Research on mobile-assisted language learning (MALL) has revealed that high rates of attrition among users can undermine the potential benefits of this learning method. To explore this issue, we surveyed 3,670 adult MALL users based on the Unified Theory of Acceptance and Use of Technology (UTAUT) and also conducted an in-depth analysis of their historical app usage data. The results of hierarchical k-means cluster analysis and recurrent event survival analysis revealed three major findings. First, three distinct profiles of learners were characterized by different MALL acceptance and engagement experiences. Second, those with greater MALL acceptance displayed more intense, frequent, and durable app usage (behavioral engagement). Lastly, high levels of MALL acceptance were associated with more frequent pauses in app usage but also (a) longer active usage, (b) shorter breaks before returning to the app, and, ultimately, (c) fewer dropouts. We argue that persistence is a multidimensional process involving cyclical phases of engagement, disengagement, dormancy, and reengagement, with each aspect, like intensity, frequency, and duration, building up cumulatively over time. Implications for promoting persistent MALL engagement are discussed.

This study was an approximate replication of Rothman (2011),examining the determiner phrase syntax of a large sample (n = 211) of L3 learners of Portuguese who spoke English and Spanish. Rothman (2011) investigated whether L3 Italian or Brazilian Portuguese speakers are differently impacted by another known Romance Language, if it was their L1 or L2. The original study concluded that groups did not perform differently on experimental tasks on the basis of a null effect, and that the typological similarity of Spanish, Portuguese, or Italian predicts transfer in the initial stages of L3 acquisition. The present replication recreated all materials, which were unavailable, and examined the same population and questions. However, rather than examining L3 Italian and L3 Brazilian Portuguese, the present work maintained a constant L3 Portuguese. Learners were divided into two groups in a mirror-image design (n = 96 L1 English-L2 Spanish, n = 115 L1 Spanish-L2 English), and data were collected online. Like the original study, there was no main effect of group in any of the two-way analyses of variance. However, results show that it should not be assumed that experimental groups behave equivalently based on a null effect: Of the four total post hoc tests of equivalence, only two were significant when the equivalence bounds were set at a small effect size (d = $ \pm $ .4). Ultimately, it is argued that determining the smallest effect size of interest and subsequent equivalence testing are necessary to answer key questions in the field of L3 acquisition.

This study closely replicates the analyses of the third research question in Römer and Berger (2019), which reported that the associations between verbs and verb argument constructions (VACs) used by German and Spanish learners of English move closer to a native usage norm as the learners’ proficiency increases. This study conducted the same correlation analyses from the original study but with a substantially expanded version of the learner corpus used therein. Additionally, we conducted zero-inflated negative binomial analyses to estimate the relationship between the frequencies of verb-VAC combinations in the British National Corpus (BNC) and in the learner subcorpora representing different proficiency levels. Our findings were consistent with the original study in showing significant positive correlations between frequencies of the verb-VAC combinations in the BNC and in the learner subcorpora but further revealed the potential effect of topic on the learners’ VAC usage. Implications for future studies are discussed.

Darcy et al. (2016) examined the relationship between language abilities and general cognition, or specifically, how inhibitory control might relate to L2 speech perception and production. Given that their findings unexpectedly indicated a stronger relationship between inhibitory control and perception in comparison to inhibitory control and production, and because inhibitory control was measured using a single, retrieval-induced inhibition task, the current study is a close replication with the inclusion of two additional tasks of intentional inhibition: the Stroop task and the Simon task. A comparison of the descriptive statistics for the tests of phonological processing and retrieval-induced inhibition between the initial study and current replication indicated negligible differences between the two participant samples. However, results of the partial correlation analyses in the current replication did not indicate clear relationships between phonological variables and inhibitory control. Possible explanations for the different patterning of results and implications for future replications are discussed.

This paper reports a complete secondary analysis of Jeon and Yamashita’s (2022) systematic review to build the second language (L2) model of the simple view of reading (SVR). The same meta-analytic methodologies were maintained, with the exception of applying meta-analytic structural equation modeling (MASEM). This study successfully replicated some of the aggregated correlations but not others, owing to (a) the recoding of the original raw data to recreate a dataset and (b) the motivated change in sample selection from a longitudinal study for MASEM. The MASEM results extended previous findings that L2 comprehension skills contribute more to L2 reading comprehension than L2 decoding skills, and together explain a large amount of variance in L2 reading comprehension. The SVR model with metalinguistic skills showed their contribution to L2 decoding and comprehension skills, but no direct impact on L2 reading comprehension, supporting the parsimonious structure of SVR in L2.

This study is a close replication of Kanwit and Geeslin (2014), a variationist investigation of the interpretation of verbal moods in adverbial clauses in Spanish. Whereas the first language (L1) of the second-language participants in the initial study was English, we explore whether Kanwit and Geeslin’s results extend to other L1 populations—Swedish and French learners of Spanish. Participants in the replication study completed the same interpretation task and grammar test as those in the initial study. Results showed, for example, that multiple factors influenced their variable interpretation of verbal moods, there was evidence of change between course levels, and there were certain differences in interpretation between the French and Swedish groups. This study contributes to knowledge about the interpretation of a variable structure by enhancing the confirmatory power of some of the initial study’s findings, while also suggesting that the learners’ L1 leads to diverging findings.

Boduch-Grabka and Lev-Ari (2021) showed that so-called “native” British-English speakers judged statements produced by Polish-accented English speakers as less likely to be true than statements produced by “native” speakers and that prior exposure to Polish-accented English speech modulates this effect. Given the real-world consequences of this study, as well as our commitment to assessing and mitigating linguistic biases, we conducted a close replication, extending the work by collecting additional information about participants’ explicit biases towards Polish migrants in the UK. We did not reproduce the original pattern of results, observing no effect of speaker accent or exposure on comprehension or veracity. In addition, the measure of explicit bias did not predict differential veracity ratings for Polish- and British-accented speech. Although the current pattern of results differs from that of the original study, our finding that neither comprehension nor veracity were impacted by accent or exposure condition is not inconsistent with the Boduch-Grabka and Lev-Ari (2021) processing difficulty account of the accent-based veracity judgment effect. We explore possible explanations for the lack of replication and future directions for this work.

The present study revisits the differential roles of form, meaning, and use aspects of phonological vocabulary knowledge in L2 listening proficiency. A total of 126 Japanese English-as-a-foreign-language listeners completed the TOEIC Listening test, working memory and auditory processing tests, the Metacognitive Awareness Listening Questionnaire, and several tasks designed to tap into three broad aspects of phonological vocabulary knowledge: (1) the ability to access phonological forms without any orthographic cues (phonologization), (2) the ability to recognize words regardless of the talker (generalization), and (3) the ability to determine the semantic and collocational appropriateness of words in global contexts in a fast and stable manner (automatization). Whereas the perceptual, cognitive, and metacognitive variables made relatively small contributions to L2 listening proficiency (0.4%–21.3%), the vocabulary factors explained a large amount of the variance (77.6%) in the full regression model (R2 = .507). These large lexical effects uniquely derived from the three different aspects of phonological vocabulary knowledge—automatization (55.3%), phonologization (20.8%), and generalization (1.5%). The findings suggest that successful L2 listening skill acquisition draws on not only various levels of phonological form-meaning mapping (phonologization, generalization) but also the spontaneous and robust retrieval of such vocabulary knowledge in relation to surrounding words (automatization).

This article proposes a conceptual and measurement framework for postpubertal, L2 speech learning aptitude that is centered around domain-general auditory processing (i.e., representing spectral and temporal characteristics of sounds). To this end, we examine the construct and reliability of a battery of auditory processing tests by presenting the results of an empirical study wherein 100 participants took the tests twice. The findings showed (a) that the tasks tapped into multiple dimensions of auditory processing ability (e.g., perceptual acuity, audio-motor integration); and (b) that test-retest reliability ranged from “fair” to “excellent” (ICC (2,2) = .4–.8). Last, we introduce an offline version of the tests (deposited at L2 Speech Tools for Researchers & Teachers; http://sla-speech-tools.com/), provide a brief user manual, and invite researchers to use these tests to examine the role of auditory processing in various aspects of L2 speech learning (naturalistic vs. classroom, perception- vs. production-based practice).