Study methods

Jons G.J. Beek, Zehava Rosenblatt, Chris Phielix, and Theo Wubbels

Introduction

In the present chapter, we describe the methods that address our research questions listed in the introduction (Chapter 1) in regard to teachers’ and principals’ accountability. We made sure that these methods best investigate the following topics, so that credible conclusions may be drawn.

• 1 Accountability as a two-dimensional concept.
• 2 Differences between accountability dispositions, divided by background characteristics such as gender and seniority.
• 3 Differences in accountability dispositions toward different audiences, in particular parents and school management.
• 4 Country differences in teachers' and principals' accountability.
• 5 Country differences in cuhural values and organizational support.
• 6 Prediction of accountability dispositions by two cultural values: individualism and collectivism.
• 7 Prediction of accountability dispositions by organizational support.
• 8 Influence of principals’ accountability on respective teachers’ accountability.

The topics of this research project are not particularly sensitive or controversial, but we were attentive to all ethical research issues regarding human subjects who were involved in this study. It was of utmost importance to us to ensure that the questionnaire items would not include any sensitive content, that respondents would have the right to refuse participation, and that they could quit the study at any time. The approval of respective Institutional Ethics Review Boards (IRBs) in the researchers’ academic institu- tions in each of the participating countries was secured.

The following sections of this chapter present the process of data collection and describe the study's instruments and measures. The last section of this chapter specifies the study's analytical approach taken to explore the study's research questions.

Data collection

We collected data from teachers and their respective principals using a questionnaire as a research tool. In total, 2,554 teachers from 185 schools and 132 principals from 117 schools completed the questionnaires. Data were collected in each participating school from both principals and teacher faculty members so that teachers were matched with their respective principals using identifying codes. This matching enabled analyses that investigated relations between principals and teachers.

Identical questionnaires were distributed in each country, using applicable translation when necessary. In two countries (Canada and China), data were collected only from teachers, not principals. Table 3.1 shows the specific provinces or regions in which the data were collected in each country and the number of teachers and principals within each sample. In most cases, the data collected were generally representative of the pertinent country in terms of educational structure and educators' background characteristics (gender, age). However, in some cases it was difficult to collect the desired representative data. For example, in the case of South Africa, only privileged schools and schools in the Pretoria area accepted our questionnaires, so the sample was comprised of predominantly (but not entirely)

Table 3.1 Provinces or Regions in the Participating Countries and Number of Teachers and Principals

white teachers and principals who were highly educated and trained, and schools were relatively wealthy.

The questionnaire data were processed first into eight separate data sets (for each of the participating countries) that included respondent identification (ID), school ID, country ID. and background variables. The separate data files were combined into three sets of data. The first consisted of teachers’ data, the second included principals’ data, and the third consisted of a combination of both teachers' and principals' data. This structure enabled a broad spectrum of analyses, with IDs used as a matching mechanism. When data were missing, these were imputed following the procedure described in Appendix З.1.¹