Behavioural insights

pattern

VicHealth - Tailored messages to recruit survey participants

Problem

Every year the VicHealth Indicators Survey asks close to 23,000 Victorians about their mental wellbeing, physical activity, diet, alcohol consumption, and other health indicators. A challenge in delivering the survey is that a large, representative number of Victorians must participate for the answers to be valid.

Solution

In 2015, VicHealth employed behavioural insights to trial a new set of SMS messages asking participants to partake in the survey. Three key elements were incorporated into the new SMS text messages, namely:

  • The Survey will improve our understanding of health' – messages highlighting the contribution of the survey to a bigger goal – collective understanding of the health of their community
  • We need more people in your area to participate' – messages intended to incorporate a sense of urgency
  • 'No change' - a control message that incorporated no behaviourally insightful alteration.

Behavioural insight

The SMS trial drew upon behavioural insights in two ways. First, it drew upon psychology and sociology to inform to what messages people would most likely respond well. We know from psychology that people can be motivated by factors such as altruism, social norms, urgency and personal gain depending on the situation. The SMS trial incorporated this by designing its SMSes to appeal to altruism or to a sense of urgency.

Second, the SMS trial drew upon the behavioural insights methodology of testing what works. The psychological literature suggested that altruism and urgency should motivate participants to take the survey. However, the only way to know for sure was to test it. Therefore, VicHealth sent to participants one of three text messages. The SMS that generated the most responses could be demonstrably shown to be more effective than the other SMS messages.

Result

Early results indicate that people who received an SMS about needing more people in their local area to take part (21 per cent) were marginally, but significantly less likely to respond to the message compared to the improving our understanding of different health contributors (23 per cent) message and the control (23 per cent) message. We hypothesise that this unexpected result was because this message inadvertently conveyed a social norm: most people in your area are not filling in this Survey, conveying a message to people that they shouldn't either.

While it was unanticipated that one of the changed messages would reduce response rates, the ability to test these messages provided VicHealth with valuable information. Based on these results, VicHealth did not roll out a campaign based on messaging that was less effective than the 'business as usual' messaging.

The project demonstrated two advantages of behavioural insights. First, VicHealth was able to design SMS messages based on what behavioural science would suggest are effective communication techniques to motivate survey participation. Second, VicHealth was able to measure the impact of different SMS message types by using a behavioural insights trial design – assigning participants at random to three groups and comparing the differences in outcomes.