Models of L2 pronunciation learning hypothesize that accurate speech perception promotes accurate speech production. This claim can be evaluated longitudinally by examining the extent to which changes in stop consonant perception predict changes in stop consonant production. Taking a time-sensitive view of the perception-production link, this study used longitudinal data to analyze perception as a time-varying predictor of production accuracy. Mixed-effects models were fit to oddity, delayed word repetition, and picture description tasks to examine how participants’ perception and production changed over time. Oddity task perception data were then decomposed into their between- and within-subjects components and integrated into the delayed repetition and picture description production models. Surprisingly, only the between-subjects predictors reached significance, and the strength of the perception-production link varied across production tasks and target phones. The methods used have implications for future research on the perception-production link.

Longitudinal research on second language (L2) sound learning demonstrates that speakers’ production of challenging L2 sounds can improve in the absence of targeted instruction. Although the growing body of longitudinal work on this topic provides insight into the rate and shape of development, the factors that shape L2 phonetic production in spontaneous speech are not yet well understood. The data described here are part of a longitudinal data set collected from native English speakers enrolled in introductory-level Spanish language courses. For the current analysis, 16 participants were selected based on the availability of data coinciding with the beginning and end of the first semester and end of the second semester. Participants completed a simplified picture description task. Voice Onset Time (VOT) was annotated and measured for all instances of Spanish /p, t, k/ in the speech samples. Participants’ instructors were recorded during two class periods. Recordings were transcribed and a frequency measure was calculated based on the resulting teacher speech data. After controlling for a range of linguistic factors known to affect VOT, modeling demonstrated that time and frequency were not significantly related to VOT.

Mixed-effects modeling is a multidimensional statistical analysis capable of modeling complex relationships between predictor and outcome variables while accounting for random variance in various dimensions of the data. Although this technique is gaining popularity in applied linguistics research, learning how to model, and how to do so in R, can be intimidating. This guide provides an introduction to fitting mixed-effects models in R (Version 3.5.3) using RStudio. It includes a written introduction describing the modeling process, a video tutorial that focuses on getting started in RStudio, a sample data set, and an R script containing code to analyze the data. By the end of this introduction, researchers should have developed a basic understanding of the modeling process and should be able to (1) read data into R and inspect its structure, (2) create a series of plots to visualize trends and/or primary variables, and (3) fit and evaluate models.

Comprehensibility, or ease of understanding, has emerged as an important construct in second language (L2) speech research. Many studies have examined the linguistic features that underlie this construct, but there has been limited work on behavioral and affective predictors. The goal of this study was therefore to examine the extent to which anxiety and collaborativeness predict interlocutors’ perception of one another’s comprehensibility. Twenty dyads of L2 English speakers completed three interactive tasks. Throughout their 17-minute interaction, they were periodically asked to evaluate their own and each other’s anxiety and collaborativeness and to rate their partner’s comprehensibility using 100-point scales. Mixed-effects models showed that partner anxiety and collaborativeness predicted comprehensibility, but the relative importance of each predictor depended on the nature of the task. Self-collaborativeness was also related to comprehensibility. These findings suggest that comprehensibility is sensitive to a range of linguistic, behavioral, and affective influences.

This study examined longitudinal changes in second language (L2) interlocutors’ mutual comprehensibility ratings (perceived ease of understanding speech), targeting comprehensibility as a dynamic, time-varying, inter action-centered construct. In a repeated-measures, within-participants design, 20 pairs of L2 English university students from different language backgrounds engaged in three collaborative and interactive tasks over 17 minutes, rating their partner’s comprehensibility at 2–3 minute intervals using 100-millimeter scales (seven ratings per interlocutor). Mutual comprehensibility ratings followed a U-shaped function over time, with comprehensibility (initially perceived to be high) being affected by task complexity but then reaching high levels by the end of the interaction. The interlocutors’ ratings also became more similar to each other early on and remained aligned throughout the interaction. These findings demonstrate the dynamic nature of comprehensibility between L2 interlocutors and suggest the need for L2 comprehensibility research to account for the effects of interaction, task, and time on comprehensibility measurements.

Researchers have increasingly turned to Amazon Mechanical Turk (AMT) to crowdsource speech data, predominantly in English. Although AMT and similar platforms are well positioned to enhance the state of the art in L2 research, it is unclear if crowdsourced L2 speech ratings are reliable, particularly in languages other than English. The present study describes the development and deployment of an AMT task to crowdsource comprehensibility, fluency, and accentedness ratings for L2 Spanish speech samples. Fifty-four AMT workers who were native Spanish speakers from 11 countries participated in the ratings. Intraclass correlation coefficients were used to estimate group-level interrater reliability, and Rasch analyses were undertaken to examine individual differences in rater severity and fit. Excellent reliability was observed for the comprehensibility and fluency ratings, but indices were slightly lower for accentedness, leading to recommendations to improve the task for future data collection.

This study investigated relationships among intelligibility, comprehensibility, and accentedness in the speech of L2 learners of Spanish who completed a prompted response speaking task. Thirty native Spanish listeners from Spain were recruited through Amazon Mechanical Turk to transcribe and rate extracted utterances, which were also coded for grammatical and phonemic errors, and speaking rate. Descriptively, although most utterances were intelligible, their comprehensibility and accentedness varied substantially. Mixed-effects modeling showed that comprehensibility was significantly associated with intelligibility whereas accentedness was not. Additionally, phonemic and grammatical errors were significant predictors of intelligibility and comprehensibility, but only phonemic errors were significantly related to accentedness. Overall, phonemic errors displayed a stronger negative association with the listener-based dimensions than grammatical errors. These findings suggest that English-speaking learners of Spanish are not as uniformly intelligible and comprehensible as FL instructors might believe and shed light on relationships among speech constructs in an L2 other than English.

This study took a dynamic approach to second language (L2) comprehensibility, examining how listeners construct comprehensibility profiles for L2 Spanish speakers during the listening task and what features enhance or diminish comprehensibility. Listeners were 24 native Spanish speakers who evaluated 2–5 minute audio clips recorded by three university-level L2 Spanish speakers responding to two prompts. Listeners rated comprehensibility dynamically, using Idiodynamic Software to upgrade or downgrade comprehensibility over the course of the listening task. Dynamic ratings for one audio clip were video-captured for stimulated recall, and listeners were interviewed to understand which aspects of L2 speech were associated with enhanced versus diminished comprehensibility. Results indicated that clips that were downgraded more often received lower global ratings but upgrading was not associated with higher ratings. Certain problematic features and individual episodes caused listeners’ impressions to converge, though substantial individual variation among listeners was evident.