A randomized control trial or randomized controlled trial (RCT) is a type of scientific experiment used primarily to determine the effect of medicines or other types of non-medical treatment on a population. The basic idea of a randomized controlled trial is to randomly offer treatment to a group of individuals (treatment group) and a placebo or nothing at all to another one (control group). The statistical properties of randomization allow both the treatment group and the control group to have identical characteristics and, in this way, the impact that the treatment has on the group that received it can be identified with respect to the one that did not receive it.
In the context of Supporting Experimentation in Innovation Agencies (INNOSUP-06-2018), grants to national and regional innovation agencies are provided to engage in policy experimentation. The tests aim to design and test new or significantly improved schemes in support of innovation in SMEs by using randomised control trials. There are not many good practices to follow or examples that could be presented when it comes to RCTs in the field of innovation and business support, since it has not been applied as much as in other fields, and the projects approved under this call started very recently.
In policymaking, the importance of RCTs is the robust evidence-based approach they offer. The adoption of new policies intervention as the best option depend on different factors. First of all, whether the stakeholders will react as one expects; second, if there will be non-foreseen consequences; and third, if there will be any rebound effect or impact on the grains. Evidence-based policymaking and scientific proof look for sustainable intervention. Relying on methods such as RCTs, the highest quality of evidence is ensured, rather than relying on perceptions collected through surveys.
The importance of randomisation is that, if strong, it is possible to establish a causal relationship between the treatment (policy intervention) and the net effect observed. Nevertheless, the main characteristics of the SMEs in both groups, control and treatment, should be comparable. Randomisation will thus ensure that similar characteristics resulting in the impact on the relationship between the intervention and the outcome will roughly be equal in both the treatment and control group, so the bias will be minimised. The European Commission explains the 9 essential steps to carry out a successful RCT:
1. Identify the policy intervention to be evaluated and compare:
-The new one against the status quo;
-Two different versions of the new one.
2. Define the expected outcome;
3. Determinate the randomisation unit
-Who or what are going to be randomised?
4. Determine how many SMEs are required to obtain robust results;
5. Randomly attribute SMEs into the control and treatment group;
6. Collect the baseline data;
7. Deploy the policy intervention for the treatment group;
8. Measure the impact and compare the results;
9. If the policy intervention worked, extend it to all SMEs
Given this information, one can check more information on the already closed call INNOSUP-06-2018 and see the list of projects approved that will conduct RCTs by following this link.
You can check this document written in 2018 to give background on RCTs. It contains further readings on the topic, such as
If you want to see an example on RCT, you can watch this video of a study conducted by UNICEF.
Written by Roser Torres, Project Officer at EURADA