"Can you hear me now?": Experimental evidence on improving public service delivery through non-electoral citizen participation

We will evaluate the impact of a series of four community meetings led by trained social organizers from the National Rural Support Programme (NRSP). These meetings help overcome both the informational and coordination constraints preventing citizens from engaging with government actors regarding policy preferences, and serve to mobilize women and men through structured activities that encourage them to identify a problem in their local public primary school, and construct and deliver a message regarding this issue to a policy actor. Our treatment will vary across the following dimensions:

Actor Type. In the politician version of the treatment, citizens will be provided with information about electoral accountability, the political system, and the contact details of their assembly member. In the bureaucrat version of the treatment, citizens will be provided with information about the education bureaucracy and the contact details of the relevant education bureaucrat.

Participant Gender. Our pilot work has revealed that women are more informed about education issues than men, but face greater obstacles in engaging with policy actors. Since our intervention helps with both informational and access constraints, we will test our intervention separately with community meetings along gender lines -- either all men or all women -- which will allow us to examine which of the two groups show greater impact.

Meeting facilitation. Based on our pilot studies, the impact of the community mobilization interventions may be enhanced if there is more explicit and observed support by socially motivated and relatively well-connected outside actors. Our implementing partner, NRSP, is one such actor. We test a facilitated version of our interventions where NRSP directly works with the local communities in constructing their message and accompanies them in the delivery of the message to the relevant policy actor. This may help citizens overcome a sense of low efficacy or fear of retribution from policy actors.

2023-06-20

2023-12-31

Citizen action on their education problem, problem resolution

We will randomize treatment -- holding of community meetings -- at the village level. Within each district, randomization is stratified by the gender of the school in the village (boys' government primary school or girls' government primary school), such that our sample has an equal number of villages with boys' schools and villages with girls' schools. Within each strata in each district, we randomly assign:
- 1/3 of villages to pure control
- 1/12 villages to female participants-politician policy actor-facilitated treatment
- 1/12 villages to female participants-politician policy actor-unfacilitated treatment
- 1/12 villages to female participants-bureaucrat policy actor-facilitated treatment
- 1/12 villages to female participants-bureaucrat policy actor-unfacilitated treatment
- 1/12 villages to male participants-politician policy actor-facilitated treatment
- 1/12 villages to male participants-politician policy actor-unfacilitated treatment
- 1/12 villages to male participants-politician policy actor-facilitated treatment
- 1/12 villages to male participants-politician policy actor-unfacilitated treatment

Not available

Randomization done in office by a computer

Village

Yes

480 villages

6,400 villagers; 1 observation pre-treatment, 2 observations post-treatment

160 villages control, 40 male/unfacilitated, 40 male/facilitated, 40 female/facilitated, 40 female/unfacilitated, 40 politician/unfacilitated, 40 politician/facilitated, 40 bureaucrat/unfacilitated, 40 politician/facilitated

Phase_I_Pre_analysis_Plan.pdf