Reciprocal and dynamic relationships of teachers' personal resources, work engagement, and instructional quality: A resource-oriented approach
Final Report Abstract
Our project focussed on three aims: (1) Expanding current models in educational research by combining key variables of different research fields (personal resources of teachers, i.e. teacher efficacy and teachers’ professional knowledge, job resources, job crafting, work engagement and instructional quality) and by addressing both between- and within-person relationships, (2) expanding research on unidirectional effects by investigating reciprocal relationships between key study variables over time and (3) generalization of the proposed mechanisms across teacher populations. In total, we reached a sample size of 159 teachers (111 biology teachers and 48 math teachers). We collected weekly self-report data on teachers' personal and job resources, job crafting behaviours, work engagement and instructional quality. After an initial survey (T0; i.e. to assess traits of study variables), participants selected a reference class for the study and received a short questionnaire once a week in the morning (referring to the past working week) and afternoon (after school, referring to the past working day). After six weekly measurements, the teachers received a closing questionnaire (T7) with the T0 variables. Professional knowledge was assessed as part of T7 with validated tests. However, the implemented planned missing design did not allow us to impute missing values, impairing test quality. Further analysis of the teachers’ professional knowledge was therefore not pursued. Due to our hierarchically structured data, we employed multilevel modelling techniques with N = 684 weekly measurements at Level 1 (within-person) and N = 111 biology teachers at Level 2 (between-person). We used multilevel structural equations models to test contemporaneous relationships and mediation effects between our constructs (Objective 1) and random-intercept cross-lagged panel models to test reciprocal relationships between study variables over time (Objective 2). We found effects of general self-efficacy on work engagement and general measures of instructional quality (classroom management and supportive learning environment). We also found effects of domainspecific self-efficacy on the domain-specific measure of instructional quality (cognitive activation), but not on work engagement. Overall, personal resources play an important role in instructional quality, but the assumed mediating role of work engagement is not as clear as expected, especially with regard to domainspecific variables. Regarding job resources, we found no effects on instructional quality and only one relation for job resources (job autonomy) to work engagement. Possibly, the null effects of supervisor support on all outcomes variables stem from the fact that teachers do not interact with their school director on a weekly basis. This raises questions regarding the role of the school principal as a major job resource compared to other situational variables such as job autonomy. However, the lack of effects for some mediation analyses may also stem from insufficient power. Surprisingly, the analysis of our weekly data with RI-CLPMs did not reveal any of the proposed reciprocal relationships over time. While we did find some contemporaneous associations between personal resources, work engagement and instructional quality, the cross-lagged effects were inconsistent. Whereas earlier studies on reciprocal relationships used traditional cross-lagged panel models (CLPMs), we employed RI-CLPMs to ensure that the estimates of lagged parameters were not confounded by trait-like associations at the between-person level. Thus, our results call into question earlier research findings regarding the magnitude of reciprocal effects between the variables under investigation in our study. Interestingly, the within-day analyses (see Objective 1) show lagged effects of morning personal resources on evening work engagement and general instructional quality. We therefore conclude that relationships exist between study variables indeed exist, but that they do not hold from one week to the next. Experiences in teaching other classes and other private or work-related experiences are likely to blur the possible weekly relationships between study variables. With regard to job crafting, we found neither a mediation effect at the within-person level, nor assocations of work engagement or job crafting with job resources. However, we did find reverse effects of work engagement on job crafting. We thus tested reciprocal relationships over a period of six weeks with a RI-CLPM approach. Model results showed no consistent reciprocal relationships between state work engagement and state job crafting over the entire period of six weeks. However, starting from Week 4, we found that state job crafting (Week 4) affected state work engagement in the following week (Week 5) which in turn was related to state job crafting one week later (Week 6). These findings provide some evidence for a positive cycles between job crafting and work engagement.
Publications
-
(2018). Der Einfluss motivationaler Orientierung auf die Unterrichtsqualität und das Wohlbefinden von Biologielehrkräften. In D. Krüger, P. Schiemann, A. Möller, A. Dittmer, J. Zabel, K. Schlüter & J. Großschedl (Eds.), Erkenntnisweg Biologiedidaktik 17 (pp. 137-148)
Milius, M. & Nitz, S.
-
(2020). Do it yourself – Job Crafting und emotionale Intelligenz für die Zukunft der Arbeit. In J. Nachtwei & A. Sureth (Eds). Sonderband Zukunft der Arbeit (HR Consulting Review, Vol. 12, pp. 268-271). VQP
Lopper, E., & Walf, J.
-
(2020). Interventions to Improve Employee Well-Being. In: L. Yang, R. Cropanzano, C. S. Daus, & V. Martinez-Tur (Eds.) The Cambridge Handbook of Workplace Affect (pp. 214–226). Cambridge: Cambridge University Press
Hülsheger, U. R., Hoppe, A., & Michel, A.
-
(2020). The Approach-Avoidance Job Crafting Scale: Development and Validation of a New Measurement. In G. Atinc (Ed.), Academy of Management Proceedings (Vol. 2020, No. 1, p. 18656)
Lopper, E., Horstmann, K. T., & Hoppe, A.
