Reliability and Validity

All empirical data needs to be quality assured. The traditional ways of doing this is to look at validity and reliability (Jacobsen, 2005).

4.9.1 Reliability

Reliability is concerned with whether or not the results of a study are repeatable (Bryman, 2008, p. 34). And so in qualitative studies it is difficult to test the reliability of the collected data. Grønmo (2005) argue that when considering reliability in qualitative research, the reliability of the research becomes influenced by the fact that the influence of the

researcher is more evident during the collection of the data than it would be in quantitative research (p. 228). Jacobsen (2005, p. 225) writes that the reliability of the data can be affected by the effect of the study taking place. This means that the data that has been gathered does not accurately reflect reality, but is created or affected by the study that is taking place. This can be caused by things such as the researcher interrupting the subject or disturbing the natural conversation or the body language of the researcher. During the course of the interviews I was aware of my potential effect on the subjects. And while I tried to stay as neutral as possible during the conversation, because the interview was semi-structured I had to have a certain presence in the conversation.

The context of the interview can also have an effect on the results (Jacobsen, 2005, 227). Jacobsen (2005) argue that in qualitative studies, the research process continue to evolve during the data collection, partially dependent on the researcher’s analysis and interpretations of the data. The interpretations made by the researcher are tied to the context in which the data collection takes place, and the research adapted to this specific context. Therefore, independent identical, collections of data based on the same research design becomes impossible; the research design and the collection of data becomes

dependent on when the study is conducted, and by whom the study is conducted (Grønmo, 2004, p.228). I preferred to meet the subject face to face as personal interviews provide data with higher reliability, and also create a greater sense of trust between the researcher and the respondent compared to for example a phone interview (Jacobsen, 2005, p. 143). It was only Hans that I ended up interviewing only over the telephone. The subjects from the Green Party were interviewed in their offices which could be considered their natural habitat, while the two other PR professionals were interviewed in cafes. Where, even as they both seemed

36 quite relaxed, the general noise and presence of other people can have had an effect on the results. At a later point, I also had to do a follow up interview with the external stakeholders in order to gather more data. These interviews were done over phone which also can have had an effect on the reliability and the quality of the data.

Another factor is whether or not the interview is surprising or planned (Jacobsen, 2005, p. 227). Before the interviews were conducted I had spoken to the subjects over the phone and I told them about my background and the subject of the study. I did not send them any of the questions I wanted to discuss prior to meeting them as I wanted the

information to flow freely without being predisposed to any reflection or potential agendas.

It can be argued that the reliability of the data is lower because the subjects were not allowed to think through the questions, but it can also be argued that the reliability is greater because the questions were not affected by any predetermined filters or agendas that could have had an effect on the data. Qualitative research can also have internal

reliability. This is when more than one observer agree about what they see or hear (Bryman, 2008, p. 376). This is not an option as I have conducted the study alone. However, Jacobsen (2009) argues that you can test internal reliability yourself by critically going through the most central phases of the research (p. 214). Jacobsen (2005, p. 217) suggests the following checkpoints: did you get the rights sources, did the sources provide correct information, when was the information gathered and how is the information extracted? For my internal stakeholder group I feel that I was able to use subject that was a good fit for what I set out to study. For the external stakeholder groups, part of the reason for why I chose to use public relations professionals was because I assumed that they could offer a professional view on communications and branding. I did, however, find that the subjects often

expressed their subjective opinions of a political party and at times failed to be professional in their views. Two of the subjects were also not very sympathetic to the Green Party, while the third, although he did not vote for them, expressed sympathy with their cause and what they are advocating. I also gathered the main bulk of my data early in the process, before having a complete overview of what data I was looking for. Because of this I later needed more information and this led to a follow up interview conducted over the telephone. When I did the telephone interview I noticed that as Jacobsen (2005, p. 143) writes, that the amount of data I was able to gather became smaller and the distance between myself and the respondent felt greater. All of the above factors may have influenced the reliability of

37 the results, in addition I found that the respondents struggled to stay objective, and that professional and personal views at times overlapped, when we discussed the Green Party.

This has a negative effect on the reliability of the analysis.

This can be because as much as this thesis is concerned with branding; politics is hard to leave completely out of the equation and politics can be a very passionate subject.

4.9.2 Transcription

The transcription of the interviews can also have an effect on the reliability of the results. To transcribe means to transform; to change form from one to another (Kvale & Brinkmann, 2009, p.187). Kvale and Brikmann write that the old saying, “translators are traitors”, also very well apples to transcription (p.187). When transcribing interviews, several problems arise. Things that are obvious in the context of the conversation can get lost: the tone of voice, irony and body language are examples. If two persons were to transcribe the same interview they would probably not end up with identical results (Kvale & Brinkmann, 2009, p.194).

In addition, a choice must be made weather to transcribe word-by-word and include things like “eeh”, pauses or laughter, or should the transcript be more formalized? I have chosen to formalize in order to make the analysis more clear and readable to the reader.

I also had an issue with cell phone interference noise on my recorded data that was done over the telephone, and while it sometimes could be difficult to hear what was being said, I feel that I was able to accurately hear and transcribe the interviews. Finally, I have conducted the entire study on my own, including the analysis of the interviews. The absence of other researchers could also therefore affect the reliability of the study.

4.9.3 Translation

The fact that my data material has been translated from Norwegian to English can also have an effect on the reliability of my findings. And as I do not believe it to be a problem in this thesis, sentences or meanings could get lost in translation and some words may not be accurate when translated.

38 4.9.4 Validity:

According to Jacobsen (2005) the validity of collected data can be examined from three different perspectives: Construct validity; internal validity and external validity.

Construct validity is about wheatear or not a study measures what the researcher sets out to measure (Jacobsen, 2005, p. 345). In my research process I have had a fairly open approach, relying on a semi structured interviews with largely open ended questions that encouraged the interviewee to answer what their understanding of a phenomenon is.

Questions can be asked if the subjects had the same understanding of the concepts we were discussing as I did. The validity of “operational definitions” is described by Lund & Haugen as how sure we can be that the terms and theories used in the data extraction have the same meaning for the interviewer and the respondent (Lund & Haugen, 2006, p.57). But as all my subjects have backgrounds and professions within communications or PR they all had a fair understanding of the concept of branding and related concepts. I also discussed the terms of branding and communication in the beginning of the conversations with the external

stakeholders in order to clarify that we were, in fact, on the same page.

Internal validity relates mainly to the issue of causality and whether or not there is a relationship between two or more variables (Bryman, 2008, p. 32). In this thesis I have not investigated the relationship between two factors so internal validity does not become relevant. However the findings can have external validity which is about wheatear or not the findings from a study can be generalized across social settings (Bryman, 2008, p. 376). I have used an intensive research design to gain in-depth knowledge from a small sample of

subjects. This makes it difficult to generalize on the results, but not impossible. Halkier (2007) writes that an intensive research design cannot fit a representative sample of a given population, but that an intensive research design still can be generalized in an analytical way (p. 113). Haliker use focus groups as an example and argue that validity in focus groups are about creating a systematic study, make it readable to others, argue that a focus group is a good methodological fit for the study, and argue in a way that is analytical and convincing to others (p. 122). Based on Halkers and his example of focus groups it can also be argued that the same would apply to qualitative interviews and that my study, based on qualitative interviews, can have validity at least in an analytical sense.

39