The role of school leader support on teacher’s engagement during intervention implementation

ABSTRACT

Teacher engagement in the implementation phase of educational interventions has been demonstrated as fundamental to the success of an intervention. However, less is known about what resources or processes can support teacher’s work engagement at this critical time period. In this study, we investigated the role of school leader support on teacher engagement during the implementation of a complex co-teaching intervention for literacy improvement. Data were collected from 150 elementary classrooms from 53 municipalities in southern Norway. Grounded in the Job-Demands Resource model, we used structural equation modeling to investigate associations between leader support, teachers’ perceived usefulness of the co-teaching intervention, and teachers’ work engagement. Contrary to our hypothesis, school leader support was not directly associated with teachers’ work engagement. However, school leader support was positively associated with teachers’ perceived usefulness of co-teaching, and the perception of usefulness was positively related to teachers’ work engagement. In summary, the perceived usefulness of co-teaching fully mediated the association between leader support and teachers’ work engagement. These findings suggest that effective school leaders should inquire and prioritize about teachers’ belief systems when launching change efforts. Ideally, we recommend that teachers be part of the conversation and decision-making in the early stages of interventions.

While school leader support is essential for teachers throughout an academic year, it is critical during periods of high work demands (Collie et al., 2018), such as the implementation of new interventions. Implementation phases require high levels of teacher work engagement and, conversely, are also risk points for teacher disengagement. By implementation, we refer to the multi-dimensional process through which an educational intervention is put into practice, including the dimensions of fidelity, quality, student responsiveness, and teacher adaptations (Lendrum & Humphrey, 2012).

Although vast resources are often invested in developing interventions when these interventions are implemented in schools, the results typically fall short – often explained by complex contextual factors (Thomas, 2021). Paradoxically, this implementation phase of interventions, while critical for success, is understudied within the life cycle of educational interventions (Lendrum & Humphrey, 2012). Co-teaching, the context of this work, is an unusual type of intervention because it is not a pedagogical intervention but instead an opportunity space where teachers can choose to adapt typical instruction due to the extra human resources. As a result, teacher engagement is particularly critical for the success of co-teaching because without intentional engagement, the status quo will persist – same instruction with two teachers.

Teachers, themselves, are always central to interventions’ success because they mediate the goals of the intervention with the constraints of the real world. First, interventions require much energy and change from teachers, so it is essential that teachers perceive the intervention as worthy or, in other words, fully buy-in. By teacher buy-in, we refer to teachers’ attitudes, receptivity, and belief that an intervention is useful, which can predict their involvement in the change effort (Datnow & Castellano, 2000). Second, teachers must have the energy and time to invest fully in an intervention. If teachers feel overwhelmed by current demands, despite positive attitudes, they may only be able to act with compliance rather than fully engage (Office for Standards in Education, 2010). Third, ongoing and intensive support systems, such as one-on-one coaching, have been shown to support teachers in implementing new programs (Reinke et al., 2014) but are not always feasible.

Yet, implementation research must not stop at the level of the classroom because teachers are nested in an ecological system created and supported partially by school leaders. Thus, we are interested in how teachers’ perceptions of support from school leaders may relate to their perceived usefulness of the intervention, and further how these factors may be associated with teacher work engagement. School leaders have the opportunity to support teachers through many avenues including professional growth, emotional well-being, resource allocation, and advocacy. Adapting Honingh and Hooge’s (2014) definition of school leader support, we define support as the extent to which teachers perceive that their school leader provides timely help, takes seriously the judgments of teachers, and appreciates and acknowledges teachers’ efforts. Perceptions of effective leadership enhance teacher satisfaction and effectiveness (Menon, 2014), which in turn benefits students and the entire school community.

Job demands-resource model (JD-R)

In addition to the empirical support connecting leader support and job engagement, it is also essential to consider this work through the lens of theory. Theory provides a foundational structure that supports in understanding and interpreting a phenomenon – in this case, teacher engagement during interventions. We draw upon organizational psychology to focus on teachers and, indirectly, school leaders through the JD-R model. The JD-R model works to explain how the environment of an organization affects employees’ performance and well-being (Tummers & Bakker, 2021). JD-R has been most commonly applied to people who work in the human services field, including characterizing teacher experiences (e.g. Dicke et al., 2018). As an example of JD-R applied to teaching, Bakker et al. (2007) demonstrated that job resources are linked to work engagement behaviors, such as enhanced dedication and going above and beyond in one’s role. This framing aligns with the perspective of Leithwood et al. (1999), who proposed that school leaders can foster a commitment to the organization’s shared goals by offering teachers opportunities for skill development and encouraging collaborative decision-making, both of which are arguably job resource provisions.

Yet most pertinent to our work is research studying educators working at periods of high demands and research during COVID-19 allows us to investigate such situations (through the lens of JD-R) and consider how resources impact well-being and job engagement. For example, during the pandemic, teacher personal resources were found to reduce the effects of increased job demands and emotional exhaustion or burnout among Italian teachers thus resulting in more engaged work behaviors (Manuti et al., 2022). Research in both the UK (Kim et al., 2022) and Germany (Stang-Rabrig et al., 2022) during COVID-19, further confirmed relations between external support from colleagues and principals and measures of teachers’ occupational well-being. This pattern extended across levels and when studying university staff during the COVID-19 pandemic (Guidetti et al., 2022), Guidetti et al, documented that supervisor support was protective against burnout for employees working in hybrid (online and in the office) arrangements. As Stang-Rabrig et al. (2022) concluded, ‘school leaders play an important role in teachers’ well-being, especially during extraordinary times’

To recap, the JD-R model conceptualizes that one’s work experiences can be categorized into two main categories: demands and resources (Demerouti et al., 2001). Both categories (demands and resources) contribute to employee engagement in work through specific pathways – demands are associated with exhaustion, whereas resources are associated with work engagement. Moreover, job resources may buffer the negative effects of high demands. Following this perspective, we recognize that teachers will naturally experience an increase in job demands during interventions so they would benefit from additional support and resources, with one obvious source being school leaders.

The current study

Grounded in the job-demands resource model (JD-R model; Bakker & Demerouti, 2007), we hypothesized that leader support of teachers is a critical job resource for boosting teachers’ perceived usefulness of the intervention, which again determines work engagement during an intervention implementation phase. While the JD-R model addresses both job demands and resources, our focus in this study is mainly on the resources part of the model referred to as ‘the motivational process’ (Bakker & Demerouti, 2007), where we focus specifically on resources provided by school leaders. Resources are particularly important when teachers are faced with high job demands (see Bakker et al., 2007). It’s important to clarify that in our research, we acknowledge the assumption that the implementation of interventions is demanding for teachers. This assumption is based on extensive research documenting the often low quality of fidelity of school interventions in general (e.g. Domitrovich & Greenberg, 2000; Elliott & Mihalic, 2004) and researchers’ subsequent calls for support systems for teachers during implementations (Domitrovich et al., 2008) as well as research on co-teaching specifically (Pratt, 2014).

To date, school leader support has consistently been shown to positively affect teachers’ nature of work and school development in general (Dicke et al., 2018) but has been minimally explored during periods of implementation. More specifically, despite school leaders’ pivotal role in shaping conditions for success, their involvement in co-teaching has been predominately ignored by research. A notable exception is in recent work by Rönn-Liljenfeldt et al. (2023), who conducted in-depth interviews with nine school leaders implementing co-teaching at their schools. Their findings underscore that while teachers primarily drive co-teaching, school leaders serve as essential facilitators in this collaborative approach. Our quantitative analysis, using data from 150 schools using co-teaching to boost student reading growth, complements their interview work, as we have captured a breadth of experiences.

Additionally, this current manuscript extends our study of a large-scale co-teaching intervention (see Solheim et al., 2017) to home specifically on the role of school leader support on teacher work engagement. This study was directly informed by themes emerging in earlier qualitative and mixed-method studies (see Gourvennec, 2021; Gourvennec et al., 2022), in which we studied teachers’ experiences when co-teaching. Most notably, high-performing teacher-dyads reported appreciating school leader support received during the co-teaching implementation, whereas other dyads reported a need for additional support. It was also made evident that school leaders varied in the amount of support and types of support offered. In summary, this current work occurs at a nexus of education, organizational behavior, and implementation science and extends our understanding from the perspective of teachers and school leaders so that we can consider intervention implementation from a more ecological perspective (Domitrovich et al., 2008).

It is important to clarify that although co-teaching is the context and not the focus of this work, there are concomitant resources and demands that must be understood. Although introducing a second teacher represents a resource, this arrangement can also be uniquely stressful for teachers (i.e. a demand) as it requires renegotiating roles within a classroom. For example, a co-teacher can be perceived to be intruding upon the homeroom teacher (Gourvennec, 2021), and such role ambiguity and role conflict have been found to contribute to burnout in team or co-teaching settings (Embich, 2001). As a final clarification, the people leading schools hold a diverse set of job titles, therefore, in this work, we use school leaders and school administrators interchangeably, indicating the individuals who are tasked with managing the operations of a school, including instructional leadership.

Literature review and theoretical frame

In the following section, we contextualize this work by first introducing our theoretical model. Next, we briefly summarize research on the implementation of the intervention in schools and consider empirical findings from the lens of the teacher and school leaders. Finally, we broaden our focus to the impact of leadership style on worker engagement in general.

Teacher work engagement and attitudes towards interventions

Interventions have been defined as ‘purposively implemented change strategies’ (Fraser & Galinsky, 2010, p. 459). Therefore, when interventions occur, teachers need to exert additional energy, as they must change how they go about their work – often learning new skills, strategies, and content. As such, teachers must fundamentally believe that an intervention is worth the demands. Not surprisingly, it has been documented that positive teacher attitude or buy-in leads to more successful interventions (e.g. Silin & Schwartz, 2003; Turnbull, 2002). Conversely, it has been long established that when initiatives are imposed in a top-down manner, a common reaction is for teachers to reject the change efforts and continue to work as they did before (Sikes, 1992).

Beyond teachers’ general attitudes toward intervention, we focus specifically on work engagement. Reflecting the reality of most current schools, the result of too many intervention demands has been acknowledged in the United Kingdom as initiative overload (Office for Standards in Education, 2010), which manifests as attitudes of cynicism and reluctance to engage efforts into new interventions (Lendrum & Humphrey, 2012). From the JD-R model, this disengagement can be viewed as a self-protection mechanism during times when individuals cannot cope with the demands and perceive that they cannot reach goals anyway (Demerouti et al., 2001).

It is also incomplete to consider teachers’ experience during intervention implementation without considering historical accountability practices. By adopting the JD-R framework to guide our approach to studying teachers’ experiences during co-teaching, we are rejecting simplistic accountability models that impose punitive consequences on schools when students fail to meet specific goals. Take, for instance, the No Child Left Behind Act of 2001 in the United States, which has faced criticism for its underlying assumption that consequences can incentivize school staff to perform better and prioritize student outcomes (Finnigan & Gross, 2007). Ironically, these educational accountability policies, originally designed to motivate, can detrimentally impact teachers’ well-being and engagement (e.g. Berryhill et al., 2009). Nonetheless, in the era of accountability, it’s essential to strike a balance between providing support and setting demands, which we aim to highlight through our work. There is already evidence that schools offering better working support for teachers, particularly regarding autonomy in the classroom, have shown the ability to better mitigate the adverse effects of accountability sanctions (Ingersoll et al., 2016).

Role of the school administrator in supporting teachers

Unfortunately, although the importance of teacher buy-in and work engagement is recognized in intervention research (Jensen & Solheim, 2020), more knowledge is needed about the extent to which school leaders’ support facilitates teachers during implementation phases. School-based research also demonstrates that teacher work engagement during interventions can be influenced by various forms of support, such as support from staff members and program developers (Turnbull, 2002). Administrators specifically have also been shown to positively influence teacher buy-in during reform efforts when they follow data-driven practices, such as attending to student assessments and behavioral reports (Yoon, 2016). This general support has been found to be negatively associated with emotional exhaustion (Ruble & McGrew, 2013) and as a protective mechanism for teacher burnout (Embich, 2001). In translating research to practice, Wasburn-Moses (2005) concluded that effective school leaders provide teachers with both emotional support (e.g. listening to individual challenges) and structural support (securing resources for continued learning).

Additionally, when we consider the intervention of co-teaching specifically, role ambiguity and role conflict predicted special education teachers’ stress levels in co-teaching situations (Crane & Iwanicki, 1986; Moss, 2015). As such, Brunsting et al. (2014) recommended that school leaders could alleviate such conflict through facilitating joint discussions and drafting clear job descriptions. Collaborative teaching can lead to stress on teachers’ self-identities as well, as each educator requires a robust yet adaptable individual teaching identity to form a common vision. Without this balance, working alongside colleagues may become more of a challenge for teachers rather than a means of support (Rytivaara, 2012). Rönn-Liljenfeldt et al., therefore challenged school leaders to take a more active and school-oriented approach to developing a culture of co-teaching rather than leaving the responsibility to co-teaching dyads.

Hypotheses

Based on convergence from our theoretical model and empirical research, we proposed and tested the following four hypotheses, starting with the simplest, direct hypothesis between leader support and teachers’ work engagement (Hypothesis 1). Then we considered the potential mediating role of teachers’ perceived usefulness of the co-teaching intervention (i.e. buy-in) within the pathway from leader support to teachers’ work engagement (Hypotheses 2–3).

Hypothesis 1: leader support is positively related to teachers’ work engagement

Leader support is an important antecedent for how leaders can help workers engage in their jobs (for meta-studies, see Crawford et al., 2010; Halbesleben, 2010) and in research specific to teachers (e.g. Hakanen et al., 2006). Leader support can take multiple forms. For example, leaders can directly increase the availability of job resources (Breevaart et al., 2014). Alternatively, and less directly, leaders may stimulate workers to make their own decisions and take control over their work, and in that way, increase their available job resources. Specifically in teaching, principal support has been shown to predict teachers’ work engagement, both directly and indirectly, via emotional pathways (Slišković et al., 2019).

Hypothesis 2: leader support is positively associated with the perceived usefulness of co-teaching

Previous research found that when leaders demonstrate support and individual consideration, workers show more positive attitudes toward change (Aarons, 2006; Farahnak et al., 2020), or in other words, workers will buy-in to a change effort. This indicates that a supportive leadership style would relate positively to workers’ attitudes to the implementation. The perceived usefulness of co-teaching, measured in the current study, is likely to reflect the teacher’s positivity or attitude toward the intervention itself. Previous research in co-teaching, indicates the need for school leaders to support teachers in visioning the possibilities in co-teaching (Mackey et al., 2018). Thus, there is reason to believe that teachers who perceive the two-teacher intervention to be useful for them may also perceive more support from their leader.

Hypothesis 3: the perceived usefulness of co-teaching relates positively to work engagement

According to the JD-R model, an increase in work environment resources will lead to higher work engagement (Bakker & Demerouti, 2008). We argue that the co-teaching intervention, where an additional teacher is added to the class, can certainly be considered a form of job resource-building intervention. However, for the two-teacher intervention to be considered a job resource, the teachers must perceive the presence of a second teacher as useful (i.e. the resources outweigh the demands). Thus, we included the teacher’s perceptions of the usefulness of being two-teaching in our model. We know that, typically, teachers are relatively isolated from their teaching colleagues (Dorman, 2003) and beyond supporting instructional tasks, an additional teacher may function as a form of social support (Jensen, 2021). Social support from colleagues can increase a sense of belonging and is essential for work engagement (Bakker & Demerouti, 2007). Additionally, more generally, research in professional development and teachers’ implementation of interventions have demonstrated that teachers’ beliefs about the intervention serve as a mitigating factor for the implementation (e.g. Fagan et al., 2017).

Hypothesis 4: leader support is indirectly associated with work engagement through the perceived usefulness of co-teaching

It has been suggested by Bakker and Demerouti (2007) that job resources may also act as mediators and interact in the prediction of engagement outcomes. Thus, in the current study, we considered it relevant to investigate whether the association between leader support and work engagement was mediated by the perceived usefulness of co-teaching Figure 1.

Figure 1. Theoretical model of the study.

Method

Context

The present study is nested within the intervention study, ‘Two Teachers’, where the main aim was to examine the effects of increased T-S ratio (which can also be described as a decreased S-T ratio) on student outcomes. One hundred fifty schools attended the project, and the schools were located in 53 different municipalities in nine counties in the southern part of Norway. Two classes at each school were randomly assigned to a treatment condition or a control condition, where the treatment condition had a second teacher present during Norwegian lessons, 8 × 45 minutes per week.

The study was approved by a third-party ethical agency in Norway, The Norwegian Social Science Data Service. Additionally, the study followed the ethical guidelines developed by the National Committee for Research Ethics in the Social Sciences and Humanities. For more detailed information on the Two Teachers study, see Solheim et al. (2017).

Sample

The sample in the current study included 150 teachers (98.7% females) in the intervention classes, with a response rate of 100%. The mean years of teaching experience were approximately 13, and 89.3% of the teachers reported having a bachelor’s degree, whereas 2.7% had a master’s degree. Moreover, 0.7% of the teachers were under 25 years old, followed by 25–29 years old (10%), 30–39 years old (27.3%), 40–49 years old (38%), 50–59 years old (18%), and over 60 years old (6%).

To collect school leader data, each participating school assigned a person responsible for conveying information between the project leadership and participating teachers, and this role could be assigned to either a principal or assistant principal. In total, 150 school leaders participated (67.8% females) with a response rate of 99%. The final sample had 55% principals, 42% assistant principals, and 3% reported that their role was ‘other’. The mean age among the school leaders was 52 years. All but one of the school leaders reported having a background as a teacher, with a mean teaching experience of 15 years. The mean years of school leadership experience were eight, and 62% of the school leaders had participated in national training for school administrators.¹ Each school received funds to support the number of hours of co-teaching, thus the co-teaching occurred at no cost for individual schools. It was voluntary for schools to participate in the research, and school leaders were encouraged to consult with classroom teachers.

Fidelity of implementation

Schools and co-teaching dyads were given autonomy to conduct co-teaching in the manner they believed best fit the needs of their students, thus fidelity was primarily collected regarding the presence of two teachers – schools reported weekly on the number of minutes of co-teaching. In total, two teachers were present in the target classrooms 88% of the time.

Measures

It should be noted that although some of the items on the measures did not specifically refer to co-teaching (e.g. Leader Support items), the surveys were distributed specifically connected to Two Teachers Project, thus we assume that teachers and leaders are referring to their experience regarding the project.

Leader support

Leader Support (Wannstrom et al., 2009) was related to teachers’ perceptions of support from their immediate supervisor. The scale consisted of three items (Cronbach’s alpha = 0.84) including,‘When needed; can you get help and support from your immediate supervisor?’ The responses were measured on a 5-point scale ranging from 1 = very seldom or never to 5 = very often or always.

Usefulness of co-teaching

The perceived usefulness of co-teaching was assessed by six items developed for the current study (Cronbach’s alpha = 0.80). The items were specifically concerned with teachers’ perceptions of how useful it was to be two teachers in the classroom in relation to assessing and observing student skills, trying out new teaching methods, keeping the class quiet, providing timely assistance for students, and planning instruction. Responses were measured on a 4-point scale from 1 = very useful to 4 = not at all useful. However, before conducting the analysis, the items were reverse coded so that higher numbers expressed higher perceived usefulness.

Work engagement

Work Engagement was measured by nine items (Cronbach’s alpha = 0.90) from the short version Utrecht Work Engagement Scale (Schaufeli et al., 2006). Sample items were: ‘At my job, I feel that I am bursting with energy; My job inspires me; I am proud of the work that I do’. Responses were assessed on a 6-point scale ranging from 0 = never to 6 = always (every day).

Statistical analysis

First, principal component analysis was applied to validate the tailor-made construct related to the perceived usefulness of co-teaching. Second, confirmatory factor analysis (CFA) with maximum likelihood estimation was applied using AMOS version 28 to establish model fit. Third, the reliability of constructs was calculated by applying Cronbach’s alpha. Fourth, we applied structural equation modeling (SEM) to assess associations between leader support, the usefulness of co-teaching, and work engagement. Fifth, to examine the indirect relation of leader support on work engagement through the perceived usefulness of co-teaching, the bias-corrected bootstrap mediation in AMOS was applied. MacKinnon et al. (2004) suggested this method to be the most appropriate method for handling skewness in the population. We applied 1000 bootstraps and 95% confidence intervals in our analysis. Additionally, we applied Mardia’s (1970) multivariate normality assumption to test for the normality of the data and applied Bollen-Stine bootstrapping (Bollen & Stine, 1993) with 1000 bootstraps to test the null hypothesis that the model is correct. Finally, means, standard deviations, and correlations of measures were calculated by applying the Statistical Package for the Social Sciences (SPSS) version 28.

A number of statistical tests were applied to determine if the model fit to the data, including the Bollen (IFI) comparative fit index (CFI) and root means square error of approximation (RMSEA). Acceptable model fit is indicated by CFI and IFI values of 0.90 or greater (Hu & Bentler, 1999), whereas Jöreskog and Sörbom (1993) suggested that an RMSEA value of < 0.08 suggests a reasonable model – data fit.

Results

Preliminary analysis

The items of the perceived usefulness of being co-teachers were created for the Two Teachers project study in particular, and the items have not been validated in previous studies. Thus, it was necessary to test the factor structure of these items before proceeding with further analysis. The Kaiser-Meyer-Olkin value exceeded the recommended value of 0.6 (0.79), and Bartlett’s test of sphericity was found to be significant (0.000), indicating that the data were suitable for factor analysis. Principal component analysis revealed that all items loaded strongest on the first component, suggesting that a one-factor solution was the most appropriate. Communalities values provide information with regard to the amount of variance explained, and low values (<0.3) could indicate a misfit of items with the other items (reference). For one of the items, the communality value was far below 0.30 (0.101). In addition, the factor loading was relatively weak at (0.318). Moreover, the wording of the item was specifically concerned with ‘how useful to be two teachers in order to hear the students read aloud’. The item may have been too specific for this scale; thus, we decided to remove it.

Confirmatory factor analysis and reliability and validity of constructs

The means, standard deviation, correlation, and reliability results for the three variables are reported in Table 1. Inter-item consistency, as measured by Cronbach’s alpha, ranged from 0.80 (Usefulness of Two Teachers) to 0.91 (Work Engagement).

Table 1. Means (M), Standard Deviations (SD), Correlations and Reliability (Cronbach’s α on the Diagonal) for the Study Variables.

Variable	n	M	SD	1.	2.	3.
1.Leader support	148	4.12	0.76	(.84)
2.Usefulness TT	147	2.17	0.41	.23**	(.80)
3.Work engagement	145	5.42	0.56	.17*	.17*	(.91)

** p < .01, * p < .05, TT = Two teachers.

Confirmatory factor analysis was performed to evaluate the fit of the data to our study model. All indexes in the study model were considered latent factors, loading on their associated manifest variables. Modification indices can be utilized to determine misfit and refer to a χ² statistics with one degree of freedom (Jøreskog & Sørbom, 1988). Considering some of the modification indices demonstrated high values, several correlations between error terms were included in the measurement model. Specifically, four correlations were included between error terms in the engagement factor, and one correlation was included between error terms in the perceived usefulness of being two teachers’ factor. Correlating error terms within the same factor is considered to be acceptable (Hooper et al., 2008). We further conducted a chi-square difference test, and the results revealed that correlating the error terms in our model significantly improved overall model fit (χ² difference = 89.315, df = 5, p < 0.001). Overall, the fit indices for our measurement model were satisfactory (χ² = 241.827; df = 127; RMSEA = 0.078, CFI = 0.923, IFI = 0.924). To test the normality distribution of the data in our measurement model, we first calculated Mardia’s (1970) multivariate normality assumption and found multivariate skewness and kurtosis. However, this can be expected with sample sizes relevant to SEM analysis (Blunch, 2012). There are various ways to deal with the non-normality of the data, and applying Bollen-Stine bootstrapping has been suggested as an acceptable method to deal with problems related to nonnormality (Blunch, 2012; West et al., 1995). Results showed that our measurement model fit better in 995 bootstrap samples but fit about equally well in 0 bootstrap samples. Moreover, the Bollen-Stine bootstrap probability of the model was .006, indicating that the model was correct.

Testing of structural relationships

We conducted structural equation modeling (SEM) to investigate the structural associations between the concepts in our model. The structural model demonstrated acceptable fit to the data (χ² = 240.418; df = 127; RMSEA = 0.077, CFI = 0.921, IFI = 0.923). Furthermore, ad hoc assessments were conducted to control for the influence of school and teacher characteristics on the outcome variable work engagement. Specifically, we included the variables of school size, teachers’ years of experience, and teachers’ formal education level. Results revealed no significant associations between the control variables and the outcome variable. Moreover, all significant paths remained significant after including the controls in our structural model, verifying the robustness of our results. Considering the inclusion of control variables reduced the model fit below the recommended thresholds, and were not significantly related to our outcome variable, we did not include these variables in our final model.

Final results revealed a positive association between leader support and the usefulness of co-teaching in the class, which was according to expectations (β = 0.23 p < 0.05), supporting hypothesis 2. Next, a positive relation was found between the usefulness of co-teaching and work engagement (β = 0.27 p = .01), supporting hypothesis 3. We also expected to see a positive relationship between leader support and work engagement, but no significant association was found (β = 0.13 ns). Thus, hypothesis 1 was not supported. Finally, we also tested whether the perceived usefulness of co-teaching mediated the association between leader support and work engagement by utilizing the bias-corrected bootstrap approach. Results showed that the perceived usefulness of co-teaching fully mediated the association between leader support and job engagement (β = 0.06, 95% [CI, 0.007, 0.18] p < .05). Conclusively, hypothesis 4 was supported Figure 2.

Figure 2. Structural model χ² = 240.418, df = 127, RMSEA = .077, CFI = .921, IFI = .923.

n = 296. ns = non-significant. *p < .05; ** p < .01.

Discussion

As we began this work, drawing upon our theoretical frame of the JD-R model, we predicted that leader support would be critical during the implementation of educational interventions on teacher perceived usefulness of the intervention and further on participating teachers’ work engagement. Implementation periods of interventions are inherently demanding because the onset of any interventions requires fundamental changes in how teachers do their work, and co-teaching can provoke particular forms of conflict. According to JD-R, when demands outweigh the level of support, as a psychological protective mechanism, teachers may disengage (Bakker & Demerouti, 2007). Therefore, we expected that leader support would be positively related to teachers’ work engagement during the implementation period. However, our findings did not indicate a simple or direct relationship. Rather the relation between support and teachers’ work engagement was fully mediated by teachers’ perceived usefulness (or buy-in) of the co-teaching intervention. Specifically, supervisory support was positively associated with teachers’ perceived usefulness of the intervention, and the perceived usefulness was positively associated with work engagement. So how do we make sense of these relationships? We will consider it by stages.

Teacher perceived usefulness and engagement

The second part of the model, in which teachers’ perceived usefulness of the intervention (i.e. buy-in) predicts their engagement in work (in general), is anticipated due to previous empirical findings. Teacher buy-in (teachers’ attitudes, receptivity, and belief that an intervention is useful) has been associated with teachers’ engagement in a change effort (see Datnow & Castellano, 2000; Timperley & Robinson, 2001). Therefore, our work reaffirms earlier work and amplifies the importance of teacher buy-in at the early stages of implementation for their work engagement.

This finding is relevant and cautionary that school leaders prioritize teachers’ belief systems when launching change efforts. For example, Timperley and Robinson (2001), using schema theory (organized knowledge and belief structures) as an explanation for the persistence of teacher beliefs, emphasized the effort in needing to change schemas. Similarly, working to understand which teachers would engage in the implementation of a school reform model, Datnow and Castellano (2000) found minimal demographic patterns predicting which teachers were supporters of the intervention, such as years of teaching or age. Instead, teacher support for an intervention was primarily connected to ideological beliefs and philosophical underpinnings. Without teachers’ belief in the value of the intervention, there is risk of disengagement or simply behavioral compliance, neither of which will lead to meaningful change.

It has been long recognized that effective leaders will take individual consideration toward the people working for them (Bass, 1985). In this case, leadership could include an administrator inquiring and understanding a teacher’s beliefs and attitudes toward intervention and using that input. In other words, teachers do not simply need resources to complete an intervention, but they also need to be part of the conversation surrounding the important decisions. As Rönn-Liljenfeldt and colleagues concluded from interviews with school leaders in schools enacting co-teaching, leaders should support teachers in developing common vision and goals for co-teaching. For example, before any collaboration begins, co-teaching dyads must discuss their epistemic views and value systems surrounding co-teaching, and these discussions may even occur at the school level, thus impacting the overall culture of teaching.

Leader support and teacher perceived usefulness

The first stage of the model, though, in which leader support predicts teachers’ perceived usefulness, is less documented in the literature to date and, therefore, is of interest. Research about teachers’ buy-in has tended to focus on what creates teacher resistance, such as that top-down mandates can result in the rejection of change (e.g. Sikes, 1992) or a culture of compliance (Blackmore, 2004). Shifting the conversation to the positive, however, we are encouraged that administrator support is associated with a greater buy-in of an intervention. Looking to our theoretical model (JD-R), at the most fundamental level, for teachers to engage in a change effort, they need to feel emotionally capable and supported (i.e. not exhausted). Perhaps the experience of early support allowed teachers to have the emotional resources available to more willingly commit to the increased demands or a general sense of optimism.

The conclusion that early support leads to perceiving the utility of co-teaching, dovetailing with Rönn-Liljenfeldt and colleagues findings (2023), in which they emphasize that leader support can and should include structural and relational support and occur early in the process of co-teaching. For example, early support may be in the form of working with teachers to partner compatible personalities and scheduling co-teaching in a manner which does not provide additional burdens for teachers. Support should also provide greater clarity of expectations through co-construction and the autonomy for teachers to make decisions that are aligned with those expectations.

In conclusion, this study underscores the pivotal role of school leaders in facilitating the effective implementation of co-teaching interventions, aligning with the established body of research emphasizing leaders’ ability to provide essential elements such as vision, support, resources, and leadership. However, our research extends this understanding by illuminating a critical pathway to fostering teacher work engagement during the time of intervention implementation which demands elevated levels of teacher efficacy for improved student outcomes. The findings emphasize that when educators perceive meaningful support in the context of co-teaching, they are more inclined to recognize the inherent value of the intervention, resulting in increased work engagement which would ultimately lead to a more successful implementation. This insight underscores the significance of cultivating a supportive environment led by school leaders, which can significantly enhance the impact of co-teaching interventions in educational settings.

Limitations & future directions

There are several limitations to our study that should not go unnoticed. First, our study is based entirely on self-reported survey data. There are important, inherent limitations associated with this type of data, such as the risk of common method biases (Podsakoff et al., 2003). However, throughout our analyses, we performed several steps to reduce problems associated with common method bias, which included developing a well-fitting measurement model, and validation of constructs (Conway & Lance, 2010). Overlap between constructs may lead to biased relationships (Brannick et al., 2010). However, confirmatory factor analysis performed in our study revealed that all constructs loaded on their respective factors, indicating that this was not a concern. Moreover, our understanding of the relationship between leader support and teacher engagement would be enhanced by qualitative data that reveals how support is enacted by leaders. Additionally, in a mixed methods approach, more objective metrics of support (e.g. the number of minutes of shared planning time per week) could be collected so that researchers could triangulate the survey data and more accurately describe the target constructs.

An additional limitation of our study is that we did not actually ask the teachers whether they perceived the intervention as demanding, implying that a specific ‘demand-variable’ was not included in our study model. Still, based on the extensive evidence confirming that all interventions put additional demands on the participants, we argue that studying how aspects of job resources such as leader support and perceived usefulness of the intervention relates to work engagement during an intervention context is valuable. Thus, in our study model, it made sense to focus mainly on the motivational process of the JD-R model. Third, as described earlier, systems approaches to implementation research should consider multiple types of individuals involved. In our work, we included the perspective of teachers but did not include the administrator report, nor the students. It is necessary to link the experiences of school leaders and teachers with student-level outcomes, particularly evidence of growth, to provide more compelling evidence of how implementation ultimately connects with school improvement. A fourth limitation relates to that data are cross-sectional, whereas to establish evidence of causal associations it is necessary to conduct longitudinal analyses. Finally, in this work, we did not divide the schools into performance categories of schools with low, middle, and/or high levels of teacher engagement. It would be illuminating to target a range of schools and identify trends within each category.

Ethical statement

The project followed the Ethical guidelines developed by the National Committee for Research Ethics in the Social Sciences and Humanities (NESH). Also, the Norwegian Social Science Data Service (NSD), a third-party ethical oversight agency in Norway, reviewed and approved the study. We obtained consent for participation from teachers and the parents of the students involved in the project.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The project is funded by The Research Council of Norway, Research Program “LÆREEFFEKT”, grant number: [256197].

Notes on contributors

Erin M. McTigue

Erin M. McTigue is a Researcher at the Norwegian National Center for Reading Education and Research, University of Stavanger, Stavanger, Norway.

Maria Therese Jensen

Maria Therese Jensen is Professor at the Norwegian National Center for Reading Education and Research, University of Stavanger, Stavanger, Norway.

Aslaug Fodstad Gourvennec

Aslaug Fodstad Gourvennec is an Associate Professor at the Norwegian National Center for Reading Education and Research, University of Stavanger, Stavanger, Norway.

Oddny Judith Solheim

Oddny Judith Solheim is a Professor at the University of Stavanger and, at the time of this research, was at the Norwegian National Center for Reading Education and Research and is currently in the Department of Education and Sport Sciences.

Notes

1. The national training for school administrators is offered school leaders in primary and secondary education, and is designed to be combined with a full-time job as a school leader. The study is part of the national leadership training for school leaders offered by the Norwegian Directorate of Education.