One of my favourite people to follow on Twitter is a man called Ben Goldacre, a doctor and writer who's best known for his Bad Science column, originally published in the Guardian. He specialises in attacking the misuse of statistics, often by the media.
I mention this because he, along with three other experts, has written a new report for the Cabinet Office on how Randomised Control Trials (RCT's) can be implemented in areas of Public Policy. In case you're new to research skills RCT's are where you basically take two (well, usually two) groups and introduce some particular stimulus (for example, participation in a particular drug rehab programme for people suffering with drug problems) for half the participants.
The 'random' part of dividing the RCT's population into groups is crucial. Without that element you'll often get a skewed result due to volunteers being positively biased to succeed to begin with. At the end of the trial you try and ascertain whether the drug programme group are healthier/happier/ than the control group who haven't taken part in the programme. In a nutshell, this supposedly simple method is what we call 'evidence based practice'.The new report looking into how RCT's can be used in the area of Public Policy is called 'Test, Learn, Adapt'. I think it's really interesting and thankfully it's pretty straightforward to comprehend. Read or download it from here.
I'll happily talk to you about statistics and research methods if you need assistance with that side of things. At the very least I can suggest some decent resources to make the issues clearer.
Right. You can get back to your summer now...
No comments:
Post a Comment