Game-Based Language Learning Research

Game Definition

• A game is a system where players engage in an artificial conflict, defined by rules, resulting in a quantifiable outcome.

Computer-Assisted Language Learning (CALL)

• Publications on CALL have significantly increased in the past decade.
• This growth should be considered compared to the overall CALL publications during the same period for a better understanding.
• Most CALL game research focuses on design issues.
• Developing game-based language learning (DGBLL) is costly and the interest in traditional tutorial CALL has decreased.
• Design focus suggests potential for integrating tutorial CALL tools into gaming environments.
• Learner perceptions, particularly regarding corrective feedback, are vital in language learning game design.
• Immersive role-playing games are generally positively viewed by learners regarding feedback.

Music-Based Language Teaching

• Music is an effective language learning tool across all levels.
• Methods include lesson planning, teaching grammar/vocabulary, or entire courses based on music.
• Many ESL teachers don't utilize music due to cost and lack of training, despite 72% expressing interest in music-based strategies.
• Music-based methods are now adapted for older learners, having initially focused on children.
• The Audio-Singual Method, using familiar tunes, aids in teaching English by creating a sense of recognition which helps overcome learners' fear and resistance.
• Music and songs are considered more effective, faster, and easier for recall than traditional drills.

Language Tests: Low Stakes & High Stakes

• Low stakes: Tests with limited impact on stakeholders, having insignificant or indirect consequences.
• High stakes: Tests directly influencing major decisions in the test-taker's life or causing extremely costly remedies if errors are made.
• Automated scoring: Raises concerns about test validity as it may not fully represent the construct being measured due to inherent limitations.
• Digital environments support sharing test items for review and adding approved items, with built-in quality checks to prevent mistakes and maintain consistency.
• A well-organized digital system collects evidence to verify test reliability and validity, including item reviews, qualifications, and feedback on items, resulting in consistent and fair scores.
• Computer-based tests offer a standardized environment, particularly when utilizing multimedia like audio and video.

Computer-Adaptive Tests

• Computer-adaptive tests utilize preexisting, pre-tested items with known characteristics to choose questions matching the test-taker's ability level efficiency.
• Adaptivity involves varying item types or numbers of items with same or different difficulty levels for high-stakes tests, which maintains consistent score interpretations.
• Adaptivity can be controlled by either the test developer (based on performance) or the test-taker (choosing items/constructs).

High Stakes Language Tests

• High-stakes language tests commonly provide control to test developers for several reasons, such as test confusion arising from allowing test-takers to control the test difficulty, leading to potentially biased results.
• Self-adaptive testing may produce inaccurate scores with capable learners scoring too low, and less capable learners scoring too high. Results vary.
• Self-adaptive tests are less reliable and have a moderate effect on reducing test anxiety.
• Computer-adaptive procedures still may not be suitable for high-stakes language tests, even when test developers maintain control.
• Computer-adaptive testing requires large, pre-tested item banks with item parameters estimated using large samples.

Technology in Language Assessment

• Technology can improve rater training and scoring through rater calibration.
• Raters score sample performances and receive feedback to ensure consistent scoring, avoiding bias.
• Test developers can use technology to monitor rating quality by comparing scores with benchmarks, aiding rater training when scores are significantly different than expected.
• Automated scoring may limit the scope of being assessed, potentially weakening the validity.
• Validity factors include various sources like expert human raters' ratings, scoring across various groups and comparison of methods, along with the ability of automated systems in identifying score manipulations.
• Stakeholder acceptance and perceptions are crucial in the implementation of automated scoring. A study demonstrated that combined automated and human scoring methods are often more trusted by test takers than automated scoring alone.
• Technology enables "gaming" the system which is why proper test security is essential, particularly for high-stakes tests.

Preventing Cheating

• High-stakes tests often face strong cheating motivations from test-takers, thus test security is emphasized.
• Test score interpretation and validity are weakened by inadequate security.
• Efforts need to be put into preventing cheating and protecting test materials to maintain score integrity.
• Cheating involves receiving or sharing information, utilizing unauthorized materials and bypassing the test procedures completely.
• Examples of how people can cheat include paying for illegally obtained tests, using technology to get help during the test, or hiring someone to impersonate the test-taker.
• Cheating can be high-tech (using technology) or low-tech (individual or collaborative cheating).
• Technological methods used for cheating include using small electronic devices to capture information or collaborating using the internet across wide geographic areas.
• Technology can also be used to prevent cheating via measures like identity validation (photos, facial recognition, multiple forms of biometric ID), on-site monitoring (video/audio recordings, digital filtering), and the development of large-scale testing programs and item banks, though exposure still makes items vulnerable to hacking.

Other Considerations

• Unexpected high scores at particular test centers can signal coordinated cheating.
• Consumer-oriented VR devices, like Oculus Rift, can enable fully immersive virtual environments accessible for authentic and interactive test creation.
• Multimodal tasks, requiring test-takers to read texts and listen to speakers as well as responding in writing or speech are increasingly used.
• Language ability is argued to stem from interaction between language and technology, whereby "reading ability" could now encompass gathering information from online and traditional resources in expanded definitions.