Conceptual overlap between stimuli increases misattribution of internal experience
Introduction
Attribution (Bem, 1972; Laird, 1974; Schwarz & Clore, 1983) is the cognitive process whereby people associate their internal experience (i.e., their thoughts and feelings) with an object that (they believe to have) elicited it (e.g., I am afraid because of the bulldog). The process of attribution involves three main components: an internal experience, the object that generated the internal experience, and the object to which the experience is eventually attributed. When the last two components refer to different objects, attribution is erroneous. Research has shown that people can misattribute internal experiences triggered by drugs, physical exercise, or scary situations (e.g., crossing a narrow bridge) to a person or a social situation (Reisenzein & Gattinger, 1982; Schachter & Singer, 1962; White, Fishbein, & Rutsein, 1981). They can misattribute somatic symptoms to WIFI radiation (Bräscher, Raymaekers, van den Bergh, & Witthöft, 2017) or to premenstrual syndrome (Schnall, Abrahamson, & Laird, 2002), and misattribute the pleasantness of a sunny day to a general satisfaction with life (Schwarz & Clore, 1983).
According to “constructivist” models of emotion, attribution is a central process in the formation of emotional experiences (Russell, 2003; Schachter & Singer, 1962). This approach commonly argues that people's internal experience has two fundamental dimensional properties, termed arousal and valence (Barrett, 2006; Russell, 2003). Emotional experience occurs when people's core affect at a given moment in time is latched on to an eliciting object or situation. For instance, when arousal is attributed to an attractive person, the resulting emotion might be infatuation. The same experience of arousal, when attributed to an annoying person, would produce the emotion irritability. Accordingly, research on misattribution has traditionally assumed that the objects of attribution are either valence (Murphy & Zajonc, 1993) or arousal (Reisenzein & Gattinger, 1982; Schachter & Singer, 1962; White et al., 1981).
More recently, research on misattribution has shown that people can misattribute semantic content (Blaison, Imhoff, Huhnel, Hess, & Banse, 2012; Deutsch & Gawronski, 2009; Hofmann & Baumert, 2010; Imhoff, Schmidt, Bernhardt, Dierksmeier, & Banse, 2011). For instance, in one study, participants judged whether Chinese ideographs seemed to evoke fear or anger, while exposed to angry and fearful faces as primes (Blaison et al., 2012). Participants in that research tended to judge targets that appeared after angry primes as angry more often than fearful, and targets that appeared after fearful primes as fearful more often than angry. Because fear and anger are both negative and arousing experiences, something other than arousal and valence caused that effect. Valence and arousal were attributed alongside specific semantic information triggered by the primes.
The semantic information that internal experiences carry might have an important role in the attribution process. A fundamental difficulty in generating accurate attributions concerns the multi-dimensionality and complexity of the situations (Quine, 1960). Any given situation contains numerous objects to which an experience can be attributed (e.g., am I excited because of this ice-cream? Because I am with the person I love?). Despite decades of research into affect attribution, scientists are still far from having a principled account of the determinants of attribution of experience to an object. In the present research, we examined whether semantic information guides attribution, leading to an increase in the effect of misattribution on judgment when the prime and the target share meaning.
Observing a kitten, for instance, might trigger positive valence, and the concepts cute and furry. The attribution process might then search for the object that activated an experience that includes both positive valence and the specific semantic information. In other words, attribution might prioritize objects that share larger conceptual overlap with the internal experience. Consequently, a larger conceptual overlap between a prime and a target would increase the likelihood that an internal experience triggered by the prime would be misattributed to the target. We define conceptual overlap, broadly, as the degree to which the prime and the target share the same superordinate categories, whether that category is based on external properties (e.g., yellow objects), a similar function (e.g., work equipment), or any other common categorization through which people understand their environment.
The degree of conceptual overlap can be roughly estimated according to the most concrete superordinate category that the two stimuli (the prime and the target) share. Daisies and lilies, for instance, are both flowers (a rather concrete category), whereas daisies and cats can only share a category as abstract as “living things”. Therefore, the conceptual overlap between daisies and lilies is certainly larger than between daisies and cats.
No previous research on misattribution has manipulated the conceptual overlap between the prime and the target. Thus, although there is evidence that semantic properties can be misattributed, it is currently unknown whether they can guide attribution toward target objects that share a high conceptual overlap with the object that triggered the internal experience. For instance, it is currently unknown whether a negative internal experience caused by an aggressive dog is more likely to be misattributed to a cat (a target that has a high conceptual overlap with a dog) than to a baby.
Interestingly, past research that investigated misattribution has often used primes and targets from very distant conceptual fields (Dutton & Aron, 1974; White et al., 1981). One possible reason for this choice is the assumption that, for misattribution to occur, participants must not suspect the prime of influencing their judgment of the target. A prime that seems similar to a target of judgment may be perceived as likely to influence judgment, and would thus raise suspicion that would eliminate misattribution (Bless & Schwarz, 2010; Wegener & Petty, 1995). However, more recent research, that measured misattribution using the Affect Misattribution Procedure (AMP)—a sequential priming task with snap judgment of multiple targets, found that misattribution does not decrease when participants are warned that the primes might influence their judgment (Payne, Cheng, Govorun, & Stewart, 2005). Therefore, at least in such judgment contexts, conceptual overlap might increase misattribution by signalling to the individual that the attribution in question is correct.
Evidence for the importance of conceptual overlap in attribution comes from recent research that investigated the applicability of the internal experience triggered by a prime to the judgment question in the AMP (Ecker & Bar-Anan, 2018). In this research, the target was kept constant, and the manipulation varied the conceptual overlap between the prime and the judgment question. The results suggested that primes had a larger influence on judgment when they had a larger conceptual overlap with the judgment question. For instance, when participants judged human faces, a dirty room prime that appeared briefly before the target had a larger influence on judgment when participants judged cleanliness than when they judged cuteness. The negative valence of the room decreased both cuteness and cleanliness ratings. However, when the judgment was specifically related to cleanliness (rather than only to the room's valence), that influence increased.
One plausible account for the increase in misattribution when the primes are more applicable to the judgment question is that, when making judgment, people expect the target to elicit mental content related to the judgment that they are trying to make. Consequently, prime content that has high conceptual similarity to the judgment signals that the attribution in question is correct. Applied to the present research question, it seems reasonable to assume that a prime stimulus that is conceptually similar to the target would elicit content that is relevant to the judgment of the target. The judgment is framed by two components: the judgment question and the judgment's target. The question “How pleasant is this object?” is unclear until the judge knows what the object is. For instance, judging the pleasantness of a food item is rather different than judging the pleasantness of a dog. Following that rationale, a prime that is similar to the target would be more likely to activate mental content relevant for the judgment of that target. Therefore, just like larger conceptual overlap between the prime and the judgment question increases the effect of misattribution on judgment, so would larger conceptual overlap between the prime and the judgment's target.
Because differences between target types (e.g., between people that belong to different social groups) in susceptibility to judgment bias are of social importance, clarifying the distinct influence of the target (as opposed to the judgment question) on misattribution is of great value. Moreover, the identity of the target might be more consequential to attribution in real life than that of the question. In real life, the judgment question is more often insinuated than expressed explicitly. People tend to extract the question from context (e.g., when asked to evaluate a colleague at the workplace, people might extract the question—“how well does she do her job?”), or default to a rather abstract judgment question of “how much do I like this person?”. The target of judgment, on the other hand, is usually clear and well-defined, and thus has a higher potential for influencing judgment.
Importantly, there are also reasons to predict that high conceptual overlap between the prime and the target would decrease, rather than increase, misattribution. Research shows that, under certain conditions, prime-target similarity can cause contrast effects—judgment bias in the opposite direction to the prime's valence or trait. Specifically, in the commonly used paradigm in impression formation research—wherein judgment is unhurried and primes appear separately from the judgment task—primes cause contrast effects when perceived as comparable to the targets (Aarts & Dijksterhuis, 2002; LeBoeuf & Estes, 2004; Moskowitz & Skurnik, 1999). Thus, although prime-judgment conceptual overlap increases the influence of misattribution on judgment, prime-target conceptual overlap might not have a similar influence.
We used the AMP (Payne et al., 2005) to examine whether people are more likely to misattribute a valenced internal experience triggered by a prime to a target of judgment when the prime and the target have high (rather than low) conceptual overlap. As a sequential priming paradigm wherein people make snap judgments about multiple targets, the AMP allows a relatively easy manipulation of the relation between the prime and the target of judgment. Early research on misattribution has mostly used procedures that involve complex social situations, rich with social detail (e.g., Dutton & Aron, 1974; White et al., 1981). Although high on ecological validity, these procedures do not allow for well-controlled variations in specific elements of the attribution process. For instance, when the prime is a narrow bridge that participants happen to cross (Dutton & Aron, 1974), physical exercise (Cantor, Zillmann, & Bryant, 1975), a sunny day (Schwarz & Clore, 1983), or a gut-wrenching movie clip (White et al., 1981), it is difficult to manipulate the similarity between that prime and the target of judgment while holding other aspects of the situation constant. This might explain why, to date, little is known about the manner in which the relations between elements in the attribution process influence attribution.
Although doubts about the mechanism behind the priming effect in the AMP have been raised (Bar-Anan & Nosek, 2012), there is currently good evidence that the effect is caused by misattribution of internal experiences triggered by the prime to the target (e.g., Gawronski and Ye, 2014, Gawronski and Ye, 2015; Oikawa, Aarts, & Oikawa, 2011). There is currently no good evidence in support of any other account. In the last decade, the AMP has improved the understanding of misattribution in the context of quick judgment. It has been shown, for instance, that focusing on gut feelings increases misattribution (De Houwer & Smith, 2013), that longer exposure to the target of judgment decreases misattribution (Payne, Hall, Cameron, & Bishara, 2010), and that a sense of uncertainty about the source of one's internal experience increases misattribution (Ruys, Aarts, Papies, Oikawa, & Oikawa, 2012). The AMP provides a well-controlled tool for a continuous and relatively reliable measurement of misattribution, in the context of a quick judgment after a brief exposure to the target of judgment. Using the AMP, we were able to investigate a fundamental, perhaps long-overdue question, namely, how the conceptual overlap between a prime and the target of judgment influences misattribution of valenced experiences.
Section snippets
Experiment 1
We measured misattribution using a modified version of the AMP (Payne et al., 2005). We manipulated both the prime and the target categories, creating two high-overlap combinations between prime and target and two low-overlap combinations. That allowed us to examine the influence of conceptual overlap beyond the particular prime or target categories. The primes in the AMP were pleasant and unpleasant people or pleasant and unpleasant landscapes. Participants judged the pleasantness of either
Experiment 2
In Experiment 2, in addition to manipulating prime and target categories, we manipulated also the target's medium. Participants judged landscape or man images that were either photographs or abstract sketches. It is less likely that people would blend visual aspects of a live image into a sketch than blend two live images. Thus, if we find an effect for prime-target overlap when the primes are photographs and the targets are sketches, it would reduce the likelihood that only perceptual
Experiment 3
We conducted this experiment as a conceptual replication of the previous experiments, using a different type of target stimuli: words. Participants judged the pleasantness of names rather than pictorial stimuli. We examined whether the effect of prime-target overlap replicates when the stimuli are semantic symbols rather than visual objects. That is a highly stringent test of the conceptual overlap hypothesis because people probably categorize word stimuli rather differently than image stimuli.
Experiment 4
Although we used different stimuli in Experiments 1–3, the categories were always landscapes and people. To generalize our results beyond these categories, in Experiment 4, we manipulated prime and target categories with three new categories: dogs, flowers, and food.
General discussion
In four experiments, we found that primes had a larger effect on judgment of target stimuli when the primes and the targets belonged to the same category than when they belonged to different categories. This result suggests that misattribution is stronger when the object that activated the internal experience and the object to which the experience is attributed share a larger conceptual overlap. We replicated this finding when the targets were words, photos and sketches, and with a diverse set
Open practices
The experiments presented in this paper earned Open Data and Open Materials badges for transparent practices. Materials and data for all experiments are available at: https://osf.io/cgydx/.
Funding
This work was supported by the Israeli Science Foundation [grant number 779/16]; the United States-Israel Binational Science Foundation [grant number 2013214]; and Project Implicit, Inc., awarded to Y. B.-A.
References (50)
- et al.
Mental construal and the emergence of assimilation and contrast effects: The inclusion/exclusion model
Advances in Experimental Social Psychology
(2010) - et al.
Are media reports able to cause somatic symptoms attributed to WiFi radiation? An experimental test of the negative expectation hypothesis
Environmental Reports
(2017) - et al.
When the method makes a difference: Antagonistic effects on “automatic evaluations” as a function of task characteristics of the measure
Journal of Experimental Social Psychology
(2009) - et al.
Perceiving an exclusive cause of affect prevents misattribution
Consciousness and Cognition
(2012) When warning succeeds: The effect of warning on success in ignoring invalid information
Journal of Experimental Social Psychology
(1993)- et al.
Illusory conjunctions in the perception of objects
Cognitive Psychology
(1982) - et al.
Category activation effects in judgment and behaviour: The moderating role of perceived comparability
British Journal of Social Psychology
(2002) - et al.
Reporting intentional rating of the primes predicts priming effects in the affective misattribution procedure
Personality and Social Psychology Bulletin
(2012) Solving the emotion paradox: Categorization and the experience of emotion
Personality and Social Psychology Review
(2006)Self-perception theory
The affect misattribution procedure: Hot or not?
Emotion
Enhancement of experienced sexual arousal in response to erotic stimuli through misattribution of unrelated residual excitation
Journal of Personality and Social Psychology
Go with your gut!
Social Psychology
Some evidence for heightened sexual attraction under conditions of high anxiety
Journal of Personality and Social Psychology
Applicability increases the effect of misattribution on judgment
Cognition & Emotion
G* power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences
Behavior Research Methods
Attentional influences on affective priming: Does categorisation influence spontaneous evaluations of multiply categorizable objects?
Cognition and Emotion
What drives priming effects in the affect misattribution procedure?
Personality and Social Psychology Bulletin
Prevention of intention invention in the affect misattribution procedure
Social Psychological and Personality Science
Relying on accessible content versus accessibility experiences: The case of processing capacity
Social Cognition
Immediate affect as a basis for intuitive moral judgment: An adaptation of the affect misattribution procedure
Cognition & Emotion
An inkblot for sexual preference: A semantic variant of the affect misattribution procedure
Cognition and Emotion
The case for motivated reasoning
Psychological Bulletin
Self-attribution of emotion: The effects of expressive behavior on the quality of emotional experience
Journal of Personality and Social Psychology
Cited by (3)
Perceived air pollution reduces consumers' self-disclosure: The role of tense arousal
2024, Psychology and MarketingUnderstanding the Impact of Technical Issues on People's Perception and Attribution of Responsibility in Videoconferencing
2022, Conference on Human Factors in Computing Systems - ProceedingsContextual positivity-familiarity effects are unaffected by known moderators of misattribution
2021, Cognition and Emotion