Online testing and cheating

Image by Alexandre Fugita via FlickrThe biggest problem with unsupervised online testing is that you can cheat. This means the result you achieve is not a credible reflection of your ability. You could use Google or Wikipedia and craft your answer from online sources. My question is... is cheating so bad?

My confession. I always copied my homework at college. To me it just made sense. Rather than create, I got to innovate. There were basically two scenarios. If my friend's solution was wrong, I got to see a partial attempt which allowed me to focus on the hardest part and learn from my friends mistakes. If the solution was right, I used copying it as a learning experience to engage with the material and simplify wherever possible.

For me the problem with getting something wrong was that it was only at the end you realise its wrong, but you already created the neural pathways to get there. That means you are more likely to get it wrong the second time. To put it another way, its just more efficient to create the correct neural pathways the first time, than rely on conditioning (e.g. learning from mistakes).

The point is that whether you cheat or not, tests can force you engage with content. Thats why with online tests, you have to differentiate between a tool which helps learning takes place and a qualification. Just because the qualification aspect has real limitations doesn't mean online testing can't be useful.

Testing is important because its a way to turn the push model into the pull model. The test will become the incentive for learning which induces a pull for knowledge. This is in contrast to the push model - learn X and you'll be tested afterwards. Test design will need to move from testing the "what" to test the "so what", and cheating (collaborating and research under time pressure) is the point. Afterall, this is what happens in real life, right?