Methods
Context of immersion
The three participants were HSs of Mexican, highland Colombian, and highland Ecuadorian varieties of Spanish (further information regarding each participant’s background is discussed in the “Case studies” section). Participants spent one calendar year serving as humanitarian volunteers with a religiously affiliated NGO in low-income communities outside of Guayaquil, Ecuador, alongside 11 other volunteers who were L2Ls of Spanish and whose data is not considered in this chapter. The NGO that hosted the volunteers emphasized community involvement and prioritized relationship building between volunteers and Ecuadorian community members. It required volunteers to live in an intentional community2 (with five or six other volunteers as housemates) and hold job assignments in educational and social-service sectors where there was substantial interaction in Spanish between the volunteers and speakers of the local dialect. The program was not tied to an academic institution, and participants received no language instruction nor academic credit during the immersion.
Quantitative methods
To collect data on the production of coda ZsZ, sociolingüístic interviews in Spanish were conducted at six different intervals during the volunteer year. Interval 0 was completed one week prior to departure to Ecuador, with each additional interval another eight to ten weeks later. 1 conducted the first and last interviews, and a native speaker of Ecuadorian Spanish -a member of one of the communities in which the volunteers lived and worked who was trained in sociolingüístic interview methodology - carried out the interviews at intervals 1-4. Following sociolingüístic methodology (Tagliamonte, 2006), the interview was designed to encourage the use of the vernacular through its progression from general, impersonal, nonspecific topics and questions to more specific, personal ones (p. 38), asking participants to share what Labov (1984) refers to as optimal topics for eliciting natural speech: narratives of personal experiences. Participants were asked about their individual backgrounds, daily routines, jobs and communities, and events that had occurred during the two months between intervals. After collection of data, a total of 533 tokens of coda ZsZ across the three participants were extracted from the corpus. Using Praat (Boersma & Weenick, 2017), each token was analyzed aurally for weakening and then confirmed visually according to the amount of periodicity in the waveform and the presence of high-frequency frication on the spectrogram. Tokens were coded as either maintained ZsZ or weakened ZsZ (aspirated or deleted).3 Figure 4.1 gives an example of variable production of ZsZ. In the figure, which represents the phrase dos pàjaros, “two birds”, the first ZsZ demonstrates high-frequency energy (8000 Hz) and aperiodicity in the waveform, signaling maintenance, whereas the second ZsZ demonstrates lower-frequency energy and more periodicity in the wave form.
In addition to the elicitation of ZsZ, the Versant Spanish Test was administered to the participants at interval 0, just prior to their arrival in Ecuador, in order to approximate their preprogram proficiency level. Based on Levelt’s (1989) model of speech production, the test uses spoken prompts in Spanish of NSs from a variety of Spanish-speaking countries to elicit oral responses from students. Within five minutes of test completion, it provides students with an overall score and an equivalent score on other proficiency tests. Although the Versant test was not designed specifically for HSs, it was chosen for this study due to its nearly instantaneous, objective, and reliable results regarding students’ abilities to speak and understand spoken Spanish (Pearson Education, Inc., 2011).
Qualitative methods
According to Merriam (2009), case studies are one way to account for the complexity of processes in language learning in immersion contexts,
Figure 4.1 Gustavo’s variable /s/ production: /s/ in dos is maintained, whereas in pájaros it is aspirated.
because they allow for the investigation of multifaceted social situations that consist of numerous variables of potential importance in the understanding of a phenomenon. Case-study methodology used in previous study abroad research has been able to highlight reasons for individual differences among participants and illuminate why some students return with greater language gains than others (e.g., Isabelli-Garcia, 2006; Kinginger, 2008; Menard-Warwick & Palmer, 2012). The case-study data for the current study is drawn from the participants’ narratives provided during their sociolingüístic interviews. These narratives were analyzed using top-down analysis and deductive coding (LeCompte & Schensul, 1999), which implies choosing a set of concepts first and then sorting out the data to determine which of the concepts it fits best. The concepts of interest were identified based on previous research on CoPs (Eckert, 2006; Lave & Wenger, 1991) and the noticing hypothesis (Schmidt, 1990, 2001). In considering CoPs, I looked for participants’ narrations of whom they spent time with and descriptions of activities they participated in both inside and outside the volunteer house. For the noticing hypothesis, 1 sought examples of how the participants characterized the Spanish of the community as well as how they described their experience with language (both Spanish and English) while in Ecuador. Data was triangulated to determine how patterns emerging from the narratives of one participant related to those of the others and how this data could account for the variation found in the quantitative analysis.
