ALTE’s 8th International Conference: Madrid, 26-28 April

ALTE’s 8th International Conference: Madrid, 26-28 April

ALTE stands for Association of Language Testers in Europe, which is an organization that brings together individuals and institutions involved in language testing in Europe.

The main goal of ALTE is to promote the development and use of high-quality language assessments that are fair, reliable, and valid. This is achieved through the collaboration and sharing of knowledge and expertise among its members.

ALTE has developed a set of guidelines and standards for language assessment, which are widely used in Europe and beyond. These guidelines cover topics such as test development, test administration, test scoring and reporting, and test evaluation.

ALTE also offers a range of services and activities to its members, including training courses, conferences, and research projects. The organization has member institutions from across Europe, including national examination boards, universities, and language schools.

Overall, ALTE plays an important role in promoting high-quality language assessment in Europe and beyond, and its guidelines and standards are widely recognized as a mark of quality in the field of language testing.

Three years ago, ALTE had to postpone, and then cancel, its International Conference in Madrid due to Covid-19. Three years later ALTE came back to Madrid. Instituto Cervantes ran and hosted the 8th International Conference, which was held at the Novotel Madrid Centre centrally located in Salamanca district.  

Much has changed in these past three years, partly due to the Covid-19 disruption, but also due to rapid advances in technology, such as the arrival of AI tools like ChatGPT, and the publication of the CERF’s Companion Volume with a renewed focus on concepts such as mediation and systemic alignment. 

ALTE’s 8th International Conference looked at how language assessment has shaped and will shape society in order to be fit for the future, fit for the Digital Age. The digital age has been with us for some years now but Covid proved a catalyst for an acceleration of all things digital which reached into language assessment and in many other areas of life. Language assessment bodies needed to change quickly while still having a longer-term view of what all this means for the industry and ultimately for the language learner themselves.

The Madrid conference took place in a hybrid format, with the encouragement to attend in person to catch up with old friends and make new ones, but with some digital provision for those who were not able to travel to Spain.

The Future of Assessment

The future of assessment is likely to be shaped by technological advancements and a growing emphasis on personalized and adaptive learning. Here are some key trends that are likely to shape the future of assessment:

1. Personalized and adaptive assessments: As technology continues to improve, we can expect to see more assessments that are tailored to the needs of individual learners. Adaptive assessments use algorithms to adjust the difficulty of questions based on a student’s previous responses, providing a more accurate and efficient measure of their knowledge and skills.
2. Artificial intelligence and machine learning: AI and machine learning can analyze large data sets to identify patterns and trends in student performance. This can help educators identify areas where students are struggling and provide targeted support to help them succeed.

Assessment in multilingual societies

The issue of assessment in multilingual societies has become increasingly significant in recent decades as multilingualism is now a legitimate reality around the world. Multilingualism is the norm rather than the exception due to ever-increasing levels of migration and globalisation. 

Assessment in multilingual societies is a complex and challenging task, as it involves evaluating language proficiency and academic achievement in a context where learners may be using multiple languages. In such societies, it is important to consider the diversity of language backgrounds and proficiency levels among learners, and to ensure that assessment tools are culturally and linguistically appropriate.

One approach to assessment in multilingual societies is to use a plurilingual approach, which recognizes and values the diversity of languages that learners bring to the classroom. This approach involves assessing learners' proficiency in multiple languages, and recognizing the ways in which learners use and transfer skills and knowledge across languages. Plurilingual assessment also involves taking into account learners’ cultural backgrounds and experiences, and considering how these may affect their language use and learning.

Assessment in multilingual societies also requires careful consideration of the language(s) of instruction and assessment. It is important to ensure that assessment tools are available in the languages that learners are using, and that assessment tasks are designed to be culturally and linguistically appropriate. It is also important to consider the potential for bias or discrimination in assessment, and to take steps to mitigate these risks.

Overall, assessment in multilingual societies requires a nuanced and flexible approach, which recognizes the diversity of learners and the complexity of language and culture in educational settings. By valuing and supporting learners’ multilingualism, and by using appropriate and culturally sensitive assessment tools, it is possible to promote equitable and effective learning outcomes for all learners.

The plenary sessions

The first plenary was delivered by Dr Nick Saville, Director of Thought Leadership, English for Cambridge University Press and Assessment, and Secretary General of ALTE.

“Language assessment is a rapidly evolving field” Dr Saville said. “With advances in technology, changes in educational policies, and increasing demands for global communication, the ways we assess language proficiency must also adapt and evolve.

One key trend that is shaping the future of language assessment is the move towards more authentic and dynamic assessments. Traditional paper-and-pencil tests are being replaced by tasks that more closely resemble real-life communication situations. This includes tasks that require test takers to produce written texts in response to authentic prompts.

Another trend is the increasing importance of technology in language assessment. Technology can be used to enhance the validity and reliability of language tests and also to increase their accessibility and efficiency. For example, computer-based testing can provide more precise estimates of language proficiency by adapting to the ability level of each test taker. Technology can also enable remote testing, which allows individuals to take tests from any location, at any time.

At the same time, we must be aware of the potential biases that can be introduced by technology. For example, automated scoring systems may be biased against certain groups of test takers, such as those with non-standard accents or those who use non-standard English. We must also be aware of the ethical implications of using technology in language assessment, particularly in terms of privacy and data protection.

Finally, we must also consider the impact of language assessment on society as a whole. Language tests have the power to shape educational policies, immigration policies, and employment opportunities. As such, it is important that language assessments are fair, transparent, and accurate, and that they do not perpetuate inequalities or discrimination.

In conclusion, language assessment must continue to evolve and adapt in response to the changing needs of society. By embracing new technologies, adopting more authentic and dynamic assessments, and ensuring fairness and transparency, we can ensure that language assessment remains fit for the future.”

Diversity and Inclusion in language testing

Diversity and inclusion are important considerations in language testing as they can have a significant impact on the validity and reliability of the tests. In language testing, it is essential to ensure that all test-takers have equal opportunities to demonstrate their language proficiency, regardless of their backgrounds, cultures, or identities.

Dr Lynda Taylor, University of Bedfordshire, a renowned expert in language testing, has been a vocal advocate for promoting diversity and inclusion in language testing. In her research and advocacy, Dr Taylor emphasizes the need for language tests to be fair, transparent, and culturally sensitive.

Overall, promoting diversity and inclusion in language testing requires a multi-faceted approach that involves stakeholders at all levels of the testing process, from test development to test administration and scoring. By prioritizing diversity and inclusion, language tests can become more accurate, reliable, and equitable for all test-takers.

The CEFR’s Companion Volume 2020

The CEFR’s Companion Volume 2020 is an updated version of the Common European Framework of Reference for Languages, which is a standard used to measure and describe language proficiency in various languages.

It provides detailed information on the CEFR’s theoretical and conceptual framework, as well as practical guidance on how to apply the CEFR in language teaching, assessment, and curriculum development.

As Dr Neus Figueras, University of Barcelona, pointed out the CEFR Companion Volume 2020 includes several updates and additions, such as:

• Revised and expanded descriptions of the CEFR levels and sub-levels, with more detailed descriptors of language proficiency at each level.
• New chapters on plurilingual and intercultural education, which emphasize the importance of developing learners’ ability to use multiple languages and understand different cultures.
• Expanded guidance on using the CEFR to assess and evaluate learners’ language proficiency, with examples of different types of assessment tasks and criteria for evaluating performance.
• Guidance on using the CEFR to develop and evaluate language learning outcomes, with examples of how to design learning objectives and assessment tasks that align with the CEFR levels.

Overall, the CEFR Companion Volume 2020 is an important resource for language teachers, assessment professionals, and curriculum developers who want to use the CEFR to enhance their language teaching and learning programmes. It provides updated and expanded guidance on how to use the CEFR to set learning goals, design assessments, evaluate language proficiency levels, and emphasizes the importance of developing learners' plurilingual and intercultural competence.

Language assessment based on the can-do statements provides a reliable and valid way to assess language proficiency, and enables learners to receive a recognized qualification.

Machine Learning Technologies

With the rapid advancement of machine learning technologies, we have seen an increasing use of automated scoring algorithms in large-scale educational assessments, including language assessment. This has prompted professional organisations to introduce guidelines for examining the validity of automated scoring in the latest standards for testing.

Language testers’ general reactions to automated scoring technologies have been described as “cautiously optimistic”. Some key issues discussed were:

• training and evaluating an automarker,
• malpractice detection, hybrid marking,
• interpretability of automated scores, and
• ethical considerations in using scoring algorithms.

The researchers’ conclusions called for more transparency in describing the general characteristics of scoring algorithms and the training data, more consistency among test developers in presenting validity and reliability evidence for automated scores, and a collaborative effort between language testers and computational linguists to improve automated scoring literacy and interpretability.

Automated Essay Scoring (AES) and the Cambridge Learner Corpus

What are the possibilities of Automated Essay Scoring (AES) to consistently, fairly and practically assess a large number of writing products? AES systems automatically score a text using machine learning and by extracting linguistic characteristics from a text. Previous research has focused almost exclusively on genre (essays) mostly written for higher education purposes.

The Cambridge Learner Corpus (CLC) is a large database of learner language that has been collected by Cambridge University Press. It contains millions of words of written and spoken learner language, covering a range of proficiency levels and language backgrounds.

One potential application of the CLC is in automated essay scoring (AES), which uses computer algorithms to evaluate and score essays based on a range of linguistic and content-related features. The idea is to provide a quick and objective way of evaluating essays, which can be useful in large-scale assessments and/or for providing instant feedback to learners.

The CLC can be a valuable resource for AES because it provides a large amount of authentic learner language data that can be used to train and test AES algorithms. By analyzing the features of successful and unsuccessful essays in the CLC, researchers can develop algorithms that can accurately predict the quality of new essays.

One common method used in machine learning-based AES is to train a model on a large dataset of human-scored essays and their corresponding features (e.g., word count, sentence length, etc.). The model then uses this information to assign a score to new essays based on their similarity to the training examples.

Marking written tasks is an expensive and time-consuming activity. Investing in machine learning technologies could save testing institutions huge amounts of money and allow them to issue results within days instead of months. AI evaluates written responses and assigns a score based on predetermined criteria, such as grammar, spelling, coherence, and relevance to the prompt.

However, there are also some limitations to using the CLC for AES. For example, the CLC may not include samples of all possible essay topics, and learners in the CLC may not represent the full range of proficiency levels and language backgrounds that may be encountered in other contexts. Additionally, the CLC may not include the types of writing tasks or prompts that are commonly used in specific educational or professional contexts.

Overall, while the CLC can be a useful tool for AES, it is important to use it in conjunction with other sources of data and to carefully consider its limitations in any specific context.

Leda Lambropoulou, Head of Assessment at LanguageCert

Remote proctoring

As Leda Lambropoulou, Head of Assessment at LanguageCert explained, remote proctoring refers to the practice of monitoring and supervising exams or assessments conducted remotely, typically through the Internet. It is designed to ensure the integrity and security of online examinations by preventing cheating or academic dishonesty.

Remote proctoring typically involves the use of specialized software or platforms that use various methods to monitor and authenticate test-takers. Common techniques used in remote proctoring include:

• Video monitoring: Test-takers are required to use a webcam during the exam. The software captures video and audio of the test-taker throughout the entire test session, allowing proctors or invigilators to monitor their behaviour in real time. This helps detect any suspicious activities or violations of the exam rules.
• Screen sharing: Remote proctoring software may require test-takers to share their screen during the exam. This allows proctors to view the test-taker's screen in real-time, ensuring that they do not access unauthorized resources or applications.
• Identity verification: Remote proctoring often includes identity verification measures to ensure that the person taking the exam is the authorized test-taker. This may involve the use of government-issued identification, facial recognition technology, or biometric authentication.
• Browser restrictions: To prevent cheating, remote proctoring software may restrict the use of external websites or applications during the exam. It can disable copy-paste functions, restrict access to specific websites, or prevent opening additional browser tabs or windows.
• AI-based monitoring: Some remote proctoring systems employ artificial intelligence (AI) algorithms to analyze the test-taker’s behaviour and detect potential cheating indicators. These algorithms can identify suspicious patterns, such as excessive eye movements, unusual mouse activity, or attempts to communicate with others during the exam.

It's important to note that while remote proctoring can help deter cheating, it does raise privacy concerns. Test-takers’ personal data, including video and audio recordings, may be collected and stored during the proctoring process. Educational institutions and proctoring service providers should adhere to privacy regulations and clearly communicate their data-handling practices to maintain transparency and trust.

Thomais Rousoulioti, Aristotle University, talking about the new assessment law for Greek citizenship

Ethics in technology

Ethics in technology refers to the moral principles and values that guide the development, use, and impact of technology on society. With the increasing role of technology in our lives, it has become imperative to consider the ethical implications of technology and its impact on society.

Some key areas of concern for ethics in technology include:

• Privacy: The use and storage of personal data by technology companies and the potential for misuse or unauthorized access to this information.
• Bias and discrimination: The potential for technology to perpetuate biases and discrimination, such as through algorithmic decision-making or facial recognition software.
• Responsibility and accountability: The ethical responsibility of technology companies to ensure the safety and security of their products and services, as well as their accountability for any harm caused by their technology.
• Sustainability: The impact of technology on the environment, including issues such as energy consumption, e-waste, and carbon footprint.

In order to address these and other ethical concerns in technology, it is important for technology companies, policymakers, and individuals to engage in open dialogue, establish ethical guidelines and best practices, and prioritize the needs and well-being of all stakeholders.

The Greeks in Madrid: From left: Korina Dourda (LanguageCert), Maria Iakovou (University of Athens), Yiannis Papargyris (LanguageCert), Anna Mouti (Aristotle University of Thessaloniki), Anastasia Spyropoulou (ELT NEWS), Paraskevi Kanistra (Trinity College, London), Thomais Rousoulioti (Aristotle University of Thessaloniki), Aggeliki Salamoura (CUP and Assessment)

Conclusion

Language assessment is an important tool to measure individuals’ language proficiency and to ensure that they have the necessary skills to function effectively in society. As we look to the future, there are several key areas that language assessment should focus on to remain relevant and effective.

• Technology-enhanced assessments: Technology has the potential to transform language assessment by providing new ways to gather data, analyze results, and deliver feedback. For example, automated scoring systems can provide rapid and accurate assessments of writing and speaking skills, and virtual reality environments can simulate real-life language use scenarios for more authentic assessment experiences.
• Multilingual assessments: In an increasingly diverse and interconnected world, language assessments should reflect the growing need for multilingualism. This means developing assessments that measure proficiency in multiple languages and recognizing the value of diverse linguistic skills.
• Performance-based assessments: Traditional language assessments often focus on discrete skills, such as grammar and vocabulary, but the ability to use language effectively in real-world situations is also important. Performance-based assessments, such as role-plays, presentations, and discussions, can provide a more comprehensive and meaningful evaluation of language proficiency.
• Contextualized assessments: Language proficiency is not just about the ability to use grammar and vocabulary correctly; it also involves cultural knowledge, social awareness, and pragmatic skills. Contextualized assessments that take into account the cultural and social context in which language is used can provide a more accurate and relevant measure of language proficiency.
• Continuous assessment: Language proficiency is not static; it can improve or deteriorate over time. Continuous assessment, such as ongoing feedback and monitoring of language development, can help learners identify areas for improvement and track their progress towards their language learning goals.
• Overall, language assessment needs to be responsive to the changing needs of society and reflect the complex nature of language proficiency. By incorporating technology, multilingualism, performance-based assessments, contextualized assessments, and continuous assessment, language assessment can remain relevant and fit for the future.