Innovator Preferences and Catastrophes: Evidence from a Climate Change Field Experiment

How to cite

Sen, Ananya; Oh, Jean; Guzman, Jorge. 2020. "Innovator Preferences and Catastrophes: Evidence from a Climate Change Field Experiment." AEA RCT Registry. AEARCTR-0005743. https://doi.org/10.1257/rct.5743-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Ananya Sen* , Jean Oh , Jorge Guzman

* Primary investigator

Registry ID
AEARCTR-0005743
DOI
https://doi.org/10.1257/rct.5743-1.0
First registered
Last updated
Status
In development
Trial dates
2020-05-05 – 2020-10-27
Countries
Private
JEL codes
O31, O35, Q01, Q10, Q54, Q55
Registry record
http://www.socialscienceregistry.org/trials/5743

Abstract

Human existence faces a series of catastrophic risks at a global level, including pandemics, climate catastrophes, or nuclear and bioterrorism. It is widely recognized that mitigating these risks include important social choices (such as, reducing economic growth), incentive design and coordination, and innovation (Nordhaus, 2019). Amongst these, innovation is particularly appealing. While the other two simply solve an externality by internalizing the costs of climate impact, innovation can directly reduce the risks, adapt to new changes, or stop them all-together. At the center of this promise stands the option of connecting innovators to climate problems, and learning who and when do they care about them. Yet, to date, little evidence exist on this topic. The purpose of this paper is to use a novel pre-registered experiment to understand the ways in which innovators display interest around several parameters of catastrophic risks, by focusing specifically on the risk of climate change.

We will perform a randomized control trial sending messages to a series of high profile innovators to get them to click to learn more about the MIT Solve challenge related to climate change. In these messages, we will vary use a two-by-two design where we include a phrase around the human cost of global warming. We will vary the time of this cost (i.e., framed as a present or future cost), and its magnitude in terms of the estimated number of human lives affected. This will allow us to separate two fundamental parameters of the human 'preferences' for solving global catastrophes, namely time preferences and the magnitude effect.

Our study will use two samples, innovators and non-innovators. The innovator sample will be individuals who have been approved by the US government for a Small Business Innovation Research grant (SBIR) or a Small Business Technology Transfer grant (STTR) in recent years. These represent highly innovative technologies that the US government agencies directly provide financial support for the subsequent development. The grantees include the inventor (principal investigator) and the manager (business contact) at each firm, who are listed publicly on the SBIR website by the U.S. government from which we have downloaded the data. We intend to email these individuals messages using one of four treatments. We will then track outcome variables reflective of their interest such as: a) whether they click, b) how long do they stay in the Solve website, and c) whether they eventually submit an application. The non-innovator sample will be a general population sample in the U.S. This sample will be recruited through Amazon Mechanical Turk and paid market price to see one of four treatment messages and respond with their interest by clicking and answering a survey questionnaire.

We expect to find that, on average, individuals discount catastrophic risks that are further into the future, and that they will be more interested as the magnitude of the risk increases. We are also interested in the heterogeneous treatment effects across different types of individuals, namely inventors, managers, and inventor-managers in the innovator sample as well as general populations in the non-innovator sample.

Intervention

The intervention for the innovator sample will involve an email encouraging individuals to learn more and apply to MIT Solve's Sustainable Food Systems Challenge. In doing so, our intervention will emphasize a specific global catastrophe, namely the impact of climate change, to garner interest in solving the problem. To elicit individuals' time preference and sensitivity to the human cost of climate change, the intervention will follow a two-by-two design where we vary the time frame of this cost (i.e. present or future), and the magnitude in terms of the estimated number of human lives at risk due to climate change. This will allow us to separate two fundamental parameters of the human 'preferences' for solving global catastrophes, namely time discounting and the magnitude effect.

The intervention for the non-innovator sample will be similar to the innovator sample in terms of messaging; the difference will be that individuals will be recruited on MTurk to see the intervention as opposed to receive it via email.

Experimental design

Our emails will emphasize two different dimensions of the catastrophic risk of climate change: timeframe and magnitude of the risk. We will randomly assign individuals with one of four treatment arms that vary along these two dimensions. The four types of subject lines are as follows. Similar messages will be included in the preview and body of the email.

1. Save 150,000 Lives in 2020 from Climate Change: Apply to MIT Solve’s Challenge
2. Save 150,000 Lives in 2050 from Climate Change: Apply to MIT Solve’s Challenge
3. Save 400,000 Lives in 2020 from Climate Change: Apply to MIT Solve’s Challenge
4. Save 400,000 Lives in 2050 from Climate Change: Apply to MIT Solve’s Challenge

Design details

Randomization method
Randomization will be done in an office by a computer.
Randomization unit
The unit of randomization is the individual. Individuals will be block randomized within type of role (inventor, manager, or inventor-manager) based on whether they are listed as the principal investigator, business contact, or both in the SBIR/STTR grant.
Clusters
30,000 individuals in the innovator sample 20,000 individuals in the non-innovator general population sample
Observations
30,000 individuals in the innovator sample 20,000 individuals in the non-innovator general population sample
Arms
7,500 individuals for each of four treatment arms in the innovator sample 5,000 individuals for each of four treatment arms in the non-innovator sample
Primary outcomes
The main outcomes of interest for the innovator sample: (1) Open rate or clicks on emails (2) Browsing on the MIT Solve website (3) Application rates based on surrogate indices (2) Time spent on website, number of pages visited We also aim to create a surrogate index to link email and website behaviors to application rates.

Cited by

No citation data available for this registration yet.

Keywords