Simon Lebus explores the potential for education in data collection, including that done in support of our unique study #AspectsOfWriting.
There is much excitement among educationalists and policymakers about the potential of big data to transform the teaching and learning process. Modern processing systems and the development of the cloud permit storage of huge volumes of data that can then be interrogated to identify the most common mistakes that students make when learning subjects such as foreign languages and maths, and permit the production of highly-focused materials that will help them navigate their way through these. We are fortunate enough in Cambridge to have a couple of long-standing projects based on this approach and I was in Japan this week talking to some of our business partners about one of these called Write and Improve. This is an online platform into which students learning English can type freeform text and receive corrections and guidance on the mistakes they have made, as well as a rating of how good their language skills are on the Common European Framework of Reference for Languages, all within a couple of seconds. The machine learning behind this application is based on the Cambridge Learner Corpus which represents a collection of samples of errors made by students taking our English language exams which has been accumulated over the last 23 years. There are about 60 million words in this, all parsed and tagged to identify the types of error that they represent and it is a good illustration of the powerful impact this type of big data can have on the learning process - but also of the reality that such data takes a long time to accumulate.
A slightly different example of this longitudinal approach to data collection has come with the launch of our latest Aspects of Writing study, an investigation of features of written English as produced by 16-year-old UK students in their GCSEs. We have been running this study roughly every 10 years since 1980, allowing us to observe changes in the way 16 year olds use English. The most striking findings in this survey (based on exam scripts from 2014) are greater use of short sentences and generally less complex sentence structures, increased use of paragraphs, and finally, a decline in the quality of spelling and use of punctuation marks among lower attaining cohorts – though this trend is not evident in the work of medium and high attaining students.
All interesting stuff, and we are fortunate to have had far-sighted research colleagues 36 years ago who started the series of studies off because of their belief that it would be an effective way of tracking changes in everyday formal written English language usage. What is particularly striking is that the 1980s study was effectively pre the computer age so that subsequent studies reveal the gradual development of a completely different method of discourse based on computers, personal phones, internet, social media, text messages and so on. That represents a major revolution in how we communicate and the amount of information that we all have to deal with in our daily lives. What’s perhaps reassuring about this study is that it shows that English language usage patterns, while responsive to changes brought about by modern communication technology , have remained fairly consistent (at least so far as written English in GCSEs is concerned) throughout. What the study suggests is that 21st century students, while acutely aware of the wide range of communicative styles available to them, remain well able to use formal written English when it is appropriate to do so.
Simon Lebus
Group Chief Executive, Cambridge Assessment
Missed the live launch of #AspectsOfWriting? Watch again, access the full report and read more audience reactions here.