Predicting Predictions with Implicit Response Testing
Our minds exist to predict. Every thought, emotion, opinion, decision, or sensory experience is a prediction about the nature of our world and our place within it.
As marketers, we try to predict how our messaging will influence the predictions of our audience and will thereby influence purchasing behavior. When we conduct market research, we try to measure that influence through controlled experiments. This can be as simple as asking a question and recording the answer or as complex as correlating EEG patterns with sensory input.
These consumer predictions might be as simple as looking at an object and predicting how it was intended to be used. Or the predictions could be more complex: seeing a face in an advertisement and predicting the emotions and intentions of the person pictured and what that says about the product. Purchasing decisions are predictions about whether a product will meet our needs, and whether we’ll feel satisfied after buying it.
Implicit response testing can provide deep insights into the predictive mechanisms we seek to understand and influence, by revealing their underlying associative networks. When respondents experience a priming image, phrase, or sound, their predictive mind activates a network of expectations based on that stimulus. The speed with which they then perform a task will depend in part on how congruent the task is with those expectations.
Show people a happy face and they will be faster at correctly sorting the word “Positive” than “Negative.” The difference may only be a fraction of a second, but that difference is measurable and consistent. Mix in a couple different happy faces, and you can measure which one yields the quicker response, even if the person isn’t consciously aware of liking one more than the other. The mind is constantly predicting based on subtle, often subconscious, cues.
When you see a screwdriver, your arm and hand are subtly prepared to grip and apply torque. This is especially true if you are holding a screw in your other hand and the screwdriver is lying on a bench before you. But it is also true to some extent when you see a picture of a screwdriver and have only memories of holding screws. You will be slightly faster at correctly sorting the word “grip” than the word “drop” after seeing that picture. You are predicting future events based on current sensory input, your model of the world, and your current goals. If you are hungry, you may be quicker to sort “grip” after viewing an image of a fork than after viewing an image of a screwdriver.
Each of us is continuously constructing and revising an evolving conceptual framework of categories and associations to process sensations and allocate bodily resources toward meeting goals. We create hypotheses based on direct interactions with our environment and on messages about the environment that we’ve received from trusted sources. In addition, the sensations we process include internal signals, such as pain, pleasure, visceral feelings, kinesthetic proprioception, and the words, music and pictures that we imagine.
We build models of the world, make predictions, and process feedback to improve the predictive accuracy of our models, employing a blend of unconscious neural calculations and conscious experiential representations of our predictions.
These predictions are embodied in a neural network that grows to include over 80 billion cells, communicating with each other in complex ever-changing subnetworks. Each neuron is a node in a biological computer that has evolved over millions of generations — 300,000 generations since we diverged from chimpanzees. At first, the predictions were as simple as guessing how to move toward food and away from threats. Over eons of trial and error, the predictions have evolved to include complex appraisals of social cues and long-range strategic planning. The biological computer we’ve inherited now configures itself in every individual by developing predictive models that never existed before and that reflect each individual’s unique perspective.
Implicit testing is uniquely qualified to provide insights into the predictive models that drive behavior, by measuring sub-second reaction-time differences between responses to carefully controlled and randomized sets of stimuli. We are slightly but measurably faster to react when the stimuli fit our predictive model, and we are slower to react when the stimuli confound our unconscious expectations and associations. By employing implicit testing in market research, you can harness this innovative technology to predict the unconscious predictions triggered by your brand, product or service.