'Millionaire' tests help kids learn

This is an interesting short report up from the New Scientist:

Replacing dry multiple-choice tests with quizzes akin to the hit TV show Who Wants To Be A Millionaire could help boost comprehension levels in children.

Tzu-Hua Wang at the National Hsinchu University of Education in Taiwan has devised a web-based multiple choice testing system with some fun elements influenced by the TV quiz. The system gives pupils the chance to "prune" away two incorrect answers from four - or, in a nod to "phone a friend", they may ask the class for help.

Unsurprisingly, children were more willing to be tested using Wang's system. But he also found kids had higher comprehension levels after using it, suggesting the system could be used for educational purposes.

This is interesting as it supports my view that DESIGN MATTERS. Engaging and holding attention is increasingly tough and just moving something online is not enough. Testing strategies really do need to move on from standard multiple choice mechanisms. This research demonstrates that different approaches can have an impact on learning effectiveness.

From Tzu-Hua Wang's paper, the use of an "Ask-Hint Strategy" turns what would otherwise be a standard web-based formative assessment into an online quiz game, called GAM-WATA (not quite as catchy a title as "Who Wants to be a Millionaire?"). The strategy consists of two elements: a ‘Prune Strategy’ and a 'Call-in Strategy'. The Prune Strategy emulates the "50/50" life line contestants are offered on the TV show, but contrary to the New Scientist report removes one option from four, not two. The Call-in Strategy provides the rate at which other test takers choose each option when answering a question. This I think is more like "Ask the Audience" than "Phone a Friend" as reported.

I'd like to understand better how they measured the improvement in comprehension amongst students taking these assessments, but it certainly illustrates that smart design does make a difference to learning outcomes.

It also shows that you should go back to the original research wherever possible as this seemed slightly misrepresented by the New Scientist report.