15  Steps in a randomised trial

You have designed a behavioural intervention. Here are the steps to test it in a randomised trial:

• Define the outcome measure: What outcomes will be used to measure impact?
• Random assignment: You need to allocate participants into a treatment and a control group.
• Data collection plan: Collecting data is complex and you need a detailed plan on how this will take place.
• Pilot: Although not essential, it is good practice to pilot your intervention and survey instrument before rolling them at scale.
• Trial Implementation: It can be useful to put in place a process evaluation to know whether the program is being implemented as planned and how well it is functioning.
• Data analysis: After you have implemented the intervention and completed your data collection, it is time to analyse the data to assess whether it achieved its core objectives.

We’ll discuss these steps in more detail over the following pages.