• No results found

3. Methodological and

3.3. Research design

3.3.3. Method for data analysis and theorization

This view will provide a template for the creation of a framework that can aid in the theoretical understanding of the factors that contribute to the destigmatization of menstrual products, which then in turn can be tested empirically and subsequently analyzed and revised (Yin, 1994). Consequently, the study pertains to a number of research domains, as previous research is sparse in the area of effects of stigma and destigmatization on fields; thus, a synthesis of relevant topics is applied. These include research on stigma, stigmatized products, the menstrual stigma, institutional fields, deinstitutionalization, social movements, institutional entrepreneurship, etc. First and foremost, I aim to contribute to the theoretical domains of product stigma and to a more limited extent, also institutional change.

becomes vital that my theorization and possible subsequent theory is dynamic and makes an effort to reflect how stigma changes over time, as well as how it affects a field and is affected by driving forces (Gioia, Corley & Hamilton, 2013).

Furthermore, considering other contextual aspects, such as my own bias as a feminist and a feminist researcher, because I believe they have different implications, the former is a part of me and something I cannot separate myself from, and that is most likely signaled solely by my having chosen to study the phenomenon in question. The latter has to do with how I view the researched world, and not just my personal world. This, I believe, I can separate more, especially as my education consists purely of business administration, wherein lies my basic theoretical standpoints. I manage these two roles, in the sense that I am open about my own feminism since I think it is quite obvious and attempting to hide it could signal insincerity and affect my relationships with my respondents and thus, my findings negatively.

Pertaining feminism, I made a point of conducting the interviews in a way that did not reveal my focus on stigma or feminist stance on the researched phenomenon until either a subsequent interview, or late in the interview. This hopefully allowed respondents to answer more freely, without the input of my bias or my “loose frame” initially, and later gave them the opportunity to respond to my more ‘suggestive’ questions. These include asking whether there are any experiences that stick out as uncomfortable or noteworthy concerning the social aspects of working with menstrual products, which proved to be very fruitful. On the other hand, I also believe that my feminism can be positive in the sense that it can allow respondents to highlight potential feminist issues, which are of course closely linked to the menstrual stigma, without feeling as though they may not be taken seriously, a common consequence both for feminists and feminine issues.

Without my feminism being noticeable, these aspects might not be articulated in interviews at all.

Furthermore, it may be noteworthy that my personal biases also bring with them underlying assumptions that are not directly connected to feminism. One such underlying assumption is that products used in intimate contact with the body, and specifically genitals and mucous membranes, are generally heavily regulated.

In absence of governmental or regional regulations, however, standards are often used to govern the safety of products. It is this underlying assumption that guides my very first study question, as well as subsequent ones.

Finally, paying attention to the context is especially important when researching complex phenomena such as processes. This is because they 1) deal with events, or “conceptual entities that researchers are less familiar with,” 2) “often involve multiple levels and units of analysis whose boundaries are ambiguous,” 3) are embedded temporally, often varying “in terms of precision, duration and relevance,” and 4) deal with data that tends “to be eclectic, drawing in phenomena such as changing relationships, thoughts, feelings and interpretations” (Langley, 1999, p.692). Addressing these difficulties is not a simple task; nonetheless, selecting an appropriate strategy for making sense of process data is imperative.

Langley (1999, p.700) suggests seven possible strategies, one of which signifies the grounded theory strategy. This strategy is especially beneficial when dealing with particular kinds of processes such as when exploring “the interpretations and emotions of different individuals or groups living through the same processes.”

She further argues that the grounded theory strategy provides the ability to analyze the data closely, while simultaneously developing dense theories.

3.3.3.2. Step 2 – incident selection

In practice, the research process was slightly more iterative than presented here.

The incident selection was done prior to the consideration of the context;

however, the analytical process is performed in this order, so I feel it makes sense to present it in this way as well. The next stage in the process was to narrow my gaze to pivotal incidents, events, or utterings highlighted in the empirics that are interesting from an empirical or theoretical perspective. Empirically interesting could mean that the information or event might be counterintuitive, which implies that there could be contradictory logics in place, or simply that respondents themselves point out certain things that bear significance to the grand narrative (Charmaz, 2006). Theoretically interesting could mean that it has been mentioned in theory I have read, it has not been mentioned and is repeated many times, sticks out as something unexpected, contradictory, etc. (Ryan & Bernard, 2003).

What is considered (un)expected, however, is very much dependent on one’s worldview as a researcher. In correspondence with the stance that I am taking on my research question and the empirical phenomenon, it is natural that I take a feminist stance. Thus, one category of what comes to my attention as interesting is that which is not in line with the feminist agenda; in this case, to further the destigmatization of menstrual products. I did, however, also try to set this agenda

aside in an attempt to be reflexive when looking for other possible interesting explanations, findings, and categories in my data, as suggested by Charmaz (2006) when employing sensitizing concepts. I practiced theoretical sampling while interviewing, hoping to go “beyond common sense tales and subsequent obvious, low-level categories that add nothing new,” as described by Charmaz (2006, p.33), which in my case could be taking what respondents say at face value and accepting that there was no budget.

3.3.3.3. Step 3 – initial coding

Next, I used a version of Charmaz’s grounded theory coding, both initial (step 3) and focused (step 4), in order to “sort data to begin an analytic accounting of them” (Charmaz, 2006, p.45). In the previous step, I selected interesting quotes and events. In this step, however, I asked material questions to determine what theoretical categories they may indicate, or “expressions of a theme [that], of course, aid us in discovering it,” as Morris Opler puts it (cited in Ryan & Bernard, 2003, p.86). Such questions included: what views and values are represented? Are there any underlying meanings? If so, what are they? How is the respondent making sense of the phenomena at hand, such as the menstrual stigma, the destigmatization of menstruation, the standardization of menstrual products, the menstrual product field? Whose point of view is this? Since I am researching a topic that is rarely discussed in the open, it is very likely that respondents do not speak candidly about them; thus, I also look for signs of shyness/shame/embarrassment/discomfort, etc., the use of euphemisms, hesitant speech or simply an absence of certain words, such as a respondents’ avoidance to say the word menstruation.

Since I have a sort of master code that is stigma, there is a risk that I will see stigma as the answer to every question I pose. I manage this potential issue by continuously questioning whether there can be other explanations and meanings that underlie the data. That way I could find some sort of extreme scenario where lots of things are about stigma, and one that is more nuanced, or perhaps completely denies the stigma. I then followed some sort of idea about what seems reasonable. That brings us to the inevitable question of reason. Is the most reasonable explanation that which coincides most with what the respondent means or my analysis of what they mean and why they say it? While there is always a risk(?) of over analysis, I chose to take my queue from Becker (1998, p.118) who states that “social scientists will be led astray if they accept the lies organizations

tell about themselves. If, instead, they look for places where the stories told do not hold up, for the events and activities those speaking for the organization ignore, cover up, or explain away, they will find a wealth of things to include in the body of material from which they construct their definitions.”

3.3.3.4. Step 4 – focused coding

Regarding this point, I came up with multiple codes without attempting to refine them; thus, I sorted them into more abstract categories, bunching together those that are significantly similar in some way (Gioia, Corley & Hamilton, 2013). This can also be described as the part where facts of an investigation are turned into theory through a process of theorization. More specifically, according to Van Maanen (1979), the first order codes do not speak for themselves. They need to be organized into patterns through concepts, which are generated by the researcher in that process. Here, it is not only interesting to group codes together, but it is also vital to pay attention to contradicting concepts; it is then that we can really say something new about the field (Maanen, 1979). Finally, I tried to identify which codes seemed most fruitful to follow up and compare to greater parts of my data. “While engaging in focused coding, we select what seem to be the most useful initial codes and test them against extensive data. Throughout the process, we compare data with data and then data with codes” (Charmaz, 2006).

In order to facilitate this process, I used the qualitative data coding software, NVivo, in which I created a data structure to more clearly visualize and sort my codes (Gioia, Corley & Hamilton, 2013). The smallest boxes represent the initial codes I deciphered from my interviews. I then grouped them together into higher level category codes, which are the focused codes in the larger boxes. Each row of larger boxes was finally categorized into mechanisms. The first row represents the factors under the reclassifying mechanism, the second represents framing, and the third represents claiming agency.

Figure 3: Data structure of focused coding