• No results found

5. Conclusion

5.2. Limitations of the Study

A number of limitations need to be considered, as there are several problematic aspects related to method and process in this study. The following section deals with some of these elements.

5.2.1. Criticism of Method and Result

Firstly, I might not have asked the two groups the same follow-up questions because they answered differently in the first place and there were only two girls, as opposed to five participants in the boys’ group. Secondly, I provided the girls with a list of characters because the girls had problems remembering names, whereas the boys did not. This may have changed whom the girls mentioned or they may have mentioned some characters that they would not have remembered if they had to speak from memory. Thirdly, I told the girls at one point what the boys had answered, but only after they seemed to have exhausted themselves on the topic.

Upon learning what the boys’ group had answered, they added to their initial response. There is reason to believe that without this input their answers would have been less nuanced.

The reliability of the study is made very much more complicated by the coincidental separation into two gender specific interviews. Due to a gender bias, I may have asked the boys more questions than the girls. Naturally, the boys would first take the male perspective, and then be encouraged to consider the female perspective, whereas I may not have asked the girls to take the male perspective, causing them to focus solely on the female aspect.

Moreover, it is relevant to think that because I am a woman, I immediately perceived the girls’ answers as true or right, and I did not ask questions more adherent to a male mind-set because the answers from the girls harmonised with my own world picture. The boys however, probably received questions whenever I was expecting a different answer from them. Thus the boys may actually have touched upon more aspects than the girls both because of my follow-up questions and because they outnumbered the girls in the first place. Had the interviewer been male, the result may have been reversed. This difference is especially evident in the question about what gives a person power: the boys were focused on the

system, but through my questions, they eventually ended up pointing to relational and personal features as well. The girls were not asked about the system, and did not mention this either.

Yet another problem relates to my role as moderator in the interview: Listening through the recordings makes two factors clear: I boldly stated that as moderator I was not to interfere and be a part of the discussion, which I then broke with much pathos. I also fear that in some situations, I very much took on the role as teacher, since my preferred method in the classroom so much resembles the talk round the table of this interview. Thus at times, I committed the crime as interviewer of interrupting the participants, I did not let the participants finish, or else I finished for them, summing up their reflections and potentially killed all the other thoughts they must have harboured on the issue. These are all typical teacher hazards.

5.2.2. Power Relationships

In the case of power relationships, there is a clear problem related to my role as teacher in the eyes of these participants: Six of the seven participants had at one point been my students and three of them were my students at the time of the interview. This may of course have influenced greatly how they answered, regardless of how strongly I emphasised that this was not school work. Both the initial talks and the interviews took place in the school building, and the participants may thus have entered the interview as students rather than participants.

In addition, their answers may have been coloured by the hand-outs I gave them (appendices 3 and 11). There is reason to believe that my dual role as teacher and researcher may have caused them to perceive these hand-outs as keys rather than resources. Some answers, especially in the more general parts hint at this. Nevertheless, I still regard the hand-outs as vital to the success of the interview, as they would have sensitised the students to the topic and thus made them better equipped for the interview.

To mitigate the power relationship between us, I did several things to minimise the bias: I carefully explained to the students the purpose of the study, and this was repeated several times during the process before, during and after the interview (McKay 55). The participants were informed about how the information would be dealt with and that answering questions was voluntary. I sometimes addressed the participants directly, but I hope that the participants were not intimidated by these questions. Still I cannot be absolutely certain that the answers are honest and a complete reflection of the students’ collective views, since there is group

dynamics at work as well. There may be other power relationships present in the participant group, which are unknown to me and which may influence the process and their willingness to speak freely. Although the participants seemed comfortable enough, this is difficult to tell.

In the process of interpreting the material I have had to consider all of the above mentioned factors. In this respect, symptomatic reading is relevant, which refers to a type of reading which takes into account both “the participants’ relationship to the topic being discussed and the social context of the interview” (McKay 55). I cannot rule out that some students gave

“clever” answers they thought I would like to hear. Therefore, the participants may say their vocabulary has expanded, because they consider this the “right” answer, one that would please me, or cause me to think they are clever, not because it is necessarily true.

A further complication occurred in the boys’ group: some of the participants were perpetually afraid of spoilers because two of the participants had already read all or most of the books. At one point this made it necessary to let three of the boys leave the room whilst the other two delved into the discussion of how power changes in the series. At other times it was enough to ask the boys to cover their ears for some seconds. The responsibility was felt heavily by those participants who had read the books: they did not want to ruin anything for the others which is recognised in the frustrated exclamation: “All I can think of are spoilers! … It is a mine field!”. It is relevant to think that these participants could have said more than they actually did out of courtesy to the other participants.

5.2.3. Reliability

Given all the implications above, it is prudent to question the reliability of the study. In qualitative research, “ [a] researcher may achieve credibility or internal validity by carefully recording an analysing all of the data gathered and presenting it in a fair and unbiased manner”(McKay 13). Triangulation is one way of achieving this (McKay 13): I used both interview and questionnaire and could thus to some extent compare the answers given since I had more than one source of data. I very carefully recorded the interview and also made relevant notes during the interview. This provided me with more answers from some of the less outspoken participants.

Internal reliability “relates to the extent to which someone else analysing the same data would come up with the same result” (McKay 12). I will argue that this is the case in my study, though the fact that the transcripts have been translated from Norwegian into English offers

the possibility of misinterpretation. This is in particular true of idiomatic expressions and puns that are not always transferable across languages. Some meaning may therefore be lost in translation, but in general this should not affect the result to a large extent. Whether there is external reliability, however, is another question, as this “deals with whether or not another researcher, undertaking a similar study, would come to the same result” (McKay 12-13). I consider this highly unlikely, as the result will depend to a large extent on the backgrounds of the students, be it their personal backgrounds, their social backgrounds, and everything which builds personality. It is the students’ personal reflections that have been investigated, not a neutral phenomenon, and given that the students had only watched parts of season 3, it will be difficult to copy this feature in a new group.

5.2.4. Validity

The internal validity of a study refers to “the extent to which the findings of one study may be generalized to a wider population” (McKay 12). Although this may be true of some aspects of the study, like the popularity of the series and the effects of reading, the entire point of the study is to show what a defined group of students thinks about power in GOT. This may be similar to what others think, but it may equally well not be, even though there seems to be consensus in this particular group of students.

When it comes to the transferability or external validity in qualitative research, which is defined by McKay as “the degree to which the findings of a qualitative study can be applied to other contexts” (13), my study does not meet the requirements. McKay specifies how researchers need to provide a “complete description of the participants and context of the research so that the readers can determine to what extent the findings may be applicable to other contexts” (13). This means very carefully describing both the context of the research and the participants. The context is already described in the methodology section, but for the sake of anonymity, information about participants has been restricted to a minimum. The group is described in general terms, in terms of age, gender, school and whether they have read the books or just watched the series. There is additional information found between the lines in the questionnaire summaries (appendices 8 and 9), and in the transcripts (appendices 5 and 6), but the answers are meant to be impossible to attribute to the different participants.

Any added information about the participants, would have further jeopardised their anonymity. It will nevertheless be impossible to copy this group both on the individual level and as far as group dynamics go, thus further information would be of limited use.

When considering the liabilities of a study listed in McKay (13), there are several problems connected to transferability: The selection of participants was carried out among a small group of students, and the researcher knew them from before, as their teacher. The interview took place in school, all of which may further have contributed to upholding the roles as teacher and students, not researcher and participants. An additional problem with transferability is the fact that these students have only seen two seasons and four episodes of season three, because that was how many episodes had been aired at the time. Consequently, it will be difficult to find a group at a later time, in which the participants have the same background and knowledge of the series. This group would necessarily have to arrive at different conclusions about power since that is one of the strong characteristics of the series;

the power balance shifts fast.

5.2.5. Dependability

Dependability relates to “the degree to which the results reported in the study can be trusted or are reliable” (McKay 14). This means providing details about procedure, which are available in the methodology section, and it requires that you catalogue your findings in a manner that makes it easy to review the data. In this case, both the transcripts are in the appendix, and there is an overview of the answers from the questionnaires as well. The “rich description of the students involved” (McKay 14) is nevertheless impossible, as is discussed above. All the other steps that were taken to carry out this study are clearly presented in the method chapter, and the necessary material is among the appendix which should make it possible to carry through the same research with a different group.

5.2.6. Generalizability

The generalizability of qualitative studies is naturally a little more difficult than in quantitative research (McKay 14), since in a focus group interview, the number of participants is very small. In addition, the researcher has less control of the data collection, since one of the points of a focus group interview is to make the participants talk as freely as possible whilst staying on the topic. The researcher is moderator, not a part of the discussion, and the aim is to find out what the members of a certain group think about a certain issue. Thus the findings will be true of this group, but that does not mean it will be true of other groups. It may be true, however, and this may be tested if someone was to carry out the study on their own.

By supporting the findings with theory, one may reach an analytic generalisation, described by McKay in the following quote: “When local conditions are not controlled and structured, as in most qualitative research, the readers themselves must determine to what extent the findings are applicable and transferable to other contexts” (15). As far as gender goes, it is obviously not possible to conclude based on the data elicited from just two individuals, but it is possible to conclude based on this group of students. In fact, if there is one generalising aspect worth mentioning, it will be the advice given by these students about the choices we are confronted with as teachers: to select material for our students, which is both intriguing enough to catch their attention and simultaneously may teach them something important.