Leveraging Pro-Social Behavior for Public Service Delivery

The COVID-19 pandemic widened pre-existing educational inequalities. School closures hit hardest the students already failing to meet minimum proficiency levels — those from socioeconomically disadvantaged backgrounds. The question of how to intervene effectively, and at scale, remains open.

One answer came from volunteer labor. The Tutoring Online Program, designed by Eliana La Ferrara (Bocconi) and Michela Carlana (Harvard), paired university student volunteers with struggling middle schoolers during Italy’s first lockdown in 2020. The pilot — 530 students, two months — produced measurable improvements in educational performance, aspirations, socio-emotional skills, and psychological well-being.

I managed the second edition, which ran from November 2020 to May 2021 and served 810 students. The core challenge was sustainability: could altruistic behavior outlast the emergency that had generated it? My prior was that the pilot’s success partly reflected a temporary convergence — genuine solidarity, surplus free time, the acute salience of collective crisis. I expected recruitment and retention to get harder as conditions normalized.

I was wrong on recruitment. Applications came in fast.

Retention was more interesting. Tutors who struggled to keep disengaged students motivated were also the ones most visibly affected by reminders of what their contribution meant. This suggests something worth investigating: intrinsic motivation, when reinforced through framing and feedback, may be more robust than standard volunteer labor models predict. But we had no variation in recruitment design, so selection effects remain unidentified.

Does the model have a role beyond the pandemic?

Three reasons to think yes.

–> The supply of high-quality tutors is geographically concentrated near universities and urban labor markets. An online model decouples supply from geography, extending access where the local tutor market is thin.

–> Even where tutors exist, access is stratified. Families with limited social capital or financial constraints are least able to procure private tutoring — and their children are the ones who would benefit most.

–> The mentorship dimension matters too. Students from disadvantaged backgrounds often make consequential decisions — secondary school track, whether to pursue university — with poor information and under the weight of low expectations set by their environment. A tutor can partially correct this.

Pro-social behavior, when channeled through a well-designed program, can deliver things neither the market nor the state provides efficiently. The conditions seem to be simple: a credible institutional frame, a legible social impact, and some mechanism for sustaining a sense of meaning over time. Whether this scales is an empirical question worth asking.