These days, implicit, System 1 consumer research is running rampant! In large part due to Daniel Kahneman’s (1) and Dan Ariely’s (2) behavioral economics work, marketers have discovered that consumers’ purchases can be influenced by automatic, nonconscious processes, and consumer researchers are racing to invent and incorporate approaches to measure them. These approaches are most often called ‘implicit’ and/or ‘System 1’ measurement tools.
In general, this is good for our industry. Implicit, System 1 processes are real, and they do influence consumer behavior, sometimes differently than explicit, System 2 processes. But implicit, System 1 processes, and approaches to measuring them, have often been misunderstood, misrepresented, and misused by marketers and consumer researchers.
This article explains what implicit, System 1 processes are, and critiques & makes recommendations for their measurement.
What is Implicit, System 1?
When people, including consumers, encounter stimuli in the environment, their brains process the stimuli vis-á-vis the current context. The current context consists of many things — the situation, the consumer’s goals, and past experiences to name a few.
Processing refers to ‘how’ the brain interprets stimuli and ‘what’ it concludes about it. The ‘how’ and ‘what’ of processing come in a variety of flavors. Among other things, ‘how’ can refer to the speed, control, conscious awareness, and degree of cognitive resources devoted to processing. ‘What’ refers to the output of processing and can be, for instance, associations or propositions (at the dimensional level), or emotions, personality traits, or functional features of the stimulus (at the discrete level).
Implicit, System 1 processing and its counterpart explicit, System 2 processing have evolved from a larger area of study in psychology called Dual-Process Theory. (For a good review see (3).) Generally speaking, Dual-Process Theory recognizes that cognitive processing of stimuli, including consumer products & services, can be divided into those that are automatic vs. those that are controlled. These roughly equate to implicit, System 1 (automatic) and explicit, System 2 (controlled).
Getting Off Track about Implicit, System 1
Kahneman’s popularization of System 1 and System 2 — pointing out that behavior can be influenced by each system independently and that System 2 often leads to ‘irrational’ behavior — lit the fire of marketers and consumer researchers looking to enhance measurement, understanding, and prediction of consumer behavior. Pursuing this goal, System 2 measures have often been positioned as ‘less truthful’ indications of consumers’ attitudes, while System 1 measures have often been positioned as ‘more accurate’ in assessing how consumers really feel. Insofar as consumers do hide their true feelings on traditional surveys, this can be true.
Unfortunately, however, this quest for better measures of consumer attitudes has led to many misinterpretations and misuses of implicit, System 1. Here as some of the most prominent.
1. Implicit and System 1 May Not Be Entirely Synonymous
Although often presented as synonymous — I’m certainly guilty of that, too — implicit and System 1 are not entirely so. Implicit is more technically synonymous with the features of automaticity (fast, unintentional, uncontrolled, nonconscious in some way, and cognitively efficient; see later in this article). However, System 1 refers more metaphorically to the concept of impulse, which relates more to fast and cognitively efficient, but not necessarily as much to control or consciousness (although these can be features).
2. System 1 and System 2 are Not Distinct Brain Regions
Although not widely, nor often literally claimed, at least sometimes System 1 and System 2 are implied to be distinct neurological brain networks. This is simply not true. The brain consists of highly diverse, degenerative networks that are quite flexible rather than specific in the functions they perform. System 1 and System 2 are admittedly (by Kahneman himself [a (1)]) metaphors for different styles of processing stimuli rather than distinct, recognizable neural systems [b].
3. Implicit, System 1 is Not Synonymous with Emotional, and Explicit, System 2 is Not Synonymous with Rational
Often marketers and researchers imply or directly state that implicit, System 1 responses are ’emotional’, while explicit, System 2 responses are ‘rational’. Again, this one-to-one correspondence, as well as the implied mutual exclusivity of emotionality and rationality, are simply not true. Implicit, System 1 responses can be non-emotional associations (e.g., Amazon can be automatically associated with ‘convenience’) and explicit, System 2 responses can be emotional (e.g., after seeing that what I originally thought was a snake is really a stick, I can feel relaxed).
4. Implicit, System 1 is Not Synonymous with Nonconscious, and Explicit, System 2 is Not Synonymous with Conscious
Although more closely related than ’emotional’ and ‘rational’ correspondences mentioned above, consciousness relationships are still not one-to-one. They come closer because one feature of implicit processing can be nonconscious in some way. However, one can be conscious of implicit, System 1 associations (e.g., I’m aware that in my mind a candy bar tastes great), but unable to control their automatic activation (e.g., I can’t stop thinking that a candy bar tastes great). (For more discussion on this issue, see Gawronski et al. (4).)
5. Implicit, System 1 Measurement is Not Synonymous with Reaction-Time Scoring
Although most implicit, System 1 measurement procedures use reaction-time to determine implicitness of response, several do not. For example, Payne’s Affect Misattribution Procedure (AMP; (5)) does not rely on reaction time in its scoring, but is still a highly-respected, truly implicit measure. (In fact, we use an adapted AMP within our IE Pro YOU® platform for measuring implicit associations.) Be wary of researchers who make System 1 measures synonymous with reaction time.
6. Implicit, System 1 Measurement Can’t Be Equated with a Particular Speed of Response
Sometimes reaction-time measures of implicit, System 1 define the process as under a certain amount of time — e.g., 300 milliseconds or less or 700 milliseconds or less. Being metaphorical in nature, System 1 doesn’t automatically map to a specific speed of response. Processing speed can differ based on the individual and topic being studied.
7. Explicit, System 2 Evaluations Can Be as Real as Implicit, System 1 Associations
We often see statements that implicit, System 1 attitudes are ‘real’ and ‘the truth’, while explicit, System 2 attitudes are not. Although explicit, System 2 responses can be conscious misrepresentations of implicit, System 1 associations, this is not necessarily true. As indicated below, System 1 associations are submitted to System 2 for evaluation. Therefore, after taking into consideration multiple System 1 associations through deliberation, System 2 evaluations can certainly be real — in fact, more real than System 1 associations because they take more of them into consideration and net with one’s ‘real’ attitude.
8. By the Way, Kahneman Didn’t Coin the Terms System 1 and System 2
Perhaps a nitpick, many who speak definitely about System 1 and System 2 credit Kahneman with these terms. However, Stanovich and West (6) actually coined the terms, as is acknowledged by Kahneman [c (1)]. References to Kahneman as the inventor of these terms should raise slight questions about the source.
Getting Back on Track Regarding Implicit, System 1
Getting back on track about implicit, System 1 processing is helped by examining the work of behavioral scientists who most closely study dual-process phenomena. Such scientists include Jan De Houwer and Agnes Moors (Ghent University), Bertram Gawronski (University of Texas, Austin), and Brian Nosek (University of Virginia) who have published extensively in this area, laying the groundwork for implicit understanding and measurement.
In particular, De Houwer and Moors (7, 8) and Gawronski (9) provide strongly pertinent views on this issue. We explore Gawronski’s work first.
Gawronski’s APE Model
Gawronski’s Associative-Propositional Evaluation Model (APE) says that implicit evaluations of stimuli are the result of associative processes, while explicit evaluations of stimuli are the result of propositional processes. These are two distinct, but interactive, ways in which the brain processes stimuli. Associative processes involve activating mental associations in memory that have similar features to the stimulus. Propositional processes involve validating associations activated by associative processes.
For example, if one encounters a stick in the woods, because features of a stick are similar to features of a snake, associative processing might automatically activate snake representations along with affective reactions (e.g., fear). Deliberating further on the stimulus invokes propositional processes which evaluate the truth of the associative representations. This truth evaluation (a.k.a., validation) will take other associations into consideration given the environment (e.g., whether the stimulus is moving or not, whether it has what appear to be eyes, etc.) and conclude whether it’s a snake or not. If the truth results in the conclusion that the stimulus is a stick, then the affective reaction changes accordingly (e.g., fear may diminish, turning into a sense of relief).
Again, Gawronski’s APE model sees implicit responses as associative in nature and explicit responses as propositional in nature. This may well serve as the fundamental difference between implicit, System 1 (associative) and explicit, System 2 (propositional) responses in consumer research studies — that is, if you believe in dual, independently operating processes. However, not all who study cognitive evaluations do.
Another Perspective from De Houwer and Moors
In fact, Jan De Houwer is a single-process believer. De Houwer believes that propositional processes can explain associative reactions, too (10). Given this, the associative-propositional distinction doesn’t hold for implicit definition. Instead, De Houwer and his colleague Agnes Moors see the nature of implicit vs. explicit responses differently.
De Houwer and Moors propose a decomposed model of implicit measurement. To them, implicit processes (what’s happening in our brain as we process stimuli) and measures (the result of implicit processes) are synonymous with being ‘automatic’. And implicit/automatic processes and measures can be defined as having one or more of a series of features [d]:
- Fast – they emerge quickly, before extensive cognitive deliberation.
- Unintentional – they begin with no conscious intention.
- Uncontrolled – they continue with no ability to stop or alter them.
- Nonconscious in some way – one can be unaware of the stimulus, its processing, or its impact on behavior.
- Cognitively Efficient – they use little cognitive resources.
De Houwer and Moors make the point that an implicit process or measure can be so-named if one or more of these automatic features apply. Given that, they see no one definition of implicitness, but, in light of this flexibility, make the strong point that researchers should define what they’re measuring by the features of automaticity that apply. For example, implicit measures that directly measure fast reactions to stimuli of interest (sometimes called ‘fast explicit’ because they are direct), can claim the implicit features of ‘fast’ and ‘cognitively efficient’, but can’t claim unintentional, uncontrolled, or a nonconscious stimulus. On the other hand, indirect implicit priming measures, like the aforementioned AMP, because of their indirectness, can claim the implicit features of unintentional, uncontrolled, nonconscious, and cognitively efficient, but not necessarily fast.
Where does all of this leave us with consideration and use of implicit, System 1 consumer research? Here are the main points:
1. When explaining implicit, System 1 in relation to corresponding research approaches:
- Don’t position implicit, System 1 as synonymous with ’emotional’ and explicit, System 2 as synonymous with ‘rational’.
- Don’t position implicit, System 1 as synonymous with ‘nonconscious’ and explicit, System 2 as synonymous with ‘conscious’.
- Don’t position implicit, System 1 as synonymous with ‘real’ and explicit, System 2 as synonymous with ‘unreal’.
- Don’t identify System 1 and System 2 as distinct brain regions.
2. Don’t position implicit, System 1 research methods as synonymous with ‘fast reaction time’.
3. When using fast reaction-time measures, don’t automatically assign a certain speed as implicit, System 1 for every study.
4. Per De Houwer and Moors, when describing your implicit, System 1 method, explain the features of implicitness/ automaticity that it captures.
(1) Kahneman, D. (2011). Thinking Fast and Slow. New York: Farrar, Straus and Giroux.
(2) Ariely, D. (2010). Predictably irrational: The hidden forces that shape our decisions. New York: Harper Perennial.
(3) Sherman, J. W., Gawronski, B., & Trope, Y. (Eds.). (2014). Dual-process theories of the social mind. New York, NY, US: Guilford Press.
(4) Gawronski, B., Hofmann, W., and Wilbur, C.J. (2006). Are “implicit” attitudes unconscious? Consciousness and Cognition, 15, 485–499.
(5) Payne, B.K., Cheng, C.M., Govorun, O., and Stewart, B.D. (2005). An inkblot for attitudes: Affect misattribution as implicit measurement. Journal of Personality and Social Psychology, 89(3), 277-293.
(6) Stanovich, K.E. and West, R.F. (2000). Individual differences in reasoning: Implications for the rationality debate? Behavioral and Brain Sciences, 23, 645-726.
(7) De Houwer, J. and Moors, A. (2007). How to define and examine the implicitness of implicit measures. In B. Wittenbrink & N. Schwarz (Eds.), Implicit measures of attitudes: Procedures and controversies. New York: Guilford Press.
(8) Moors, A., Spruyt, A., and De Houwer, J. (2010). In Search of a Measure That Qualifies as Implicit: Recommendations Based on a Decompositional View of Automaticity. In B. Gawronski and B.K. Payne (Eds.), Handbook of Implicit Social Cognition: Measurement, Theory, and Application. New York: The Guilford Press.
(9) Gawronski, B. and Bodenhausen, G.V. (2014). Implicit and Explicit Evaluation: A Brief Review of the Associative–Propositional Evaluation Model. Social and Personality Psychology Compass, 8/8, 448–462.
(10) De Houwer, J. (2014). A Propositional Model of Implicit Evaluation. Social and Personality Psychology Compass, 8/7, 342–353.
[a] From Thinking Fast & Slow, Kahneman says “…I describe mental life by the metaphor of two agents, called System 1 and System 2….”
[b] Having said this, when System 2 is in action, ‘thinking’ is taking place, which can be indicated by activity in the frontal cortex. But this activity is not solely associated with the more metaphorical System 2.
[c] From Thinking Fast & Slow, Kahneman says “I adopt terms originally proposed by Keith Stanovich and Richard West, and will refer to two systems in the mind, System 1 and System 2.”
[d] De Houwer and Moors’ features of automaticity are more intricate than summarized here. For more detail, see articles (7) and (8).