A few years ago a new website was launched by Ben Goldacre to help facilitate the kind of randomized control trials (RCT) that are at the heart of evidence based decision making. The site, called Randomise Me, was designed to make setting up a RCT in as easy a way as digital natives can use platforms like Google Optimizer.
Unfortunately the site failed to really take off and the domain has since lapsed. The Behavioural Insights Team (BIT) hopes to do better with a new tool that they’ve calling Test+Build. The tool was born out of the extensive experimentation BIT do as a core part of their work, and they wanted to see if they could develop a tool that automates much of the work behind a RCT.
Test and build
The tool is specifically designed to help with the improvement of tax collection by using behavioral science. It supports jurisdictions in the creation of randomized control trials by guiding people through four stages of the TEST methodology developed by BIT (Target, Explore, Solution and Trial).
At each of these stages the user is supported via guides, case studies and videos to showcase compliance and enforcement. BIT researchers are also available to provide their advice as well as conducting randomizations and analyzing the results. The hope is that it will encourage organizations to run their own trials.
“By letting users work through the process themselves, with support from BIT researchers at key points along the way, we’ve significantly reduced the cost to organisations of running a BIT trial – by about 50 per cent,” the team say.
True to their ethos, the BIT team tested Test+Build themselves on a number of users and projects. They found that it significantly reduced the cost of running a trial by around 50%. They also tested it in the field with Medway Council, and the tool allowed staff to test a couple of messages to encourage council taxpayers to use Direct Debit to settle their account.
“The Test+Build platform was so easy to use and so informative. It helped us clearly define our project objective and design the interventions. Additionally, the Test+Build team provided excellent support, sage advice and carried out the complex analysis giving us confidence that the excellent results were real and not just a mere coincidence,” the council say.
RCTs are crucial if we are to have evidence based decision making, so hopefully Test+Build will fare better than RandomiseMe did before it. If you’d like to try it out then you can do so via the url linked to above.
