Innovating teaching and instruction in turbulent times: The dynamics of principals’ exploration and exploitation activities

Abstract

In turbulent environments, schools have to adapt to constantly changing conditions. According to ambidexterity theory, whether they are successful in this primarily depends on their leaders and how they manage the tension between the use of current knowledge (exploitation) and the search for new knowledge (exploration). Through unique top-down and bottom-up pathways, they thus influence the innovation outcome of a school. However, it is so far unclear whether these assumptions are correct. Using data from a panel of principals who are representative of Germany and were surveyed before and during the COVID-19 pandemic, we therefore investigate if and how school leaders adapted to the turbulent environment caused by the pandemic and evaluate the extent to which this had an impact on their schools’ innovations in teaching and instruction. The results demonstrate that principals’ exploration activities increased markedly during the pandemic, while their exploitation activities decreased noticeably. Further, a focus on the use and refinement of existing knowledge in comparatively predictable (pre-COVID-19) environments harmed principals’ readiness to explore new knowledge in increasingly uncertain environments. Nevertheless, exploitation had positive consequences for the innovativeness of schools, and exploration goes along with more radical innovations in teaching and instruction. Our research suggests that schools that innovatively addressed the COVID-19 pandemic had school leaders who were able to quickly shift between the two modes of exploitation and exploration. A capacity to transition seamlessly between these modes of thinking and working thus appears to be vital for the longevity of schools.

Full text

Vol.:(0123456789)
Journal of Educational Change (2023) 24:549–581
https://doi.org/10.1007/s10833-022-09458-2
1 3
Innovating teaching andinstruction inturbulent times:
The dynamics ofprincipals’ exploration andexploitation
activities
MarcusPietsch1 · PierreTulowitzki2· ColinCramer3
Accepted: 21 April 2022 / Published online: 24 May 2022
© The Author(s) 2022
Abstract
In turbulent environments, schools have to adapt to constantly changing conditions.
According to ambidexterity theory, whether they are successful in this primar-
ily depends on their leaders and how they manage the tension between the use of
current knowledge (exploitation) and the search for new knowledge (exploration).
Through unique top-down and bottom-up pathways, they thus influence the innova-
tion outcome of a school. However, it is so far unclear whether these assumptions are
correct. Using data from a panel of principals who are representative of Germany
and were surveyed before and during the COVID-19 pandemic, we therefore inves-
tigate if and how school leaders adapted to the turbulent environment caused by the
pandemic and evaluate the extent to which this had an impact on their schools’ inno-
vations in teaching and instruction. The results demonstrate that principals’ explo-
ration activities increased markedly during the pandemic, while their exploitation
activities decreased noticeably. Further, a focus on the use and refinement of exist-
ing knowledge in comparatively predictable (pre-COVID-19) environments harmed
principals’ readiness to explore new knowledge in increasingly uncertain environ-
ments. Nevertheless, exploitation had positive consequences for the innovativeness
of schools, and exploration goes along with more radical innovations in teaching
and instruction. Our research suggests that schools that innovatively addressed the
COVID-19 pandemic had school leaders who were able to quickly shift between
the two modes of exploitation and exploration. A capacity to transition seamlessly
between these modes of thinking and working thus appears to be vital for the lon-
gevity of schools.
Keywords Ambidexterity· COVID-19· Exploitation· Exploration· Innovation·
Knowledge· Principals
* Marcus Pietsch
pietsch@leuphana.de
Extended author information available on the last page of the article
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

550
Journal of Educational Change (2023) 24:549–581
1 3
Introduction
The COVID-19 pandemic has posed a massive challenge for education systems
across the globe. During the first peak period in April 2020 alone, schools were
closed nationwide in 190 countries, with more than 90% of the world’s students
affected by these closures as a consequence (UNESCO, 2020). Never before have
schools had to deal with so much uncertainty and faced such unexpected and
unique challenges on a global scale. Accordingly, Audrey Azoulay, Director-Gen-
eral of UNESCO, called the COVID-19 crisis “the most unprecedented disruption
in the history of education” (UNESCO, 2020, p. iii).
It would be hard to describe more aptly the situation that schools faced during
this pandemic, particularly as the abrupt suspension of face-to-face teaching and
learning in the classroom could have led to schools, or at least their core pro-
cesses of teaching and learning, ceasing to operate entirely (Viner etal., 2020).
This was particularly so because the move to remote learning environments was
unfamiliar to many schools, was often not an exigency for which schools had pre-
pared, and exposed wide gaps in access to technology (Kuhfeld etal., 2020). In
this respect, the pandemic and the resulting school closures constituted extreme
turbulence for schools (Beabout, 2012), an environment that had the potential to
cause “structural damage to the institution’s normal operation” (Gross, 2014, p.
260).
All professionals in school had to deal with these extremely challenging cir-
cumstances, and it was particularly incumbent on principals to navigate their
schools through these turbulent times and ensure their continued functioning
(Harris & Jones, 2020). As leaders, they had to balance reducing uncertainty for
schooling on the one hand (Weiner et al., 2021) and immediately creating and
launching new ways of teaching and learning on the other (Harris & Jones, 2020).
In order to achieve both aims, principals had to ensure that educational processes
continued to operate while applying a “messy, trial-and-error, butterflies-in-the-
stomach leadership” (Harris, 2020, p. 324; see also Munby, 2019, p. 2).
In the management and organizational literature, some scholars use the term
exploitation–exploration paradox for this kind of tension (Andriopoulos & Lewis,
2008), with exploitation referring to incremental improvements in and refinement
of school activities and their leaders and exploration relating to experimentation
and radical innovation (Bingham & Burch, 2019; Pietsch etal., 2020). Ambidex-
terity theory assumes that organizations and their leaders need to shift between
these two complementary, mutually affecting knowledge strategies on an ongoing
basis to secure the functioning and survival of the organization—and thus to act
ambidextrously—particularly in dynamic and turbulent environments (Benoliel
& Schechter, 2017; Bingham & Burch, 2019; Da’as, 2021, 2022; Pietsch etal.,
2020).
To date, however, no study has examined the relationship between exploration
and exploitation and the implementation of innovations in learning and teach-
ing, either before or during the COVID-19 pandemic. As a result, although many
studies in other research fields point to a relationship between the exploitation
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

551
1 3
Journal of Educational Change (2023) 24:549–581
and exploration activities of leaders and innovation (Gieske etal., 2020; Guisado-
González etal., 2017; Jansen etal., 2005; Rosing & Zacher, 2017), there is cur-
rently no evidence to draw corresponding conclusions in the school context.
Moreover, even in general ambidexterity research, the relationship between indi-
vidual ambidexterity and the innovation performance of organizations has hardly
been investigated so far (Pertusa-Ortega et al., 2021). Thus, in this paper, we
explore the question of whether and to what extent the exploitation and the explo-
ration activities of principals before and during the pandemic fostered innovation
in teaching and instruction during pandemic-related school closures.
For this purpose, we analyze data from the Leadership in German Schools (LineS)
study, a principal panel (N = 493) that is representative of Germany. The study aims
to research the careers of school leaders in Germany for the first time using a repre-
sentative random sample on a longitudinal basis. In order to investigate the disrup-
tion that accompanied the COVID-19 pandemic and the role of school leaders in
this context, a special survey was conducted at short notice at the beginning of the
pandemic as part of the panel. Hence, the data used were collected at two meas-
urement points within one school year; in autumn 2019, about 6months before all
schools were closed in Germany due to COVID-19, and in spring 2020, during that
nationwide school closure. Based upon this data, we seek to understand (a) if princi-
pals’ exploitation and exploration activities were affected by the turbulent environ-
ment caused by COVID-19 and (b) if the dynamic duality between exploitation and
exploration is associated with a schools’ innovativeness on the level of teaching and
instruction. One aim of our article therefore is to interrogate the dynamics of princi-
pals’ exploitative and explorative activities over time and to investigate whether and
how principals adapted their activities to the changed environment. A second aim is
to investigate possible longitudinal effects of principals’ exploitation and exploration
on innovation of teaching and instruction in the extremely turbulent environment
caused by COVID-19.
Background andconceptual grounding
Conceptualizing exploitation, exploration, andambidexterity
Exploitation and exploration are two different types of organizational and individual
adaptation to the environment (Lavie etal., 2010; March, 1991). Exploitation refers
to activities that capitalize on knowledge and competencies already available to the
organization or individual; in other words, it works within and refines the familiar
frame of reference (March, 1991). Exploration encompasses activities like experi-
menting and innovating in pursuit of new knowledge. Explorative activities are
therefore prone to uncertainty and failure but can also lead to disruptive innovations
that can change the status quo, even in organizations with comparatively limited
resources (Christensen etal., 2015). On the individual level, explorative activities
are characterized by thinking outside the current frame of reference and beyond the
currently accepted ways of doing things (Good & Michel, 2013).
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

552
Journal of Educational Change (2023) 24:549–581
1 3
Thereby, the exploration–exploitation distinction exhibits common features with
many other concepts in the field of organizational learning, but points to the tension
between different forms of learning and the corresponding dynamics (Papachroni
etal., 2015). For example, some authors (e.g., Brix, 2019; Lam, 2019; Papachroni
etal., 2015) equate exploitation with single-loop or adaptive learning and explo-
ration with double-loop or generative learning in the sense of Argyris and Schön
(1978) and Senge (1990). However, while classical concepts assume that organiza-
tions and their leaders usually have to make either/or decisions about which learn-
ing strategy to use March’s (1991) concept of exploration and exploitation views
the different learning modes as dualities that influence each other but need to be
actively managed on an ongoing basis (Gibson & Birkinshaw, 2004; Lam, 2019;
Lewis, 2000). Therefore, the focus here is on “managerial and organizational flex-
ibility” (Lavine, 2014, p. 191) and on the “continuous developing and changing of
the exploration–exploitation configuration” (Krause-Söhner, 2021, p. 32) in ever-
evolving contexts.
Accordingly, exploitation and exploration are fundamentally different logics
that create tension because they require different modes of operation and different
resource allocations and compete for scarce resources (March, 1991). On the indi-
vidual level, leaders must be able to manage tensions between these two knowledge
strategies and (repeatedly) reallocate individual resources accordingly (O’Reilly &
Tushman, 2011). Hence, dealing with these contradictory demands may lead to cog-
nitive strain (Keller & Weibler, 2015) and stress in them (Hunter etal., 2017). Both
modes can be considered dynamic capabilities (O’Reilly & Tushman, 2008) and
complementary, mutually affecting forces (Raisch & Zimmermann, 2017) that need
to be carefully balanced to achieve organizational success over time (Raisch etal.,
2009). The corresponding theoretical concept is called ambidexterity, which can be
understood as the ability of an organization or an individual within an organization
to pursue exploitation and exploration simultaneously (Mom etal., 2009; O’Reilly
& Tushman, 2004).
The dynamics ofexploitation andexploration
Ambidexterity has been reported to be vital for the longevity of an organization
(Gibson & Birkinshaw, 2004), particularly in competitive, more dynamic, and unpre-
dictable contexts where the likelihood of a disruptive change is higher (Tushman
& O’Reilly, 1996). The original understanding of ambidexterity was being able to
manage these concurrent processes effectively by balancing the competing demands
of exploration and exploitation (March, 1991). In March’s view (1991, p. 105), “the
basic problem confronting an organization is to engage in sufficient exploitation to
ensure its current viability and, at the same time, devote enough energy to explo-
ration to ensure its future viability” and thus is about continuously balancing two
ends of a continuum. However, Gupta etal. (2006) proposed a more sequential per-
spective, viewing ambidexterity as being able to shift rapidly between exploration
and exploitation. Viewed in this vein, principal ambidexterity can be understood as
dynamic dualism between these two knowledge strategies, “whereby stability may
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

553
1 3
Journal of Educational Change (2023) 24:549–581
enable change, and change may enable stability” (Papachroni & Heracleous, 2020,
p. 17) and consequently one activity can strengthen the other if both are connected
through learning (Cao etal., 2009). For instance, members of an organization might
come up with new ideas even when working within familiar knowledge frames (Cao
etal., 2009). By the same token, exploration might be able to strengthen exploitation
by generating additional complementary resources (Cao etal., 2009). For example,
an innovation stemming from exploration might not only lead to new ways of doing
things in an established area (exploitation), but also in another, unexpected and pos-
sibly unrelated one. In organizational research, the dominant view is that exploita-
tion is more favored by management due to its promise of short-term gains (Lev-
inthal & March, 1993). This risks an organization becoming “stuck” on a path that
revolves mostly around exploitation, which is known as exploitation bias:
The competition for resources is asymmetric, with exploitation innovations
harming exploration innovations but not the other way around. The result is
that performing exploitation does not merely improve an organization’s exploi-
tation routines and increases the likelihood that exploitation will be performed
again; it also reduces the resources available for exploration (Greve, 2007, p.
953).
While an exploitation bias is more common, any focus that becomes too dominant
and rigid can put an organization at risk (Keller & Weibler, 2015). For example,
leaders in an organization that seems to perform well might focus more on maintain-
ing and optimizing the current state of affairs, which becomes the dominant area of
attention, with hardly any attention paid to innovation and generating new knowl-
edge. This is referred to as the competency trap (Levinthal & March, 1993). Such an
organization may not be able to adapt to a disruptive change. The counterpart to the
competency trap is the failure trap, in which leaders focus too much on exploration,
constantly chase new ideas, thus creating (in the worst case) a cycle of failed innova-
tions without reaping any benefit (Levinthal & March, 1993). The core issue with
both these extremes is the inability to change modes dynamically, which is referred
to as path dependency. A major challenge therefore is not only achieving but also
maintaining ambidexterity, staying dynamic and mindful of both exploration and
exploitation in ever-evolving contexts. Ambidexterity as a dynamic capability allows
one “to overcome inertia and path dependencies [and] is at the core of dynamic
capabilities” (O’Reilly & Tushman, 2008, p. 187). Overcoming both exploitation
bias and path dependency is crucial for being able to adapt in the face of the unex-
pected (Andriopoulos & Lewis, 2008).
Innovating education inturbulent times
Already in his foundational work, March (1991) linked the tension between
exploration and exploitation to innovation, noting that exploitation is primarily
associated with efficiency and refinement, while exploration is primarily associ-
ated with innovation and experimentation. Consequently, both modes are associ-
ated with change, with exploitation leading to gradual, cumulative change and
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

554
Journal of Educational Change (2023) 24:549–581
1 3
exploration leading to discontinuous, radical change (Maclean etal., 2021). In
this understanding, all innovation is change, but not all change involves innova-
tion (Osborne & Brown, 2005). Accordingly, an organization’s innovation out-
come is the result of exploration and exploitation and can be measured by the
extent to which an innovation differs from existing alternatives in its degree of
newness and novelty (Damanpour & Aravind, 2012), more specifically its innova-
tion radicalness (Johannessen etal., 2001). This corresponds to the understanding
of the OECD, which defines innovation in organizations in its guidelines for col-
lecting, reporting, and using data on innovation as follows:
An innovation is a new or improved product or process (or combination
thereof) that differs significantly from the unit’s previous products or pro-
cesses and that has been made available to potential users (product) or
brought into use by the unit (process). (OECD & Eurostat, 2018, p. 60).
In ambidexterity research, leaders are seen as the key drivers in addressing the
tension between exploration and exploitation and its relationship to organizations’
innovation (Mom etal., 2019; Rosing & Zacher, 2017; Smith & Tushman, 2005;
Zimmermann etal., 2018). Ambidexterity is considered here as dynamic manage-
rial capability (Papachroni & Heracleous, 2020) that enables leaders to manage
complex organizations (Smith & Lewis, 2012; Smith etal., 2010) and promote
ambidexterity at both team and organizational levels (Jansen etal., 2016) through
interactions across organizational levels (Mom etal., 2019) ultimately leading to
positive effects in organizational performance (Eisenhardt etal., 2010), such as
an organization’s innovation outcome (de Visser & Faems, 2015).
Research in business organizations indicates that “turbulent environments
favor organizations that can promptly take advantage of emerging opportunities
and abandon expiring certainties” (Lavie etal., 2010, p. 119), that such organiza-
tions allocate more resources toward exploration activities during turbulent times
(Lant & Mezias, 1992), and that the rate and radicalness of innovations can sig-
nificantly increase as a consequence (Germain, 1996). In particular, leaders have
been shown to drive more radical changes in increasingly uncertain environments
(Koberg etal., 2003). Consequently, the ambidexterity of leaders is particularly
useful for organizational performance and innovativeness in unpredictable and
highly dynamic contexts (Good & Michel, 2013).
Compared to business organizations, however, schools have repeatedly been
characterized as rather resistant to fundamental change; Tyack and Tobin (1994)
identified a “grammar” of schooling in long-standing structures (e.g., subject-
based instruction, age-based classes, fixed lesson schedules) that influences many
aspects of schooling and effectively “absorbs” many innovative efforts. The rea-
sons for this are varied; structural characteristics of schools are themselves are
limiting, schools serve multiple constituents making changes hard to plan and
predict and they are responsible for passing down civic and cultural knowledge
and thus have a certain obligations to preserve the past (Tye, 2000). The conti-
nuity of the school system therefore can also be viewed as a way to protect the
current strengths of the system. Some scholars have started to reject the notion
of ‘continuity vs. innovation’ and argued that schools and school systems can
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

555
1 3
Journal of Educational Change (2023) 24:549–581
be considered hybrids, maintaining “stability by adopting incremental changes”
(Cuban, 2020, p. 669).
When it comes to challenging the grammar of schooling and driving innovation
on the classroom level, school leaders play a key role (Hubbard & Datnow, 2020).
As they can influence key characteristics of the school—the vision, structures and
processes, and working conditions and staff capacity (Hallinger, 2011; Leithwood
etal., 2020)—they can act as drivers of educational change (Harris etal., 2013).
It has been shown that principals’ ambidexterity does affect teaching and learning
by promoting a climate that supports teacher creativity and consequentially teach-
ers’ classroom practices (Da’as, 2021). Furthermore, the pandemic highlighted the
general relevance of school leaders in creating conditions of psychological safety
and for innovation in times of crisis (McLeod & Dulsky, 2021; Weiner etal., 2021).
In a similar vein, Beabout argues that in times of crisis, “change is also dependent
on that system having enough stability for members to safely experiment with new
ways of doing things while remaining grounded in the safety of a recognizable sys-
tem” (Beabout, 2010, p. 419).
The present study
Hypotheses
Based on the above literature review and the lack of empirical studies in the educa-
tional field in this regard, we test the following hypotheses.
Change ofresource allocation inturbulent times
H1a Exploitative activities of principals decrease over the course of the COVID-19
pandemic.
H1b Explorative activities of principals increase over the course of the COVID-19
pandemic.
Path dependency
H2a Exploitative activities of principals prior to the COVID-19 pandemic are posi-
tively associated with exploitative activities of principals during the COVID-19
pandemic.
H2b Explorative activities of principals prior to the COVID-19 pandemic are posi-
tively associated with explorative activities of principals during the COVID-19
pandemic.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

556
Journal of Educational Change (2023) 24:549–581
1 3
Competency andfailure trap
H3a Exploitative and explorative activities of principals are associated over time,
meaning that exploitative activities prior to the COVID-19 pandemic improve or
reduce exploration activities during the COVID-19 pandemic.
H3b Explorative and exploitative activities of principals are associated over time,
meaning that explorative activities prior to COVID-19 pandemic improve or reduce
exploitation activities during the COVID-19 pandemic.
Association ofexploration andinnovativeness
H4 Explorative activities of principals prior to and/or during the COVID-19 pan-
demic are positively associated with a school’s innovativeness, specifically the crea-
tion of new teaching and instruction processes, during the COVID-19 pandemic.
Association ofexploration andinnovation radicalness
H5 Explorative activities of principals prior to and/or during the COVID-19 pan-
demic are positively associated with their schools’ process innovation radicalness,
specifically the degree of novelty in teaching and instruction processes, during the
COVID-19 pandemic.
Study context
How a school system deals with the challenges posed by the COVID-19 pan-
demic depends to a large extent on structures that have evolved in national or
state school systems. In this respect, the German school system can be described
as rather conservative with regard to innovation in comparison with its Scandi-
navian neighbors (Groß Ophoff & Cramer, 2022). This applies not only to the
use of research for innovation in schools and teaching but also to the slow pace
of equipping schools with up-to-date digital infrastructure. About €5 billion in
federal funds for digitizing schools were made available already before the pan-
demic began. These funds were initially only drawn down to a small extent by
schools, which had to submit an application justifying their need for and intended
use of the funds (Drahmann etal., 2020). With the advent of the pandemic, many
schools appeared unable to cope with the increased demands on their digital
infrastructure and subsequently failed to be in direct contact with students digi-
tally or to teach synchronously using video tools. But even at the beginning of the
pandemic, the funds were drawn down only hesitantly. Obviously, the potential of
digital media for distance learning was initially underestimated or schools at first
relied on digital equipment in the private households of families. These recent
experiences exemplify the fact that school leaders in Germany have traditionally
tended to focus on their administrative tasks, i.e., ensuring schooling by using the
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

557
1 3
Journal of Educational Change (2023) 24:549–581
already available means and resources (exploitation), which are associated with a
high workload, because they are finally not suitable for achieving this goal. Radi-
cal innovations (exploration) have often been missed, such as a consistent expan-
sion of the digital infrastructure of schools, because space for innovation is rare,
which in the pandemic finally was proving to be a problem (Pietsch etal., 2020)
and could even further increase exploitative activities to compensate for earlier
omissions (exploitation bias).
In the Federal Republic of Germany, education is the responsibility of 16 federal
states, so the structures and organization of school systems and teacher education
differ (Terhart, 2019). After compulsory education at primary schools for students
from ages 6 to 10 in all states, students of different abilities are tracked into one of
several types of school, which usually differ in both duration and curriculum. While
German states traditionally have three different secondary school types in lower sec-
ondary education and an additional upper secondary school, educational reforms
have led most states to introduce at least one comprehensive secondary school. In
this still highly differentiated school system, the COVID-19 pandemic is only a cata-
lyst that makes more virulent the problems that arose from the school systems and
teacher education of past decades.
The German school system has consistently been oriented to face-to-face teach-
ing. Except for homework and exam preparation, there is hardly any experience with
distance learning and even less with using digital media. For example, the Interna-
tional Computer and Information Literacy Study (ICILS) found that students in Ger-
many spend very little time learning with digital media and that relevant equipment
and competences are lacking by international comparison (Eickelmann etal., 2019;
Fraillon etal., 2020). Depending on their parents’ social background, the students
do not always have the necessary digital infrastructure with fast internet and suit-
able digital devices, and their schools do not provide them with all the appropri-
ate devices as a matter of course. Even teachers are generally not provided comput-
ers and are left on their own when it comes to digital media. In teacher education,
too, systematic engagement with content such as digital teaching and learning has
only recently found its way into the compulsory curriculum in university courses.
Although all students use digital devices in their studies, didactic issues surrounding
the use of digital media and tools in schools and classrooms continue to play only a
marginal role in teacher education.
Deficits in the digital infrastructure of schools pose a particular challenge under
pandemic conditions. In Germany, as the number of infections increased, all schools
were completely closed for attendance in mid-March 2020, with only a few chil-
dren offered emergency care. As a result, distance learning at home had to be imple-
mented nationwide and monitored. Because education with school attendance is
compulsory for all children, home schooling (also known as home education) is
a novelty in Germany; indeed, it is legally prohibited and a matter of controversy
(Spiegler, 2009). It was not until mid-May 2020 that the schools gradually reopened
to face-to-face teaching, alternating small groups of students. The inadequate digital
infrastructure, especially for socially disadvantaged children, was an enormous chal-
lenge, and not all schools had the necessary number of digital devices to equip all
children. As a result, a number of students in home schooling were unable to follow
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

558
Journal of Educational Change (2023) 24:549–581
1 3
synchronous digital learning via video conferencing, so assignments and work mate-
rials also had to be sent by regular mail or picked up and dropped off at schools.
In view of these failures, the extent to which schools in Germany have so far
been able to build up a digital infrastructure and use digital media in their everyday
work depends on their individual commitment. School leadership is of particular
importance in this context, as it has a significant influence on the extent and imple-
mentation of digitization at a given school. There are likely to be significant differ-
ences between the degree to which individual school leaders advanced digitization
and thus the conditions for distance learning before the pandemic and how quickly
they were able to respond to the demands posed by COVID-19. In particular, school
leaders with the ability to anticipate the necessary innovations and respond quickly
to requirements are likely to have advantages in dealing with the pandemic-related
challenges. However, the qualification paths of school leaders in Germany are very
diverse and unsystematic (Tulowitzki etal., 2019), and preparation for digital chal-
lenges has thus far played only a minor role (Cramer etal., 2019).
Sample andprocedure
Our study relies on a randomized and nationally representative panel of German
principals who responded to online questionnaires during two waves within a single
school year (thus far). The underlying population for the data consists of all prin-
cipals in Germany working at schools of all types. The data from the first wave,
the primary sample, were gathered between August and November 2019 by forsa
GmbH, a leading German survey firm, using a piloted and standardized online ques-
tionnaire and comprised N = 405 principals. The data from the second wave were
gathered in the same school year, between mid-April and mid-May 2020, during the
period when all schools in Germany were closed due to the pandemic. All panel
members were recruited using a multi-stage random process within forsa’s daily
omnibus survey, in which a sample of 1000 people over the age of 14, representative
of Germany, is randomly interviewed by telephone on various topics every working
day. Thus, in a first step, within the framework of this survey, a sub-sample of school
principals was determined by means of screening and then given an individualized
link to the online survey. In a second step, this random sample of school principals,
also representative of Germany, answered the questionnaire we developed and made
available online by forsa.
To handle potential panel attrition, a refreshment sample (Deng et al., 2013;
Hirano etal., 2001; Taylor etal., 2020) of N = 88 (> 20% of the primary sample)
principals was sampled by forsa in wave two, applying the same criteria and the
same procedure as in wave one. This was also due to the fact that the second survey
was not originally planned at this time and therefore took place at short notice and
unexpectedly for the participants. N = 218 of the principals completed the question-
naires during both waves, N = 187 principals provided information during only the
first wave, and N = 88 principals, the refreshment sample, answered questions only
during the second wave. To minimize common method biases, we followed the pro-
cedural suggestions of Podsakoff etal. (2012) during both waves. Thus, for example,
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

559
1 3
Journal of Educational Change (2023) 24:549–581
we varied item wordings and scale properties across different scales and scrambled
and personalized both individual items and item blocks throughout the surveys.
In our total sample, 55.8% of the principals were female, 43.8% were male and
0.2% did not provide gender information. The mean age was 53.54years (SD: 7.72).
On average, principals worked as teachers for 15.11years (SD: 7.25) before becom-
ing principals and had been in a leadership position for 9.94years (SD: 7.31) at the
time of the survey. In addition, 91.5% of the principals worked in public schools and
8.5% worked in private schools. Respondents indicated that they have changed jobs
2.86 times (SD: 1.94) so far and that they work a total average of 48.72h per week
(SD: 9.05), of which they teach 11.27h per week (SD: 5.77) in addition to their
leadership activities.
Measures
We measured exploitation and exploration by applying items and scales developed
by Mom etal. (2009). Thus, regarding the first wave, the exploration scale deter-
mines the extent to which a principal engaged in exploration activities during the
previous year, while the exploitation scale determines the extent to which the prin-
cipal engaged in exploitation activities during the previous year (Base question: “To
what extent did you, during the last 12months, engage in work-related activities
that can be characterized as follows?”). In the context of the longitudinal study, four
items of the original six-item scale (see ‘Appendix 2’) were applied during both
waves as anchor items, two per dimension. Thus, during the first wave the principals
answered two items measuring exploitation behaviors or activities characterized by
focusing attention on refining existing knowledge and skills and implementing exist-
ing plans (e.g., “Activities which you can properly conduct by using your present
knowledge”). Here, our indicator of internal consistency, McDonald’s Omega (ω,
McDonald, 1999), was ω = .69. They then answered two items indicating explora-
tion (ω = .76), behaviors or activities in the school context associated with fewer
certainties and a higher risk for failure (e.g., “Activities requiring you to learn new
skills or knowledge”).
Regarding the second wave, the exploration scale determines the extent to which
a principal engaged in exploration activities since his or her school was closed by the
COVID-19 pandemic, while the exploitation scale determines the extent to which
a principal engaged in exploitation activities during that same period (Base ques-
tion: “To what extent did you, since your school closed due to COVID-19, engage in
work-related activities that can be characterized as follows?”). The internal consist-
ency at the second measurement point was ω = .75 for exploitation and ω = .75 for
exploration. All items were measured on a four-point Likert scale ranging from “a
very small extent” to “a very large extent” of engagement in either explorative or
exploitative activities.
Innovation was measured by adapting items and scales from the European Com-
munity Innovation Survey (CIS, Behrens etal., 2017), which is based on the afore-
mentioned definition from the OECD’s & Eurostat (2018) Oslo guidelines for col-
lecting, reporting, and using data on innovation. As Arundel etal. (2019) state, the
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

560
Journal of Educational Change (2023) 24:549–581
1 3
OECD definition encompasses a broad range of innovations, from minor incremen-
tal improvements to disruptive or transformative innovations that completely alter
or replace processes or services. The guidelines also distinguish between product,
process, marketing, and organizational innovations.
Accordingly, with regard to teaching and instruction, which we understand to be
the core processes of schooling, we provided the following description to the prin-
cipals in terms of process innovations: “Process innovations are new or noticeably
changed processes with regard to the pedagogical work of the school (e.g., instruc-
tion and/or teaching).” Next, the principals were asked if their school had introduced
such new or significantly improved processes since school closure was implemented
(item: “Did your school introduce new or significantly improved processes since the
school closed?”), with answers binary coded as 0 = no, 1 = yes. Subsequently, prin-
cipals were asked in an open-ended question to name and describe these innova-
tions (item: “What were the main innovations in this area during the school closure?
Please give a maximum of three examples”). Finally, they had to specify the innova-
tion radicalness of these innovations (item: “Are these changes incremental (improv-
ing and/or supplementing and/or adapting what already exists) or radical (introduc-
ing something completely new) for your school?”) on a ten-point Likert scale.
Because several contextual factors could influence the ambidexterity of principals
and the innovation capacity of their schools, we use the also collected as part of the
survey information below to control for possible confounding effects:
School type applies the International Standard Classification of Education (ISCED;
UNESCO Institute for Statistics, 2012), which distinguishes education systems
according to uniform criteria: ISCED 1 refers to “primary education” and covers
the 1st to 4th school years in Germany, ISCED 2 refers to “lower secondary educa-
tion” and covers the 5th to the 10th years, and ISCED 3 refers to “higher second-
ary education” and covers the 11th to 13th years. Thus, in our study we differenti-
ate between primary schools, secondary schools, special needs schools, and other
schools (mainly schools with both primary and secondary branches). We constructed
four dummy-coded variables (coded 0 and 1) and defined primary schools as the ref-
erence group. Within our sample, 51.3% are leaders of primary, 38.9% are leaders of
secondary, 6.7% are leaders of special needs, and 3.0% are leaders of other schools.
School size is measured by the total number of students enrolled in a school. This
variable was added to our analyses partly because school size may affect interper-
sonal distance and organizational structures (Bush, 2010), which may be relevant
to a principal’s choice of management and leadership practices. In addition, the size
of a school’s student body might be associated with its innovation capacity (Pres-
ton etal., 2012). Within our sample, school sizes ranged from 25 to 2000 students
enrolled, with a mean of 360.83 (SD = 299.64).
School location or rural–urban split refers to the urban or rural character of the
area in which a school is situated. We control for this variable because urban and
rural schools may differ with regard to infrastructure, especially internet connec-
tions, and other factors that might affect innovation (Bouck, 2004). To survey the
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

561
1 3
Journal of Educational Change (2023) 24:549–581
urban–rural profile, we applied an item from PISA 2012 (OECD, 2013): “Which
of the following definitions best describes the community in which your school is
located?” Within our sample, 87 schools (17.6%) were in a village, hamlet or rural
area (fewer than 3000 people), 160 (32.5%) in a small town (3000 to about 15,000
people), 158 (32.0%) in a town (15,000 to about 100,000 people), 66 (13.4%) in a
city (100,000 to about 1,000,000 people), and 21 (4.3%) in a large metropolitan city
(over 1,000,000 people).
Analytical strategy
As we are interested in the dynamic effects of principal ambidexterity on innova-
tion in teaching and instruction during the COVID-19 pandemic and potential
path dependencies of principal exploration and exploration, we scrutinized latent
cross-lagged panel models (CLPMs, Zyphur etal., 2020a) in MPlus 8.3 (Muthén &
Muthén, 2017). These models provide two types of coefficients: First, autoregres-
sive paths that provide information about inter-individual differences in one variable
over time, in this case the stability of exploitation and exploration between the two
measurement points, and second, cross-lagged paths that provide information about
the relation of two (or more) different variables over time and that make it possible
to examine whether a predictor variable accounts for a change in another longitudi-
nal observed variable. As these associations enact a temporal order, the panel coef-
ficients can be interpreted as causal influences (Little, 2013).
When CLPMs are fitted, invariance for modeled factors, here exploration and
exploitation, over time is assumed (Xu etal., 2020); thus, it is important to ensure
that the cross-lagged relationships investigated are not biased by the instability of
the factor structure of latent variables across time points (Widaman etal., 2010).
Hence, we successively tested for factorial measurement invariance (Meredith,
1993), as longitudinal measurement invariance can be evaluated at four levels, rang-
ing from weak to strong (Widaman etal., 2010): configural, metric, scalar, and strict
invariance. Configural invariance (i.e., factor structures are the same over time) is
the weakest, while strict invariance (i.e., factor loadings, thresholds, and residuals
are the same across time points) is the most restrictive. Configural invariance means
that constructs are indicated by the same items over time. Metric invariance indi-
cates that factors over time have the same meaning, that their units and intervals
are comparable, as factor loadings are equal across time points. Scalar invariance
occurs when, in addition, item intercepts are equal and thus all items indicate the
same differences in latent means over time. Strict invariance, finally, indicates that
residual variances are the same over time in addition to the equality of factor load-
ings and item intercepts.
To assess the fit of the models, the comparative fit index (CFI), root mean square
error of approximation (RMSEA), and standardized root mean square residual
(SRMR) as provided by MPLUS are all reported. Generally, acceptable fit is indi-
cated by a CFI over .900, an RMSEA below .080, and an SRMR less than .080
(Hu & Bentler, 1999; Marsh etal., 2004). Regarding the evaluation of invariance,
we investigated changes in the CFI, RMSEA, and SRMR at each stage of testing.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

562
Journal of Educational Change (2023) 24:549–581
1 3
Here, a difference greater than .010 in the CFI (ΔCFI ≥ .010), a difference greater
than .015 in the RMSEA (ΔRMSEA ≥ .015), and a difference greater than .030 in
the SRMR (ΔSRMR ≥ .030) values between less constrained and more constrained
models suggest a lack of invariance (Chen, 2007; Cheung & Rensvold, 2002).
Additionally, we assumed invariance existed between models if changes in the CFI
(ΔCFI) were less than or equal to .002 (ΔCFI ≤ .002), as Meade etal. (2008) have
shown that ΔCFI is not sensitive to sample size and is adequate for sample sizes of
400 or more.
Unit non-response or panel attrition (i.e., principals that participated in wave
one but not wave two of our study) between waves in our data was 46.2%; item
non-response for our measures during wave one was 0.70% and 3.60% during wave
two. To handle unit non-response and cross-sectional missing data (item non-
response), we followed Akande etal. (2021), Deng etal. (2013), and Hirano etal.
(2001) and thus combined refreshment with a multiple imputation approach (i.e.,
P+R approach, see Deng etal., 2013). Consequently, at each stage of analysis we
generated a completed data set that included all N = 493 cases from the panel and
refreshment sample, imputed the data 100 times, and used these data for estimating
our CLPM and all other reported coefficients and statistics (see ‘Appendix3’ for an
exemplary Mplus input).
Results
Descriptive statistics, correlations, andunivariate analyses
As shown in Table1, exploitation and exploration at both measurement points are
correlated with each other. It is notable that both modes of principal ambidexterity
are negatively correlated. Thus, a trade-off is observable and is clearly more pro-
nounced at the second measurement point, during COVID-19-related school closures
(r = − .703, p < .001), than during the first measurement point (r = − .454, p < .001),
about 6months before the COVID-19 pandemic led to school closures across Ger-
many. Frequent exploitation is thus at the expense of exploration, as school leaders
have to make an either/or decision and allocate their available resources accordingly
(Gibson & Birkinshaw, 2004). Moreover, all measures are significantly related with
one another over time, indicating a potential path dependency for both explora-
tion (r = .387, p = .001) and exploitation (r = .432, p < .001), as well as a persistent
trade-off between exploration and exploitation over time (r = − .249, p = .019 and
r = − .414, p < .001). Further, principals spent more time on explorative activities
(m = 2.94) than on exploitative activities (m = 2.64) during the school closure. Thus,
the proportion of the first measurement point were virtually reversed compared to
the first measurement point, meaning that about 6months before COVID-19, princi-
pals spent far more time executing exploitative activities (m = 3.22) than explorative
activities (m = 2.55). Both changes are statically significant (p < .001), so H1a and
H1b are accepted, as we found an increase in explorative and a decrease in exploita-
tive activities of principals over time.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

563
1 3
Journal of Educational Change (2023) 24:549–581
Table 1 Means, standard deviations and latent correlations of exploration, exploitation and innovation
Mean SD Correlations
1 2 3 4 5 6
r p r p r p r p r r
1. Exploitation t0 3.22 0.63 1
2. Exploration t0 2.54 0.66 − 0.454 0.000 1
3. Exploitation t1 2.64 0.71 0.432 0.000 − 0.249 0.019 1
4. Exploration t1 2.94 0.69 − 0.414 0.000 0.387 0.001 − 0.703 0.000 1
5. Process innovativeness 0.83 0.38 0.198 0.032 − 0.135 0.241 0.002 0.980 − 0.122 0.198 1
6. Process innovation radicalness 5.30 2.33 − 0.053 0.653 0.189 0.123 − 0.167 0.091 0.345 0.001 – 1
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

564
Journal of Educational Change (2023) 24:549–581
1 3
With regard to process innovations during the COVID-19-related school clo-
sures, 82.7% of the surveyed principals reported that such innovations had been
implemented at their school during the school closure. In total, the surveyed prin-
cipals reported 28 different types of innovations in teaching and instruction. This
broad spectrum of innovations can be clustered into five areas: (a) digitalization,
(b) supervision and support of students, (c) tasks and formats, (d) classroom-related
student–teacher and student–student interaction, and (e) stakeholder (students, par-
ents, teachers, principal) feedback. The five innovations most frequently mentioned
with regard to teaching and instruction during the school closures were (a) the use of
video conference systems (15.8%), (b) the introduction of digital learning platforms
(13.3%), (c) the production of explanatory videos (11.2%), (d) the application of
other kinds of digital learning approaches (9.1%), and (e) the introduction of weekly
schedules for learning (7.1%).
Table 1 also reveals that the innovativeness of schools during school closure
was significantly related to the exploitative activities of principals prior to closure
(r = .198, p = .032) but not to their explorative activities, whether before or during
the pandemic (r = − .135, p = .241 and r = − .122, p = .198), or to exploitative activi-
ties during school closure (r = .002, p = .830). The radicalness of innovations in
teaching and instruction (m = 5.30) is significantly related to the principals’ explora-
tive (r = .345, p = .001) and exploitative (r = − .167, p = .091) activities during school
closure but apparently not to exploitative (r = − .053, p = .653) or explorative activi-
ties (r = .189, p = .123) prior to closure.
Evaluation ofmeasurement invariance
Measurement invariance is necessary to ensure that the measurement properties
of our latent variables, exploration and exploitation, are stable over time and that
changes are not a consequence of a change in the meaning and/or measurement of
the measures. Hence, a series of successively more constrained models was con-
ducted to evaluate the extent to which model assumptions are met. Table2 shows
the goodness-of-fit indexes of these models assessing longitudinal invariance.
Model 1 tests configural invariance, or whether the items show the same pat-
tern of loadings on our constructs across measurement points and demonstrated a
good fit (χ2 = 5.36; df = 9; CFI = 1.000; SRMR = .017; RMSEA = .000). Model 2
(χ2 = 7.37; df = 9; CFI = 1.000; SRMR = .019; RMSEA = .000) tested for metric
invariance and thus whether the meaning of our latent variables were the same over
time. No relevant differences in CFI (ΔCFI = 0), RMSEA (ΔRMSEA = 0), or SRMR
Table 2 Fit indexes for invariance across time of exploration and exploitation measures
CFI ΔCFI RMSEA ΔRMSEA SRMR ΔSRMR χ2(df) Δχ2(df)
Configural invariance 1.000 – 0.000 – 0.017 – 5.36 (9) –
Scalar invariance 1.000 0 0.000 0.000 0.019 0.002 7.37 (9) 2.01 (0)
Metric invariance 1.000 0 0.000 0.000 0.035 0.016 10.55 (11) 3.18 (2)
Strict invariance 1.000 0 0.000 0.000 0.040 0.005 13.72 (15) 3.17 (4)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

565
1 3
Journal of Educational Change (2023) 24:549–581
(ΔSRMR = .002) were found. Further, ΔCFI did not exceed the .002 threshold,
providing evidence of metric invariance. Model 3 tested for scalar invariance and
thus assumed that intercepts are equivalent across time points. This model also fit-
ted the data well (χ2 = 10.55; df = 11; CFI = 1.000; SRMR = .035; RMSEA = .000).
Both ΔRMSEA and the ΔCFI were 0, with the latter thus below the .002 thresh-
old, while ΔSRMR reached .016 and was thus well below the .030 threshold,
suggesting that scalar invariance was supported. Finally, in Model 4, we tested
for strict invariance and thus whether the residual variance was equivalent across
time points. The model also fitted the data well (χ2 = 13.72; df = 15; CFI = 1.000;
SRMR = .040; RMSEA = .000), and we found no relevant differences with regard to
CFI (ΔCFI = 0), RMSEA (ΔRMSEA = 0), or SRMR (ΔSRMR = .005). Here as well,
ΔCFI did not exceed the .002 threshold. Overall, these results provide substantial
evidence for a very good level of invariance regarding our latent variables and allow
for comparisons across measurement occasions.
Cross‑lagged panel study
Next, we scrutinized the basic longitudinal CLPM (Model 1, see Fig. 1). For
this purpose, we established autoregressive and predictive cross-lagged asso-
ciations of principal exploitation and exploration over time. Further, we
Fig. 1 Latent cross-lagged panel model for principal exploration and exploitation with two time points.
t0 = time point 1; t1 = time point two; standardized regression coefficients, standard errors in parentheses;
non-significant paths grayed out
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

566
Journal of Educational Change (2023) 24:549–581
1 3
introduced latent correlations between exploitation and exploration for both meas-
urement points in our model. This model fitted the data well (χ2 = 11.83; df = 14;
CFI = 1.000; SRMR = .026; RMSEA = .000), and we found significant stand-
ardized path coefficients for all paths (βexploration t0 -> exploration t1 = .250, p = .087,
βexploitation t0 -> exploitation t1 = .389, p = .001, βexploitation t0 -> exploration t1 = − .264,
p = .048), with the exception of the path between explorative activities prior to
COVID-19-related school closures and the exploitative activities during that time
(βexploration t0 -> exploitation t1 = − .072, p = .564). The results demonstrate that principals’
ambidextrous activities depend upon their previous exploitative and explorative
activities and that principals reproduce their familiar learning patterns even in times
of crisis. The results further point to an exploitation bias, as principal exploitation
before COVID-19 increased the likelihood that exploitation was performed again
during the school closures and hindered potential increases in explorative activities
during that time. Thus, hypotheses H2a, H2b, and H3b are confirmed and H3a is
rejected.
In a next step, we introduced process innovativeness (Model 2a) and process
innovation radicalness (Model 2b) during the school closure in the previous model
to test whether principal ambidexterity was associated with process innovations in
schools during the COVID-19 pandemic. Thus, we scrutinized autoregressive and
cross-lagged associations and regressions between exploitation and exploration
at both times and a school’s innovativeness respectively innovation radicalness in
teaching and learning during the school closure in the previous model. Regard-
ing innovativeness (Model 2a, χ2 = 14.66; df = 18; CFI = 1.000; SRMR = .026;
RMSEA = .000), we only found a statistically significant association of principals’
exploitative activities prior to the COVID-19 pandemic with the innovativeness of
schools (βexploitation t0 -> innovativeness t1 = .243, p = .025). Hence, schools where princi-
pals focused mainly on efficiency and refinement before the pandemic began were
more likely to innovate teaching and instruction during the pandemic. All other
paths (βexploration t1 -> innovativeness t1 = .050, p = .780, βexploitation t1 -> innovativeness t1 = −.147,
p = .360, βexploration t0 -> innovativeness t1 = .061, p = .594) were not related to the innova-
tiveness of schools. As we observed no associations between exploration and exploi-
tation during the COVID-19 pandemic and the innovativeness of schools during that
time, we did not test for indirect (longitudinal) effects.
Regarding the innovation radicalness in teaching and instruction (Model 2b,
χ2 = 26.93; df = 18; CFI = .971; SRMR = .033; RMSEA = .032), we again discovered
one statistically significant association. With regard to the degree of novelty of the
innovations created during school closure, we found a significant association between
explorative activities during the second measurement point and innovation radical-
ness in teaching and instruction (βexploration t1 -> innovation radicalness t1 = .393, p = .053).
Here, too, all other direct paths (βexploitation t0 -> innovation radicalness t1 = .170, p = .256,
βexploration t0 -> innovation radicalness t1 = .084, p = .526, βexploitation t1 -> innovation radicalness t1 = .081,
p = .683) were not significantly related to the innovativeness of schools during the pan-
demic. As exploration activities during the pandemic were associated with innovation
radicalness during that time and we observed significant relations with both explora-
tive and explorative activities of principals before the pandemic started, we evalu-
ated possible indirect effects in this regard, following Preacher and Kelley (2011).
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

567
1 3
Journal of Educational Change (2023) 24:549–581
Although the coefficients were comparatively large, no statistically significant rela-
tionship could be demonstrated (βexploitation t0 -> exploration t1 -> innovation radicalness t1 = − .102,
p = .200, βexploration t0 -> exploration t1 –> innovation radicalness t1 = .096, p = .229). Conse-
quently, H4 was rejected and H5 confirmed.
In a final step, we controlled all model variables—exploitation at t0 and t1,
exploration at t0 and t1, and innovativeness (Model 3a) and innovation radicalness
(Model 3b)—by school size, school type, and urban–rural character to rule out other
possible causes for the observed relationships (Hamaker etal., 2015; Little, 2013).
Model 3a (see Fig. 2) also fitted the data well (χ2 = 33.70; df = 38; CFI = 1.000;
SRMR = .025; RMSEA = .000). When controlled for those contextual variables,
all standardized path coefficients show tendencies similar to the earlier model
(βexploitation t0 -> innovativeness t1 = .248, p = .028, βexploitation t1 -> innovativeness t1 = − .157,
p = .445, βexploration t0 -> innovativeness t1 = .050, p = .661, βexploration t1 -> innovativeness t1 = .061,
p = .776). Again, we did not calculate any indirect effects. Regarding our control
variables, we found that school size (rschool size -> exploration t0 = − .232, p < .001) and
rural–urban split were significantly negatively associated with principals’ explora-
tive activities (rrural urban split -> exploration t0 = − .114, p = .048) and significantly posi-
tively associated with their exploitative activities (rschool size -> exploration t0 = .121,
p = .037, rrural urban split-> exploration t0 = .141, p = .013) prior to school closure. We did
not discover statistically significant associations between school type and any model
variable or associations between the control variables and exploitation, exploration,
or innovativeness during the pandemic (p > .10).
Model 3b also fitted the data well (χ2 = 47.15; df = 38; CFI = .976; SRMR = .028;
RMSEA = .022; see Fig. 3) and demonstrated the overall stability of the model
parameters. Even when we controlled for potential contextual confounders, prin-
cipal exploration during the COVID-19 pandemic had a statistically significant
Fig. 2 Latent associations of principals’ exploration and exploitation activities with schools’ innova-
tiveness in teaching and instruction. Note t0 = time point 1; t1 = time point two; standardized regression
coefficients, standard errors in parentheses; non-significant paths grayed out. All variables controlled for
school size, school type and rural–urban split
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

568
Journal of Educational Change (2023) 24:549–581
1 3
effect on innovation radicalness in teaching and instruction during school closure
(βexploration t1 -> innovation radicalness t1 = .432, p = .044), whereas we still could not detect
any significant effects on this for exploitation (βexploitation t0 -> innovation radicalness t1 = .132,
p = .469) and exploration (βexploration t0 -> innovation radicalness t1 = .077, p = .591)
prior to the pandemic or for exploitative activities during the pandemic
(βexploitation t1 -> innovation radicalness t1 = .151, p = .392). All covariates included in the
model were not statistically significantly related to innovation radicalness (p > .100).
We also found no indirectly mediated effects in this model that would have with-
stood a significance test (βexploitation t0 -> exploration t1 -> innovation radicalness t1 = − .133,
p = .166, βexploration t0 -> exploration t1 -> innovation radicalness t1 = .110, p = .181). Thus, even
with the addition of time-invariant control variables, H4 was rejected and H5 was
confirmed.
Limitations
Despite the strengths of our study, which featured a representative panel of prin-
cipals based on a random sample, a few limitations deserve mention: First, we are
aware that it was impossible in our study to control properly for unit effects and esti-
mate robust autoregressive terms, as we only applied two measurement occasions,
although we integrated time-invariant control variables in our analyses (Zyphu etal.,
2020a, 2020b). Second, our analyses rely on principal self-reports and thus misre-
porting cannot be completely ruled out. In this context, responses of school leaders
to the survey given during the spring of 2020 could reflect a general societal state
that was marked by uncertainty and volatility. In terms of measuring innovation, we
think we have gone a good distance in applying the OECD’s Oslo guidelines for col-
lecting, reporting, and using data on innovation in an educational setting. However,
Fig. 3 Latent associations of principals’ exploration and exploitation activities with schools’ innovation
radicalness in teaching and instruction. Note t0 = time point 1; t1 = time point two; standardized regres-
sion coefficients, standard errors in parentheses; non-significant paths grayed out. All variables con-
trolled for school size, school type and rural–urban split
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

569
1 3
Journal of Educational Change (2023) 24:549–581
even if such surveys are considered state of the art by some (OECD, 2014), non-
reactive measures would be helpful in determining which innovations actually
emerge in the field. Hence, future research would do well to explore innovations
through measurements decoupled from surveys. Third, although the concepts of
exploitation, exploration, and ambidexterity are fairly a well-researched topic out-
side the educational sector, there is a lack of studies that would allow us to draw
conclusions about the extent to which the concept is generalizable and, more impor-
tantly, transferable to the school setting. Since we believe this is a concept that can
help provide new insights into how school leaders affect educational change, espe-
cially in light of the increasing dynamics and uncertainties in education, further
studies that follow our lead would be desirable.
Discussion andconclusion
The central purpose of this study was to investigate whether and how school leaders
in Germany adapt their exploitation and exploration activities to a turbulent envi-
ronment and whether and how these two complementary, mutually affecting knowl-
edge strategies were associated with school-wide innovation efforts in teaching and
instruction in German schools during the COVID-19 pandemic.
The findings broadly show that the concepts of exploitation, exploration, and
ambidexterity can be applied to the school setting. Thus, our study confirms assump-
tions and findings that have already been demonstrated in other fields of research:
Leaders dynamically adapt their knowledge strategies to the context (Germain,
1996; Koberg etal., 2003), they increasingly use explorative strategies in times of
crisis (Lavie etal., 2010), and those activities have an impact on the innovativeness
of their institutions (de Visser & Faems, 2015). In this respect, we were able for
the first time to demonstrate path dependency, exploration bias and a relationship
between a school leader’s exploration and exploitation and school-wide innovation
and innovation radicalness in teaching and instruction.
Further, we were able to show that the school leaders in our sample used a dif-
ferent approach in turbulent, uncertain environments than in secure and certain
environments. We found that these principals’ strong focus on “refinement, choice,
production, efficiency, selection, implementation and execution” (March, 1991, p.
71) in more secure environments (pre-COVID-19) seemed to inhibit creativity, flex-
ibility, risk-taking, and experimentation in uncertain times (COVID-19). For more
profound innovations to emerge that have the potential to bring radical change, our
analysis suggests that the “pursuit of new knowledge” (Levinthal & March, 1993, p.
105) is needed.
Accordingly, our research suggests that those schools that innovatively addressed
the COVID-19 pandemic were schools whose leaders were able to quickly shift
between the two modes of exploitation and exploration, or, as Tushman and
O’Reilly (1996, p. 11) put it, who proved to be jugglers of knowledge when schools
were closed. In this respect, our study makes it clear that during the pandemic, it
was necessary “to navigate a different course, to create new pathways through the
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

570
Journal of Educational Change (2023) 24:549–581
1 3
disruption” (Harris & Jones, 2020, p. 246) in other words to have the capacity to
quickly shift modes.
In terms of the (changing) role of school leaders over the course of COVID-19,
our findings are consistent with those of other studies and show that flexibility, crea-
tivity and changing priorities were central to school leaders in the early stages of the
pandemic (Beauchamp etal., 2021; Huber & Helm, 2020; Longmuir, 2021; McLeod
& Dulsky, 2021; Thornton, 2021). However, unlike many other studies, we had the
advantage of longitudinal data, with the first measurement point prior to the pan-
demic outbreak, so we were able to assess changes over time and consequently eval-
uate the concept of ambidexterity as a dynamic managerial capability (Papachroni &
Heracleous, 2020) of school leaders.
In this way, we were able to explore how school leaders dealt with the funda-
mental tension between efficiency and flexibility, and clarify the microfoundations
of innovative performance of schools in dynamic environments (Eisenhardt etal.,
2010). Results are unambiguous: while previous studies suggested that the “capa-
bilities of school principals to foster conditions that support effective teaching and
learning practices … are at the core of effective school leadership” (Lai, 2015, p.
70), our research makes clear that a certain degree of flexibility is also required to
apply those capabilities in dynamic, changing contexts.
In summary, our investigation demonstrates in many ways the practical relevance
of the exploitation-exploration distinction in relation to school leadership and inno-
vation in teaching and instruction during the COVID-19 pandemic in the German
context. It makes clear, that schools and their leaders must continuously maintain
and improve upon the status quo while always being prepared for the unexpected.
Hence, for a school to thrive in challenging circumstances, thinking outside the box
and being able to dynamically switch modes and adapt appear to be crucial skills in
a principal.
Although it is currently rather unclear how to promote individual ambidexterity
(Turner etal., 2015), the key prerequisite for successfully dealing with the demands
of exploration and exploitation seems to be a paradoxical mindset (Smith & Tush-
man, 2005) that enables school leaders to use these two knowledge strategies not
as disjunct either/or trade-offs but rather as interwoven both/and approach (“How
can you do A without letting B be?”, Smith etal., 2016). Therefore, in our opinion,
in practice a special demand must be placed on the formal qualifications of school
leaders, who must constantly be aware of the relevance of exploration despite the
strong pull toward exploitation that is often exerted by their day-to-day tasks and
administrative demands.
In terms of research, first and foremost it should be pointed out that, in the entire
field of research on ambidexterity, our study is only one of two to investigate the
presumed relationship between individual ambidexterity and an organization’s inno-
vation outcome (Pertusa-Ortega etal., 2021) and the only one so far to empirically
examine this in the context of schools. Thus, in summary, it thus appears worth-
while to explore alternative and novel ways of researching educational leadership
and school improvement and innovation in turbulent times. Since ambidexterity
research is scarce in the field of education, there are a number of questions that need
to be answered in the future. As our study suggests that a schools’ innovativeness
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

571
1 3
Journal of Educational Change (2023) 24:549–581
and innovation radicalness is significantly affected by a school leaders’ capability
to handle the paradox of exploitation and exploration, the most pressing question is
probably the one about the modes and mechanisms of action. Theoretically as well
as empirically, the following question has to be addressed: How and in what ways
does the ambidextrous behavior of school leaders contribute to changing schools?
Here, it seems particularly important to focus on the role of teachers, as both
research on leadership in schools (Leithwood etal., 2020) and research on ambi-
dexterity (Mom etal., 2019) have shown that it is the individuals in a school who
contribute to its success through unique top-down and bottom-up pathways, and fur-
thermore, that the central position of the principal in a school’s social network is an
important factor in fostering innovation in schools (Moolenaar etal., 2010). Given
that our study covers only a short, albeit disruptive and highly dynamic, period in
which, moreover, the social contacts of school staff, e.g., as a result of home offices,
were often of a different nature than in the pre-COVID-19 period, it would be
worthwhile to explore the extent to which our findings can be replicated in non-
turbulent periods, taking into account the dynamics of principal ambidexterity and
educational innovation as presented here as well as the multilevel nature of school
leadership (Boyce & Bowers, 2018; Da’as, 2021; Pietsch etal., 2019).
Appendix1
Items used tomeasure principal exploitation andexploration
For measuring principal exploitation and exploration, we use and adept items devel-
oped by Mom etal. (2009).
t0: To what extent did you, during the last 12months, engage in work-related
activities that can be characterized as follows:
t1: To what extent did you, since your school closed due to COVID-19, engage in
work-related activities that can be characterized as follows:
Exploitation
Activities which you can properly conduct by using your present knowledge.*
Activities of which it is clear to you how to conduct them.*
Activities that serve to ensure the smooth running of everyday business.1
Activities which you carry out as if it were routine.2
Exploration
Activities requiring you to learn new skills or knowledge.*
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

572
Journal of Educational Change (2023) 24:549–581
1 3
Activities of which the associated yields or costs are currently unclear.*
Activities focusing on strong renewal of products/services or processes.1
Activities requiring quite some adaptability of you.2– All items were measured on a
four-point scale (1 = to a small extent to 4 = to a large extent).
*Anchor Items, 1only used at t0, 2 only used at t1.
Appendix2
Items used tomeasure innovativeness andinnovation radicalness
For measuring innovativeness and innovation radicalness, we adapted and expanded
the survey design of the European Community Innovation Survey (CIS, Behrens
etal., 2017).
Introduction
Many schools have had to make changes and innovations in everyday school life as
a result of school closures. Corresponding innovations can affect the school’s pro-
cesses, organization or social structure, among other things:
– Process innovations comprise new or noticeably changed processes with regard
to the pedagogical work of the school (e.g., instruction and/or teaching);
– Organizational innovations include the structural development or redesign of the
school’s internal work processes or work organization (e.g., with regard to con-
ferences and staff meetings);
– Social innovations include the creation of new or improved conditions or meas-
ures for school employees (e.g., with a view to cooperation within the teaching
staff);
– Service innovations include the provision of new and/or optimized services (e.g.,
changed contact options for students and parents with regard to questions).
Measurement ofinnovativeness
Were any innovations introduced at your school during the school closure in the fol-
lowing areas?
(a) Process innovations
(b) Organizational innovations
(c) Social innovations
(d) Service innovations
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

573
1 3
Journal of Educational Change (2023) 24:549–581
– All items were measured on a binary scale (0 = no; 1 = yes).
Measurement ofconcrete innovations (respectively, for2a, 2b, 2c and2d)
What were the main innovations in this area during the school closure? Please give a
maximum of three examples.
– Open item format.
Measurement ofinnovation radicalness (respectively, for2a, 2b, 2c and2d)
Are these changes incremental (improving and/or supplementing and/or adapting
what already exists) or radical (introducing something completely new) for your
school?
– Item measured on a ten-point scale (1 = incremental to 10 = radical).
Appendix3
Exemplary Mplus input
Title: Latent Cross-Lagged Panel Model with Regression(s) on Innovation
Radicalness
Data: File is ambdext0t1.dat; ! file name
Format is f4.0, 11f2.0, f4.0; ! format statement
Variable: Names are ! Variable names
id ! principal ID
f10_1 f10_2 ! indicators for exploitation t0
f10_4 f10_5 ! indicators for exploration t0
t1f9_1 t1f9_2 ! indicators for exploitation t1
t1f9_4 t1f9_5 ! indicators for exploration t1
t1f8b_1 ! innovation radicalness
f2 f3 nf4; !control variables
Use variables are ! Names of analysis variables
f10_1 f10_2 f10_4 f10_5t1
t1f9_1 t1f9_2 f9_4 t1f9_5
t1f8b_1 f3 nf4 ws foe as;
ID Variable is id; ! Name of principal ID variable
Missing are f10_1-t1f9_5 (99) ! Missing codes
f2 (99) f3 (99) nf4 (9999) ! Missing codes
t1f8b_1 (99); ! Missing codes
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

574
Journal of Educational Change (2023) 24:549–581
1 3
!Recoding the school type variables into dummy variables
DEFINE:
gs = 0; ws = 0; foe = 0; as = 0; ! gs = primary, ws = secondary, foe = special needs,
as = other school
if (f2 eq1) then gs = 1; if (f2 eq2) then ws = 1; if (f2 eq3) then ws = 1; if (f2
eq4) then ws = 1;
if (f2 eq5) then ws = 1; if (f2 eq 6) then ws = 1; if (f2 eq 7) then foe = 1;if (f2
eq8) then bs = 1;
if (f2 eq9) then as = 1;
!Imputation model
Data Imputation:
Impute =
f10_1 f10_2 f10_4 f10_5 t1f9_1 t1f9_2 t1f9_4 t1f9_5 t1f8b_1 f3 nf4 ws(c) foe(c)
as(c);
Ndatasets = 100;
Save = ambdex1*.dat;
!Latend cross-lagged model with regression on innovation radicalness
Model:
exploit0 by f10_1 f10_2; ! Factor exploitation t0
explort0 by f10_4 f10_5; ! Factor exploration t0
exploit1 by t1f9_1 t1f9_2; ! Factor exploitation t1
explort1 by t1f9_4 t1f9_5; ! Factor exploration t1
explort0 with exploit0; ! Correlation of exploitation and exploration t0
explort1 with exploit1; ! Correlation of exploitation and exploration (residuals) at t1
exploit0 with f3 nf4 ws foe as; ! Correlation of exploitation and control variables
at t0
explort0 with f3 nf4 ws foe as; ! Correlation of exploration and control variables at t0
explort1 on explort0 (A); ! Autoregression between exploration t0 and t1
explort1 on exploit0 (B); ! Cross-lagged regression between exploitation t0 and
exploration t1
explort1 on ws foe as f3 nf4; !Regression of control variables on exploration t1
exploit1 on exploit0; ! Autoregression between exploitation t0 and t1
exploit1 on explort0; ! Cross-lagged regression between exploration t0 and
exploitation t1
exploit1 on ws foe as f3 nf4; !Regression of control variables on exploitation t1
t1f8b_1 on exploit0 explort0 exploit1 ! Regression of exploitation t0, t1 and
exploration t1 on
!innovation radicalness t1
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

575
1 3
Journal of Educational Change (2023) 24:549–581
t1f8b_1 on ws foe as f3 nf4; ! Regression of control variables on innovation
radicalness t1
t1f8b_1 on explort1(C); ! Regression of exploration t1 on innovation radicalness t1
!Estimation of standardized indirect effects
!Formula: path1*path2*SDx/SDy, see Preacher and Kelley (2011)
Model constraint:
new (plor ploi);
Plor = A*C*0.358/2.330; ! Path exploration t0 on innovation radicalness via
exploration t1
Ploi = B*C*0.505/2.330; ! Path exploitation t0 on innovation radicalness via explo-
ration t1
Output: Tech1 Tech4 Tech8 TECH9 Standardized; ! Output command
Funding Open Access funding enabled and organized by Projekt DEAL. This work was supported by
Deutsche Forschungsgemeinschaft [Grant No. 451458391 (PI 618/4-1)].
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as
you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com-
mons licence, and indicate if changes were made. The images or other third party material in this article
are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is
not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission
directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen
ses/ by/4. 0/
References
Akande, O., Madson, G., Hillygus, D. S., & Reiter, J. P. (2021). Leveraging auxiliary information on
marginal distributions in nonignorable models for item and unit nonresponse. Journal of the
Royal Statistical Society: Series A (Statistics in Society). https:// doi. org/ 10. 1111/ rssa. 12635
Andriopoulos, C., & Lewis, M. W. (2008). Exploitation-exploration tensions and organizational ambi-
dexterity: Managing paradoxes of innovation. Organization Science, 20(4), 696–717. https://
doi. org/ 10. 1287/ orsc. 1080. 0406
Argyris, C., & Schön, D. (1978). Organizational learning: A theory of action perspective.
Addison-Wesley.
Arundel, A., Bloch, C., & Ferguson, B. (2019). Advancing innovation in the public sector: Aligning
innovation measurement with policy goals. Research Policy, 48(3), 789–798. https:// doi. org/ 10.
1016/j. respol. 2018. 12. 001
Beabout, B. R. (2010). Leadership for change in the educational wild west of post-Katrina
New Orleans. Journal of Educational Change, 11, 403–424. https:// doi. org/ 10. 1007/
s10833- 010- 9136-8
Beabout, B. R. (2012). Turbulence, perturbance, and educational change. Complicity: An International
Journal of Complexity and Education. https:// doi. org/ 10. 29173/ cmplc t17984
Beauchamp, G., Hulme, M., Clarke, L., Hamilton, L., & Harvey, J. A. (2021). ‘People miss people’: A
study of school leadership and management in the four nations of the United Kingdom in the early
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

576
Journal of Educational Change (2023) 24:549–581
1 3
stage of the COVID-19 pandemic. Educational Management Administration & Leadership, 49(3),
375–392. https:// doi. org/ 10. 1177/ 17411 43220 987841
Behrens, V., Berger, M., Hud, M., Hünermund, P., Iferd, Y., Peters, B., Rammer, C., & Schubert, T.
(2017). Innovation activities of firms in Germany—Results of the German CIS 2012 and 2014:
Background report on the surveys of the Mannheim Innovation Panel Conducted in the Years 2013
to 2016. In ZEW Dokumentationen (No. 17–04; ZEW Dokumentationen). ZEW - Leibniz Centre
for European Economic Research. https:// ideas. repec. org/p/ zbw/ zewdok/ 1704. html
Benoliel, P., & Schechter, C. (2017). Promoting the school learning processes: Principals as learning
boundary spanners. International Journal of Educational Management, 31(7), 878–894. https://
doi. org/ 10. 1108/ IJEM- 02- 2016- 0023
Bingham, A. J., & Burch, P. (2019). Reimagining complexity: Exploring organizational ambidexterity as
a lens for policy research. Policy Futures in Education, 17(3), 402–420. https:// doi. org/ 10. 1177/
14782 10318 813269
Bouck, E. C. (2004). How size and setting impact education in rural schools. The Rural Educator, 25(3),
38–42. https:// doi. org/ 10. 35608/ rural ed. v25i3. 528
Boyce, J., & Bowers, A. J. (2018). Different levels of leadership for learning: Investigating differences
between teachers individually and collectively using multilevel factor analysis of the 2011–2012
Schools and Staffing Survey. International Journal of Leadership in Education, 21(2), 197–225.
https:// doi. org/ 10. 1080/ 13603 124. 2016. 11391 87
Brix, J. (2019). Ambidexterity and organizational learning: Revisiting and reconnecting the literatures.
The Learning Organization, 26(4), 337–351. https:// doi. org/ 10. 1108/ TLO- 02- 2019- 0034
Bush, T. (2010). Theories of educational leadership and management (4th ed.). Sage Publications.
Cao, Q., Gedajlovic, E., & Zhang, H. (2009). Unpacking organizational ambidexterity: Dimensions, con-
tingencies, and synergistic effects. Organization Science, 20(4), 781–796. https:// doi. org/ 10. 1287/
orsc. 1090. 0426
Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural
Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504. https:// doi. org/ 10. 1080/ 10705
51070 13018 34
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement
invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9(2), 233–255. https:// doi.
org/ 10. 1207/ S1532 8007S EM0902_5
Christensen, C. M., Raynor, M., & McDonald, R. (2015). What is disruptive innovation? Harvard Busi-
ness Review, 93, 44–53.
Cramer, C., Johannmeyer, K., & Drahmann, M. (2019). Fortbildungen von Lehrerinnen und Lehrern in
Baden-Württemberg. Universität Tübingen. https:// doi. org/ 10. 25656/ 01: 16567
Cuban, L. (2020). Reforming the grammar of schooling again and again. American Journal of Education,
126(4), 665–671. https:// doi. org/ 10. 1086/ 709959
Da’as, R. A. (2021). The missing link: Principals’ ambidexterity and teacher creativity. Leadership and
Policy in Schools. https:// doi. org/ 10. 1080/ 15700 763. 2021. 19176 21
Da’as, R. A. (2022). Principals’ attentional scope and teacher creativity: The role of principals’ ambidex-
terity and knowledge sharing. International Journal of Leadership in Education. https:// doi. org/ 10.
1080/ 13603 124. 2022. 20275 25
Damanpour, F., & Aravind, D. (2012). Organizational structure and innovation revisited: From organic
to ambidextrous structure. In M. Mumford (Ed.), Handbook of organizational creativity (pp. 483–
513). Elsevier.
De Visser, M., & Faems, D. (2015). Exploration and exploitation within firms: The impact of CEO s’
cognitive style on incremental and radical innovation performance. Creativity and Innovation Man-
agement, 24(3), 359–372. https:// doi. org/ 10. 1111/ caim. 12137
Deng, Y., Hillygus, D. S., Reiter, J. P., Si, Y., & Zheng, S. (2013). Handling attrition in longitudinal stud-
ies: The case for refreshment samples. Statistical Science, 28(2), 238–256. https:// doi. org/ 10. 1214/
13- STS414
Drahmann, M., Soğuksu, A. F., & Cramer, C. (2020). Teacher education in times of migration and digi-
talization: Comparative examples from Germany and Turkey. In K. Pushpanadham (Ed.), Teacher
education in the global era: Perspectives and practices (pp. 33–48). Springer Nature. https:// doi.
org/ 10. 1007/ 978- 981- 15- 4008-0_3
Eickelmann, B., Bos, W., Gerick, J., Goldhammer, F., Schaumburg, H., Schwippert, K., Senkbeil, M.,
Vahrenhold, J., & Waxmann Verlag. (2019). ICILS 2018 #Deutschland Computer- und informa-
tionsbezogene Kompetenzen von Schülerinnen und Schülern im zweiten internationalen Vergleich
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

577
1 3
Journal of Educational Change (2023) 24:549–581
und Kompetenzen im Bereich Computational Thinking. Waxmann. http:// nbn- resol ving. de/ urn: nbn:
de: 0111- pedocs- 181664
Eisenhardt, K. M., Furr, N. R., & Bingham, C. B. (2010). Microfoundations of performance: Balanc-
ing efficiency and flexibility in dynamic environments. Organization Science, 21(6), 1263–1273.
https:// doi. org/ 10. 1287/ orsc. 1100. 0564
Fraillon, J., Ainley, J., Schulz, W., Friedmann, T., & Duckworth, D. (2020). IEA international computer
and information literacy study 2018—Technical report. IEA.
Germain, R. (1996). The role of context and structure in radical and incremental logistics innovation
adoption. Journal of Business Research, 35(2), 117–127. https:// doi. org/ 10. 1016/ 0148- 2963(95)
00053-4
Gibson, C. B., & Birkinshaw, J. (2004). The antecedents, consequences, and mediating role of organi-
zational ambidexterity. The Academy of Management Journal, 47(2), 209–226. https:// doi. org/ 10.
2307/ 20159 573
Gieske, H., Duijn, M., & van Buuren, A. (2020). Ambidextrous practices in public service organizations:
Innovation and optimization tensions in Dutch water authorities. Public Management Review,
22(3), 341–363. https:// doi. org/ 10. 1080/ 14719 037. 2019. 15883 54
Good, D., & Michel, E. J. (2013). Individual ambidexterity: Exploring and exploiting in dynamic con-
texts. The Journal of Psychology, 147(5), 435–453. https:// doi. org/ 10. 1080/ 00223 980. 2012. 710663
Greve, H. R. (2007). Exploration and exploitation in product innovation. Industrial and Corporate
Change, 16(5), 945–975. https:// doi. org/ 10. 1093/ icc/ dtm013
Groß Ophoff, J., & Cramer, C. (2022). The engagement of teachers and school leaders with data, evi-
dence and research in Germany. In C. Brown & J. R. Malin (Eds.), The Emerald international
handbook of evidence-informed practice in education (pp. 175–195). Emerald. https:// doi. org/ 10.
1108/ 978-1- 80043- 141- 62022 1026
Gross, S. J. (2014). Using turbulence theory to guide actions. In C. M. Branson & S. J. Gross (Eds.),
Handbook of ethical educational leadership (pp. 246–262). Routledge. https:// doi. org/ 10. 4324/
97802 03747 582. ch16
Guisado-González, M., González-Blanco, J., & Coca-Pérez, J. L. (2017). Analyzing the relationship
between exploration, exploitation and organizational innovation. Journal of Knowledge Manage-
ment, 21(5), 1142–1162. https:// doi. org/ 10. 1108/ JKM- 01- 2017- 0039
Gupta, A. K., Smith, K. G., & Shalley, C. E. (2006). The interplay between exploration and exploitation.
Academy of Management Journal, 49(4), 693–706. https:// doi. org/ 10. 5465/ amj. 2006. 22083 026
Hallinger, P. (2011). Leadership for learning: Lessons from 40 years of empirical research. Journal of
Educational Administration, 49(2), 125–142. https:// doi. org/ 10. 1108/ 09578 23111 11166 99
Hamaker, E. L., Kuiper, R. M., & Grasman, R. P. P. P. (2015). A critique of the cross-lagged panel model.
Psychological Methods, 20(1), 102–116. https:// doi. org/ 10. 1037/ a0038 889
Harris, A. (2020). COVID-19—School leadership in crisis? Journal of Professional Capital and Com-
munity, 5(3/4), 321–326. https:// doi. org/ 10. 1108/ JPCC- 06- 2020- 0045
Harris, A., Day, C., Hopkins, D., Hadfield, M., Hargreaves, A., & Chapman, C. (2013). Effective leader-
ship for school improvement. Routledge.
Harris, A., & Jones, M. (2020). COVID 19—School leadership in disruptive times. School Leadership &
Management, 40(4), 243–247. https:// doi. org/ 10. 1080/ 13632 434. 2020. 18114 79
Hirano, K., Imbens, G. W., Ridder, G., & Rubin, D. B. (2001). Combining panel data sets with attrition
and refreshment samples. Econometrica, 69(6), 1645–1659. https:// doi. org/ 10. 1111/ 1468- 0262.
00260
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conven-
tional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal,
6(1), 1–55. https:// doi. org/ 10. 1080/ 10705 51990 95401 18
Hubbard, L., & Datnow, A. (2020). Design thinking, leadership, and the grammar of schooling: Implica-
tions for educational change. American Journal of Education, 126(4), 499–518. https:// doi. org/ 10.
1086/ 709510
Huber, S. G., & Helm, C. (2020). COVID-19 and schooling: Evaluation, assessment and accountability
in times of crises—Reacting quickly to explore key issues for policy, practice and research with the
school barometer. Educational Assessment, Evaluation and Accountability, 32, 237–270. https://
doi. org/ 10. 1007/ s11092- 020- 09322-y
Hunter, S. T., Cushenbery, L. D., & Jayne, B. (2017). Why dual leaders will drive innovation: Resolv-
ing the exploration and exploitation dilemma with a conservation of resources solution. Journal of
Organizational Behavior, 38(8), 1183–1195. https:// doi. org/ 10. 1002/ job. v38.8
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

578
Journal of Educational Change (2023) 24:549–581
1 3
Jansen, J. J. P., Kostopoulos, K. C., Mihalache, O. R., & Papalexandris, A. (2016). A socio-psychologi-
cal perspective on team ambidexterity: The contingency role of supportive leadership behaviours.
Journal of Management Studies, 53(6), 939–965. https:// doi. org/ 10. 1111/ joms. 12183
Jansen, J. J. P., Van den Bosch, F. A. J., & Volberda, H. W. (2005). Exploratory innovation, exploita-
tive innovation, and ambidexterity: The impact of environmental and organizational antecedents.
Schmalenbach Business Review, 57(4), 351–363. https:// doi. org/ 10. 1007/ BF033 96721
Johannessen, J., Olsen, B., & Lumpkin, G. T. (2001). Innovation as newness: What is new, how new, and
new to whom? European Journal of Innovation Management, 4(1), 20–31. https:// doi. org/ 10. 1108/
14601 06011 03655 47
Keller, T., & Weibler, J. (2015). What it takes and costs to be an ambidextrous manager: Linking lead-
ership and cognitive strain to balancing exploration and exploitation. Journal of Leadership &
Organizational Studies, 22(1), 54–71. https:// doi. org/ 10. 1177/ 15480 51814 524598
Koberg, C. S., Detienne, D. R., & Heppard, K. A. (2003). An empirical test of environmental, organi-
zational, and process factors affecting incremental and radical innovation. The Journal of High
Technology Management Research, 14(1), 21–45. https:// doi. org/ 10. 1016/ S1047- 8310(03) 00003-8
Krause-Söhner, E. (2021). Dynamics of organizational ambidexterity: Studies from a processual con-
structivist perspective. Springer. https:// doi. org/ 10. 1007/ 978-3- 658- 34127-5
Kuhfeld, M., Soland, J., Tarasawa, B., Johnson, A., Ruzek, E., & Liu, J. (2020). Projecting the potential
impact of COVID-19 school closures on academic achievement. Educational Researcher, 49(8),
549–565. https:// doi. org/ 10. 3102/ 00131 89X20 965918
Lai, E. (2015). Enacting principal leadership: Exploiting situated possibilities to build school capacity
for change. Research Papers in Education, 30(1), 70–94. https:// doi. org/ 10. 1080/ 02671 522. 2014.
880939
Lam, A. (2019). Ambidextrous learning organizations. In A. Örtenblad (Ed.), The Oxford handbook of
the learning organization (pp. 163–180). Oxford University Press.
Lant, T. K., & Mezias, S. J. (1992). An organizational learning model of convergence and reorientation.
Organization Science, 3(1), 47–71. https:// doi. org/ 10. 1287/ orsc.3. 1. 47
Lavie, D., Stettner, U., & Tushman, M. L. (2010). Exploration and exploitation within and across organi-
zations. The Academy of Management Annals, 4(1), 109–155. https:// doi. org/ 10. 1080/ 19416 52100
36912 87
Lavine, M. (2014). Paradoxical leadership and the competing values framework. The Journal of Applied
Behavioral Science, 50, 189–205. https:// doi. org/ 10. 1177/ 00218 86314 522510
Leithwood, K., Sun, J., & Schumacker, R. (2020). How school leadership influences student learning: A
test of “The Four Paths Model.” Educational Administration Quarterly, 56(4), 570–599. https://
doi. org/ 10. 1177/ 00131 61X19 878772
Levinthal, D. A., & March, J. G. (1993). The myopia of learning. Strategic Management Journal, 14,
95–112. https:// doi. org/ 10. 1002/ smj. 42501 41009
Lewis, M. W. (2000). Exploring paradox: Toward a more comprehensive guide. Academy of Management
Review, 25(4), 760–776.
Little, T. D. (2013). Longitudinal structural equation modeling. Guilford Press.
Longmuir, F. (2021). Leading in lockdown: Community, communication and compassion in response to
the COVID-19 crisis. Educational Management Administration & Leadership. https:// doi. org/ 10.
1177/ 17411 43221 10276 34
Maclean, M., Harvey, C., Golant, B. D., & Sillince, J. A. A. (2021). The role of innovation narratives in
accomplishing organizational ambidexterity. Strategic Organization, 19(4), 693–721. https:// doi.
org/ 10. 1177/ 14761 27019 897234
March, J. G. (1991). Exploration and exploitation in organizational learning. Organization Science, 2(1),
71–87.
Marsh, H. W., Hau, K.-T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis-testing
approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s
(1999) findings. Structural Equation Modeling: A Multidisciplinary Journal, 11(3), 320–341.
https:// doi. org/ 10. 1207/ s1532 8007s em1103_2
McDonald, R. P. (1999). Test theory: A unified treatment. Lawrence Erlbaum.
McLeod, S., & Dulsky, S. (2021). Resilience, reorientation, and reinvention: School leadership during the
early months of the COVID-19 pandemic. Frontiers in Education. https:// doi. org/ 10. 3389/ feduc.
2021. 637075
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

579
1 3
Journal of Educational Change (2023) 24:549–581
Meade, A. W., Johnson, E. C., & Braddy, P. W. (2008). Power and sensitivity of alternative fit indices in
tests of measurement invariance. Journal of Applied Psychology, 93(3), 568–592. https:// doi. org/
10. 1037/ 0021- 9010. 93.3. 568
Meredith, W. (1993). Measurement invariance, factor analysis and factorial invariance. Psychometrika,
58(4), 525–543. https:// doi. org/ 10. 1007/ BF022 94825
Mom, T. J. M., Chang, Y.-Y., Cholakova, M., & Jansen, J. J. P. (2019). A multilevel integrated framework
of firm HR practices, individual ambidexterity, and organizational ambidexterity. Journal of Man-
agement, 45(7), 3009–3034. https:// doi. org/ 10. 1177/ 01492 06318 776775
Mom, T. J. M., van den Bosch, F. A. J., & Volberda, H. W. (2009). Understanding variation in managers’
ambidexterity: Investigating direct and interaction effects of formal structural and personal coordi-
nation mechanisms. Organization Science, 20(4), 812–828. https:// doi. org/ 10. 1287/ orsc. 1090. 0427
Moolenaar, N. M., Daly, A. J., & Sleegers, P. J. (2010). Occupying the principal position: Examining
relationships between transformational leadership, social network position, and schools’ innova-
tive climate. Educational Administration Quarterly, 46(5), 623–670. https:// doi. org/ 10. 1177/ 2F001
3161X 10378 689
Munby, S. (2019). Imperfect leadership: A book for leaders who know they don’t know it all. Crown
House Publishing.
Muthén, L. K., & Muthén, B. O. (2017). Mplus user’s guide (8th ed.). Muthén & Muthén.
O’Reilly, C. A., & Tushman, M. L. (2004). The ambidextrous organization. Harvard Business Review,
82(4), 74–81.
O’Reilly, C. A., & Tushman, M. L. (2008). Ambidexterity as a dynamic capability: Resolving the
innovator’s dilemma. Research in Organizational Behavior, 28, 185–206. https:// doi. org/ 10.
1016/j. riob. 2008. 06. 002
O’Reilly, C. A., & Tushman, M. L. (2011). Organizational ambidexterity in action: How managers
explore and exploit. California Management Review, 53(4), 5–22. https:// doi. org/ 10. 1525/ cmr.
2011. 53.4.5
OECD. (2013). PISA 2012 assessment and analytical framework: Mathematics, reading, science,
problem solving and financial literacy. OECD. https:// doi. org/ 10. 1787/ 97892 64190 511- en
OECD. (2014). Measuring innovation in education—A new perspective. OECD Publishing. https://
doi. org/ 10. 1787/ 20769 679
OECD, & Eurostat. (2018). Oslo manual 2018: Guidelines for collecting, reporting and using data on
innovation (4th ed.). OECD Publishing. https:// doi. org/ 10. 1787/ 97892 64304 604- en
Osborne, P., & Brown, K. (2005). Managing change and innovation in public service organizations.
Routledge.
Papachroni, A., & Heracleous, L. (2020). Ambidexterity as practice: Individual ambidexterity through
paradoxical practices. Journal of Applied Behavioral Science, 56(2), 143–165. https:// doi. org/
10. 1177/ 00218 86320 913048
Papachroni, A., Heracleous, L., & Paroutis, S. (2015). Organizational ambidexterity through the lens
of paradox theory: Building a novel research agenda. The Journal of Applied Behavioral Sci-
ence, 51(1), 71–93. https:// doi. org/ 10. 1177/ 00218 86314 553101
Pertusa-Ortega, E. M., Molina-Azorín, J. F., Tarí, J. J., Pereira-Moliner, J., & López-Gamero, M. D.
(2021). The microfoundations of organizational ambidexterity: A systematic review of individ-
ual ambidexterity through a multilevel framework. Business Research Quarterly, 24(4), 355–
371. https:// doi. org/ 10. 1177/ 23409 44420 929711
Pietsch, M., Tulowitzki, P., & Cramer, C. (2020). Principals between exploitation and exploration:
Results of a nationwide study on ambidexterity of school leaders. Educational Management
Administration & Leadership. https:// doi. org/ 10. 1177/ 17411 43220 945705
Pietsch, M., Tulowitzki, P., & Koch, T. (2019). On the differential and shared effects of leadership for
learning on teachers’ organizational commitment and job satisfaction: A multilevel perspective.
Educational Administration Quarterly, 55(5), 705–741. https:// doi. org/ 10. 1177/ 00131 61X18
806346
Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in social sci-
ence research and recommendations on how to control it. Annual Review of Psychology, 63(1),
539–569. https:// doi. org/ 10. 1146/ annur ev- psych- 120710- 100452
Preacher, K. J., & Kelley, K. (2011). Effect size measures for mediation models: Quantitative strate-
gies for communicating indirect effects. Psychological Methods, 16(2), 93–115. https:// doi. org/
10. 1037/ a0022 658
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

580
Journal of Educational Change (2023) 24:549–581
1 3
Preston, C., Goldring, E., Berends, M., & Cannata, M. (2012). School innovation in district context:
Comparing traditional public schools and charter schools. Economics of Education Review,
31(2), 318–330. https:// doi. org/ 10. 1016/j. econe durev. 2011. 07. 016
Raisch, S., Birkinshaw, J., Probst, G., & Tushman, M. L. (2009). Organizational ambidexterity: Bal-
ancing exploitation and exploration for sustained performance. Organization Science, 20(4),
685–695. https:// doi. org/ 10. 1287/ orsc. 1090. 0428
Raisch, S., & Zimmermann, A. (2017). Pathways to ambidexterity: A process perspective on the
exploration–exploitation paradox. In W. K. Smith, M. W. Lewis, P. Jarzabkowski, & A. Langley
(Eds.), The Oxford handbook of organizational paradox (Vol. 1, pp. 315–332). Oxford Univer-
sity Press.
Rosing, K., & Zacher, H. (2017). Individual ambidexterity: The duality of exploration and exploita-
tion and its relationship with innovative performance. European Journal of Work and Organiza-
tional Psychology, 26(5), 694–709. https:// doi. org/ 10. 1080/ 13594 32X. 2016. 12383 58
Senge, P. (1990). The fifth discipline: The art and practice of the learning organization. Doubleday/
Currency.
Smith, W. K., Binns, A., & Tushman, M. L. (2010). Complex business models: Managing strategic
paradoxes simultaneously. Long Range Planning, 43(2/3), 448–461. https:// doi. org/ 10. 1016/j.
lrp. 2009. 12. 003
Smith, W. K., & Lewis, M. W. (2012). Leadership skills for managing paradoxes. Industrial & Organiza-
tional Psychology, 5(2), 227–231. https:// doi. org/ 10. 1111/j. 1754- 9434. 2012. 01435.x
Smith, W. K., Lewis, M. W., & Tushman, M. L. (2016). “Both/and” leadership. Harvard Business
Review, 94(5), 62–70.
Smith, W. K., & Tushman, M. L. (2005). Managing strategic contradictions: A top management model
for managing innovation streams. Organization Science, 16(5), 522–536. https:// doi. org/ 10. 1287/
orsc. 1050. 0134
Spiegler, T. (2009). Why state sanctions fail to deter home education: An analysis of home education in
Germany and its implications for home education policies. Theory and Research in Education,
7(3), 297–309. https:// doi. org/ 10. 1177/ 14778 78509 343738
Taylor, L. K., Tong, X., & Maxwell, S. E. (2020). Evaluating supplemental samples in longitudinal
research: Replacement and refreshment approaches. Multivariate Behavioral Research, 55(2), 277–
299. https:// doi. org/ 10. 1080/ 00273 171. 2019. 16286 94
Terhart, E. (2019). Teacher education in Germany. In Oxford research encyclopedia of education (pp.
1–20). Oxford University Press. https:// doi. org/ 10. 1093/ acref ore/ 97801 90264 093. 013. 377
Thornton, K. (2021). Leading through COVID-19: New Zealand secondary principals describe their real-
ity. Educational Management Administration & Leadership, 49(3), 393–409. https:// doi. org/ 10.
1177/ 17411 43220 985110
Tulowitzki, P., Hinzen, I., & Roller, M. (2019). Die Qualifizierung von Schulleiter*innen in Deutschland
ein bundesweiter Überblick. Die Deutsche Schule, 111(2), 149–170. https:// doi. org/ 10. 31244/ dds.
2019. 02. 04
Turner, N., Maylor, H., & Swart, J. (2015). Ambidexterity in projects: An intellectual capital perspective.
International Journal of Project Management, 33(1), 177–188. https:// doi. org/ 10. 1016/j. ijpro man.
2014. 05. 002
Tushman, M. L., & O’Reilly, C. A. (1996). Ambidextrous organizations: Managing evolutionary and rev-
olutionary change. California Management Review, 38(4), 8–29. https:// doi. org/ 10. 2307/ 41165 852
Tyack, D., & Tobin, W. (1994). The “Grammar” of schooling: Why has it been so hard to change? Ameri-
can Educational Research Journal, 31(3), 453–479. https:// doi. org/ 10. 3102/ 00028 31203 10034 53
Tye, B. B. (2000). Hard truths: Uncovering the deep structure of schooling. Teachers College Press.
UNESCO. (2020). Global education monitoring report 2020: Inclusion and education—All means all.
UNESCO.
UNESCO Institute for Statistics. (2012). International standard classification of education: ISCED 2011.
UNESCO Institute for Statistics. http:// www. uis. unesco. org/ Educa tion/ Docum ents/ isced- 2011- en.
pdf
Viner, R. M., Russell, S. J., Croker, H., Packer, J., Ward, J., Stansfield, C., Mytton, O., Bonell, C., &
Booy, R. (2020). School closure and management practices during coronavirus outbreaks includ-
ing COVID-19: A rapid systematic review. The Lancet Child & Adolescent Health, 4(5), 397–404.
https:// doi. org/ 10. 1016/ S2352- 4642(20) 30095-X
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

581
1 3
Journal of Educational Change (2023) 24:549–581
Weiner, J., Francois, C., Stone-Johnson, C., & Childs, J. (2021). Keep safe, keep learning: Principals’
role in creating psychological safety and organizational learning during the COVID-19 pandemic.
Frontiers in Education. https:// doi. org/ 10. 3389/ feduc. 2020. 618483
Widaman, K. F., Ferrer, E., & Conger, R. D. (2010). Factorial invariance within longitudinal structural
equation models: Measuring the same construct across time. Child Development Perspectives,
4(1), 10–18. https:// doi. org/ 10. 1111/j. 1750- 8606. 2009. 00110.x
Xu, J., Zhang, Q., & Yang, Y. (2020). Impact of violations of measurement invariance in cross-lagged
panel mediation models. Behavior Research Methods, 52(6), 2623–2645. https:// doi. org/ 10. 3758/
s13428- 020- 01426-z
Zimmermann, A., Raisch, S., & Cardinal, L. B. (2018). Managing persistent tensions on the frontline:
A configurational perspective on ambidexterity. Journal of Management Studies, 55(5), 739–769.
https:// doi. org/ 10. 1111/ joms. 12311
Zyphur, M. J., Allison, P. D., Tay, L., Voelkle, M. C., Preacher, K. J., Zhang, Z., Hamaker, E. L., Sham-
sollahi, A., Pierides, D. C., Koval, P., & Diener, E. (2020a). From data to causes I: Building a
general cross-lagged panel model (GCLM). Organizational Research Methods, 23(4), 651–687.
https:// doi. org/ 10. 1177/ 10944 28119 847278
Zyphur, M. J., Voelkle, M. C., Tay, L., Allison, P. D., Preacher, K. J., Zhang, Z., Hamaker, E. L., Sham-
sollahi, A., Pierides, D. C., Koval, P., & Diener, E. (2020b). From data to causes II: Comparing
approaches to panel data analysis. Organizational Research Methods, 23(4), 688–716. https:// doi.
org/ 10. 1177/ 10944 28119 847280
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional affiliations.
Authors and Aliations
MarcusPietsch1 · PierreTulowitzki2· ColinCramer3
1 Leuphana University ofLueneburg, Universitätsallee 1, 21335Lüneburg, Germany
2 University ofApplied Sciences andArts Northwestern Switzerland, Bahnhofstrasse 6,
5210Windisch, Switzerland
3 Eberhard Karls University ofTübingen, Wilhelmstrasse 31, 72070Tübingen, Germany
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center
GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers
and authorised users (“Users”), for small-scale personal, non-commercial use provided that all
copyright, trade and service marks and other proprietary notices are maintained. By accessing,
sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of
use (“Terms”). For these purposes, Springer Nature considers academic use (by researchers and
students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and
conditions, a relevant site licence or a personal subscription. These Terms will prevail over any
conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription (to
the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of
the Creative Commons license used will apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may
also use these personal data internally within ResearchGate and Springer Nature and as agreed share
it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not otherwise
disclose your personal data outside the ResearchGate or the Springer Nature group of companies
unless we have your permission as detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial
use, it is important to note that Users may not:
use such content for the purpose of providing other users with access on a regular or large scale
basis or as a means to circumvent access control;
use such content where to do so would be considered a criminal or statutory offence in any
jurisdiction, or gives rise to civil liability, or is otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association
unless explicitly agreed to by Springer Nature in writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a
systematic database of Springer Nature journal content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a
product or service that creates revenue, royalties, rent or income from our content or its inclusion as
part of a paid for service or for other commercial gain. Springer Nature journal content cannot be
used for inter-library loans and librarians may not upload Springer Nature journal content on a large
scale into their, or any other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not
obligated to publish any information or content on this website and may remove it or features or
functionality at our sole discretion, at any time with or without notice. Springer Nature may revoke
this licence to you at any time and remove access to any copies of the Springer Nature journal content
which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or
guarantees to Users, either express or implied with respect to the Springer nature journal content and
all parties disclaim and waive any implied warranties or warranties imposed by law, including
merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published
by Springer Nature that may be licensed from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a
regular basis or in any other manner not expressly permitted by these Terms, please contact Springer
Nature at
onlineservice@springernature.com