The achievement gap and AI augmented online tutoring

5 July 2018

The achievement gap between students who come from different socio-economic backgrounds is a well-known and persistent problem in education.

Disparities in achievements between high and low socio-economic groups are evident in children as young as age 3 years and seem to be a problem the world over. Despite pupils’ overall attainment scores rising over the last decade or so, the gap between students from different socio-economic groups remains intractably present and widespread.

Family finances play a big part amongst the various reasons for this disparity. Pupils from low socio-economic backgrounds (SEBs) are often only able to attend a few, if any of the extra-curricular activities enjoyed by their more affluent peers. Access to good schools for pupils from SEBs is often reduced as is their access to the educational and occupational aspirations which can impact children’s academic achievement.

Moreover, with the achievement gap starting at a very early age, the lack of prior achievement decreases the likelihood of students with low SEBs ever catching up with students from high SEBs.

One of the biggest factors in this attainment gap is an inability by low-income families to access private tutoring. As a recent research report reveals, pupils with the same levels of achievement from high SEBs receive 2.5 hours more additional instruction per week than less advantaged students. Similarly, disadvantaged pupils across the UK complete less additional school work. Furthermore, only half of 15-year-olds from disadvantaged SEBs regularly receive help with homework from their parents, compared to more than two-thirds of those from the most advantaged background. In short, bright pupils from low SEBs receive much less support than their better-off peers.

The recent Reform report into the use of EdTech as a tool for improving social mobility raised some important and timely issues. The EDUCATE programme, based at UCL Institute of Education, and part-funded by the ERDF and other partners, aims to address some of these issues by supporting EdTech start-ups and entrepreneurs to develop their products with a strong focus on evidence of what works. This is what allowed one of our early cohorts, My Tutor, to contribute to the report with its own findings (pages 19-20) to demonstrate the impact of its reasonably-priced, affordable on-line tutoring company. From Bloom’s seminal “the 2-sigma problem” paper to the most recent best evidence syntheses on the topic we know that teaching makes a significant difference to learner achievement and one-to-one teaching approaches are effective for learning. However, the cost of one-to-one tuition is very high.

Although, on-line tuition can provide a cost-effective alternative to traditional face-to-face approaches evidence shows that the quality of teaching is crucial for its effectiveness. Therefore, the main question that we all should be focussing on is: how can we ensure the quality of tutoring in online one-to-one tutoring systems?

Technology has the potential to help ensure the quality of on-line tutoring sessions, but discussions surrounding EdTech often assume that technology can automatically enhance human teaching and learning practices. The evidence showing whether this is in fact the case or not is all too rarely considered.

If instead we use the learning sciences: what we know about how people learn, to inform the design of technology for tutoring, then positive outcomes are far more likely. In addition to our work with MyTutor, UCL Institute of Education are also working with an industrial partner called Matr to design technology to monitor and ensure the quality of tutoring in online settings by augmenting labour-intensive human input, with Artificial Intelligence (AI). The project is identifying the main signifiers of online tutor success as reported in the learning sciences research literature and as observed in real-world online one-to-one tutoring. Online tutoring sessions are then tagged using these signifiers and processed by AI algorithms to dramatically reduce the human evaluator input that is required. In addition to improving the efficiency of the evaluation process, the AI can also be used to individualise continuing professional development for online tutors to ensure that they are constantly developing their skills individualised according to their personal strengths and weaknesses

We believe that high quality online tutoring that is also affordable could contribute to bridging the long-lasting challenge of achievement gaps in education. We are looking forward to sharing our results in the near future.

Co-written by Professor Rosemary Luckin, Professor of Learner Centred Design, UCL Knowledge Lab and Director, EDUCATE and Dr Mutlu Cukurova, Lecturer in Digital Technologies in Education and Learning Scientist, UCL Knowledge Lab

Professor Rosemary Luckin

Dr Mutlu Cukurova

Comments

Comments

No comments yet.