Across three experiments college students were given refutational messages or other treatments encouraging signing drivers' licenses to donate bodily organs in case of accidental death. An attitudinal measure (willingness to sign) predicted observed signing behavior, but relatively weakly. A measure of behavioral expectation to sign or not to sign predicted more strongly. A discussion with family was encouraged, the impact of that discussion also predicted signing but less strongly so. In Study 1 groups given a refutational message signed at no greater rates than a group given a control message. In Study 2 a refutational message group and a group asked to think about donating for 2 min signed at higher rates than a true control group. In Study 3 groups asked to decide whether or not to sign by a certain date and time signed at a higher rate than a refutational message group as in earlier studies. These procedures are recommended for organ donation signing campaigns.