Free cholesterol testing as a motivation device in blood donations: evidence from field experiments.

Background: Health tests are often seen as promising donor incentives to improve the supply of blood. However, systematic behavioral evidence on donor recruitment is scarce.

Study Design And Methods: To study the effectiveness of a free cholesterol test in attracting new donors and motivating previous donors, two field experiments were conducted. In Study 1, 2825 nondonors were randomly assigned to one of three treatments: a solicitation letter, a solicitation letter plus an appeal, or a solicitation letter plus an appeal and the offer of a free cholesterol test. In Study 2, 8269 previous donors were randomly assigned to one of three treatments: a standard invitation, an invitation plus an appeal, or an invitation plus an appeal and a cholesterol test. Marginal effects from probit estimations were calculated to study the effects of the treatments on donors' response.

Results: In Study 1, only 0.6 percent reacted to the solicitation letter. There were no significant differences in the response rates between the three treatments. In Study 2, 45.3 percent of the invited previous donors came to donate. The appeal (marginal effect, -0.5%; standard error [SE], 1.9%) and offering a cholesterol test (marginal effect, 1.6%; SE, 1.8%) did not significantly increase the probability of a donation relative to the standard invitation. The treatment effects for the cholesterol test did not systematically differ between frequent and infrequent donors and female and male donors. There is some evidence that young donors responded relatively most positive to the cholesterol test (marginal effect, 4.4%; SE, 2.2%).

Conclusions: Contrary to conclusions from survey studies, free cholesterol testing did not significantly increase donations from nondonors and previous donors during a 3-month campaign. The two studies show that field experiments are an important method to evaluate donation incentives, because measuring donors' intentions alone can lead to significantly different conclusions.