Experimental Design Details
In this study, we collaborate with a Japanese company that provides a smartphone app for citizens to collect, post, and record species information, and conduct a field experiment with its users. The app was launched in May 2017, and the number of users exceeded 400,000 in April 2022.
When users take photos of plants and animals they see outdoors and post them to the app, the app determines their names and types based on the AI. At the same time, the app records its species information as well as the time and place of taking and posting the photos. Citizens collecting species information and sharing it to the public will lead to a better understanding for what kind of, where, and how many species, including rare and invasive ones, are living. Ascertaining this actual situation is essential for biodiversity conservation.
We collect participants for this field experiment in the following ways: We post a recruitment ad on the front page of the app. On the transition page, we obtain opt-in participation consent from users. The target number of participants is around 900, based on our research budget, etc.
After obtaining their participation consent, we then ask them to complete the first survey. The survey includes a set of questions to ascertain their socioeconomic demographics, as well as their use of the app, daily steps, mental health, reasons for participating in this event, etc.
Next, we randomly assign them to either of the control group, prosocial treatment group, or financial treatment group (300 participants x 3 groups). We set a two-week treatment period, during which those assigned to the prosocial treatment group, or financial treatment group receive the following interventions, respectively:
Prosocial treatment: Based on the number of species information posted by each user, the experimenter donates to activities related to biodiversity conservation.
Financial treatment: Based on the number of species information posted by each user, the experimenter gives them financial rewards that can be used for their online shopping.
After the treatment period, we conduct the second survey to again ascertain their daily steps, mental health, etc.
Using randomization and difference-in-differences approach, we measure the impact of each treatment on changes in the number of posted species information and that of posted species photos. We also measure the impact of each intervention on changes in daily steps and mental health.
Among the outcome variables, the number of posted species information and that of posted species photos are automatically recorded by the app. The analysis uses for the variables the difference between the baseline period prior to participation and the treatment period. Daily steps and mental health are collected in a self-report format in two surveys, as described above. The analysis uses for the variables the difference between the two surveys.