UW Today

This is an archived article.

September 17, 1996

UW psychologists develop first reproducible method showing subliminal messages can influence behavior, thought processes

Ever since a New York motivational researcher claimed 40 years ago that he could persuade drive-in theater patrons to purchase popcorn and Coca-Cola with “hidden” visual messages, psychologists have been searching for reproducible evidence that subliminal visual messages can influence human behavior, thought processes and decision-making.
Finding such evidence has proven elusive, but researchers from the University of Washington, writing in tomorrow’s issue of the journal Science, report that they have developed the first reproducible method demonstrating that subliminal messages do affect human cognition. The techniques, developed by a UW research team headed by psychology professor Anthony Greenwald, provide the clearest window yet into the operation of the unconscious mind.

“What we see indicates that unconscious cognition is capable of only simple mental operations compared to the powers of conscious thought,” said Greenwald. “In addition, the influence of a subliminal message is fleeting, lasting only a brief flicker of time, perhaps as little as one-tenth of a second.”

“The techniques we are using in our laboratory are not that different from the highly edited images put on television. People have been afraid that those images might contain subliminal messages that they couldn’t detect. This is something that advertisers have sometimes been accused of, even though it’s rarely done — and, when done, is fully detectable on videotape.”

Greenwald and co-authors Sean Draine, a UW doctoral candidate, and Richard Abrams, a former UW graduate student, developed a “subliminal sandwich” technique to test unconscious cognition on more than 300 volunteer subjects. The subjects were asked to identify nearly 500 target words as either male or female names or as pleasant or unpleasant. The words briefly flashed on a computer monitor one at a time, every few seconds, and subjects identified them by pressing keys on a keyboard.

Just prior to seeing each target word, the subjects were exposed to the subliminal sandwich, which also flashed on the monitor. Each sandwich was composed of a string of 15 consonants, then a priming word — again either male/female or pleasant/unpleasant — and then another string of 15 consonants. On some trials the priming and target words agreed — such as two female names. On others they didn’t — a female priming name preceded a male target name.

The researchers controlled the amount of time subjects were exposed to the subliminal priming word and the interval between the priming word and the target word. Most subjects were able to correctly identify target words in five-tenths to six-tenths of a second with an average error rate of about 5 percent. To test the mind’s ability to receive a subliminal message, Greenwald and his associates obliged the subjects to make faster choices and identify words in four-tenths of a second.

Using this method, which Greenwald calls the response window technique, he was able to measure the effect of the priming word on correctly identifying the target word by calculating the error rates when the two words agreed or disagreed. When subjects were time-pressured, the error rate was sharply increased if the prime word differed in meaning from the target word. This showed that the subjects were perceiving the priming words unconsciously and that this was affecting their identification of the target words

Greenwald said similar work by researchers in the past has occasionally showed some subliminal influence, but those findings have proved nearly impossible to reproduce. Time is the key element in the new research, making it possible to easily replicate the effects of subliminal activation in any laboratory.

“It’s both the time between the prime and the target, which we’ve found has to be a tenth of a second or less, and the time pressure on the subject to respond rapidly. These two time ingredients have been lacking in previous research,” Greenwald explained.

The new findings are sure to add fuel to the on-going debate between cognitive psychologists and psychoanalytical researchers about the nature of unconscious cognition. Cognitive psychologists such as Greenwald, whose earlier research showed that subliminal self-help tapes to improve memory or self-esteem have little effect, believe that unconscious cognition is “a pretty dumb but fast processor that scans sensory input and directs our attention.” Sigmund Freud and other psychoanalysts, on the other hand, view the unconscious as a very smart and powerful entity that protects a person from psychic threats, such as knowledge of a fatal illness or personal incompetence, that are considered to be too anxiety-provoking for the more fragile conscious mind to cope with.

The new research also confirms that James Vicary did concoct a hoax in 1957 when he reported being able to influence people to buy more popcorn and soft drinks by flashing “invisible messages” on the screen of a Ft. Dix, N.J. drive-in. Greenwald says his research indicates that Vicary’s two-word messages of “eat popcorn” and “drink Coca- Cola” are too complex to be deciphered by unconscious thought. In addition, Vicary’s claim of having produced an observable effect minutes after showing his supposed subliminal message also conflicts with the new evidence which shows the influence of a single subliminal word is gone in about one-tenth of a second.

For additional information, contact Greenwald at (206) 543-7227 or by e-mail at .

Other psychologists willing to discuss the new findings include:

John Kihlstrom, professor of psychology at Yale University and authority on unconscious cognition, at (203) 432-2596 or at .

Eliot Smith, professor of psychology at Purdue University who has independently replicated the methods developed by Greenwald and colleagues, at (317) 494-7709 or at (esmith@psych.purdue.edu).