2013-06-28

Co.Exist

Do I Play Chess Better In A Dress? Test Any Theory With This Randomized Trial Tool

You can now set up your own (semi) clinical trials and perform ridiculous research, with the help of Randomise Me.

Ever wondered if cheese really does give people nightmares? Whether you’re more efficient checking email at set times, or replying constantly? Or, if men play chess better in a dress?

If so, you can now test the theory easily online. The site is called Randomise Me, and basically it’s about using the trusted methods of High Science to resolve everyday questions.

A randomized control trial is simply when researchers split volunteers into two groups, testing an intervention--a new drug, say--with one group, but not the other. By randomizing selection, and comparing the results, you can see whether the intervention works as supposed.

To set up an experiment at Randomise Me, you start by setting a question and choosing a test style. You can measure results with a score (say, 0 to 10 for happiness intensity), with a yes/no answer (yes dressing like a girl does improve my game), or by counting (how many times did you dream of Barry Manilow last night?). You then decide how many observations you need for the research to be "statistically significant," and set some instructions for participants. At that point, you can hit the "randomise" button and start assigning tasks. The site will automatically send out reminders (e.g. "check the color of your urine this morning"), and will help you report the results in a coherent manner.

The site was created by Ben Goldacre, a columnist for the U.K. Guardian, and his colleague Carl Reynolds. Goldacre writes:

Our ambition is that Randomise Me will enable anyone to set up and turn their own trial, helping to robustly test different issues and interventions in an agile and low-cost way. This is a great leap forward for anyone wanting to know the impacts of something in their lives.

Indeed. The site is another example of the democratization of science, and a potentially very useful way of demystifying vexing questions. Time to go random!

Add New Comment

1 Comments

  • Josh

    Yes cool and everything, but still doesn't resolve the age-old correlation/causation conundrum - in fact, it seems almost to obscure it behind a veil of the ease, and therefore validity, of 'democratised science'