News & Events

Building evidence through natural experiments or events

September 13, 2018

Many large-scale social programs and policies are thought to affect health, but rigorous evidence demonstrating such impact is rare. Changes in social programs or policies can present unique opportunities to establish their impacts on health outcomes. The purpose of this announcement is to highlight the opportunity for funding research that takes advantage of natural events or experiments to rigorously evaluate the health impacts of such interventions.

Demonstrating causal effect generally requires baseline data gathered before a change goes into effect from populations who will be affected by the new policy or program and from comparison groups who will not be affected. Traditional funding mechanisms may not have the flexibility required to respond to funding requests in time to collect relevant baseline data within the implementation timeline, or to encourage research or implementation design features that would strengthen evaluation. As a rolling submission program, Evidence for Action (E4A) can more readily respond to research requests that involve natural experiments or events, such as those anticipated as a result of a policy or program decision. 

Scenarios that occur under real world “natural” conditions offer valuable opportunities to understand the impact of shifts in practice on population outcomes. Specific implementation features, such as lotteries when there are insufficient resources to serve all eligible individuals, are ideal to deliver convincing evidence on how a policy or program affects health. However, settings where there is an essentially comparable comparison group and the timing of a new policy is arbitrary can also provide convincing evidence (e.g., through a difference in difference or similar quasi-experimental design). 

Examples of anticipated changes in policies or practices that could be considered natural experiments or events and that have important implications for population health, well-being, and equity are described below. These examples are not meant to be prescriptive or exclusionary, but merely to illustrate potential opportunities to investigate relationships between impending naturally occurring interventions and health.

Example: Public benefit work requirements. Public benefit programs such as Temporary Assistance for Needy Families (TANF), the Supplemental Nutrition Assistance Program (SNAP), and Medicaid are part of the social safety net for low-income populations. Both TANF and SNAP require participation in a minimum level of employment-related activities, or “work requirements,” for certain groups such as able-bodied adults; and states were recently granted the option to introduce work requirements for Medicaid enrollment. Currently, federal, state, and local authorities are considering additional new work requirements for some programs and reinstating requirements that had been temporarily suspended during the economic recession. Variation among states and counties in the timing or enforcement of work requirements could be used to examine the impacts of work requirements on health of participants and their households. 

Example: School start times. As evidence on the impact of sleep on adolescent health and academic outcomes has mounted, school districts across the country have begun considering changing school start times. Some places, like Fairfax County VA, Cherry Creek, CO, and Durham, NC, have already implemented later start times for middle and high schools; and the California state legislature recently passed a bill that would impose start times of no earlier than 8:30am for all public middle and high schools across the state. While there is already substantial evidence about adolescent sleep and health, less is known about how later school start times for older children impact broader family and community outcomes and dynamics. Variation in the timing of when districts implement new start times creates the opportunity for matched comparison sites or longitudinal data analysis. Learn more about an evaluation E4A has already funded about this issue.

Example: Bail reform. A handful of states, including New Jersey, Kentucky, New Mexico, and California have recently adopted changes to their criminal justice systems, including reducing or eliminating the role of money (or cash) bail in determining whether a defendant can be released or detained in jail while awaiting trial. Such changes are intended to reduce mass incarceration among individuals and groups who pose no elevated risk to the public but do not have the ability to post bail. However, there is insufficient evidence that alternative systems to assess risk are effective or equitable. The variations in new bail reform measures across states, as well as timing of implementation, may enable investigators to identify the causal impacts of such practices on the health of individuals, families, and communities.

How to apply
Those interested in applying for funding to evaluate a natural experiment or event may submit a Letter of Intent through the E4A Call for Proposals.There is no deadline for submission. Applications are accepted on a rolling basis and will be reviewed as part of our normal competitive process. Applicants should indicate if a proposal is time-sensitive based on a naturally occurring event or experiment. Visit the E4A website for further guidance and instructions.