Finding questions that Google can't answer (updated)

The old reason for banning mobile phones and the use of 'always on' internet-enabled devices in schools was that children 'cheated'. We're beginning in some places to see over the top of that particular mountain, but how about this for a contentious question: should we allow smartphones and internet-enabled computers into examinations?

I'd argue it's worth thinking about. I was a French and German teacher, subjects which, when I was at school, did not allow the use of a dictionary in the examination. For some time now, students have been able to use dictionaries, something that tends to bring lower results to students who have not been taught well in specific dictionary and reading skills.

If we were to teach students how to effectively use the web, search, social search and shared bookmarking techniques within a pressure environment, in much the same way as we've done for decades in languages and dictionary tuition, what would we be left with?

My guess is that many educators and examination bodies would still not be happy, since too many of the answers sought could be machine programmable or searchable.

So, we need to change the way we ask questions, we need to change the way we test and assess. The remaining question is therefore: how?

Well, in a lovely example or two of truly higher order thinking, analytical and creative thinking, Bill Boyd mentioned two things.

First Dan Meyer's talk on how we should teach mathematics to be "less helpful", and construct a creative rub against which students can learn. His talk is a superb 20 minutes for any teacher:

But further still, and totally new for me, is the concept of Fermi Questions. These are questions named after the Nobel Prize-winning Italian physicist Enrico Fermi, who was well known for solving problems which left others baffled. There is no searchable answer, and no one way of answering them. They are the true meaning of "there is no right way to answer this".

Update: Had they not already written a blog post which is now Googleable, I'd have said that finding out how to play the world's shortest possible game of Monopoly would make a great Fermi question. But they did, and it is, so it's not.

Related articles by Zemanta

• The 21st-century textbook (radar.oreilly.com)

Comments

[Alan Paxton]

I've never heard them called Fermi Questions, but they (the piano tuners is canonical) are heavily used in Software Engineering interviewing. I think Microsoft started it. They're nowadays losing out to simulated real software development problems, but the key is surely to illustrate that you have the relevant tools to perform a complex task. I'm reminded of the old exam instruction (probably specific to maths and physical sciences) to "Show your working!" because it's the working that matters.

[Waltatek]

And "The little book of Thunks" is right by my elbow on the bookshelf. Bet Google can't answer many of these!
Great resource for getting the brain juices flowing... (or the neurons firing or whatever)!

[Liz Dorland]

Found this buried on a tab in Firefox and I'm going to push back a tiny bit. The assumptions in the so-called "iconic" Fermi Question "How many piano tuners are there in NYC?" question are very ad hoc. One family in five owns a piano? Where does that come from?

If you google "Piano Tuner" "New York City" and count dots on the google maps that come up (or look at hits - 1300+ - and take off bogus ones I bet you would get a more accurate answer. Most piano tuners (and music shops that sell pianos) would advertise online and show up these days. I vote that this question be taken out of the Fermi Question test bank. ;-)

Liz D.

[Joe]

Love the TED Talks, I always find myself browsing through them on a weekly basis, so much interesting stuff in there.

[Ewan McIntosh]

Good point, Liz, although there is at least some manual investigation required to get to the answer, rather than just expecting someone else to have crafted a neat one-liner. Half in, half out? ;-)

[Philippa Cross]

The 18 sets of 10 questions in the quiz set by King Williams's College - http://tiny.cc/likgr - are deliberately cryptic and often nigh on impossible to find answers to immediately. Set to the college boys the day before the Christmas holidays (scores generally 2 out of 180) and then again on their return after having had time for research, this has always seemed to me to be a great way of building not only research skills, but those of analysis, connection making, and collaboration. My family always have a go, and do much better when we all sit and ponder it together - and we find not everything is easy to locate on google even when we have the theme - it forces reference to other sources.

We've seen it get more obscure over the years - the quiz setter alledgedly googles the questions and changes anything that is too easily discovered. Of course that is all undermined by the fact that within hours of it being published, other people have published their answers...