While the ELM is most widely understood and researched as a persuasion theory, it can be used to understand all human social cognition including processes like choice, decisions, stereotyping, planning, economic behavior, and on and on. When people are “thinking” in any reasonable sense of the word, then WATTage is in the scene, and ELM directs the play. And since virtually all interesting, volitional, and changeable human thought and action requires WATTage, please consider the ELM as a flexible theory for explaining why and how humans think and act. To illustrate this flexibility I’m going to look at an older set of experiments that looked at stereotyping, circadian rhythms, and the Kahneman and Tversky model of cognition.
Galen Bodenhausen conducted two experiments in 1990 that employ variables not typically associated with persuasion and the ELM, but deliver results that are entirely predictable and consistent with the theory. Bodenhausen had people show up to a lab during the morning, afternoon, or evening to complete a standard person perception task related to stereotyping. For example, you’d read a list of attributes about a person that are stereotyped as common qualities. Kahneman and Tversky describe a Similarity Heuristic where people can carelessly link ideas, objects, or people into the same category because they share superficial similarities. It’s a top of the head heuristic. Now, oftentimes similar things can be identical, so the Similarity Heuristic often works.
You can see how this applies with the person perception task on stereotyping. With this heuristic, people make probability estimates based on the apparent similarity of the event or characteristic being judged to a representative stereotype. People show a strong tendency to believe that the conjunction of a representative (stereotypic) and an unrepresentative element is more probable than the probability of the unrepresentative element in isolation. Stated another way, when people are using a stereotype as a Peripheral Route Cue they will assimilate a nonstereotyped attribute into that stereotype anyway. Here’s how Bodenhausen handled this. He had people read a list of attributes like this.
- Bill likes sharp pencils.
- Bill enjoys working with numbers.
- Bill is precise.
- Bill is neat.
- Bill understands decimals.
After reading this list of stereotyped attributes, everyone would be given a descriptor that illustrates a heuristic from the work of Kahneman and Tversky. With our example, they’d read:
- Bill is an accountant who plays saxophone as a hobby.
If you are using the Similarity Heuristic, you are more likely to believe the conjunction. Thus, people who are working top of the head with the Similarity Heuristic would be more likely to believe that accountants play the saxophone simply because the two (unrelated) attributes are read together at the same time while the person is that Low WATT, Peripheral Route processing, Cue-ing along with the stereotype.
Now, recall that Bodenhausen had people take this person perception task either during the morning, afternoon, or evening. In addition to the stereotyping test, Bodenhausen also had the participants complete a self report measure of circadian rhythm to establish whether the person was either an “early bird” or a “night owl.” He then crossed when the person did the experiment with that circadian score so that he had two categories: You took the stereotyping test when you were “in” your rhythm (early bird taking the test in the morning, for example) or when you were “out” of your rhythm (a night owl taking the test in the morning).
Realize that this crossing of time (of day) and rhythm (of circadia) manipulates WATTage. On a fairly large scale of abstraction, when night owls are processing at night, they are Higher WATT than during the day, especially morning. And, when early birds are processing in the morning, they are Higher WATT than later in the day, especially evening. Does this actually happen? Here’s what Bodenhausen found.
94% of subjects with “morning” personalities committed the conjunction fallacy during an evening experimental session (n = 16), while only 71% did so in the morning (n = 14). Conversely, subjects with “evening” personalities were more likely to commit this fallacy during a morning experimental session (92%, n = 12) than during the evening (70%, n = 17). In an analysis of error rates, the interaction of personality type and time of testing was significant, F(1, 55) = 4.55, p < .05.
The effect size here is a Small plus Windowpane, about 40/60.
Bodenhausen ran a second study on stereotyping for a court case that manipulated perceive race with a person’s name (Robert Garner versus Roberto Garcia, for example). Again, he used the circadian rhythm and time of day to manipulate WATTage as participants read a transcript then assigned guilt ratings. As the following Result demonstrates, the predicted interaction between WATTage and stereotyping heuristic obtained.
For morning types perceived stereotyped targets to be more likely to be guilty in the afternoon and evening than the morning, F(1, 44) = 5.16, p < .05. For evening types, perceptions of the stereotyped targets’ guilt were significantly greater in the morning than in the afternoon or evening, F(1, 47) = 4.39, p < .05.
The effect size here is a Medium Windowpane, about 35/65. How do you interpret these findings? Here’s how Bodenhausen put it.
The results obtained in these studies have a number of interesting implications. First and foremost, they support the view that stereotypes function as judgmental heuristics and, as such, are likely to be more influential under circumstances in which people are less motivated or less able to engage in more systematic and careful judgment strategies (Bodenhausen & Lichtenstein, 1987; Chaiken et al., 1989).
I’d now like to drive a theoretical truck through this empirical opening. WATTage, Arguments, and Cues can be understood as highly abstract constructs that apply not just to standard persuasion variables (involvement, arguments about Senior Comprehensives, cues about source credibility), but all types of human cognition and action. WATTage is the crucial variable in all this, with thoughtfulness, that Long Conversation in the Head, as the determinant of Route of Processing and all That implies.
Now, I’m not the first guy to see this basic Dual Process Model. Shelly Chaiken and Yaacov Trope edited a fabulous book on Dual Process Theories (1999) that presented dozens of these ideas together in one book. Here’s just a quick, abbreviated, and short list.
- Impressions: Piecemeal vs Category (Fiske)
- Persuasion: Central vs Peripheral (Petty and Cacioppo)
- Behavior: Automatic vs Controlled (Bargh)
- Behavior: Mindful vs Mindless (Langer)
- Persuasion: Heuristic vs Systematic (Chaiken)
- Att Activation: Deliberate vs Spontaneous (Fazio)
- Memory: Elaboration vs Activation (Anderson)
- Decisions: System 1 vs System 2 (Kahneman and Tversky)
There’s much more in the book and if you truly pant for the weariness of great knowledge, this book may slake your thirst. Best of all, most of the chapter authors in this collection smartly observed the generality of two processes controlled through what I call the WATTage switch.
Now, the nuance. Yes, there are large differences between the theories and you can go places with Susan Fiske’s model of person perception you cannot with a General ELM. And, yes, you could pick a different theory and make it the General over the others with interesting advantages. And, yes, I am a cheerleader for the ELM who cannot see through my blind eyes which means I’m just a shouter. Yes. Yes. Yes.
But, a General ELM provides an extremely effective practical blueprint for thinking about, planning, executing, and evaluating any and all real time, real world, here-and-now human thought and action. If you are a scientist, this is a weak Argument, but I’m not writing about persuasion here just for scientists. Isn’t that what you do in the peer-review literature? I’m trying to make persuasion science work in everyday life.
General ELM is the best way for that.