3.3 Sampling and recruitment
3.3.2 Research sample and sampling techniques
The exploratory nature of this dissertation and the lack of previous research projects about the same phenomenon called for a sample where the phenomenon is represented with as many facets as possible. In order to achieve this, Merkens (2004, 167) cites Patton (1990) to propose that sampling should cover extreme cases, typical cases, and critical cases. In phenomenological research, sampling is influenced by the two crucial requirements that participants must meet: 1) they have to have had experienced the phenomenon and 2) they have to be able to provide full and sensitive descriptions of the phenomenon (Morgan 2012, 12; Polkinghorne 1989, 47). It is also necessary to be aware that the diversity among individuals can potentially become a difficulty in phenomenological projects. Although Creswell (2008, 122) notes this problem, I welcomed this challenge and embraced its role at problematizing and enriching the fixed and stereotyped notions that surround the comics reading experience.
My final sample consists of seventeen participants, nine female and eight male, from seventeen to twenty five years of age. In this section I offer some basic
information about the participants. For further background information, in the analysis section I have included a series of personal portraits that focus on their relationship with reading, and comics reading in particular. The following sections describe and reflect the way I carried out the process of selection, access, and recruitment.
Pseudonym Age Recruitment site Familiarity with the medium
Date
Devi 19 MIT - Talk Beginner 05/03/2010
HunterS 20 MIT - Snowball Beginner 10/03/2010
Preacher 18 MIT - Talk Intermediate 17/03/2010
Oracle 23 MLIS - Talk Intermediate/ Expert
09/04/2010
Shade 24 Snowball Expert 09/04/2010
Lorraine 19 MIT - Snowball Beginner 27/04/2010
Walker 23 MIT - Snowball Beginner 21/04/2010
Shalmanaser 20 Library - Random Intermediate 28/04/2010
Jacob 17 Librarian Beginner 01/03/2011
Templesmith 20 Heroes Intermediate/ Expert
05/04/2011
Marian 24 Library - Random Intermediate/ Expert
18/04/2011
Kalo 24 MLIS - Talk Intermediate 11/05/2011
Daniel Feireday 23 Comics store staff Expert 06/09/2011
Promethea 21 Public market - Random Beginner 08/09/2011
Selina 24 MLIS - Random Beginner/Inter mediate
14/10/2011
Baa 17 Library - Random Intermediate 12/12/2011
Table 4: Recruitment information
Although my personal interest and experience with youth were highly influential for deciding the parameters of this research, my main criteria reflected two different but related needs. First, is the need for more information about the experience of reading from young readers themselves that was indicated and justified in the Literature Review. Second, is the general need for research about youth, especially research involves and actively listens to youth. In response to these needs I established the age range from fifteen to twenty five years of age. Discussing the difficulty of defining youth, Heath et al. agree with the United Nations; definition of youth as aged fifteen to twenty four years; this same definition also guided my choice (UN 2012; Heath et al. 2009, 4). Finally, my methodological choice also helped determine the participants’ age range since I was looking for youth who could share their experiences with reading comics and engage in semi-structured, and in many cases, lengthy
interviews.
From the planning stages of the project, I sought to establish a sample that reflected different experiences, including beginning readers, occasional but interested readers, and expert readers. As I will explain later in this section, the recruitment tools I developed along with the support and commitment of gatekeepers and participants were crucial to achieve this objective. Although I did not include any reference to gender-related aspects in my research, I was aware that it might become a factor to
study. The balanced gender composition of the sample is an anomaly in average comics readership where male readers and producers are often more predominant. However, this situation could be explained because of the characteristics sought in the sample. Female readership is more evident when reading tastes are expanded to include alternative comics or manga (Lopes 2009, 156; Brown 2001, 98; Pustz 1999, 13, 86). Although I was unable to identify studies that look specifically at the influence of gender in the recruitment process, feminist researchers have discussed the influence of gender, especially in the interviewing process (e.g., Broom, Hand and Tovey 2009; Finch 1993).
In order to construct a rich sample, I planned to combine three sampling techniques: purposeful sampling, theoretical sampling, and snowball sampling. These techniques complement each other in the process of selecting the cases that offer the richest experience. Examining the differences between purposeful sampling and theoretical sampling, Coyne resolves that in qualitative research all sampling is
purposeful sampling because “the sample is always intentionally selected according to the needs of the study” (1997, 629). Since I was looking for participants who share an experience, purposeful sampling supported my search for “people who offer a picture of what is like to be themselves as they make sense of an important experience” (Cohen, Kahn and Steeves 2000, 50). At the same time, theoretical sampling is a responsive approach rooted in grounded theory that reinforces discovery (Corbin and Strauss 2008, 143-157). In an iterative process, sampling is intrinsically connected to the analysis of the data, permitting the inclusion of new or relevant concepts in the search for new participants (Corbin and Strauss 2008, 145). It is also a type of sampling recommended for exploratory projects (Merkens 2004, 169). One of the most
important contributions of theoretical sampling to my data collection hypotheses was the active exclusion of readers who were just focused on superhero comics. The discussion in the Literature Review explains the need to research reading comics beyond the fan experience. In relation to this, I must recognize the success of the recruitment tools and the works of gatekeepers and participants themselves since they
acted as successful filters. My research project also offers another salient example of how crucial theoretical sampling can be for the development of a rich sample,
especially as a way of counteracting some of the weaknesses of snowball sampling, especially the potential of finishing with a heterogenic sample. After interviewing seven participants, five of whom were undergraduate students in the same program, I
realized from the data analysis that their experiences were rather similar, although clearly contrasted with the experiences of my other two participants. At that moment I decided that I would not accept any more participants with that background, with the exception of Daniel Feireday who was also a staff member at a comics store.
Snowball sampling is a technique that facilitates access to closed or marginalized groups. In my case, it worked less effectively among participants; however, attendees to my presentations and gatekeepers became very active at connecting me with possible participants. With the participants’ verbal permission, I informed the gatekeepers about the adequacy of their suggestions, thus helping their own selective process. A weakness of this technique is that it privileges clusters and participants who might share many characteristics. However, this effect was mitigated through the implementation of two measures. One is connected to the recruiting sites of the project; I purposefully planned to recruit in three different sites to try to access a diverse population. This strategy is supported by Hays who recommends initiating the snowball sample within several groups of potential participants rather than just
focusing on one group (2011, 170). As part of my interview routine, I invited participants to share the research and their experience with anyone who they thought might be appropriate. I always had flyers or professional cards to give them. In connection with the idea of participants opting into the research and respecting confidentiality, I waited for possible participants to contact me. To counteract the difficulties that arose from adopting this position, I tried to make my presence noticeable at the different sites with regular visits and interactions with gatekeepers, as I also tried to make myself available to past participants for questions or recommendations. In relation to snowball
the comics store staff, who went well beyond allowing me to access their work spaces; they also actively publicized my research and constantly referred potential participants.
Creswell comments on the general difficulty of defining the sample size for qualitative study (2007, 126-128). The saturation point guided the size of my sample. Saturation is often explained as the moments where “all concepts are well defined and explained” (Morgan, 2012, 12-13; Corbin and Strauss 2008, 145; Lincoln and Guba 1985, 202). Kvale and Brinkmann connect the number of participants to the purpose of the study and identify the points of saturation with the formula: 15 +/- 10 (2009, 113). Morgan explains that usually a group of ten to twelve participants produces saturation and that the researcher should reflect about the number of participants in relation to finding “an adequate variety of perspectives” (2012, 13). Seidman highlights the complexity of determining a number of participants for phenomenological
interviewing; he mentions both sufficiency and saturation as criteria to determine how many are “enough” (2006, 54-55). The only number provided in his discussion is twenty- five (as determined by Douglas as cited by Seidman, 2006, 55). My dissertation work maps onto Morgan’s formula; I started to detect saturation around my thirteenth interview. Morgan advises gathering at least two more accounts when the researcher suspects saturation has been reached. I followed her recommendation, especially because the participants who contacted me could potentially increase the diversity of the experience I was researching.
Recruiting was the most difficult and time-consuming part of the research project. Heath et al. comment on the difficulty of convincing young people to participate in research projects. This difficulty has been linked to “long term social trends such as social atomisation, increasing personal contact from a range of research organizations (social survey, market research and direct marketing organizations), and a general decline in people’s willingness to participate for the greater good rather than for individual gain” (2009, 142). Despite this, I managed to locate and work with a group of teenagers and young adults who were highly committed to sharing their experience
of reading; they reassured me of the value and relevance of the project and, in some cases, found the interview process rewarding for their own personal growth.