Kh. Smith et M. Rogers, EFFECTIVENESS OF SUBLIMINAL MESSAGES IN TELEVISION COMMERCIALS - 2 EXPERIMENTS, Journal of applied psychology, 79(6), 1994, pp. 866-874
Subjects watched television commercials, monitoring them for a message
(the words choose this) presented at different levels of contrast wit
h the background. Following each commercial, subjects rated their inte
ntion to respond positively to it. In Experiment 1, neither detected n
or undetected messages affected these ratings compared with a no-messa
ge control condition. In Experiment 2, incentives produced higher rati
ngs with detected messages; with undetected messages the effect was st
atistically ambiguous and only one-tenth as large. Unexpectedly, in bo
th experiments commercials containing undetected messages were subsequ
ently less likely to be remembered than commercials with no messages.
The tiny effects of subliminal (undetected) messages compared with the
large effects of supraliminal (detected) ones are discussed in relati
on to the hazards and difficulties of placing them.