Projects

Title: Foreign Language (FL) students’ perception of spoken language variation

Collaborator: Tania Ferronato
The primary purpose of the present study is to explore if FL learners at the intermediate and advanced level are aware of different accents and spoken variations in the target language and if so, what their attitude towards them is and what their learning preferences in terms of motivation are, for example. Additionally, the study aims to measure intelligibility, comprehensibility and acceptability of different FL accents in order to find out if there are FL language accents that should be prioritized in the FL class according to students’ needs and if so, which one(s) they are. Pedagogical implications will be drawn not only for FL teachers, but also for textbook editors and FL curriculum designers.

Title: Impact of accent familiarity training on listener processing of L2-accented speech

Collaborator: Okim Kang
In communication between first (L1) and second (L2) language users, the default has long been for L2 speakers to assume most, if not all, responsibility for adjusting their speech to accommodate their interlocutor. In order to foster greater equity in linguistically diverse communication, L1 listeners should be trained to develop the appropriate communicative tools to be active partners with their L2 interlocutors (Subtirelu & Lindemann, 2016). Implicit, explicit, and structured intergroup contact training approaches have demonstrated success in improving attitudes toward L2-accented speech (e.g., Kang et al., 2015). However, little is known about how such trainings impact listeners’ cognitive processing of L2-accented speech.

The current study examined the effectiveness of exposure-based online listener training by measuring listeners’ processing and perception of L2-accented speech with speakers from three different L1s (Mandarin, Russian, US English). Seventy-one L1 English listeners were randomly assigned to a two-week training program focusing on either a) L2-accented English varieties (i.e., an accent familiarity experimental group, n = 41) or b) traditional L1 English listening comprehension tasks (i.e., an L1 English-only comparison group, n = 30). Before and after training, listeners completed psycholinguistic measures of response time, processing stability, and response accuracy, as well as speech ratings for comprehensibility and accentedness. Listeners also commented on perceived training effectiveness. Quantitative data were analyzed in a series of linear mixed-effects models, while qualitative data were thematically coded.

Quantitative findings indicated statistically significant improvements for both groups’ cognitive processing as a result of training, while experimental participants had an added benefit in accentedness ratings. Qualitative findings highlighted the importance of training for attentional allocation. These results inform future approaches to listener training and integration of psycholinguistic methods in listener training research. They also offer important pedagogical implications within and beyond higher education as universities prepare students to enter a globalized workforce.

Title: “Communication entails both listening and speaking”: Willingness to listen as part of willingness to communicate

In equitable oral communication among first (L1) and second (L2) language users, listeners can and should share responsibility for effective communication, rather than relying on speakers to adjust their speech, which may disproportionately affect L2 speakers (Derwing & Munro, 2009). When training listeners as active communication partners, it is important to understand how listener factors, including their willingness to communicate (WTC), impact their perception and comprehension of L2-accented speech (Yaw, 2022). A potential limitation of using WTC measures for listener training, however, is their heavy focus on language production (i.e., speaking) as evidence of WTC while neglecting language reception (i.e., listening). In other words, the extent to which WTC instruments capture an interlocutor’s willingness to listen (WTL) in communicative interactions remains underexplored.

The current study examined the relationship between listeners’ WTC and WTL across situational contexts (e.g., public speaking, conversations) and receiver/interlocutor types (e.g., strangers, friends), along with the predictive role of Big Five personality traits on both WTC and WTL. One hundred fifty-two respondents completed a) a WTC measure asking them what percentage of time they would choose to communicate in English in 20 different situations (McCroskey, 1992), b) a researcher-adapted WTL version of this measure, and c) a personality trait inventory. Participants also commented on these measures. Quantitative data were analyzed with correlation and linear mixed-effects models, while qualitative data were coded thematically.

Quantitative findings demonstrated a weak relationship between WTC and WTL, with approximately 43% of respondents indicating that they imagined communication in the WTC measure involving speaking only. Additionally, responses showed greater WTL than WTC across all context and receiver types. Qualitative findings highlighted limitations to using questionnaires to measure listener behaviors. Results offer implications for how listening fits into WTC, and inform future approaches to measuring WTL and accounting for this in listener training programs.

Title: Exploring listener individual differences as predictors of cognitive processing and perception of L2-accented speech

Collaborator: Okim Kang
Training L1 listeners to be prepared for interactions with L2 interlocutors can complement the communicative benefits of intelligibility-based approaches to L2 pronunciation training, encouraging listeners to share responsibility for communicative success (Derwing & Munro, 2009). Indeed, previous approaches to listener training, including implicit, explicit, and structured intergroup contact methods, have demonstrated attitudinal benefits among listener trainees (e.g., Derwing et al., 2002; Kang et al., 2015; Lindemann et al., 2016). One area that remains underexplored, however, is the impact of listener individual differences on the training process, which has implications for both personalization of training design and interpretation of differential training outcomes.

The present study examined four areas of listener individual difference as potential predictors of listeners’ processing and perception of L2-accented speech. As part of a larger listener training study, 71 L1 English listeners responded to instruments measuring accent familiarity (adapted from Ockey & French, 2016), personality (Goldberg, 1992), linguistic stereotyping (Zahn & Hopper, 1985), and willingness to communicate/listen (McCroskey, 1992; Gallagher & Robins, 2015). They also rated speech produced by speakers from three different L1 backgrounds (Mandarin, Russian, US English) for comprehensibility and strength of accent, and completed psycholinguistic measures of response time, processing stability, and response accuracy. Individual difference measures were treated as fixed factors in a series of linear mixed-effects models with processing and perception measures as the dependent variables.

Results showed accent familiarity and linguistic stereotyping as statistically significant predictors of both perception and processing, while emotional stability emerged as a predictor of comprehensibility and intellect of response accuracy. Particularly promising is the overall impact of accent familiarity, as this can be built through exposure and training. These findings inform the development of more targeted listener training approaches that account for listeners’ individual differences and offer recommendations for future research in this area.

Title: “We’re all a bit guilty of this, aren’t we?”: Investigating researcher perceptions of Questionable Research Practices

Collaborators: Scott Sterling, Luke Plonsky, Tove Larsson, Merja Kytö
Applied linguistics tends to not focus on research ethics beyond the needs associated with ethical review boards (Sterling & Gass, 2017; Yaw et al., 2023; Wood et al., under review). This reduced focus is problematic given that every step of the research process contains decision points that impact the quality and ethicality of research. Often, the decisions that scholars make fall between actions deemed fully ethical and those that clearly violate ethical norms (Artino et al, 2018; Fanelli, 2009). Surveys have found that an overwhelming majority of applied linguists have engaged in questionable research practices (QRPs) in their careers (Isbell et al., 2022; Larsson et al., in press). However, the degree of (un)ethicality of QRPs may be dependent on the context in which they are made. For example, space requirements in journals might dictate the level of methodological discussion allowed. That is, QRPs by their nature exist in a grey zone.

In a prior research phase, we surveyed 230 scholars in the quantitative humanities, including applied linguistics, to understand how scholars rated a list of QRP items in terms of their frequency and severity. The current qualitative project investigates the feedback respondents left on the survey, including their overall thoughts and reactions to each QRP item. Using a bottom-up coding scheme, several themes emerged. These themes related to justifying prior research decisions, discussing contexts in which a decision is more or less questionable, and commenting on specific items in a way that indicated that scholars potentially did not understand, or may have used different definitions for, terms in the items (e.g., cherry-picking, salami publications, p-hacking, and HARKing). Results from this study show how context specific QRPs are and thus emphasize the need to train scholars in methods of ethical and methodological thinking that go beyond simple procedural ethical needs.