OSHA currently markets its On-Site Consultation program (OSC) in part by sending brochures explaining the program’s benefits to eligible businesses.  Working in collaboration with the Department of Labor and academic marketing experts, we designed a study to test whether a redesign of the brochure and/or different modes of communication could increase the number of requests for enrollment. Instead of simply comparing different marketing strategies wholesale, the study will determine what underlying factors cause some approaches to work better than others. For example, do fear-based appeals work better than appeals to a person’s feelings of autonomy and competence? What about appealing to a company’s financial interests?   Does the format of a marketing brochure (an “exemplar”) influence the results? And finally, regardless of the message, are recipients more likely to sign up for the program if we also send them an email with a link to an enrollment form?

We use a factorial research design to test the impact of each message, exemplar, and delivery mode. For this study, we randomly assigned more than 34,000 establishments to 18 treatment arms, with each establishment being mailed brochures in spring 2014. Our analysis will compare the impact of each marketing strategy on program sign-ups, in comparison to both the current OSC brochure and a no-marketing status quo. Our goal is to develop knowledge that is applicable not only to OSHA’s OSC program, but also to other voluntary government initiatives.

Drawing on lessons from the OSHA OSC study, this presentation will provide insight into the process of working with government agencies and academic experts to test behavioral theories of marketing. We discuss the challenging tradeoffs that arise when designing a study whose findings will apply to a broad array of government programs, but where the materials being tested must appropriately reflect the goals of the specific program—all without overpromising results or inadvertently harming the agency’s public image.  We will also discuss the implementation and analytic challenges inherent to a factorial study with 18 treatment arms and 100,000 total study participants nationwide.