Edexcel's turn to suffer a scandal

Examination scandals, usually reserved for announcement of results, have come early this year. The problem this time is that Edexcel deliver the exam papers to centres ahead of time to offset the risk of delays in the post. Enterprising young North London tealeaves exploit this window of vulnerability to help themselves and their classmates to higher marks than they might otherwise have got.

All three exam boards are well aware of the potential to solve this problem with electronic delivery of exam papers, because each question needn’t leave the exam board until the very moment the first student needs to read it. Whilst they are experimenting with electronic delivery, they are taking it very slowly. The reason: they fear even greater scandals if it goes wrong. No exam board wants to be the first to suffer headlines claiming that their computer error had ruined the lives of thousands of young people. When I first presented Paperless School to the boards, almost three years ago, a panel member from one board put it to me thus:

“We will move to electronic delivery when and only when we judge the probability of a scandal to be lower through electronic delivery than it is through paper delivery.”

Chalkface now has considerable experience of delivering mixed-mode assessments (i.e. not just multiple-choice) through the web, so it might be useful to look at our experience of what can go wrong.

  • Peaky system loads. It’s quite normal for the load on Paperless School’s servers to fluctuate by 2000% in the space of 5 minutes because so many schools have the same class times. It’s the actual process of students logging on that puts the biggest loads on the server.
  • Hosting vulnerabilities. Every hosting company claims to be invulnerable; none are. Inside the data centre itself they may have reduced the probability of a fatal error to something very small, but what about their connection to the outside world? At some point there’s a cable that’s very, very vulnerable to a drunk with a JCB. Our solution has been to look at the physical location of the data centre and to try to find one with the shortest route to a major exchange from which your data can flow in multiple directions. An exam board would do well to go further, and have duplicate systems running in two different data centres.
  • Local difficulties. How good is the centre’s network and connection to the outside world? There are frequently teething problems caused by rogue proxy servers or over-zealous security software. Using a proven delivery system is vital; I’d recommend several full-scale rehearsals in each centre.
  • Terminal failures. Individual candidates’ computers will sometimes crash in the middle of an assessment. This needn’t be a huge problem, provided the software is designed to save the candidate’s work to the server as often as possible, and there’s a procedure in place to allow a little extra time to the candidate in compensation.

Mitigate these risks and you are on track for successful, electronically-delivered exams. You’ve probably achieved a good tradeoff between the risks of paper and electronic delivery.

I would like to see the exam boards go much further and embrace the potential of electronic delivery by moving from synchronous to asynchronous delivery. But more on that in a future post.

Mobile phones may yet improve education

I’ve quoted Berthold Weidmann on students using mobile phones to access lessons before, and will doubtless do so again, because I’m intrigued by this largely-ignored scenario.

Now, this excellent overview from Reuters puts a little flesh on the bones. It explains why we’ve not got there already with 3G, and covers the technical choices that Europe now faces.

This introductory paragraph could not fail to set me thinking:

Third-generation mobile phone services are finally here after a mammoth effort that cost the industry at least $123 billion, but new systems that operate much faster already threaten to consign 3G to history.

$123 Bn!!! What could we have achieved if we’d decided that it was an economic imperative to invest an extra $123Bn into educating Europe’s children?

When you think about where that money actually went, you discover this need not be an idle daydream. Most of the cost was the operating licenses auctioned by governments across Europe. Our governments. Effectively, that $123Bn is a tax on phone calls that mobile users will be paying off for the next decade.

And what are taxes for? That’s right. Education.