1 of 14

Subliminal Messaging Experiment

By Adrian A. Berrigan

SI 340

2 of 14

What is a subliminal message?

A subliminal messaging is a message that the mind is not consciously aware of reading

Classic Example: Word appearing on screen for mere milliseconds

3 of 14

Motivation

Past Panic & Coverage

Recent Media Coverage

James Vicary famously caused a scare back in the 1950s when he claimed he conducted an experiment that increased the sale of Coke and popcorn using subliminal advertising.
Regardless of its accuracy it left a lasting impression on the public

In 2022 the EU proposed an AI act that prohibits systems that use “subliminal techniques”
Occasionally, panic provocting blog/news posts appear that overblow the potential for manipulation by subliminal advertising

4 of 14

Motivation Continued

The average American is being exposed to an unprecedented amount of media and information:

According to the Kaiser Family Foundation, kids ages 8-18 now spend around 7.5 hours in front of a screen
Citizens of some countries get as high as 9.5 hours which is around 60% of the time they are awake

Naturally this equates to more incentive for companies to market their products towards these digital users.

Increase in Screen Time

5 of 14

Who does this effect?

Motivation Continued

It has been estimated that worldwide digital advertising spending was 522.5 billion dollars

Advertisers are always looking for new ways to influence consumers, therefore, even if the effect is small it is worth conducting further research into the effects of subliminal messaging

Why should I care?

While we more or less consider it ethical for advertisements to be plastered all over our media it becomes far less so when we don’t realize we are being advertised to

Therefore, it is important to understand the possible impact these types of messages can have on you (even in lab controlled/fringe cases) so that we can decided whether or not this is something we will allow

6 of 14

Literature Review

Karremans et al. (2006). Beyond Vicary’s fantasies: The impact of subliminal priming and brand choice. Journal of Experimental Social Psychology

Study proving that in a lab controlled experiment that subliminal messages promoting a certain drink brand could affect participant choice when the participant was thirsty.

Verwijmeren et al. (2011). The workings and limits of subliminal advertising: The role of habit. Journal of Experimental Social Psychology

A confirmation of Karrenman’s experiment. This study identified habitual consumption as another important factor. It was found that when a drink was habitually consumed subliminal messages promoting it did not increase participant’s rate of choosing it. However, subliminal messaging did increase consumption of the non-habitual option when applied.

Verwijmeren et al. (2013). Warning: You are being primed! The effect of a warning on the impact of subliminal ads. Journal of Experimental Social Psychology

This study indicates that warning participants about the effects of subliminal advertising decreased the effects on choice. For the purposes of my experiment (or any future research) it is therefore necessary to keep the experiment controlled and not allow participants to know exactly what is being studied from the start.

7 of 14

Open Questions

Literature Review Continued

Why is subliminal advertising only effective under specific conditions (like thirst)?

A large portion of research focuses on drinks choice. Can subliminal messaging prove effective with other products as the choice?

Things like “thirst” in these experiments have relied on self-reported data. I would like to see more work done involving more directly controlling these factors

Their approach involves framing the activity as a visual detection task rather than a more applied method (like the classic example of watching a movie)

My Approach

Theorize that certain physiological urges (thirst, hunger, sleep, etc.) make us more susceptible to the effects of subliminal messaging/advertising

Randomized, controlled lab experiment with setup similar to the one conducted by Verwijmeren et al. , but changing the activity away from visual detection task. Replace thirst with hunger as the physiological state of being. Chips replace soft drinks.

8 of 14

Experiment Design

Control: Participants are not hungry and receive no subliminal messaging

Treatment 1: Participants receive subliminal messaging corresponding to a non-habitual brand and are hungry

Treatment 2: Participants receive subliminal messaging corresponding to a habitual brand and are hungry

Treatment 3: Participants receive subliminal messaging corresponding to a non-habitual brand and are not hungry

Treatment 4: Participants receive subliminal messaging corresponding to a habitual brand and are not hungry

9 of 14

Experiment Design - Details

Participants will be asked to eat/not eat before the study depending on what group they have been assigned to

Participants will fill out a questionnaire about brand preferences and level of hunger

Participants will then begin watching a series of videos (similar to Instagram Reel or TikTok) and will be subliminally primed with the word related to their corresponding treatment.

The primed word will appear for 17 ms between each clip. In order to keep things consistent with Verwijmeren’s research this will be preceded + followed by a random letter string for 500 ms to mask the prime.

Participants will then be offered the choice of two chip brands (Habitually Consumed & Generic)

10 of 14

Experiment Design - Hypotheses

H₀: The presence of hunger does not influence the effect of subliminal messaging on brand choice

Hₐ: The presence of hunger does influence the effect of subliminal messaging on brand choice

H₀: When hungry, the presence of subliminal messaging promoting a non-habitual brand will not result in an increase in that brand being chosen by participants compared to the habitual brand.

Hₐ: When hungry, the presence of subliminal messaging promoting a non-habitual brand will result in an increase in that brand being chosen by participants.

11 of 14

Experiment Design - Sample

2(1.96 + 0.84)^2 * (2)^2 = ~ 63 participants per group

This is for a medium effect size. This leads to a total of 315 participants. Prior research

(Verwijmeren et al.) had 146 participants , but less treatment groups. Interestingly of those 146

participants 119 of them were women. Given that 81% of them were women in this prior research

this is certainly something to keep in mind.

12 of 14

Experiment Design - Continued

Independent Variables

Hunger
Subliminal Message
Habits

Dependent Variable

Product Choice

13 of 14

Limitations

Still reliant on self-reported data and participants following instructions (starving participants is probably frowned upon). Additionally, hunger levels fluctuate over time differently person-to-person and this is something to consider.

This experiment is changing a lot of things at once. Thirst => Hunger,

Visual Detection Task => Video, etc.

Based on prior research (Verwijmeren et al. 2013), if participants figure out they are being exposed to subliminal messages its effect will be greatly decreased. There is a good chance that this could occur and figuring out a way to probe participants to see if they know what the study is about might be a good way to control for this.

14 of 14

Thank you!

Adrian Berrigan