizes mentoring as either
knowledge transmissionor knowledge transformation. Second, the
model proposed byFeiman-Nemser (1998, 2001) introduced the concept
of educativementoring, which it distinguished from conventional
mentoring.
Cochran-Smith and Paris (1995) distinguished two approachesto
mentoring, which they labeled knowledge transmission andknowledge
transformation (Wang & Odell, 2002). According to theknowledge
transmission model, mentors perceive their role asexpert teachers
and transmit their knowledge within a hierar-chically structured
relationship. In this learning environment,novices are socialized
into the prevailing culture of schooling,which manifests the status
quo. In contrast, the knowledge trans-formation model assumes an
asymmetrical but collaborative rela-tionship with the mentor
teacher, in which knowledge aboutteaching is mutually generated.
This approach facilitates the ex-change and generation of ideas and
may support change andinnovation in classroom practice.
Feiman-Nemser (1998, 2001) coined the term educative men-
D. Richter et al. / Teaching and Te168toring, as distinguished
from conventional mentoring. In educativementoring, which can be
traced back to Dewey (1938), mentorsprovide opportunities that
foster growth and development. Theyinteract with their novices in a
way that supports inquiry and thatenables them to learn in and from
their practice (Feiman-Nemser,2001). In contrast, conventional
mentoring focuses on situationaladjustment to the new school
environment, technical advice, andemotional support. Moreover,
conventional approaches view thebeginning teacher as a recipient of
knowledge and the mentor asthe expert teacher.
The two approaches proposed by Cochran-Smith and Paris(1995) and
Feiman-Nemser (1998, 2001) show similarities to twoparadigms of
learning theory (Sfard, 1998). The knowledge trans-mission model
and conventional mentoring are based on behav-iorist theories of
learning, which conceptualize learning as theaccumulation of
knowledge provided by experts. From thisperspective, learning is a
unidirectional process in which learnersare passive recipients of
information. We label the mentoring stylethat is in line with these
ideas of learning transmission-orientedmentoring.
The knowledge transformation model and educative mentoringreect
a constructivist learning theory. According to this theory,learners
construct their own knowledge by connecting new in-formation to
their prior knowledge (Shuell, 2001). Learning is anactive process
that takes place in a social community (Brown,Collins, &
Duguid, 1989). Lave and Wenger (1991) described ap-prenticeships as
an example of this form of learning. In appren-ticeships, novices
are introduced into a community through activeparticipation in
authentic tasks. Novices acquire mastery in skills asthey gradually
become more involved in the community and itsactivities. A strict
hierarchical relationship between expert andnovice is not assumed.
In line with these ideas of learning we labelthis mentoring style
constructivist-oriented mentoring.
In sum, both constructs, constructivist- and
transmission-oriented mentoring, can be used to distinguish the
quality ofmentoring. They are related to different paradigms of
learningtheory (constructivist and behaviorist learning theory) and
are inline with other approaches that describe teacher
mentoring.Moreover, both approaches can be linked to the goals of
mentoringintroduced in the previous section. In particular, the
descriptions ofconstructivist- and transmission-oriented mentoring
illustrate twodifferent forms of instructional support because the
primary goal ofmentoring is to foster professional knowledge and
skills. However,emotional support and role modeling of mentors is
not explicitlyincluded in both models. Therefore, both approaches
characterizeonly one part of all the functions mentors
exercise.
Although both constructs have not yet been measured
empiri-cally, the theoretical literature on teacher mentoring
suggests thatconstructivist-oriented mentoring is best suited to
foster beginningteachers development. Carter and Francis (2001, p.
260) pointedout that mentoring relationships that promote
collaborative in-quiry, cooperative practice and reection are
fundamental toworkplace learning for beginning teachers. Similarly,
Edwards(1998) argued that mentors should help to transform
thelearner instead of simply providing knowledge. Finally, Wang
andOdell (2002, p. 490) suggested that mentor[s] should guide
nov-ices discovery with principles rather than simply providing
arepertoire of teaching knowledge.
It can be hypothesized that constructivist mentoring representsa
particularly benecial learning environment for beginningteachers
that helps them develop professional competence, i.e.,professional
characteristics which in turn affect their work-relatedbehavior,
well-being and classroom practice (see Desimone, 2009).A
comprehensive model of professional competence proposed byBaumert
and Kunter (2006) distinguishes between professional
er Education 36 (2013) 166e177knowledge, beliefs, motivational
orientation, and self-regulation
-
each(see also Kunter et al., 2013). Following the theoretical
assumptionspointed out above (Carter & Francis, 2001;Wang &
Odell, 2002) andempirical evidence from student learning (e.g.,
Hmelo-Silver,Duncan, & Chinn, 2007; Schwartz, Lindgren, &
Lewis, 2009) weassume that mentoring following constructivist
principles oflearning may foster the development of all four
aspects: knowl-edge, beliefs, motivation and self-regulation. In
particular, weexpect that successful self-regulation of ones
resources may sup-port the well-being of teachers. Empirical
studies have alreadypointed out that the support of mentors may
foster the develop-ment of knowledge and skills (e.g., Fletcher
& Barrett, 2004;Hudson, 2013), professional beliefs (e.g.,
Haser & Star, 2009) andindividual well-being (e.g., Ballantyne
et al., 1995; Kessels et al.,2008). However, evidence that shows
what kind of mentoringmay be relevant for a successful development
of professionalcompetence is still lacking.
1.3. The present investigation
The study investigates whether constructivist-
andtransmission-oriented mentoring provides teachers at the start
oftheir career with adequate opportunities to acquire the
pre-requisites for successful classroom instruction. In this study,
wefocus on a set of characteristics that are located in the
professionalcompetence model introduced earlier: Teacher efcacy and
teacherenthusiasm represent indicators of teacher motivation. In
particular,teacher efcacy belongs to this domain because it
regulates the timeand effort teachers invest in tasks. Transmissive
and constructivistbeliefs about learning can be characterized as
teacher beliefs. Thetwo remaining aspects of teacher competence,
professionalknowledge and self-regulation, are not investigated in
this study.Moreover, we examine teachers well-being because the rst
yearsof teaching are often perceived as highly challenging
(Fantilli &McDougall, 2009; Veenman, 1984). In particular, we
chose to focuson teachers emotional exhaustion as one indicator of
burnout(Maslach, Schaufeli, & Leitner, 2001) and job
satisfaction as an in-dicator of work-related well-being (Locke,
1969). The four con-structs of teacher competence and the two
constructs of well-beinghave been shown to predict teachers
instructional quality and canbe expected to be malleable in the rst
years of the teaching career(Dubberke, Kunter,McElvany, Brunner,
& Baumert, 2008; Klusmann,Kunter, Trautwein, Ldtke, &
Baumert, 2008; Kunter et al., 2007).
1.3.1. Teacher efcacyBandura (1986, p. 391) initially dened
self-efcacy in general
terms as peoples judgments of their capabilities to organize
andexecute courses of action required to attain designated types
ofperformances. Transferred to the teaching domain, it
describesteachers individual beliefs about their capabilities to
teach difcultor unmotivated students even in the presence of
obstacles(Tschannen-Moran & Hoy, 2001). It has been shown that
teacherefcacy is associated with teacher characteristics including
beliefsabout instruction (Woolfolk, Rosoff, & Hoy, 1990), job
satisfaction(Caprara, Barbaranelli, Borgogni, & Steca, 2003),
and teacherburnout (Skaalvik & Skaalvik, 2007), and is
positively related tostudents sense of efcacy (Anderson, Greene,
& Loewen, 1988),motivation (Midgley, Feldlaufer, & Eccles,
1989), and achievement(Ashton & Webb, 1986; Guo, Connor, Yang,
Roehrig, & Morrison,2012). Despite the numerous studies, one
needs to acknowledgethat teacher efcacy is a subjective measure of
competence thatmay deviate from assessments based on tests.
Only a few studies investigated the development of teacher
ef-cacy of beginning teachers. Woolfolk Hoy and Spero (2005)
foundthat teacher efcacy increased during student teaching but
D. Richter et al. / Teaching and Tdeclined signicantly during
the rst year of teaching. Theyspeculated that mentoring programs
may provide support thatprotects and builds teacher efcacy. Another
study, conducted byTschannen-Moran and Woolfolk Hoy (2007),
investigated theimpact of different sources of teacher efcacy for
novice andexperienced teachers and found that the support of
colleaguesnegatively predicted teacher efcacy. The authors argued
thatteachers who struggle early in the career depend heavily on
thesupport of their colleagues. Our study thus further
investigateswhethermentoring fosters the development of teacher
efcacy.Wehypothesize that constructivist-oriented mentoring
positivelypredicts the development of teacher efcacy because
feedbackprovided in a constructivist mentoring setting may promote
apositive sense of mastery.
1.3.2. Teacher enthusiasmTeacher enthusiasm can be described as
the degree of enjoy-
ment, excitement and pleasure that teachers typically experience
intheir professional activities (Kunter et al., 2008). Enthusiasm
iswidely regarded as an important characteristic of teacher
quality(Brophy & Good, 1986; Dunkin & Biddle, 1974; Kunter,
Frenzel,Nagy, Baumert, & Pekrun, 2011) and it has been shown to
bepositively related to instructional quality, job satisfaction
(Kunter etal., 2008), student enjoyment (Frenzel, Goetz, Ldtke,
Pekrun, &Sutton, 2009), and students motivation (Patrick,
Hisley, &Kempler, 2000). However, research has not yet
determined whichfactors support the development of teacher
enthusiasm. The Kunteret al. (2008) described teacher enthusiasm as
a stable and trait-likecharacteristic. However, there is reason to
believe that enthusiasmfor teaching is pronounced at the beginning
of the career and de-creases thereafter. We hypothesize that mentor
teachers can helpbeginning teachers to maintain their enthusiasm by
providing op-portunities for reection and the freedom to explore
differentteaching methods. We therefore predict that
constructivist-oriented mentoring fosters teacher enthusiasm.
1.3.3. Transmissive and constructivist beliefs about
learningBeliefs can be dened as psychologically held un-
derstandings, premises or propositions about the world that
arefelt to be true (Richardson, 1996, p. 104). Pajares (1992)
pointedout that teacher beliefs lter teachers perceptions and thus
in-uence their instructional practice and classroom
behavior.Teacher beliefs are therefore regarded as an aspect of
teachercompetence (Baumert & Kunter, 2006) and teacher
quality(Richardson, 1996). In particular, we are interested in
teacherstransmissive and constructivist beliefs about learning,
which arealso reected in different mentoring approaches. Teachers
withtransmissive beliefs view learning as a process in which
studentsaccumulate knowledge that is presented by the
teacher(Dubberke et al., 2008). In contrast, teachers with
constructivistbeliefs understand learning as individual process
that dependson individuals prior knowledge and characteristics of
the envi-ronment (Staub & Stern, 2002). Research on the effects
of teacherbeliefs has demonstrated that constructivist beliefs are
positivelyrelated to students achievement gains in mathematics
(Staub &Stern, 2002), whereas transmissive beliefs are
negativelyrelated to instructional quality (Dubberke et al., 2008;
Stipek,Givvin, Salmon, & MacGyvers, 2001). It has been shown
thatteachers adopt constructivist beliefs during their university
ed-ucation, but often revert to transmissive beliefs after they
startteaching (Hoy & Woolfolk, 1990; Veenman, 1984). This
effect hasalso been called the reality shock. We hypothesize
thatconstructivist-oriented mentoring reduces the reality shock
bysupporting beginning teachers constructivist beliefs. At the
sametime, we expect transmission-oriented mentoring to foster
the
er Education 36 (2013) 166e177 169development of transmissive
beliefs.
-
gitudinal sample were on average 27.5 years old (SD 3.8 years)at
the beginning of the study and the majority was female(65.9%).1
Participation in the study was voluntary; beginningteachers who
participated in both assessments received a mon-etary reward of
V100.
2.3. Measures
2.3.1. Independent variables2.3.1.1. Constructivist- and
transmission-oriented mentoring.We have assessed constructivist-
and transmission-oriented men-toring by asking beginning teachers
to rate their interaction withtheir mentor teacher. This does not
include the content of theirinteractions. Constructivist-oriented
mentoring was measured byitems 1e4 in Table 1;
transmission-orientedmentoring by items 5e7. The assessment format
was a 6-point scale ranging from (1)strongly disagree to (6)
strongly agree. A larger set of items was
Table 1Factor loadings from the exploratory and conrmatory
factor analysis (two-factormodel).
Item Item Wording Results of the EFA Results of the CFA
Factor 1 Factor 2 Factor 1 Factor 2
My mentor. l (SE)
do differently
1 Comparisons between individuals participating only at the rst
or both mea-surement points, respectively, indicated no signicant
differences with respect togender and socio-economic status but
individuals participating at both assessmentswere approximately 2
years younger. Furthermore, we found no statistically sig-nicant
differences between both groups for the outcome variables teacher
efcacy,teacher enthusiasm, transmissive and constructivist beliefs
and emotional
acher Education 36 (2013) 166e177The data were collected in the
COACTIV-R study, which wasdesigned to assess the development of
beginning teacherscompetence during the period of practical
training (Referendariat)that is obligatory for obtaining a
permanent teaching license inGermany (Lenhard, 2004). The
Referendariat follows directly onfrom university training and lasts
between 18 and 24 months.During this time, beginning teachers
attend theoretical courses at ateacher training institution
(Studienseminare) and gain practicalteaching experience in a
regular school, which includes observingother teachers, guided
teaching (under the supervision of experi-enced teachers), and
independent teaching. In addition, all begin-ning teachers are
assigned to more senior mentor teachers whosupervise some of their
instruction and provide feedback andadvice over the Referendariat
period (Jones, 2000). In general, thementor teachers teach the same
subjects as their novices and arenot directly involved in grading
their nal exams. They are usuallyselected by the principal based on
their experience and profes-sional expertise. Mentor teachers can
therefore be regarded as apositive subset of the schools teachers,
though they do notgenerally receive formal training for their
mentorship role.
2.2. Study design and sample
The study used a pre-test/post-test study design with two
co-horts and two points of measurement. Beginning teachers in
cohort1 were assessed at the beginning and at the end of the rst
year ofpractical training. Their counterparts in cohort 2 were
assessed atthe beginning and at the end of their second year of
practicaltraining. Thus, both cohorts were assessed at an interval
of oneof instruction to be more adequate (Klusmann et al., 2008).
It istherefore important to facilitate teachers occupational
well-being,in particular emotional exhaustion and job satisfaction,
during theinduction process. Research on mentoring has shown that
begin-ning teachers value the emotional support they receive from
theirmentors (Ballantyne et al., 1995; Hall et al., 2008; Odell
& Ferraro,1992). However, only one study has found positive
relationshipsbetween mentor support and beginning teachers
well-being(Kessels et al., 2008). Therefore, more research is
needed to docu-ment the longitudinal effects of mentoring on
well-being. More-over, research has not yet determined which
approaches tomentorenovice interaction are supportive. We therefore
examinedthe extent to which constructivist- and
transmission-orientedmentoring predict the development of beginning
teachersemotional exhaustion and job satisfaction. We expect
mentoringbased on collaborative inquiry to create a learning
environment inwhich novices and mentors can reect on instructional
problemstogether, thus alleviating beginning teachers stress.
Therefore, wehypothesize that constructivist-oriented mentoring is
positivelyrelated to job satisfaction and negatively associated
with emotionalexhaustion.
2. Methods
2.1. Setting1.3.4. Emotional exhaustion and job satisfactionThe
rst years of classroom teaching are a particularly
demanding time in the teaching career (Fantilli & McDougall,
2009;Veenman, 1984). The experience of stress has negative
conse-quences for teachers health (Guglielmi & Tatrow, 1998)
and class-room practice (Klusmann et al., 2008). Students whose
teachersexperience lower levels of emotional exhaustion and higher
jobsatisfaction perceive instruction to bemore structured and the
pace
D. Richter et al. / Teaching and Te170school year. This design
enables us to investigate the developmentof teacher characteristics
and the relevance of mentoring in thisperiod of induction.
The sample was recruited from randomly selected teachertraining
institutions (Studienseminare) in four federal states inGermany
(Bavaria, Baden-Wuerttemberg, North Rhine-West-phalia, and
Schleswig-Holstein). All beginning teachers in thesample will be
teaching mathematics at secondary level. The rstcohort of beginning
teachers included 546 individuals that wereon average 27.4 years
old (SD 4.0 years) and predominantlyfemale (65.4%). The second
cohort consisted of 210 individualswith a mean age of 29.3 years
(SD 5.1 years) and 63.3% females.In total, 756 beginning teachers
of both cohorts participated atthe rst measurement point and 551 of
them (72.9%) continuedat the second measurement point. The
participants in the lon-
in lessons.
Note. Items 1 through 4 represent constructivist-oriented
mentoring; items 5through 7 indicate transmission-oriented
mentoring. EFA Exploratory FactorAnalysis. CFA Conrmative Factor
Analysis. *p < .05.1 helps me to improveindependently.
.85* (.03) .02 (.01) .88* (.03) e
2 supports me in trying outdifferent teaching methods.
.82* (.03) .08 (.05) .75* (.03) e
3 gives me the opportunityto draw my own conclusions.
.66* (.04) .29* (.05) .76* (.03) e
4 has ideas that promptself-reection.
.70* (.04) .03 (.05) .63* (.04) e
5 tells me what I need toimprove.
.15* (.06) .83* (.04) e .81* (.04)
6 has specic ideas abouthow I shouldteach the lesson
content.
.11 (.06) .70* (.04) e .73* (.04)
7 tells me what I have to .00 (
-
eachused to assess constructivist- and
transmission-orientedmentoringin a pilot study (16 and 15 items
respectively); for reasons ofparsimony, a subset of these items was
used in the present study.
Because this instrument was newly developed for the purposeof
this study, we carried out additional analyses to test its
reliabilityand construct validity (Messick, 1989). The scale
measuringconstructivist-oriented mentoring revealed an internal
consistencyof a .84; the scale tapping transmission-oriented
mentoring, areliability of a .80. Construct validity was examined
throughcross-validation, with half the data being used in an
exploratoryfactor analysis (EFA) and the other half in a conrmatory
dataanalysis (CFA). Analyses were conducted with Mplus (Muthn
&Muthn, 1998e2007), using oblique rotation (geomin) to allowfor
correlations between factors. The results of the EFA demon-strated
high factor loadings for the items designed to measure
therespective constructs and low cross-loadings for all other
items(jlj < .29). Based on this evidence, we tested a two-factor
modelusing a conrmatory approach with the second half of the data.
Theresults of the CFA demonstrated a model t of CFI .97 andRMSEA
.08. According to the recommendations of Hu and Bentler(1999), the
CFI value demonstrates a good model t and the RMSEAvalue suggests a
reasonable model t. The high factor loadings ofthe two-factor model
demonstrate that all items are suitable in-dicators of the
construct under investigation (see Table 1). Thecorrelation between
the two factors was nonsignicant at r .05(p .52). In sum, these
results suggest that our instrument assessestwo independent factors
that reect constructivist- andtransmission-oriented mentoring.
To ensure that beginning teachers ratings were an
accuraterepresentation of mentoring quality and agreed with their
mentorteachers perceptions of beginning teachers learning, we
obtainedadditional validity data from a second data source. For
this study,we conducted a supplementary study with mentor teachers
inNorth Rhine-Westphalia whose mentees participated in COACTIV-R.
All mentees in the participating state North Rhine-Westphaliawere
invited to contact their mentor teacher and provide themwith a
questionnaire about the mentoring relationship. Among all231
beginning teachers in North Rhine-Westphalia, 33 mentorteachers
completed this questionnaire. The mentor teachers whoparticipated
in this study were on average 44.1 years old (SD 9.1years) and
63.6% of themwere female. The questionnaire included a10-item scale
tapping the mentor teachers constructivist beliefsabout learning to
teach (example item: Beginning teachers learn toteach best by
experimenting with their own solutions to theproblems arising in
lessons.). The assessment format was a 6-pointscale ranging from
(1) strongly disagree to (6) strongly agree. Scalescores were
aggregated for the analysis (unweighted mean).Overall, the scale
had an internal consistency of a .70.
Results showed thatmentor teachers constructivist beliefs
werenegatively correlated with transmission-oriented mentoring
asrated by the mentees (r .40, p < .05) and positively
correlatedwith constructivist-oriented mentoring as rated by the
mentees,although the latter association was not statistically
signicant dueto the small sample size (r .32, p > .05). This
nding illustratesindicates that beginning teachers ratings
overlapped with theirmentors beliefs about learning to teach, which
suggests that thementee ratings are useful indicators of the
quality of mentoring.
2.3.1.2. Frequency of mentoring. In addition to the quality of
men-toring, we assessed its quantity in terms of the frequency of
in-teractions between mentor and beginning teacher. Frequency
wasassessed by a single item On average, how often did you talk to
yourmentor teacher in the rst year of your practical classroom
training?Responses were made on a 6-point scale ranging from (1)
less than
D. Richter et al. / Teaching and Tonce a month to (6) every day.
We dichotomized this item for thepurpose of data analysis, because
the scale is not equidistant andtherefore cannot be treated as
continuous variable. Responses thatindicated mentorementee
interactions several times a week orevery day were recoded as 1 and
all other responses that indi-cated less frequent interactions were
recoded as 0.
2.3.2. Dependent variablesTeacher efcacy was assessed with an
established teacher ef-
cacy scale (Schwarzer & Jerusalem, 1999). Beginning teachers
wereasked to rate their agreement with 10 statements on a 4-point
Likertscale ranging from (1) strongly disagree to (4) strongly
agree (e.g., Iknow that I am able to equip even the most
problematic studentswith the knowledge they need for the exam.).
Reliability wassatisfactory at both measurement points (a1 .75; a2
.77).
Enthusiasm for teaching was assessed with a scale developed
inthe COACTIV study (Kunter et al., 2008), the precursor to
theCOACTIV-R study. Prompted by the instruction How much do
youenjoy your work?, beginning teachers were asked to rate
theiragreement with 6 statements (e.g., I teach with enthusiasm.)
on a4-point Likert scale ranging from (1) strongly disagree to
(4)strongly agree. Reliability was good at both measurement
points(a1 .85; a2 .84).
Constructivist- and transmission-oriented beliefs about
teach-ing were measured with two scales developed in the
COACTIVstudy (Dubberke et al., 2008). Items tapping constructivist
beliefsabout teaching measure the conviction that student learning
re-quires cognitively activating tasks and opportunities for
students toconverse about and nd alternative solutions to tasks.
The scaleconsists of 10 Likert-type items, which were rated from
(1) stronglydisagree to (4) strongly agree (e.g., Students can nd
solutions tomany mathematics problems without guidance.).
Reliability wasgood at both points of measurement (a1 .79; a2
.82).
The items tapping transmission-oriented beliefs about
teachingassess the conviction that teachers need to provide
examples andsolutions to tasks that students can pick up and
practice. The scaleconsists of 10 Likert-type items which were
rated from (1) stronglydisagree to (4) strongly agree (e.g.,
Teachers should equip studentswith detailed procedures for solving
problems.). Reliability wasgood at both points of measurement (a1
.78; a2 .83).
Emotional exhaustion was measured by a German adaptation(Enzmann
& Kleiber, 1989) of the Maslach Burnout Inventory(Maslach,
Jackson, & Leitner, 1996). The scale comprised 5 Likert-type
items that were rated from (1) strongly disagree to (4)strongly
agree (e.g., I often feel exhausted at school.). Reliabilitywas
good at both points of measurement (a1 .77; a2 .82).
Teachers job satisfaction was assessed with a short
Germanversion (Baumert et al., 2008) of the work satisfaction scale
from theJob Diagnostic Survey (Hackman & Oldham,1975). This
scale focusedon the overall satisfaction with work rather than on
certain aspectsof the job. Beginning teachers rated 9 items on a
4-point Likert scaleranging from (1) strongly disagree to (4)
strongly agree (e.g., Giventhe choice, I would denitely become a
teacher again.). Reliabilitywas high at both points of measurement
(a1 .89; a2 .91).
2.4. Statistical analyses
Researchers seeking to predict interindividual differences
inchange must bear in mind that unreliability of the measured
vari-ables can distort the parameter estimates (Kaplan, 2000).
Wetherefore used latent variable models to determine the effect of
thedifferent mentoring approaches on the development of
beginningteachers. These models analyze interindividual differences
at thelatent level rather than at the observed level. They offer
theadvantage of distinguishing structural relationships from
error-of-
er Education 36 (2013) 166e177 171measurement components.
-
Fig. 1 presents the generic model that was estimated
separatelyfor each of the six outcome variables. It consists of a
measurementmodel that denes the latent constructs and a structural
model thatdenes the relationships among the latent variables
(Bollen, 1989).Themeasurementmodel species
theoutcomesvariablesmeasuredat both assessments (shown as v1 and
v2) and constructivist- andtransmission-oriented mentoring (cm and
tm) assessed at the sec-ond point of measurement. The factor
loadings, intercepts, and re-sidual variances were constrained to
be invariant over time, inaccordance with the assumption of
measurement invariance(Vandenberg & Lance, 2000). As
recommended by Marsh and Hau(1996; see also Jreskog, 1979),
correlated uniquenesses wereincluded for thematching items
administered at bothmeasurementpoints. The structural part of the
model denes regression re-lationships predicting the dependent
variable (v2) at the secondpoint of measurement. The score of this
construct is controlled forthe score of the same construct (v1) at
therst point ofmeasurementand additionally regressed on the
predictor variables transmission-oriented mentoring (tm),
constructivist-oriented mentoring (cm),and frequency of interaction
with the mentor.
The analyses were conducted with the Mplus 5.1 softwarepackage
(Muthn & Muthn, 1998e2007) using the Full Informa-tion Maximum
Likelihood (FIML) estimator. This estimator uses allavailable data
for each person, estimatingmissing information fromrelations among
variables in the full sample (Schafer & Graham,2002).
Hypothesis testing was conducted at signicance level ofp < .05.
The correlations of the latent and manifest variables in
thisanalysis are provided in the appendix.
3. Results
3.1. Descriptive statistics
Table 2 provides the means and standard deviations of
themanifest variables for each cohort at bothmeasurement points.
Theresults show statistically signicant change in different
variables
Table 2Descriptive statistics of the independent and dependent
manifest variables.
Outcome Cohort N Time 1 Time 2 t p
M1 SD1 M2 SD2
Teacher efcacy 1 404 3.09 .37 3.04 .36 2.59 .012 132 3.11 .32
3.17 .38 1.97 .05
Teaching enthusiasm 1 405 3.55 .44 3.53 .47 1.02 .312 125 3.55
.39 3.60 .35 1.65 .10
Transmissive beliefs 1 407 2.36 .46 2.42 .49 2.75 .012 127 2.34
.47 2.37 .53 .86 .39
Constructivist beliefs 1 406 3.40 .37 3.34 .41 3.50 .002 127
3.45 .39 3.42 .41 .93 .36
Emotional exhaustion 1 409 2.04 .61 2.20 .74 4.82 .002 127 2.31
.64 2.06 .63 4.98 .00
Job satisfaction 1 409 3.40 .50 3.30 .64 4.13 .002 127 3.35 .54
3.44 .48 2.60 .01
Transmission-orientedmentoring
1 405 e e 3.33 1.31 e e2 127 e e 3.44 1.27 e e
Constructivist-orientedmentoring
1 407 e e 4.53 1.18 e e2 127 e e 4.75 1.20 e e
Frequency of interaction 1 407 e e .56 .50 e e2 127 e e .72 .45
e e
D. Richter et al. / Teaching and Teacher Education 36 (2013)
166e177172Fig. 1. Structural equation model investigating the
impact of constructivist- and transmission-oriented mentoring.
-
and cohorts. Teacher efcacy decreased signicantly (t 2.59,p .01)
in cohort 1 (rst year of training) and increased, though
notsignicantly, in cohort 2 (second year of training). There was
sig-nicant change in beginning teachers transmissive and
construc-tivist beliefs in cohort 1, but in opposing directions:
transmissivebeliefs about learning increased over the rst year of
training(t 2.75, p < .05), whereas constructivist beliefs
decreased(t 3.50, p < .05). Emotional exhaustion and job
satisfactionshowed signicant changes in both cohorts. Cohort 1
demonstratedincreased emotional exhaustion (t 4.82, p < .05) and
decreasedjob satisfaction (t 4.13, p< .05). In cohort 2, in
contrast, emotionalexhaustion decreased (t 4.98, p < .05) and
job satisfactionincreased (t 2.60, p < .05). The effect sizes of
the statistically
The regression coefcients of the model predicting the
devel-opment of teacher efcacy show that the baseline measure
stronglypredicted teacher efcacy at the second measurement point(b1
.64, p< .05), indicating high stability of the construct
betweenthe two measurement occasions. Moreover, beginning
teacherswhose mentors exhibited a constructivist-oriented mentoring
styleshowed a signicant increase in teacher efcacy over the year(b3
.14, p < .05). No signicant changes in teacher efcacy
wereobserved for beginning teachers who experienced a
transmission-oriented mentoring style or interacted frequently with
theirmentor teacher. The model explained a total of 45% of the
varianceand exhibited reasonable model t (RMSEA .04).
Similarly,teachers enthusiasm for teaching was strongly predicted
by their
e se
Tra
.7
.0.0.0
.5
.0
.0
D. Richter et al. / Teaching and Teacher Education 36 (2013)
166e177 173signicant differences were small (Cohen, 1988), ranging
betweend .14 and d .39. In sum, the results indicate that
beginningteachers developed differently in their rst and second
year oftraining which may reect differences in the training
provided ineach year.
Table 2 also shows descriptive statistics for the two
mentoringapproaches. Ratings of transmission-oriented mentoring
weresignicantly below the theoretical mean of 3.5 in cohort 1(t
2.59, p < .05), whereas ratings of
constructivist-orientedmentoring signicantly exceeded the
theoretical mean in bothcohorts (cohort 1: t 17.73, p < .05;
cohort 2: t 11.72, p < .05).
3.2. Predicting beginning teachers development
The descriptive results showed distinct developmental
patternsfor beginning teachers in cohorts 1 and 2. Due to these
patterns, weexamined whether the structural equation models (see
Fig. 1) needto be estimated separately for both cohorts. To check
for invariance,we compared an unrestricted model (i.e., different
parameters inboth cohorts) with a restricted model that constrained
all param-eters to be equal in both cohorts. The comparison was
conductedwith a likelihood ratio (LR) test, the results of which
showed thatthe unrestricted model provided a better t to the data.
However,given that the LR test is strongly affected by sample size,
we usedthe alternative indicators DRMSEA and DCFI to compare the t
ofboth models (Cheung & Rensvold, 2002). The results of
theinvariance tests showed that the restricted model produced
onlysmall changes in the overall model t. The DRMSEA estimates
didnot exceed .005; the DCFI estimates did not exceed .02.
Therestrictedmodel thus provides an appropriate representation of
thedata. Therefore, we can collapse the data from cohorts 1 and 2
andobtain model estimates that are valid for both groups.
Table 3 shows the regression coefcients from the six
structuralequation models (represented in the columns) and their t
statis-tics. The predictors were the baseline measure of the
outcomevariable as assessed at the rst measurement point and the
threevariables indicating the quality and frequency of
mentoring.
Table 3Results of structural equation models predicting
beginning teachers outcomes at th
Predictors Teacher efcacy Teaching enthusiasm
Baseline measurea b1 .64* .63*Transmission-oriented mentoring b2
.05 .05Constructivist-oriented mentoring b3 .14* .08*Frequency of
interaction b4 .04 .00Variance explainedR2 .45 .43Model tRMSEA .04
.0490% Conf. interval .036, .044 .035, .047Note. *p < .05.a
Baseline measure of the outcome variable assessed at the rst
measurement point (baseline enthusiasm (b1 .63, p < .05) and a
constructivist-oriented mentoring style (b3 .08, p < .05). The
model explained43% of the variance in teacher enthusiasm and it
provided a good tto the data (RMSEA .04).
The models investigating the development of beginning teach-ers
beliefs showed strong stability between the rst and
secondmeasurement point (transmissive beliefs: b1 .70, p <
.05;constructivist beliefs: b1 .63, p < .05). However, whereas
atransmissive style of mentoring positively predicted the
develop-ment of transmissive beliefs (b2 .09, p< .05), a
constructivist styleof mentoring was not signicantly associated
with beginningteachers constructivist beliefs. Only frequent
interactions with thementor negatively predicted constructivist
beliefs (b4 .10,p< .05) when all other variables in the model
were controlled. Thisnding indicates that beginning teachers who
develop construc-tivist beliefs interact less frequently with their
mentor teacher. Themodel explained 53% of the variance in
transmissive beliefs and42% of the variance in constructivist
beliefs. The goodness-of-tindices indicated acceptable model t
(transmissive beliefs:RMSEA .05; constructivist beliefs: RMSEA
.05).
The models predicting beginning teachers well-being alsoshowed
high stability coefcients (emotional exhaustion:b1 .56, p < .05;
job satisfaction: b1 .65, p < .05). Whenbaseline levels were
controlled, beginning teachers who experi-enced
constructivist-oriented mentoring showed a statisticallysignicant
decline in emotional exhaustion (b3 .20, p < .05)and a
statistically signicant increase in job satisfaction (b3 .13,p <
.05). This is clear evidence that constructivist-orientedmentoring
supports beginning teachers well-being. The modelexplained 40% of
the variance in emotional exhaustion and 48%of the variance in job
satisfaction. Both models showed accept-able model t (emotional
exhaustion: RMSEA .05; job satis-faction: RMSEA .06).
In addition to investigating the main effects of quality and
fre-quency of mentoring, we were interested in testing
whetherbeginning teachers who experience a particular
mentoringapproach benet frommore interaction with their mentor. In
other
cond measurement point.
nsmissive beliefs Constructivist beliefs Emotional exhaustion
Job satisfaction
0* .63* .56* .65*9* .03 .09 .052 .08 .20* .13*4 .10* .01 .04
3 .42 .40 .48
5 .05 .05 .0641, .049 .041, .049 .044, .058 .053, .061v1).
-
achwords, we were interested in the interaction effects between
thefrequency of interactions and the two mentoring approaches.
Wetherefore estimated additional structural equation models
thatincluded all mentoring variables as main effects as well as
twovariables representing the interaction between frequency of
in-teractions and the twomentoring approaches. However, therewereno
signicant interaction terms, which suggests that beginningteachers
do not draw additional benet from more frequentmentorementee
interactions, regardless of the mentoringapproach they
experience.
4. Discussion
This study distinguished two mentoring approaches:
construc-tivist- and transmission-oriented mentoring. In contrast
to previ-ous research, we developed a theoretically driven
instrument tomeasure both approaches and evaluated its quality.
Having estab-lished the reliability and validity of our measure, we
predictedbeginning teachers development by reference to qualitative
andquantitative characteristics of the mentoring experience. We
wereparticularly interested in examining whether empirical
evidencecould be found for the theoretically predicted positive
impact ofconstructivist mentoring. To this end, we used a
pre-test/post-teststudy design to investigate whether mentoring
predicts begin-ning teachers teacher efcacy, enthusiasm for
teaching, beliefsabout learning, emotional exhaustion, and job
satisfaction whenthe respective baseline levels were
controlled.
4.1. Experience of different mentoring approaches
Our preliminary analyses showed that most beginning teachersin
our sample experienced constructivist-oriented mentoringinvolving
opportunities for reection, experimentation withdifferent teaching
methods, and autonomous decision making. Inother words, many mentor
teachers evidently provide a learningenvironment that supports
individual learning and development.This tendency toward
constructivist mentoring might be consid-ered surprising, because
the teachers who serve as mentors are notprofessionally trained
teacher educators, but regular classroomteachers selected by the
principal. Thus, mentor teachers seem tonaturally select helpful
supervision strategies despite their lack offormal training.
Further the ndings of the descriptive analyses showed that
thetwo mentoring approaches are not related. In other words,
trans-mission- and constructivist-oriented mentoring represent
twoqualitatively different approaches that cannot be described as
thetwo poles of a continuum. Both forms of mentoring make
differenttheoretical assumptions about the nature of learning.
Therefore, itwould not be warranted to exclude either of them from
empiricalinvestigations (Sfard, 1998).
4.2. Constructivist mentoring explains beginning
teachersdevelopment
The study provides evidence that beginning teachers
whoexperience constructivist mentoring show higher levels of
efcacy,teaching enthusiasm, and job satisfaction and lower levels
ofemotional exhaustion after one year of training compared
toteachers without constructivist mentoring. This conrms
previousndings that mentoring is a crucial source of support for
beginningteachers (see overview in Ingersoll & Strong, 2011).
More impor-tantly, however, it showed that different mentoring
approachesdifferentially predict beginning teachers development.
Our nd-ings supported the theoretical prediction that mentoring
based on
D. Richter et al. / Teaching and Te174collaborative inquiry and
critical reection is benecial for thedevelopment of beginning
teachers motivation and well-being(Carter & Francis, 2001;
Hudson, 2004; Wang & Odell, 2002). Incontrast, although
transmission-oriented mentoring increasedbeginning teachers
transmissive beliefs, it did not signicantlyaffect any other
outcome examined in this study. This suggests thatmentors who
supervise their mentees closely and convey theirideas of teaching
to their mentee do not successfully foster begin-ning teachers
competence and well-being. Findings also showedthat beginning
teachers who interacted frequently with theirmentors showed less
constructivist beliefs when the quality ofmentoring (i.e.,
constructivist and transmissive mentoring) wastaken into account.
It may be hypothesized that beginning teacherswho receive
constructivist mentoring may not need close guidanceand frequent
interaction. The development of the other outcomeswas not
associated by the frequency of interaction.
The effects of mentoring observed in this study are small in
size(Cohen, 1988). When interpreting these effects, we need to bear
inmind that beginning teachers are exposed to many other
learningopportunities (e.g., those provided in teacher training
institutions,independent teaching, etc.) that may also inuence
their develop-ment during their practical training. Further,
mentors have limitedtime for intensive collaboration with the
mentee because theirteaching load is generally not reduced to
accommodate their su-pervisory duties (Jones, 2000). Thus, it is
important to highlightthat mentors do facilitate beginning teachers
development despitethe limited resources available. It is also
important to rememberthat the results of this study represent the
effects of just one year. Intotal, practical training lasts between
1.5 and 2 years. Thus, cu-mulative effects of mentoring may be even
larger than the effectsshown in this study.
4.3. Limitations and future research
Several limitations of the study warrant attention. First,
thestudy assessed the mentoring of beginning teachers in
mathe-matics. All participants in our study also obtained a
teaching licensein one or two additional subjects, andmay have
receivedmentoringin these as well. However, we did not assess this
aspect in thepresent study. The combined effect of all mentors may
be largerthan that indicated by the results of this study. Second,
the men-toring variables were assessed through self-report. We
chose thisformat because we were interested in beginning teachers
per-ceptions and how these relate to their development. Data from
oursmall validation study showed that mentors and mentees
in-terpretations of the mentoring situation overlap. However,
forfurther validation, it would be interesting to collect
additional dataabout the mentoring relationship from mentor
teachers andexternal observers because they may perceive the
mentoringrelationship in a different way. This additional
information could beused to generate additional indicators of
mentoring quality. Third,the instrument was newly developed for the
purpose of this studyand was applied in a German sample of
beginning mathematicsteachers. Although reliability and validity
were established, futurestudies need to conrm that the instrument
can be used in othersubject areas (beyond mathematics) and other
cultural contexts(beyond Germany). Fourth, beginning teachers
reported on thementoring they received only once, at the end of
their rst orsecond year of practical training. Ballantyne et al.
(1995) haveindicated that mentoring may change throughout the rst
year ofpractice; therefore, multiple assessments scattered across
theschool year may be better suited to understand the dynamics
ofmentoring relationships. Our study design with two points
ofmeasurement did not allow for multiple assessments. Fifth,
ourstudy focused on teacher mentoring as one important learning
er Education 36 (2013) 166e177opportunity and did not
investigate other potential sources such as
-
to investigate change in beginning teachers over one year,
thusgoing beyond the limitations of cross-sectional analyses that
arepresent in many similar studies. The pre-test/post-test study
designwas used to examine the extent to which mentoring predicts
thedevelopment of beginning teachers motivation, beliefs, and
well-being. Our main nding was that, over and above the frequencyof
mentoring, constructivist mentoring supports beginning teach-ers
development in the rst two years of teaching. Future studies,
References
Anderson, R. N., Greene, M. L., & Loewen, P. S. (1988).
Relationships among teachersand students thinking skills, sense of
efcacy, and student achievement. AlbertaJournal of Educational
Research, 34(2), 148e165.
Ashton, P. T., & Webb, R. B. (1986). Making a difference:
Teachers sense of efcacy andstudent achievement. New York, NY:
Longman.
Ballantyne, R., Hansford, B., & Packer, J. (1995). Mentoring
beginning teachers: aqualitative analysis of process and outcomes.
Educational Review, 47(3), 297e307.
http://dx.doi.org/10.1080/0013191950470306.
(2008). Professionswissen von Lehrkrften, kognitiv aktivierender
Mathemati-
5
1.6
7. Transmissive beliefs (time 1) .23 L.20 .10 .10 L.16 L.19 18.
Transmissive beliefs (time 2) .26 L.15 L.13 L.19 L.11 L.20 .77 19.
Constructivist beliefs (time 1) L.16 .18 .19 .20 .20 .24 L.58 L.48
110. Constructivist beliefs (time 2) .08 .16 .17 .30 .19 .32 L.42
L.54 .68 111. Emotional exhaustion (time 1) .07 L.12 L.45 L.37 L.53
L.36 .17 .11 .09 L.13 112. Emotional exhaustion (time 2) .13 L.26
L.37 L.51 L.41 L.43 .02 .06 .06 .09 .67 113. Job satisfaction (time
1) .04 .20 .50 .37 .69 .49 L.11 .09 .12 .15 L.63 L.46 114. Job
satisfaction (time 2) .09 .25 .39 .55 .55 .67 .08 .08 .14 .18 L.45
L.69 .70 115. Frequency of interaction (time 2) .02 .16 .01 .03 .03
.03 .04 .04 .02 .01 .01 L.07 .03 .05 1
Note. Values printed in bold are signicant at p < .05.
eacher Education 36 (2013) 166e177 175however, should consider
study designs with multiple measure-ments over the course of one
year and if possible of multiple yearswithin the induction
period.
The study has implications for the theory on teacher
mentoringand practice in schools. Constructivist- and
transmissive-orientedmentoring can be mapped on already existing
mentoring con-cepts and can also be aligned to different paradigms
of learningtheory. The proposed constructs can therefore serve as
an inte-grative framework to evaluate previous research and set up
futurestudies. Due to the diversity of theoretical models currently
used inthis area of research (see overview in Wang & Odell,
2002), it isdifcult to compare study ndings and make informed
decisions inthe eld of teacher induction. Thus, these constructs
can be used asa framework to successively build up a coherent
knowledge base inthis content area.
In regard to the practical implications, the ndings suggest
thatmentor teachers should be carefully selected and should
receivetraining about successful supervision of mentees. Our
resultsindicate that mentors could be chosen on the basis of their
super-workshops, teaching practice or feedback from colleagues.
There-fore, future studies should take into account the complex
system oflearning opportunities and analyze their unique effects.
Sixth, wehave limited our study to examining the type of
mentorementeeinteraction. We did not, however, assess the content
of their dis-cussions and the initial skills of beginning teachers.
Therefore,future research needs to take a closer look at the topics
of thementorementee-dialogs and mentees individual
prerequisitesthat may affect the support that mentors provide.
Finally,constructivist and transmissive mentoring is closely
related to theinstructional support of mentoring rather than to the
emotionalsupport and role modeling. Therefore, this study cannot
provideinformation as to whether mentors actually provided support
withrespect to the other two goals of mentoring.
5. Conclusions and practical implications
The results of this study inform the discussion about the
effec-tiveness of teacher mentoring. The study provided the
opportunity
Variables 1 2 3 4
1. Transmission-oriented mentoring (time 2) 12.
Constructivist-oriented mentoring (time 2) .04 13. Teacher efcacy
(time 1) .08 .17 14. Teacher efcacy (time 2) .01 .24 .74 15.
Teaching enthusiasm (time 1) .04 .20 .58 .496. Teaching enthusiasm
(time 2) .03 .21 .45 .58
D. Richter et al. / Teaching and Tvision practice. Mentors who
provide their mentees withkunterricht und die Entwicklung von
mathematischer Kompetenz (COACTIV):Dokumentation der
Erhebungsinstrumente [Professional competence ofteachers,
cognitively activating instruction, and development of
students'mathematical literacy (COACTIV): scale documentation].
Materialien aus derBildungsforschung, 83. Berlin: Max Planck
Institute for Human Development.
Baumert, J., & Kunter, M. (2006). Stichwort: Professionelle
Kompetenz von Lehrk-rften [Teachers professional competence].
Zeitschrift fr Erziehungswissen-schaft, 9(4), 469e520.
http://dx.doi.org/10.1007/s11618-006-0165-2.
Bollen, K. A. (1989). Structural equations with latent
variables. New York, NY:Wiley.
Borko, H., & Mayeld, V. (1995). The roles of the cooperating
teacher and universitysupervisor in learning to teach. Teaching and
Teacher Education, 11(5),
501e518.http://dx.doi.org/10.1016/0742-051X(95)00008-8.
Bradbury, L. U., & Koballa, T. R., Jr. (2008). Borders to
cross: identifying sources oftension in mentoreintern
relationships. Teaching and Teacher Education, 24(8),2132e2145.
http://dx.doi.org/10.1016/j.tate.2008.03.002.
Brophy, J., & Good, T. L. (1986). Teacher behavior and
student achievement. InM. C. Wittrock (Ed.), Handbook of research
on teaching (3rd ed.). (pp. 328e375)Bandura, A. (1986). Social
foundations of thought and action: A social cognitive
theory.Englewood Cliffs, NJ: Prentice-Hall.
Barnett, B., Hopkins-Thompson, P., & Hoke, M. (2002).
Assessing and supporting newteachers. Lessons from the southeast.
Chapel Hill, NC: The Southeast Center forTeaching Quality.
Baumert, J., Blum, W., Brunner, M., Dubberke, T., Jordan, A.,
Klusmann, U., et al.opportunities to practice and support them with
help if needed,seem to be more helpful than mentors who closely
guide theirmentees. In addition, mentors should participate in
trainings tofurther develop professional competence as teacher
educator.
Acknowledgments
The COACTIV-R research project at the Max Planck Institute
forHuman Development was funded by the Max Planck SocietysStrategic
Innovation Fund (2008e2010). We wish to thank themembers of the
COACTIV team, Susannah Goss and Ricarda Kleinfor language editing,
and Thilo Kleickmann and Yee Lee Shing forhelpful comments on an
earlier draft of this article.
Appendix. Correlation matrix of the latent and manifestvariables
used in the structural equation models
6 7 8 9 10 11 12 13 14 15
6 1New York, NY: Macmillan.
-
achBrown, J. S., Collins, A., & Duguid, P. (1989). Situated
cognition and the culture oflearning. Educational Researcher,
18(1), 32e42.
Caprara, G. V., Barbaranelli, C., Borgogni, L., & Steca, P.
(2003). Efcacy beliefs asdeterminants of teachers job satisfaction.
Journal of Educational Psychology,95(4), 821e832.
http://dx.doi.org/10.1037/0022-0663.95.4.821.
Carter, M., & Francis, R. (2001). Mentoring and beginning
teachers workplacelearning. Asia-pacic Journal of Teacher
Education, 29(3), 249e262.
http://dx.doi.org/10.1080/13598660120091856.
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating
goodness-of-t indexes fortesting measurement invariance. Structural
Equation Modeling, 9(2),
233e255.http://dx.doi.org/10.1207/S15328007SEM0902_5.
Cochran-Smith, M., & Paris, C. L. (1995). Mentor and
mentoring: did Homer have itright. In J. Smyth (Ed.), Critical
discourses on teacher development (pp. 181e202).London, UK:
Cassell.
Cohen, J. (1988). Statistical power analysis for the behavioral
sciences. Hillsdale, NJ:Lawrence Erlbaum Associates.
Desimone, L. M. (2009). Improving impact studies of teachers
professional devel-opment: toward better conceptualizations and
measures. EducationalResearcher, 38(3), 181e199.
Dewey, J. (1938). Experience and education. New York, NY:
Collier Books.Dubberke, T., Kunter, M., McElvany, N., Brunner, M.,
& Baumert, J. (2008). Lern-
theoretische berzeugungen von Mathematiklehrkrften: Einsse auf
dieUnterrichtsgestaltung und den Lernerfolg von Schlerinnen und
Schlern[[Mathematics teachers beliefs and their impact on
instructional quality andstudent achievement]]. Zeitschrift fr
Pdagogische Psychologie, 22(3e4), 193e206.
http://dx.doi.org/10.1024/1010-0652.22.34.193.
Dunkin, M. J., & Biddle, B. J. (1974). The study of
teaching. New York, NY: Holt,Rinehart and Winston.
Edwards, A. (1998). Mentoring student teachers in primary
schools: assisting stu-dent teachers to become learners. European
Journal of Teacher Education, 21(1),47e62.
http://dx.doi.org/10.1080/0261976980210106.
Enzmann, D., & Kleiber, D. (1989). Helfer-Leiden: Stress und
Burnout in psychosozialenBerufen. Heidelberg: Asanger.
Fantilli, R. D., & McDougall, D. E. (2009). A study of
novice teachers: challenges andsupports in the rst years. Teaching
and Teacher Education, 25(6),
814e825.http://dx.doi.org/10.1016/j.tate.2009.02.021.
Feiman-Nemser, S. (1998). Teachers as teacher educators.
European Journal ofTeacher Education, 21(1), 63e74.
Feiman-Nemser, S. (2001). Helping novices learn to teach:
lessons from an exem-plary support teacher. Journal of Teacher
Education, 52(1), 17e30.
http://dx.doi.org/10.1177/0022487101052001003.
Feiman-Nemser, S., & Buchmann, M. (1987). When is student
teaching teachereducation? Teaching and Teacher Education, 3(4),
255e273. http://dx.doi.org/10.1016/0742-051X(87)90019-9.
Fletcher, S. H., & Barrett, A. (2004). Developing effective
beginning teachers throughmentor-based induction. Mentoring and
Tutoring: Partnership in Learning, 12(3),321e333.
http://dx.doi.org/10.1080/030910042000275936.
Frenzel, A. C., Goetz, T., Ldtke, O., Pekrun, R., & Sutton,
R. E. (2009). Emotionaltransmission in the classroom: exploring the
relationship between teacher andstudent enjoyment. Journal of
Educational Psychology, 101(3), 705e716.
http://dx.doi.org/10.1037/a0014695.
Glazerman, S., Doln, M., Bleeker, M., Johnson, A., Isenberg, E.,
Lugo-Gil, J., et al.(2008). Impacts of comprehensive teacher
induction: Results from the rst year of arandomized controlled
study (NCEE 2009-4034). Retrieved from
http://ies.ed.gov/ncee/pdf/20094035.pdf.
Gold, Y. (1996). Beginning teacher support: attrition,
mentoring, and induction. InJ. Sikula (Ed.), Handbook of research
on teacher education (pp. 548e594). NewYork, NY: Macmillan.
Guglielmi, R. S., & Tatrow, K. (1998). Occupational stress,
burnout, and health inteachers: a methodological and theoretical
analysis. Review of EducationalResearch, 68(1), 61e99.
Guo, Y., Connor, C. M., Yang, Y., Roehrig, A. D., &
Morrison, F. J. (2012). The effects ofteacher qualication, teacher
self-efcacy, and classroom practices on fthgraders literacy
outcomes. Elementary School Journal, 113(1), 3e24.
http://dx.doi.org/10.1086/665816.
Hackman, J., & Oldham, G. R. (1975). Development of the job
diagnostic survey.Journal of Applied Psychology, 60, 159e170.
http://dx.doi.org/10.1037/h0076546.
Hall, K. M., Draper, R. J., Smith, L. K., & Bullough, R. V.,
Jr. (2008). More than a place toteach: exploring the perceptions of
the roles and responsibilities of mentorteachers. Mentoring and
Tutoring: Partnership in Learning, 16(3),
328e345.http://dx.doi.org/10.1080/13611260802231708.
Haser, C., & Star, J. R. (2009). Change in beliefs after
rst-year of teaching: thecase of Turkish national curriculum
context. International Journal ofEducational Development, 29(3),
293e302. http://dx.doi.org/10.1016/j.ijedudev.2008.08.007.
Hawkey, K. (1997). Roles, responsibilities, and relationships in
mentoring: a liter-ature review and agenda for research. Journal of
Teacher Education, 48(5), 325e335.
http://dx.doi.org/10.1177/0022487197048005002.
Hennissen, P., Crasborn, F., Brouwer, N., Korthagen, F., &
Bergen, T. (2008). Mappingmentor teachers roles in mentoring
dialogues. Educational Research Review,3(2), 168e186.
http://dx.doi.org/10.1016/j.edurev.2008.01.001.
Hmelo-Silver, C., Duncan, R. G., & Chinn, C. A. (2007).
Scaffolding and achievement
D. Richter et al. / Teaching and Te176in problem-based and
inquiry learning: a response to Kirschner, Sweller, andClark
(2006). Educational Psychologist, 42(2), 99e107.Hobson, A. J.,
Ashby, P., Malderez, A., & Tomlinson, P. D. (2009). Mentoring
begin-ning teachers: what we know and what we dont. Teaching and
Teacher Edu-cation, 25(1), 207e216.
http://dx.doi.org/10.1016/j.tate.2008.09.001.
Hoy, W. K., & Woolfolk, A. E. (1990). Socialization of
student-teachers. AmericanEducational Research Journal, 27(2),
279e300. http://dx.doi.org/10.3102/00028312027002279.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for t
indexes in covariance structureanalysis: conventional criteria
versus new alternatives. Structural EquationModeling, 6(1), 1e55.
http://dx.doi.org/10.1080/10705519909540118.
Hudson, P. (2004). Specic mentoring: a theory and model for
developing primaryscience teaching practices. European Journal of
Teacher Education, 27(2), 139e146.
http://dx.doi.org/10.1080/0261976042000223015.
Hudson, P. (2013). Mentoring as professional development: growth
for bothmentor and mentee. Professional Development in
Education.
Huling-Austin, L. (1990). Teacher induction programs and
internships. InW. Houston (Ed.), Handbook of research on teacher
education (pp. 535e548).Reston, VA: Association of Teacher
Educators.
Ingersoll, R. M. (2001). Teacher turnover and teacher shortages:
an organizationalanalysis. American Educational Research Journal,
38(3), 499e534. http://dx.doi.org/10.3102/00028312038003499.
Ingersoll, R. M., & Kralik, J. M. (2004). The impact of
mentoring on teacher retention:What the research says. Denver, CO:
Education Commission of the States.
Ingersoll, R. M., & Strong, M. (2011). The impact of
induction and mentoring pro-grams for beginning teachers: a
critical review of the research. Review ofEducational Research,
81(2), 201e233.
Jones, M. (2000). Becoming a secondary teacher in Germany: a
trainee perspectiveon recent developments in initial teacher
training in Germany. European Journalof Teacher Education, 23(1),
65e76. http://dx.doi.org/10.1080/026197600411634.
Jreskog, K. G. (1979). Statistical estimation of structural
models in longitudinal-developmental investigations. In J. R.
Nesselroade, & P. B. Baltes (Eds.), Longi-tudinal research in
the study of behavior and development (pp. 303e351). NewYork, NY:
Academic Press.
Kaplan, D. (2000). Structural equation modeling: Foundations and
extensions. Thou-sand Oaks, CA: Sage.
Kessels, C., Beijaard, D., Veen, K. v.,& Verloop, N. (April
2008). The importance ofinduction programs for the well-being of
beginning teachers. In Paper presentedat the annual meeting of the
American Educational Research Association, NewYork, NY.
Klusmann, U., Kunter, M., Trautwein, U., Ldtke, O., &
Baumert, J. (2008). Teachersoccupational well-being and quality of
instruction: the important role of self-regulatory patterns.
Journal of Educational Psychology, 100(3), 702e715.
http://dx.doi.org/10.1037/0022-0663.100.3.702.
Klusmann, U., Kunter, M., Voss, T., & Baumert, J. (2012).
Emotional exhaustion andjob satisfaction of beginning teachers: the
role of personality, educationalexperience and professional
competence [[Beruiche Beanspruchung ange-hender Lehrkrfte: Die
Effekte von Persnlichkeit, pdagogischer Vorerfahrungund
professioneller Kompetenz]]. Zeitschrift fr Pdagogische
Psychologie, 26,275e290.
http://dx.doi.org/10.1024/1010-0652/a000078.
Kukla-Acevedo, S. (2009). Leavers, movers, and stayers: the role
of workplaceconditions in teacher mobility decisions. Journal of
Educational Research, 102(6),443e452.
http://dx.doi.org/10.3200/JOER.102.6.443-452.
Kunter, M., Baumert, J., Blum, W., Klusmann, U., Krauss, S.,
& Neubrand, M. (2013).Cognitive activation in the mathematics
classroom and professional competence ofteachers. Results from the
COACTIV project. New York, NY: Springer.
Kunter, M., Frenzel, A., Nagy, G., Baumert, J., & Pekrun, R.
(2011). Teacher enthu-siasm: dimensionality and context specicity.
Contemporary Educational Psy-chology, 36(4), 289e301.
http://dx.doi.org/10.1016/j.cedpsych.2011.07.001.
Kunter, M., Klusmann, U., Dubberke, T., Baumert, J., Blum, W.,
Brunner, M., et al.(2007). Linking aspects of teacher competence to
their instruction: results fromthe COACTIV project. In M. Prenzel
(Ed.), Studies on the educational quality ofschools. The nal report
of the DFG priority programme (pp. S. 39eS. 59).Mnster:
Waxmann.
Kunter, M., Tsai, Y. M., Klusmann, U., Brunner, M., Krauss, S.,
& Baumert, J. (2008).Students and mathematics teachers
perceptions of teacher enthusiasm andinstruction. Learning and
Instruction, 18(5), 468e482.
http://dx.doi.org/10.1016/j.learninstruc.2008.06.008.
Lave, J., & Wenger, E. (1991). Situated learning. Legitimate
peripheral participation.Cambridge, UK: Cambridge University
Press.
Lenhard, H. (2004). Zweite Phase an Studienseminaren und Schulen
[[The secondphase of teacher education in schools and teacher
training institutions]]. InS. Blmeke, P. Reinhold, G. Tulodziecki,
& J. Wildt (Eds.), Handbuch Lehrerbildung(pp. 275e290). Bad
Heilbrunn: Klinkhardt.
Lindgren, U. (2005). Experiences of beginning teachers in a
school-based mentoringprogram in Sweden. Educational Studies,
31(3), 251e263. http://dx.doi.org/10.1080/03055690500236290.
LoCasale-Crouch, J., Davis, E., Wiens, P., & Pianta, R.
(2012). The role of the mentor insupporting new teachers:
associations with self-efcacy, reection, and quality.Mentoring and
Tutoring: Partnership in Learning, 20(3), 303e323.
http://dx.doi.org/10.1080/13611267.2012.701959.
Locke, E. A. (1969). What is job satisfaction? Organizational
Behavior and HumanPerformance, 4(4), 309e336.
http://dx.doi.org/10.1016/0030-5073(69)90013-0.
Luft, J. A., & Cox, W. E. (2001). Investing in our future: a
survey of support offered tobeginningsecondaryscienceandmathematics
teachers.ScienceEducator,10(1),1e9.
er Education 36 (2013) 166e177Marable, M. A., & Raimondi, S.
L. (2007). Teachers perceptions of what wasmost (andleast)
supportive during their rst year of teaching. Mentoring and
Tutoring: