Can computers really mark exams? Benefits of ELT automated assessments

ÃÛÌÒapp Languages
Hands typing at a laptop with symbols

Automated assessment, including the use of Artificial Intelligence (AI), is one of the latest education tech solutions. It speeds up exam marking times, removes human biases, and is as accurate and at least as reliable as human examiners. As innovations go, this one is a real game-changer for teachers and students.?

However, it has understandably been met with many questions and sometimes skepticism in the ELT community ¨C can computers?really?mark speaking and writing exams accurately??

The answer is a resounding yes. Students from all parts of the world already take AI-graded tests.??and?Versant?tests ¨C for example ¨C provide unbiased, fair and fast automated scoring for speaking and writing exams ¨C irrespective of where the test takers live, or what their accent or gender is.?

This article will explain the main processes involved in AI automated scoring and make the point that AI technologies are built on the foundations of consistent expert human judgments. So, let¡¯s clear up the confusion around automated scoring and AI and look into how it can help teachers and students alike.?

AI versus traditional automated scoring

First of all, let¡¯s distinguish between traditional automated scoring and AI. When we talk about automated scoring, generally, we mean scoring items that are either multiple-choice or cloze items. You may have to reorder sentences, choose from a drop-down list, insert a missing word- that sort of thing. These question types are designed to test particular skills and automated scoring ensures that they can be marked quickly and accurately every time.

While automatically scored items like these can be used to assess receptive skills such as listening and reading comprehension, they cannot mark the productive skills of writing and speaking. Every student's response in writing and speaking items will be different, so how can computers mark them?

This is where AI comes in.?

We hear a lot about how AI is increasingly being used in areas where there is a need to deal with large amounts of unstructured data, effectively and 100% accurately ¨C like in medical diagnostics, for example. In language testing, AI uses specialized computer software to grade written and oral tests.?

How AI is used to score speaking exams

The first step is to build an acoustic model for each language that can recognize speech and convert it into waveforms and text. While this technology used to be very unusual, most of our smartphones can do this now.?

These acoustic models are then trained to score every single prompt or item on a test. We do this by using human expert raters to score the items first, using double marking. They score hundreds of oral responses for each item, and these ¡®Standards¡¯ are then used to train the engine.?

Next, we validate the trained engine by feeding in many more human-marked items, and check that the machine scores are very highly correlated to the human scores. If this doesn¡¯t happen for any item, we remove it, as it must match the standard set by human markers. We expect a correlation of between .95-.99. That means that tests will be marked between 95-99% exactly the same as human-marked samples.?

This is incredibly high compared to the reliability of human-marked speaking tests. In essence, we use a group of highly expert human raters to train the AI engine, and then their standard is replicated time after time.??

How AI is used to score writing exams

Our AI writing scoring uses a technology called . LSA is a natural language processing technique that can analyze and score writing, based on the meaning behind words ¨C and not just their superficial characteristics.?

Similarly to our speech recognition acoustic models, we first establish a language-specific text recognition model. We feed a large amount of text into the system, and LSA uses artificial intelligence to learn the patterns of how words relate to each other and are used in, for example, the English language.?

Once the language model has been established, we train the engine to score every written item on a test. As in speaking items, we do this by using human expert raters to score the items first, using double marking. They score many hundreds of written responses for each item, and these ¡®Standards¡¯ are then used to train the engine. We then validate the trained engine by feeding in many more human-marked items, and check that the machine scores are very highly correlated to the human scores.?

The benchmark is always the expert human scores. If our AI system doesn¡¯t closely match the scores given by human markers, we remove the item, as it is essential to match the standard set by human markers.

AI¡¯s ability to mark multiple traits?

One of the challenges human markers face in scoring speaking and written items is assessing many traits on a single item. For example, when assessing and scoring speaking, they may need to give separate scores for content, fluency and pronunciation.?

In written responses, markers may need to score a piece of writing for vocabulary, style and grammar. Effectively, they may need to mark every single item at least three times, maybe more. However, once we have trained the AI systems on every trait score in speaking and writing, they can then mark items on any number of traits instantaneously ¨C and without error.?

AI¡¯s lack of bias

A fundamental premise for any test is that no advantage or disadvantage should be given to any candidate. In other words, there should be no positive or negative bias. This can be very difficult to achieve in human-marked speaking and written assessments. In fact, candidates often feel they may have received a different score if someone else had heard them or read their work.

Our AI systems eradicate the issue of bias. This is done by ensuring our speaking and writing AI systems are trained on an extensive range of human accents and writing types.?

We don¡¯t want perfect native-speaking accents or writing styles to train our engines. We use representative non-native samples from across the world. When we initially set up our AI systems for speaking and writing scoring, we trialed our items and trained our engines using millions of student responses. We continue to do this now as new items are developed.

The benefits of AI automated assessment

There is nothing wrong with hand-marking homework tests and exams. In fact, it is essential for teachers to get to know their students and provide personal feedback and advice. However, manually correcting hundreds of tests, daily or weekly, can be repetitive, time-consuming, not always reliable and takes time away from working alongside students in the classroom. The use of AI in formative and summative assessments can increase assessed practice time for students and reduce the marking load for teachers.

Language learning takes time, lots of time to progress to high levels of proficiency. The blended use of AI can:

  • address the increasing importance of?formative assessment?to drive personalized learning and diagnostic assessment feedback?

  • allow students to practice and get instant feedback inside and outside of allocated teaching time

  • address the issue of teacher workload

  • create a virtuous combination between humans and machines, taking advantage of what humans do best and what machines do best.?

  • provide fair, fast and unbiased summative assessment scores in high-stakes testing.

We hope this article has answered a few burning questions about how AI is used to assess speaking and writing in our language tests. An interesting quote from Fei-Fei Li, Chief scientist at Google and Stanford Professor describes AI like this:

¡°I often tell my students not to be misled by the name ¡®artificial intelligence¡¯ ¡ª there is nothing artificial about it; A.I. is made by humans, intended to behave [like] humans and, ultimately, to impact human lives and human society.¡±

AI in formative and summative assessments will never replace the role of teachers. AI will support teachers, provide endless opportunities for students to improve, and provide a solution to slow, unreliable and often unfair high-stakes assessments.

Examples of AI assessments in ELT

At ÃÛÌÒapp, we have developed a range of assessments using AI technology.

Versant

The Versant tests are a great tool to help establish language proficiency benchmarks in any school, organization or business. They are specifically designed for placement tests to determine the appropriate level for the learner.

PTE Academic

The ?is aimed at those who need to prove their level of English for a university place, a job or a visa. It uses AI to score tests and results are available within five days.?

More blogs from ÃÛÌÒapp

  • A woman reading outside with a book

    Three ways English students can learn to read at home

    By Kate Fitzpatrick
    Reading time: 4 minutes

    Read more at home if you want to learn English faster: three ways to get into the habit

    I can hear parents, ELT learners and teachers all over the world thinking to themselves: We know it helps if English students learn to read for pleasure at home, but how are we supposed to find the time? Who is buying these materials? What if a parent doesn¡¯t speak English themselves?

    That¡¯s the beauty of reading at home, also called extensive reading: it¡¯s completely autonomous and parents can be as involved as much or little as they like. There¡¯s more good news as it¡¯s never been cheaper to assemble a selection of extensive readers for your children or students, meaning that cost is no longer so much of a barrier to reading at home.

    Is extensive reading really that critical to learning English?

    Stephen D. Krashen¡¯s [1]??offers a marvellous summary and critique of extensive reading studies around the world, concluding that:

    "When [second language learners] read for pleasure, they can continue to improve in their second language without classes, without teachers, without study and even without people to converse with.¡¯"?(Krashen 1993 p. 84)

    Philip Prowse¡¯s excellent article, ¡°What is the secret of extensive reading?¡±[2]?agrees with Krashen¡¯s conclusion.

    Prowse goes into more detail about efficacy studies at primary, secondary and adult levels. This body of evidence finds that reading for pleasure improves results in grammar, writing, speaking and fluency, as well as comprehension and vocabulary ¨C both alongside and instead of traditional textbooks.

    So, we know it works. As with so many education-related things, the question is how to implement them. Christine Nuttall talks about the virtuous circle of reading ¨C once a learner begins to enjoy reading, they are more likely to read more and benefit more from it, so they learn to read more, and so on.

    The reverse is also true. The questions then follow: how do we motivate our Instakids to read at home in English, if they won¡¯t read in their first language? How do we carve out time between travel, work, school and homework? Here are three ways you can form the habit of reading at home:

    1. Learners need access to extensive reading material at home to use it

    Krashen establishes this common-sense fact based on five studies from 1983 to 2003.[2]?It can be a reading app, an online library subscription or a pile of readers in the corner ¨C whatever it is, it has to be the right level for the student and it has to be a topic they¡¯re interested in, or they¡¯ll never learn to read for pleasure.

    Negative reading habits can happen simply because there isn¡¯t much available to the learner: Worthy and McCool studied 11 sixth-graders in 1998 who "hated to read", and found a direct correlation between those students and the lack of reading material at home.[3] Thankfully, we now have more options than we used to:

    For extensive reading online, the Extensive Reading Foundation offers good-quality, free materials, in audio and print, at its .?These text resources and audiobooks tend to be quite basic and the stories are largely classics. You can choose by level and genre, and there is also a publisher directory.

    • You can purchase full ÃÛÌÒapp English Readers and other publishers¡¯ Kindle editions on the Kindle store, iBookstore and Google Play, and read them on an e-reader, phone or tablet using the Kindle app. These are finely-graded, contemporary, relevant e-books with titles like , , , , , , and .
    • An e-book library subscription can be a cost-effective way to get access to a lot of e-books online through your browser. is a Japanese-run online library which offers hundreds of full-text graded readers, from reputable publishers, and charges about $19 per year.
    • For print readers, cost can be an issue. If you can't buy readers at your local bookshop from a publisher like ÃÛÌÒapp, you can buy first- or second-hand readers cheaply from Amazon or the Book Depository, or you can ask your school to let you know when they¡¯re upgrading their readers library, as you may be able to take some of the older books home.

    2.?Make the most of the commute or the school run

    The key here is routine ¨C give it a try and see if it works for you. Reading doesn¡¯t just happen on a page. Today¡¯s English learners have multiple ways to read for pleasure on their various devices as well as in print, all of which are well-adapted for reading and listening on the train/on the bus/in the car/on foot.

    I listen to podcasts on my commute by train and, to this day, I know my times tables thanks to a tape my mother used to play in the car on the way to primary school.

    • Download a podcast or audiobook. Ideally, an English learner would both read and listen, but one or the other is better than nothing. Audible.com has plenty of English extensive readers in audiobook format, and a year¡¯s membership is $10 per month, or you can buy individual audiobooks. There are classic extensive reading podcasts available on iTunes for $4.99 each.
    • Never underestimate your public library.??is an online service that finds your local library for you, wherever you are in the world. You can also search by title and see which libraries carry that particular book. Just think: you could create an instant, extensive reading library at your home for free that changes every month.

    3. Consider the power of rewards

    You can reward your child or reward yourself for building a reading habit. Remember, we are talking about starting a virtuous circle: persuading a learner to begin a new habit of reading in English for pleasure. Reward mechanisms can be very effective.

    This idea should be explored on a case-by-case basis ¨C it depends on what you or your child responds to best. In my opinion, starting a reading habit is well worth a glass of wine, a chocolate treat, or an extra half-hour playing video games.

    References

    [1]?Krashen, Stephen D. (2004)?,?p57

    [2]? Prowse, Philip: ¡°¡±?

    [3] Worthy, J. and McKool, S. (1996): ¡°¡± in Ibid, p61

  • A student writingon a paper with other students doing the same in the background

    More commonly misspelled English words

    By Charlotte Guest
    Reading time: 3 minutes

    Spelling can often feel like navigating a maze. But fear not, for you are not alone in this quest. Whether you're a fluent speaker or learning English as a second language, the challenge of spelling is universal.

    Yet, just as heroes rise to conquer their foes, you too can triumph over misspelled words. With dedication and the right tools, you'll soon find yourself spelling with confidence and ease. Carrying on from our previous post, 'The most commonly misspelled words in English', let's explore more commonly misspelled words and empower ourselves with the knowledge to spell them correctly.

  • a teacher stood with two students sat a desk

    What¡¯s it like to teach English in Spain?

    By Steffanie Zazulak
    Reading time: 2 minutes

    Tim Marsh has been teaching English since 1985 and has taught over 3,500 students, with ages ranging from six to 65. He is therefore well placed to describe teaching English as a ¡°difficult and demanding¡± job, as well as to share the five lessons he¡¯s learned during his impressive time in his career¡­

    1. Know your stuff

    ¡°The Spanish expect paid professionals to know everything about their expertise but there are few teachers of the English language who do know everything. We should prepare lessons adequately when teaching aspects we¡¯re not entirely confident about.

    Many CELTA tutors say that if you are asked a question that you cannot answer confidently, you shouldn¡¯t panic but instead inform the student that you will check and give them the detailed answer at the following lesson. This may be useful when you first start out, but it shouldn¡¯t happen frequently, as your honesty will not always be appreciated!¡±

    2. Expect the unexpected

    ¡°Teaching English is very rewarding and can be full of surprises. As a result, it¡¯s not a good idea to try to follow a rigid teaching plan. Write a plan that¡¯s flexible enough to allow for a good dose of spontaneity to enter into proceedings. I can honestly say that not one single day is the same as another.

    If a Spaniard is not in the mood for working on a particular skill, as will happen from time to time, then be prepared to change that lesson at the drop of a hat. It¡¯s always a good idea to keep four or five ¡®favourite¡¯ lessons filed within easy reach for just such occasions ¨C preferably skills lessons that can be easily adapted to the theme that you are currently working on.

    Whatever you had planned for this week can always be done next week. The customer is always right and, when living in Spain, big lunches, high temperatures, Barcelona against Real Madrid and the after-match party can bring about very unexpected lessons!¡±

    3. Stick to what you¡¯re being paid to do

    ¡°The Spanish are extremely friendly people who love to talk and are happy to share ¨C sometimes in great detail ¨C the problems in their working and even private lives. In an effort to establish friendly relationships, they often create an intimacy: what is referred to in Spanish as ¡®confianza¡¯.

    This is much the same kind of trust and confidence that we have with our doctors or lawyers, so, unless you¡¯re careful, you can find yourself doubling as teacher and therapist, which will alter the dynamic of the classroom.

    A teacher of English teaches English. Stick to what you know, stick to what you¡¯re being paid to do and create a professional framework in which to do your best as a teacher and not as a therapist.¡±

    4. Do not offer guarantees

    ¡°The busiest time of the year is often during the summer, when language schools begin to fill up as state-school exam results come in. Parents enroll their children on intensive or exam revision courses so that they can take their resits in September.

    English courses are often expensive and parents will expect a guarantee that their child will pass the school English exam at the end of the summer. Never offer a guarantee! There are usually a number of reasons why the child has failed in the first place and it is better to lose a client than to make promises you can¡¯t keep.¡±

    5. Have a good pair of shoes

    ¡°Many years ago, the famous soprano Rita Hunter was asked what she considered to be the most important requirement when singing opera. She answered, ¡°A good pair of shoes.¡± She went on to say that when she was appearing in a Wagner opera that started at 5.30 pm and didn¡¯t finish until 11 pm, the most important thing to look after was her feet.

    I¡¯ve always tried to avoid institutions that insist on a uniform or on wearing a shirt and tie. Students often feel uncomfortable in a classroom where the teacher is formally dressed. I have always found the working environment much more relaxed when dressed in a similar way to my students. This and the fact that in Spain the temperature can hit the 30s in June and stay there into September mean that I dress casually, often in shorts. And I always wear a good pair of shoes.¡±