75 thus  a  series  of  non-parametric  tests  will  be  used  since  the  data  obtained  are  in
non-continuous  scale.  Furthermore,  as  Pallant  2011  suggested,  when  there  is only a small number of sample obtained, non-parametric statistics will work best.
The Wilcoxon signed-rank test was the ideal test for comparing the mean score of non-continuous data, which in this case was ordinal data. Therefore, the Wilcoxon
signed-rank  test  was  used  to  compare  median  scores  from  the  pre-  and  post- questionnaires responses.
b. Qualitative data analysis
All  the  three  forms  of  qualitative  data  serving  to  answer  the  second research  question
,  i.e.  interview  transcripts,  students’  reflective  journals,  and researcher’s observation field notes were analyzed together since the information
they  provided  was  interrelated  and supporting to each other. Following Mbato’s
2013  advice,  the  qualitative  data  were  analyzed  both  iteratively  during  the research and after the research was completed in order that the evidence becomes
more  organized  and  ideas  are  clarified.  The  steps  of  analyzing  those  qualitative data were organizing, describing, and interpreting the data Best  Kahn, 2006.
At  the  stage  of  organizing,  the  data  were  grouped  and  coded  into categories. Each qualitative data form was coded differently in order to assist the
researcher  in  finding  the  commonality  within  the  data  set  as  well  as  in  building themes  subsequently.  There  are  two  forms  of  codes  as  coined  by  Gibson  and
Brown  2009,  p.141,  namely  apriori  codes,  which  are  created  prior  to  the  data examination,  and  empirical  codes,  which  are  generated  through  the  data
76 examination.  Based  on  that  notion,  the  codes  generated  in  this  research  were
empirical codes. There were some coding guidelines created in order to assist the researcher
in  organizing  the  data  and  help  the  readers  understand  the  data  sources.  The following table presents the coding guidelines.
Table 3.3. Data coding guideline
Coded Items Codes
Interview transcripts INT
Students’ reflective journals REFL
Researcher’s observation field notes FN
Student ST
To follow the full code, an example is given for a data coded ‘P ST01-INT 5’. The
first letter preceding the code indicates the appendix, so it means that the data is located in Appendix P. ST01 means student 1. INT 5 means the 5
th
conversation from  an  unnumbered  interview  transcript.  This  is  because  each  student  only
contributed  one  interview  transcript.  Another  example  is ‘Q ST05-REFL 2.1’. It
means the data is located in Appendix Q, student 5. Meanwhile, REFL 2.1 shows that the data was from reflection number two, statement one.
As  they  were  organized,  the  data  were  reduced  into  a  manageable  form Mills,  2011.  Once  the  data  were  organized,  each  data  category  was  described
and  explained.  Then,  the  data  were  interpreted  in  order  to  answer  the  second research question.
Triangulation was also used in the qualitative data analysis. Citing Glesne 2006 and Bogdan  Biklen 2007, Mertler 2009 suggested that triangulation
is an important technique since it is  “a process of relating multiple sources of data PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI