Why is it so hard to study school leadership in cross-cultural research?

Jelena Veletić

Evidence shows that some of the key aspects of educational policies, like school leadership, are very sensitive and shaped by the culture and context where they are executed.  Nevertheless, we frequently want to compare and contrast such concepts across different cultures and contexts. To do so, our instrument (e.g. test, questionnaire) must operate equally across different groups. In more simple words, we must ensure that our instrument measures the same thing in all groups of interest. This is called measurement invariance. You may wonder why it is important to establish measurement invariance. Well, consider some of the challenges we faced in our study of school leadership.   

The study

Our study focuses on a specific school leadership conceptualization – instructional leadership. Strong instructional leaders are principals, who emphasise high-quality instruction, give instructional feedback to teachers and support the use of assessment in the classroom. As previously mentioned, leadership is enacted within a context of culturally embedded values and worldviews, and comparisons between different groups may be challenging. Therefore, we use data from the Teaching and Learning International Survey (TALIS) from a set of Nordic and Baltic countries sharing geographical, cultural and linguistic characteristics. Although the selected countries are different in many respects, the selection is relevant because the Baltic countries have been influenced by their Nordic neighbours over the recent decades. However, it turned out that modelling instructional leadership, even in seemingly homogenous countries, was far from being straightforward.

What was challenging?

The different interpretation of feedback in diverse countries led to an inability to compare models across them. Giving feedback is an important element in the process of managing instruction in schools, which is an important dimension of instructional leadership. Across countries, does that mean that effective leadership assumes a certain amount of feedback at a certain time? Does it assume that feedback is equally perceived across groups? A simple first look at the TALIS data reveals that teachers in some countries receive significantly more feedback than in others. However, when modelling these responses as a latent variable, it becomes clear that the interpretation of feedback differs across countries. In more technical terms: the measure was not invariant. In our case, this invariance appeared as a cross-cultural difference between the Nordic and Baltic countries. Several cultural dimensions could be relevant to understand why this is so. Some cultural values, such as power distance and collectivism, might be relevant to explain the differences commonly seen between eastern and western countries. Power distance refers to the extent to which the community accepts and endorses authority. For example, in high power distance cultures, feedback and interactions between different hierarchical levels (in the context of school: principals, teachers, students) may be regarded as activities where power and authority are established and maintained. As such, feedback received from principals might connote control or monitoring. Conversely, in low power distance countries, feedback might be welcomed and represent support and a mechanism where leaders and teachers collaborate and engage in joint meaning-making. Furthermore, in individualism-oriented countries, in contrast to more collectively oriented countries, feedback appears to be necessary to ensure that each individual’s efforts work in unison with the larger aims and goals of the organisation (i.e., the school). Thus, feedback as one of the instruments of instructional leadership practice appears to differ across these cultures in its sign or value direction (positive-negative), perceived reliability and accuracy as well as the form.

Sensitivity to such differences is essential in making conclusions and giving adequate policy recommendations. Otherwise, any differences we observe between groups may be due to the instrument used and not due to meaningful differences in the constructs of interest.