How Output Outweighs Input and Interlocutors Matter for Study‐Abroad SLA: Computational Social Network Analysis of Learner Interactions

This data‐driven study framed in the interactionist approach investigates the influence of social graph topology and peer interaction dynamics among foreign exchange students enrolled in an intensive German language course on second language acquisition (SLA) outcomes. Applying the algorithms and me...

Full description

Saved in:
Bibliographic Details
Published in:The Modern language journal (Boulder, Colo.) Colo.), 2022-12, Vol.106 (4), p.694-725
Main Authors: PARADOWSKI, MICHAŁ B., CIERPICH–KOZIEŁ, AGNIESZKA, CHEN, CHIH–CHUN, OCHAB, JEREMI K.
Format: Article
Language:English
Subjects:
Algorithms
community/cluster detection
Computational Linguistics
computational social network analysis
English
Exchange Programs
Foreign Students
German
German as a second language
Intensive Language Courses
interactionist approach
Interpersonal relations
Language
Language Acquisition
Linguistic Input
Native Speakers
Network Analysis
peer learner network
Peer Relationship
Predictor Variables
Second Language Instruction
Second Language Learning
social graph
Social network analysis
Social Networks
Students
Study Abroad
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This data‐driven study framed in the interactionist approach investigates the influence of social graph topology and peer interaction dynamics among foreign exchange students enrolled in an intensive German language course on second language acquisition (SLA) outcomes. Applying the algorithms and metrics of computational social network analysis (SNA), we find that (a) the best predictor of target language (TL) performance is reciprocal interactions in the language being acquired, (b) the proportion of output in the TL is a stronger predictor than input (Principle of Proportional Output), (c) there is a negative relationship between performance and interactions with same‐first‐language speakers, (d) a significantly underperforming English native‐speaker dominated cluster is present, and (e) there are more intense interactions taking place between students of different proficiency levels. Unlike previous study abroad social network research concentrating on the microlevel of individual learners’ egocentric networks and presenting an emic view only, this study constitutes the first application of computational SNA to a complete learner network (sociogram). It provides new insights into the link between social relations and SLA with an etic perspective, showing how social network configuration and peer learner interaction are stronger predictors of TL performance than individual factors such as attitude or motivation, and offering a rigorous methodology for investigating the phenomenon. Graphical :
ISSN:0026-7902
1540-4781
DOI:10.1111/modl.12811