Association for Public Policy Analysis & Management

We conduct an experiment to test whether active choice framing impacts individual behavior in decision-making for the takeup of a high-stakes, intergenerational education benefit. The specific context we study is the Post-9/11 GI Bill, which provides a generous education benefits package, currently valued at $200,000 - $300,000, that the service member can either use himself or transfer to a dependent family member, though transfer must occur while the service member is still on active duty. Our intervention is email-based and delivers information that may influence the takeup decision; we focus on a sample of 97,000 service members who are eligible to transfer benefits but had not yet done so at experiment launch.

We randomize subjects into one of four groups: control, information only, active choice framing, and active choice + planning prompt. The control group receives no communication from us. Each subject in the information only treatment receives an email informing him that he can use education benefits himself or transfer to a family member. In the active choice treatment, the email frames education benefits use as a choice between own use and transfer. Finally, a subject in the active choice + planning prompt treatment receives an email with active choice framing and a planning prompt to assist in the commitment to making a decision. Across all treatments, we observe how subjects interact with the email – including click behavior to pursue information – as well as level of engagement with the customized website in which we locate information relevant to benefits own-use or transfer. We also observe post-experiment, in individual-level data retrieved from the Veteran’s Administration (VA), whether the service member initiated benefits transfer to a family member.

Early analysis of experimental results suggests three main findings. First, the use of active choice framing significantly increases the likelihood that a subject seeks out information relevant to GI Bill use or transfer; subjects receiving an email with active choice framing were 60% more likely to click thru the email to the information website than were subjects receiving an “information only” email. Next, we find no overall impact of any treatment on short-term GI Bill transfer to a family member. Of course, we will want to follow the subjects over time to observe own-use rates of the benefit: whether service members are saving the GI Bill benefits for themselves and own-use rates increase. Finally, we separately observe that active choice treatment does increase benefits transfer for officers and highly educated individuals – higher-SES subgroups from within our sample. These results may suggest the need for more intensive intervention strategies to support lower-SES service members to make informed, active choices about their GI Bill benefits.