Experimental Design Details
As part of this randomized control trial, we have developed a custom-built mobile phone application called AirPulse, which will service as a one-stop shop for us to conduct surveys, implement interventions, and pay respondents for their time. We argue that this is a valuable way to conduct our research because (i) the population we are targeting will be smart-phone savvy, (ii) it reduces the need for enumerators and complications with payouts to participants, (iii) it enables us to easily conduct follow-up surveys and randomized interventions with participants, and (iv) it allows us to measure private pollution avoidance behavior by observing users’ movements in space (with their consent).
Our sample size and treatment group size will depend on the size of recruitment through social media channels (like Google ads, Meta ads, and civic Whatsapp groups). The study will be advertized through two types of ads including a general ad to recruit participants for a study about local issues and a targeted ad that mentions recruitment for an air pollution study. We will stratify treatment by the type of ad that attracted participants to test for selection and inherent tendencies.
At the beginning of the experiment, we will elicit a baseline estimate of the respondents’ usage of protective investments like face masks and air purifiers. We will also ascertain their willingness to pay for direct (as in, not delivered through policy changes) improvements in air quality through incentivized offers for air purifiers and face masks that protect from air pollution, their hypothetical willingness to support policies that reduce air pollution but impose taxes, and preferences for hypothetical politicians with different priorities concerning air pollution. Participants will receive “credits” for answering our surveys that can be cashed out at the end of the study period or whenever they choose to quit the study.
Our information interventions are designed to meaningfully increase respondents’ willingness to engage in civic action by solving possible market failures that could stand in their way. Importantly, this allows us to assess whether technological advances aid widespread information dissemination that can change the demand for clean air, facilitating the alignment of policy priorities with citizen’s environmental preferences. All our treatment arms will examine potential information barriers:
1. To test for the possibility that individuals are poorly informed about air pollution, this group will be provided with periodic reminders of the current level of air pollution as well as its health effects.
2. To test for the possibility that respondents don’t think that there are viable paths to pollution reduction by engaging with policy, this arm will describe several concrete policy changes and local state actions that could substantially reduce air pollution and benefit from their engagement.
3. To measure the effect of only providing local air quality information on mobilizing demand for air quality as a yardstick for our other treatments
All treatment groups will receive information on different actions they can take to directly engage with civic agencies and local governments – privately or publicly, or indirectly through petitions. For this, we will examine participants’ willingness to participate in a range of civic actions targeted to improve air quality. We will give participants information about the different channels through which they can advocate for improvements in air quality. These avenues are: 1) the Delhi government’s Green Delhi app that allows citizens to share their pollution grievances privately, 2) links to publicly available social media accounts of nearby air quality advocates and regulators 3) signing online petitions to improve Delhi’s air quality, and 4) joining a Facebook group called "Help Delhi Breathe" that keep citizens informed and organized on air pollution policy. All these avenues are accessible to citizens even in the absence of our study. Our experiment will only nudge them by providing information about these platforms. We will vary the effort and time cost (but not monetary costs) of engaging in these actions through our app by embedding the ease with which these action buttons are accessible for a random subset of our participants to see if the users are willing to act despite the added friction introduced by the app. This is to test whether respondents are willing to send costless signals but that willingness to participate quickly declines as the cost of the action rises in terms of effort or time. Such nudges will be introduced regularly (likely, weekly). Additionally, every day we will ask citizens about how they are feeling about the local air quality and twice a week if they have observed any government actions towards reducing air pollution in their neighborhood recently (example: smog guns, action against diffused polluters etc.).
The endline survey will cover additional questions on shifting beliefs and preferences (like vignettes for preferences for political candidates, stated preferences on willingness to pay for private and public pollution mitigation actions) as well as revealed willingness to pay for private adaptation to air pollution by entering a lottery to win an air purifier examine the degree to which these interventions served to shift beliefs and preferences for pollution regulation policies relative to a control group that didn’t receive any treatments. The study period will last for 3 months. We are planning to recruit respondents and conduct the baseline in January 2024. The winter is always the most polluted time of the year in New Delhi – so we are planning to time our study period to overlap with that.