The data collection took place in January-February 2021 when the questionnaire for secondary school teachers was posted to three groups on a popular social media platform and the participants of the interview study were asked to distribute it among their colleagues. As indicated at the beginning of the questionnaire (see in Appendix F), its completion was voluntary, and it took approximately 20-25 minutes to fill it in. After data collection, data were transferred to SPSS 22 and subjected to analysis.

 

In all cases, the Cronbach’s Alpha value reached the .6 threshold, which was an important first step, because as it was suggested by Dörnyei and Csizér (2012), above that threshold the scale is considered reliable. After calculating the Cronbach’s Alpha reliability coefficients, a principal component analysis (PCA) was run to test whether the grouped items measure the same dimension. In the case of the first three scales, it was not sensible to run this test, however by running factor analysis on them, there could be some new groups of topics observed, which teachers think similarly about. The items on scales 4, 5, and 6 seemed to be on the same dimension, so based on their reliability analysis and the PCA, they were deemed reliable and ready to be used for the main study. The seventh scale turned out to be problematic and the reason was that there should have been constructs created with an almost equal number of items so that this scale could work. A Varimax rotation factor analysis was run as an exploratory tool to identify the items that loaded together and constituted separate dimensions. Eight dimensions were identified in this way, however, after testing them again with PCA and computing their Cronbach’s Alpha, none of them emerged as reliable constructs. What became evident after the reliability check was that the variables influencing the teachers’ decisions about incorporating global content in their lessons cannot be analysed as one dimension as there are some underlying constructs and these constructs needed to be properly developed, with enough items. As detailed in the previous section (3.4.2.1), seven new constructs were created for the secondary school context (for 30 items) and eight constructs were created for the university context (for 33 items), and the modified questionnaire was considered as the final questionnaire for the main study.