Learning through behvavioural science research – our new report

Authors: Nathalie E.J. Dijkman, James Kengah, Gideon Too

Find full report here.

The battle is half-won.

Using behavioral science to improve citizen feedback for better public service delivery.

For the few people deep enough in the geekdom, GI Joe will forever be known for breaking the boundary of “dolls for boys” by being the first ever action figure – and introducing the term ‘action figure’ to the market. The rest of us might not be too familiar with all this – or even care about dolls and action figures, but odds are we have come across the phrase ‘knowing is half the battle.’ It was brought about a series of public services announcements done by the company in 1985 to educate children on general safety. The videos generally set up an unsafe situation, encouraging children to pause, think and reflect on situations rather than taking arbitrary action. They always ended with a child going “oh, now I know!” To which the protagonist (GI Joe or one of his crew) would respond “and knowing is half the battle.” 

With challenges such as lack of adequate capacity and insufficient data, governments are often trying to get services delivered to their people amidst a sea of uncertainty. Feedback from the people allows governments to know what part of their service delivery needs improvement. Like the children in the GI Joe videos, governments often need the support of its partners with the ‘knowing’ part of this battle. 

SEMA is one organization that helps the government in this way. Based in Kampala, SEMA is a civic tech organization that collects, synthesizes and reports citizen feedback on public services such as police stations and health centers. Citizens are encouraged to rate their latest experience at a public office by using a custom-made feedback device with five buttons: from happy to sad. In addition, mobile and in-person surveys are held with citizens about their experiences at these public offices. Their previous work shows that  public offices where citizen feedback was collected and reported improved citizen’s satisfaction for the first year post collection. We were brought in to use behavioral science to help figure out how this level of citizen satisfaction can be sustained and/or surpassed in subsequent years as well as how citizen engagement with feedback tools can be improved.

To  build a holistic view of people’s experiences engaging with SEMA’s feedback mechanisms and reporting structures across different key offices, we held in-depth interviews and focus group discussions with 51 citizens, citizen feedback collectors and public officers. We then tested the impact of having police officers acting as champions by pushing for increased use of feedback tools within their police station. The champions were trained on using the SEMA custom-made feedback device and then mandated with sharing knowledge to their fellow police officers on how to use the device. Our idea was simple – having champions within the police station was likely to increase the number of police that know how to use the SEMA feedback device. By increasing the knowledge and awareness of the police we were hoping to improve the quantity and quality of feedback delivered by citizens through the feedback device as well as increase the awareness of SEMA among the police officers.

We recruited a total of 88 police officers from 4 police stations in Kampala, randomizing the stations into either a treatment group or control group. A total of 38 police officers were in the treatment group and 50 police officers in the control group. Each group had 2 police stations. After observing the intervention for 2 weeks we deployed phone surveys with police officers to capture outcomes of interest. 

Barriers to effective citizen feedback

Faced by citizens

Faced by public officers

Mistrust of public offices: Citizens perceive that their welfare is not a priority for public offices. This not only demotivates citizens from giving feedback on the services provided but also from accessing these services in the first place. 

Lack of systematic feedback mechanisms: Public offices do not have clear systems or structures to encourage citizen feedback on their services. New tools like SEMA’s are therefore needed. 

Fear of repercussions: Citizens have an instilled fear of consequences from raising any concerns on public services, especially when their anonymity is not assured at the point of feedback.

Feedback is not processed properly: This research showed that officers at different levels of rank are interested in hearing what citizens have to say. However, feedback doesn’t always reach all officers, and there are varied levels of engagement with feedback reports like SEMA’s. The format of how feedback is presented, to whom it is presented and how often, can have a big impact on the reaction of public servants

Corruption/ Bribery: the process was undermined by both citizens and police officers who would prefer to create an environment conducive to corruption.  

Lack of resources: This cuts across budget constraints and understaffed offices both of which limit the extent to which a public office can implement some of the recommendations stated in citizen feedback reports.

Inaccessible options for feedback: Design of the feedback mechanisms rarely take language barriers and people with disabilities into consideration.

Bureaucracy: There is little decentralization in the public sector resulting in long unstructured processes to approve even small budgets and initiatives. Multiple requests stream upwards to senior management (sometimes beyond the institution itself) for consideration causing long avoidable delays


Surprisingly, most officers knew how to use the feedback device, but the champions did not have an impact on this. The existing knowledge could probably be related to previous SEMA visits at the stations.

Stations with a champion had more valid feedback than stations without a champion. This shows that public service champions can increase the issuance of valid feedback at public offices. Once training has happened, having a feedback champion may lead to an increase in the uptake of citizen feedback in public offices. However, having a champion does not prevent the misuse of feedback mechanisms, i.e. fraudulent use of feedback tools, such as having officers giving themselves positive feedback. 

While an internal champion was found to be effective, the research further showed that public officers prefer being encouraged to offer better public services by an external person, rather than a fellow public officer. They also prefer that this reminder happens regularly. This means that there’s a real need for CSOs to maintain regular contact with public officers, reminding them of the importance of good public service delivery.

Where do we go from here?

Governments are, ultimately, working to serve the needs of their people. Effective feedback mechanisms ensure that the governments know, and are responding accurately to, these needs. Work like this allows governments to begin to tackle the barriers to effective feedback. Beyond barriers, CSOs need support to provide additional learnings on the process of efficient public service delivery. Further research is also needed into the effect of citizen feedback tools and reports on the perceptions of citizens on public service delivery.  Using behavioral science CSOs can help governments understand and influence the relationships between civil servants and citizens, ultimately gaining insight into how to increase trust in public service providers – and knowing is half the battle. 

Contact SEMA for more information.

info@talktosema.org / www.talktosema.org

Connect with SEMA on Facebook, Twitter and LinkedIn.

Contact Busara for more information

contact@busaracenter.org / www.busaracenter.org   

Connect with Busara on Facebook, Twitter and LinkedIn.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>