Mobile-assisted Language Learning: A Comprehensive Study Guide
Mobile-assisted Language Learning: A Duolingo Case Study
Abstract
Growing availability of mobile technologies has led to an increase in mobile-assisted language learning (MALL).
MALL allows learners to autonomously study a second language (L2) anytime, anywhere (Kukulska-Hulme, Lee & Norris, 2017; Reinders & Benson, 2017).
Research on the effectiveness of MALL has been limited, particularly for commercial apps such as Duolingo.
A commissioned study showed favorable language learning outcomes (Vesselinov & Grego, 2012).
However, independent research has highlighted issues regarding learner persistence, motivation, and program efficacy (Lord, 2015; Nielson, 2011).
This study investigates the semester-long experiences and results of nine participants learning Turkish via Duolingo.
Participants showed improvement in L2 measures, with a positive correlation between time spent on Duolingo and learning gains.
General perceptions included positive views on flexibility and gamification, but noted variability in motivation and frustration with materials.
Keywords
Duolingo, mobile-assisted language learning (MALL), self-directed language learning, Turkish
1. Introduction
Impact of Mobile Technologies on L2 Learning
Mobile technologies have revolutionized the study and learning of L2 (Kukulska-Hulme et al., 2017).
Learners can make autonomous choices on where, when, and how to study L2 (Reinders & Benson, 2017).
Popularity of apps like Duolingo (200 million active users as of 2018) emphasizes interest in technology for L2 study.
Need for Effective Evaluation of Language Learning Programs
Heift and Chapelle (2012) emphasized the need to understand new conditions for second language acquisition (SLA) shaped by technological capabilities.
Plonsky and Zeigler (2016) highlighted the necessity to explore how technology can enhance optimal language learning opportunities.
1.1 Mobile-Assisted Language Learning (MALL)
Definition: MALL involves using mobile technologies for educational purposes, allowing learning at flexible times and locations (Burston, 2015; Duman et al., 2015).
Key Components of MALL:
Flexibility in study time and location.
Continuity across devices (phones, tablets, computers).
Easy access to information.
Adaptability to individual study habits (Duman et al., 2015).
Popularity and Usage: Despite some limitations, MALL apps are widely used for autonomous language learning, often outside of formal education.
1.2 Effectiveness of MALL
Research on MALL has shown advantages for reading, listening, and speaking skills (Burston, 2015).
Commercial Programs:
Commissioned studies have found favorable outcomes, while independent research raises concerns over learner persistence (Lord, 2015; Nielson, 2011).
1.3 Duolingo
Overview: Duolingo, launched in 2011, is a free, mobile-accessible language education platform that syncs progress across devices.
Effectiveness Claims:
The platform claims to teach language material equivalent to a first semester college language course within an average of 34 hours (Vesselinov & Grego, 2012).
Mixed Research Findings:
Studies suggest Duolingo may serve as a cost-effective alternative but highlight pedagogical limitations, especially in terms of authentic language context (Cunningham, 2015; Rachels & Rockinson-Szapkiw, 2018).
2. Method
Participants: Nine individuals from Michigan State University (3 female, 6 male), all with prior L2 experience but no prior exposure to Turkish.
Graduate students and a professor comprised the participant group, with varying L1 backgrounds.
Materials:
Duolingo app and additional testing instruments (Duolingo Progress Test and a Turkish 151 Test).
Participants maintained journals to document their experiences.
Procedure: Four phases of study over 12 weeks were followed by language assessments (Duolingo Progress Test and Turkish 151 Test).
2.1 Materials
Duolingo Features:
Lessons divided into skills and presented in linear order.
Activities: L1–L2/L2–L1 translations, multiple-choice, dictation, and oral repetition (available on desktops only).
Gamification elements: setting goals, earning experience points (XP), and competition through leaderboards.
2.2 Procedure Phases
Turkish Study & Journaling: Participants began with a minimum target of 34 hours of study using only Duolingo.
Language Assessment: All completed language tests after 12 weeks.
Continued Study: Participants could study further to meet the 34-hour target.
Final Assessment: Retake the tests (July 2016).
3. Results
3.1 Learning Outcomes - RQ1
Study Time: Participants spent an average of 29 hours studying; mobile app and web combat methods were analyzed for time spent.
Test Scores:
Duolingo Progress Test average: 0.63/5.00 with a range from 0.35-1.78.
Turkish 151 Test average: 48% with a range of 23%-76%.
Subscore Performance:
Better performance in reading, writing, and lexicogrammar compared to speaking and listening.
3.2 Experiences - RQ2
Flexibility of Use: Many students appreciated the ability to learn anywhere, although some felt pressured to fit in study time.
Wavering Motivation: Initial motivation decreased as time progressed.
Perceived Progress: Doubts regarding actual language acquisition were prevalent despite completing lessons.
Individual Approaches: Varied each participant's study style, some prioritizing review over new lessons.
4. Discussion
Participants benefited somewhat from Duolingo, but outcomes were mixed.
Low overall passing rates challenge the claims made about the platform's effectiveness.
Motivation variations affected learning gains; students lacked intrinsic investment in Turkish as a language.
Importance of pairing technology with formal education to optimize outcomes.
5. Limitations
Small sample limits generalizability.
Participants' previous L2 experience may not reflect typical Duolingo users' outcomes.
6. Conclusion
Duolingo can improve L2 knowledge, but claims regarding effectiveness may be overstated. Incorporating a wider range of activities and making improvements based on SLA research is recommended.
References
Extensive references are provided for further reading, including works by Burston, Krashen, and Kukulska-Hulme, among others.
Acknowledgments
Authors acknowledge contributors to the project and adherence to ethical research standards.