How personalized text messages increased fine repayments by 30%

Nudges: Interventions where behavioral architects change aspects of the choice architecture—as opposed to aspects of the choices themselves—to lead people to choose in a certain way.

Randomized Controlled Trial (RCT): an experiment or study which is designed to isolate the effects of an intervention (personalized text messages) on an outcome (fine repayment) in order to evaluate the impact of the intervention.

The Courts Service in the UK wanted to find a way to encourage people to pay their court fines before sending a bailiff to their homes. Bailiff interventions can be relatively costly and time consuming, so the Courts Service wanted to see if they could identify a more efficient way to collect fine repayments.

The basics of RCTs

Randomized controlled trials can help isolate the effects of an intervention to more accurately test the effectiveness of a new program. RCTs do this by obtaining a sample of individuals or groups and randomly assigning them into a control group and intervention group, which will eliminate the possibility of the observed results being affected by external factors. This means that any observed effect should only be caused by the interventions that each group receives. There can be more than one intervention group, and each receives one type of intervention. The control group does not receive any intervention. Think of the control group as a placebo condition in a clinical drug trial.

Text messaging to encourage fine repayments

The RCT run in the United Kingdom by the Courts Service and the Cabinet Office’s Behavioural Insights Team wanted to test the effectiveness of nudges, in the form of text messages, to encourage people to pay their overdue court fines. This was in comparison to the current situation, in which no text reminders were sent.

The team tested multiple variations of text messages to find the one that was most effective. The trial consisted of five different groups. There were four intervention groups, each receiving one of the following four interventions:

• A standard text 
• A standard text with the specified amount
• A standard text with the recipient’s name
• A standard text with the specified amount and the recipient’s name

The fifth group, the control group, were not sent text messages. Randomization was used to assign each participant an intervention group.

All four interventions increased the likelihood of citizens repaying their overdue fines, suggesting that prompts were highly effective. The standard text resulted in a response rate of 23%, while there was a 25% response rate from the standard text with the specified amount. For texts sent with the recipient’s name, there was a 33% response rate, followed by a 29% response rate for texts sent with the recipient’s name and the specified amount. In contrast, the control group saw a response rate of 5%.

Following this RCT, a second trial was conducted with more participants to further identify the specific aspects of personalised text messages which were most effective at increasing fine repayments. Findings from both trials show that the most effective intervention was the standard text message with the citizen’s name, resulting in an increase of more than 30% in the average value of fine repayments.

If rolled out, sending a personalized text message instead of a standard text message is estimated to bring in £3 million annually in addition to reducing 150,000 bailiff interventions annually. 

• The results of the study were significant, highlighting the positive effects of personalized text messages in increasing fine repayments.
• The intervention’s plan to monitor the validity of the results was impressive, as the RCT, which was robust and well designed, was able to isolate the effects of the different variations of the text message intervention against the control group.
• There was no mention of the diversity or heterogeneity of the participants of the intervention, in terms of gender, age, or wealth etc.

Nudges Work (and practitioners know exactly how well): Dr. Linos talks about an investigation done with Dr. DellaVigna on the effectiveness of nudging based on data provided by two of the largest Nudge units in North America.

Helping the City of Rome improve citizen engagement: With fewer people than expected participating in the city’s campaign to encourage citizens to vote in Rome’s participatory budget, the municipal administration seeked to launch an internal behavioral insights unit, with the goal of increasing citizen engagement through more personalized messaging.

How Behavioral Science Informs Policymaking: Dr. Ball talks about how research findings have raised questions about whether the focus on the method of evaluation, such as RCTs, was influencing the types of projects these teams were working on, therefore shaping how behavioral insights were informing the design and implementation of policy.