'Subliminal' advertising experiments
This is a report of an experiment that compares embedded imagery in real ads with modified versions of the same ads. The results were presented at the International Interdisciplinary Conference on Figuring Addictions/Rethinking Consumption. University of Lancaster, United Kingdom, 4-5th April, 2002.
As can be noted by reviewing the literature on so-called subliminal advertising. or reading one of the few accurate summaries e.g. as produced by Ji-Young Hong at the University of Texas at Austin, little attention is actually paid to the practice of advertising. Most of the commentary draws upon studies into subliminal perception and have little to say about advertising. The study reported in the Powerpoint Presentaton is an attempt to rectify this situation and place 'subliminal' advertising in an appropriate perspective.
Psychologists, Marketing and Advertising specialists have managed to bamboozle the public (and possibly themselves to some extent) by inappropriate generalisations from related fields such as subliminal perception and the extrapolation of conclusions relating to subliminal perception to the world of advertising. Further confounding the matter has been disingenuous and inappropriate use of the term subliminal. This is perhaps not surprising given the lack of evidence regarding 'subliminal' advertising.
The present study used extracts from a number of recent adverts incorporating secondary or 'subliminal' elements. Three sets of images were produced for the purposes of comparison. Set 1 comprised original imagery extracted from ads. Each extract contained an element that, in the opinion of the author and his colleague Jonathan Ling, would be considered 'subliminal' ie not part of the overall context of the ad and with a distinctly different meaning. Set 2 comprised the same images with the 'subliminal' component modified so that it no longer was 'recognisable' as a separate entity. Set 3 modified a different area of the original extract as a means of ensuring that any judgements arising from a comparisoin of Sets 1 and 2 would not simply be due to changing the image. Psychologists would call this third set a 'control condition'. The rationale of this approach is as follows. If changes in judgements were due simply to change having been made to the original extract then judgements made regarding Sets 2 and 3 should be similar. If the results were due due to removal of the embedded, 'subliminal' image rather than simply change per se, then the results for Sets 1 and 3 should be similar - these both contained the 'subliminal' image. With such a result it could thus be inferred that any differences were due to the meaning attributable to the embedded image.
It was suggested to the author that a more appropriate control condition would have been the substitution of another 'image' in place of the 'images' identified by the author and Jonathan Ling. In some respects this would be true. However, the validity of such a substitution would hinge on whether there were, in fact, images embedded in the ad extracts. As there is currently no evidence to determine that this is so, there is simply the views of the two authors. So, prior to any study making use of changes in imagery (which ought to focus on the meaning associated with the embedded content) it is necessary to first of all demonstrate that the embedded imagery does exist. This can be achieved by demonstrating that the images referred to (regardless of what they may represent) does impact on the judgement of viewers.
The results of the study supported the author's view that secondary or 'subliminal' elements existed in the ad extracts and that these secondary elements were meaningful to viewers. One of the subjects reported, when asked, that they thought the study concerned subliminal advertising. This subject had previously attended a talk by one of the authors on the subject of subliminal advertising. His results were removed from the data before analysis. Data from the remaining viewers, even though they did not consciously perceive the embedded secondary elements, indicated that they did respond to them in a manner that indicated preconscious preception.
The judgements of the viewers fell into the anticipated pattern. Results were similar for Sets 1 and 3, containing the secondary 'subliminal' imagery, and Sets 1 and 3 differed from the judgements for Set 2, in which a 'neutral' aspect of the imagery had been changed. One can therefore conclude from these results that 'subliminal' advertising can influence viewers, even although they may not consciously perceive the contents of the ad.
The authors study is not an isolated one, nor is the conclusion unique, although it is the only one that has taken secondary elements from adverts and examined them in an experimental situation. However, in the few empirical studies of simulated advertising reported in the literature, when these have shown notable results, they have been downplayed on the basis that 'they only influence judgements, not behaviour'. This type of conclusion is grossly misleading as it implies that advertising does influence behaviour in the short term. This is not true.
Most advertising is designed to keep customers aware of products. It is only over time, rather than at the point of viewing, that advertising can be proven to influence behaviour - and then on the basis of sales figures rather than experimental outcomes. The exceptions are sale announcements and special offers: these are likely to have customers beating an immediate path to their local store. Advertising thus functions pretty much the same way as 'subliminal' elements in ads and influences judgements, views and values. The impact of ads when compared with 'subliminal' thus differ only in degree of influence.
If 'subliminal' elements in ads can influence the judgements of the relatively few viewers in the authors study (and equivalent numbers in undergraduate studies), it seems likely that the long term and unethical use of 'subliminal' advertising by major companies has 'seduced' a financially worthwhile response from consumers. If one wishes to be even more critical, where such ads emanate from tobacco and alcohol companies, one can translate commercial success into ill health, preventable disease and the premature death of some consumers.
The subjects in the study only viewed a sections of ads, rather the full ads. It has often been stated that (if 'subliminal' advertising exists) the impact of 'secondary' imagery will be 'washed out' by the overwhelming impact of more salient aspects of ads. This is an issue that needs further research. However, the authors contend that, in carefully executed ads, where the 'secondary' content is relevant to the subject matter being advertised, the influence of 'secondary' images is unlikely to be 'washed out'. Additionally, one should note that adverts can and do present the same message repeatedly, sometimes for many years. This is rarely the case in the best of experimental studies, where stimuli are generally presented once only - at best, on a few occasions. Such long term exposure is likely to ensure that preconscious processing will occur on at least some occasions when consumers view an ad.
A second experiment into the nature and impact of 'subliminal' ads carried out by an undergraduate student addressed this issue and made use of full page and double page ads for Marlboro cigarettes. This study used reproductions of Marlboro ads containing 'subliminal' elements and compared these with similar quality reproductions of the same ads with the 'subliminal' element removed. The outcome showed that, despite the complexity of the whole ads, viewers still responded to the 'subliminal' element.
On the basis of such studies, it therefore seems unreasonable to accept standard pronouncements that 'subliminal' advertising is ineffective. When used appropriately but, needless to say, unethically, such ads are likely to be commercially effective. But the extent to which they influence consumers over time can only be determined by analysis of sales figures associated with different advertising campaigns ie information that is virtually always considered commercially confidential (see the Coke pages for the views of one company on this matter).
The evidence presented above - and the unwillingness of companies making use of 'subliminal' techniques to engage in dialogue over their practices - would seem to justify closer scrutiny of 'subliminal' advertising. In particular, tobacco and alcohol advertising over the past decade need to be examined. The relevant companies and their ad agencies also need to disclose whether or not their use of 'secondary' imagery was effective. Additionally, where health issues are involved, the public needs to know who such ads influence and to what extent they are influential.
Click for the Download page for a copy of the Powerpoint presentation slides reporting this experiment.
Click for thumbnails, links to full size copies of the images used in the experiment and links to commentary and illustrations of the ads from which the experimental extracts were taken.
Click to view Expert.htm for alternative sources of information and other views on the issue of whether 'subliminal' ads influence consumers.
Last Revised: 3rd January, 2003