Paul Shafer, Erika Franklin-Fowler, Laura Baum and Sarah Gollust *, **have a new paper out that evaluates the change in information seeking-behavior when Kentucky went from a state government and governor that strongly supported the ACA to a governor and state government that strongly opposed the ACA. Governar Beshear wanted to make the ACA work and local advertising was part of that. Governor Bevins did not want the ACA to work, and as soon as he took power, he directed that all advertising for kynect was to be cut-off.
Colloquially, it measures the Give a Damn factor of elite support for expanded health insurance coverage. How much does supportive local messaging change enrollment?
The marginal effects of kynect advertising during open enrollment are positive and significant for page views, visits, and unique visitors. Specifically, each additional kynect ad airing per week during open enrollment was associated with approximately 8000 additional page views (7972.9, P=.001), 400 additional visits (390.2, P=.003), and 400 additional unique visitors (387.5, P<.001) to the kynect website. There is no evidence of a dose-response effect for advertising by healthcare.gov, insurance companies, and insurance agencies for these outcomes (all marginal effects are not statistically significant for these sponsor types). In contrast, advertising by nonprofits seemed to drive traffic away from kynect resources (calls: −197.6, P=.02; page views: −30,061.5, P=.03). Advertising by other state governments was associated with increased calls (215.9, P<.001) to kynect but no change in Web traffic.
Kynect was trusted. The ACA was not well supported and Obamacare was hated. Advertising drove people to Kynect where quite a few were found to be Medicaid eligible. Healthcare.gov advertising that leaked over the border from all of the surrounding states was indifferent in effect.
Elite political support mattered. The manifestation of that support in advertising a trusted, local brand drove action towards enrollment. Elite political opposition mattered. The cut-off of advertising decreased enrollment action. This is a good way to look at isolating the incremental impact of elite leadership giving or not giving a damn.
We’ve looked at that before in a cruder manner. Over at the Health Affairs Blog, Andrew Sprung and I looked at enrollment on-Exchange between California and Healthcare.gov states during the 2018 Open Enrollment period. We assumed California was operating at the frontier of elite giving a damn about covering people while Healthcare.gov in 2018 was far from that frontier. Using our crude analysis, we believe that giving a damn led to a 6% change in enrollment.
Shafer, Fowler-Franklin, Baum and Gollust show in a different context that giving a damn as manifested by advertising or not advertising also led to changes in behavior.
Remember this next Tuesday.
* Shafer PR, Fowler EF, Baum L, Gollust SE. Television Advertising and Health Insurance Marketplace Consumer Engagement in Kentucky: A Natural Experiment (Preprint) 2018. doi:10.2196/preprints.10872.
** I am collaborating with Paul on a forthcoming paper that uses the Obama/Trump transition to look at the same type of question on a national scale without incorporating the advertising data. I am collaborating with this entire crew (and others) on advertising behavior papers (abstract submitted for NHPC-2019)
dnfree
You are prolific! And informative.
David Anderson
@dnfree: This is what happens when the Patriots are in a field goal fest