Page 71 [71]
STUDY ONE
The interviews were conducted in Hungarian, the mother tongue of all the
participants, and the length of the interviews was approx. 15-20 minutes. The
reason behind the preference for Hungarian was that it is wiser to carry out
interviews in the mother tongue of the participants as they are not limited by
their English language proficiency (Dérnyei 2007). The instrument used for
data collection was also devised in Hungarian, although the interview proto¬
col provided in Appendix A has been translated into English.
4.4 METHODS OF DATA ANALYSIS
The reason for the number of interviews (12 interview sessions altogether) was
that in qualitative research and, therefore, in interview studies, two key con¬
cepts have to be taken into consideration. One of these concepts is iteration,
and the other one is saturation (Dôrnyei 2007). Iteration refers to the cyclical
process of collecting data (conducting interviews) and analyzing the obtained
data and doing this repeatedly until data saturation is achieved. This means
that data collection lasts as long as the newly conducted interviews reveal new
patterns and themes. As soon as the interviews start yielding data that simply
repeats what previous participants have already shared in the previous inter¬
views, data is saturated, i.e., new themes do not emerge (Corbin—Strauss 2015).
The data analysis was conducted based on the Constant Comparative Meth¬
od of the Grounded Theory (Corbin-Strauss 2015, Glaser-Strauss 1967).
After reaching data saturation, all recordings were transcribed verbatim, and
the transcriptions were read through, followed by the initial coding of the
texts. At this stage, emerging themes were searched for, and when all possible
new themes had emerged, all transcripts were reread and checked in order to
ensure that no important information was left out. Initial coding was followed
by second-level coding, wherein the transcripts were read again, the codes
were compared, and similar or closely related codes were clustered together
under a new label. Furthermore, in order to ensure quality control, two of my
colleagues were asked to take the role of co-coders and code certain segments
of the transcript. Finally, participants’ utterances provided as quotations in
this chapter were translated into English.
4.5 THE FINDINGS OF STUDY ONE
After the final (axial) coding, all of the codes were collected and resulted in
the following emerging main themes summarized in Table 8.
+ 69 +