Confident Postures

People who feel confident act confidently (e.g., having more erect postures, making expansive use of physical space, expressing their opinions in public), and people who lack confidence act as if they are doubtful (e.g., showing compressive postures and more hesitations). In one relevant study examining the impact of confident postures, Briñol, Petty, and Wagner (2009) asked participants to think about and write down their best or worst qualities while sitting with their backs erect, chests inflated (i.e., powerful posture associated with high confidence) or while sitting slouched forward (i.e., low-power posture associated with low confidence). Then, participants completed a number of measures, including self-esteem. It was predicted and found that the self-relevant thoughts generated affected self-evaluation more in the confident, powerful posture. Importantly, changes in self-esteem were mediated by differences in participants’ confidence in the self-beliefs (thoughts) they had listed.

Another pose potentially related to confidence is when people make aggressive or threatening facial expressions. Briñol, Petty, and Requero (2017) showed that the effect of the direction of thoughts (positive vs. negative) on self-evaluation was greater after people engaged in a behavior associated with threat to others (i.e., having participants show their canine teeth as if they were about to attack, which is a posture associated with high confidence) than after engaging in a control embodiment induction (i.e., participants covering their teeth with their lips in a neutral pose).

Relevant to the present article, it is important to note that the observed effects of confident postures on thought reliance are dependent on various situational factors. First, the research shows that the very same action (pushing the chest out) can have different effects depending on the mental content that is currently active in people’s minds. Confident postures increased self-esteem when people were thinking about their strengths but decreased self-esteem when people were thinking about their weaknesses. This research suggests an important caveat to the recent trend of encouraging “power posing” as a means of becoming more successful across different domains of life (Lammers et al. 2013). Rather than being inherently positive, the confidence that comes from body postures magnifies whatever its mental target is (i.e., both good and bad), at least when it operates through a self-validation mechanism.

Second, the self-validation mechanism requires a level of elaboration that is sufficiently high for individuals to both generate thoughts and to consider their validity. Instead, if elaboration is low, power is more likely to serve as a simple cue or heuristic allowing people to evaluate something as positive or negative (Briñol et al. 2007, experiment 4). For example, under low ability and motivation to think, the positive feelings that emerge from power can lead people to believe that they are more correct than others or even that they are more attractive. This moderating condition helps clarify how a meta-cognitive process such as self-validation fits within the other ELM processes that emphasize primary cognition.

Third, body and facial postures are more likely to influence consumer evaluation by self-validation when induced following or during (rather than preceding) the generation of thoughts. When body postures associated with power precede thinking, they are more likely to serve in another role such as determining the extent of thinking. In an early example of this possibility, Petty et al. (1983) asked undergraduate students to wear and evaluate headphones. Some participants were told to stand up straight while testing the headphones (a posture associated with high confidence), whereas others were told to lie down. Consistent with the hypothesis that confident postures (such as standing up) can affect the amount of thinking, Petty et al. (1983) found that reclining participants were differentially persuaded by the strong and weak arguments, but standing participants were not. The confident posture may have led individuals to believe that they were already correct in their views and did not need to process the opinions of others. This was not the case for the more vulnerable supine individuals. Thus, along with elaboration, timing is another boundary condition for self-validation process to operate.

Fourth, for most people in most situations, it seems that feeling powerful and confident would have a clear positive association (e.g., validity, rightness, pleasantness). However, the meaning associated with power can vary across individuals, situations, and cultures (Cesario and McDonald 2012; Garrison, Tang, and Schmeichel 2016). For example, the experience of power can include appraisals of negative valence (e.g., power is bad when it is associated with corruption and abuse) and uncertainty (e.g., powerful people are sometimes wrong due to incompetence or carelessness), and an avoidance orientation (e.g., power can paralyze when it is associated with ambivalence and conflict during decision making or when perceived as undesirable or as illegitimate; Durso, Briñol, and Petty 2016). Similarly, although making an attacking face might often be associated with power and confidence, the effects of this induction would be quite different if the facial expression is viewed differently (e.g., as a symptom of worry or tension).

In an illustration of this spontaneous variation in meanings, Schubert (2004) found that performing a single physical display of power (making a fist) increased feelings of power (high confidence/validity) in men while activating feelings related to frustration and loss of power in women (low confidence/validity). As noted, our proposition is that if the meaning and the basis associated with a physical display changes, then the effect of that display on subsequent attitudes could also change. For example, if people feel guilty about possessing power or they do not like it (Hays and Goldstein 2015), this negative association might reduce persuasion through a variety of process such as serving as a simple negative cue (under low thinking conditions) or biasing thinking in an unfavorable way (under high thinking).

Cleaning Our Minds from Biases

We have seen that research clearly demonstrates that body postures such as those associated with confidence can affect judgments by various processes such as giving people more confidence in their thoughts and thereby polarizing judgments based on the valence of those thoughts. These effects of posture tend to occur naturally in that people are presumably not aware of the impact that their postures are having. However, if people became aware that their bodily responses were inappropriately influencing their judgments (e.g., do I like this advertisement just because my face is in a smile?), they might adjust their judgments in a direction opposite to the expected bias, just as they do for many other perceived biasing factors such as the likability of a message source (Petty, Wegener, and White 1998).

Importantly, not only can people correct their attitudes if they perceive physical actions to be an unwanted contaminating factor, but they can also do the opposite; that is, sometimes physical actions can be used to metaphorically clean or correct their judgments (Lee and Schwarz 2011). In a line of research examining this possibility, Paredes, Briñol, and Petty (2016) tested a very subtle action-oriented way of inducing bias correction. Participants were first exposed to a persuasive message about a new brand of detergents presented by a credible versus noncredible source and then were told to engage in a task that involved completing mathematical operations of addition versus subtraction. It was predicted and found that credible sources were more persuasive than noncredible ones for participants primed with adding compared to subtracting. However, the traditional effect of source credibility on persuasion was reversed (showing more favorable attitudes for low credibility) for participants assigned to the subtraction condition.

These findings are conceptually consistent with prior research showing that cleansing actions (e.g., washing our hands, removing dirt) can reduce not only the impact of negative responses (e.g., guilt associated with a previous transgression) but can also reduce the impact of positive responses (Lee and Schwarz 2011; Florack et al. 2014). In line with the proposition outlined throughout this article, we argue that if the meaning of the cleansing action varies, the subsequent effect of this action can also change. In an initial examination of this possibility, Kim et al. (2017) had participants thinking about a recent time they did something wrong and then gave them the opportunity to wash their hands. When the action of washing was framed as removing dirt, the results showed that guilt decreased, replicating the original effect of hand washing. In contrast, when the same action was framed as adding spiritual preparation to purify the body and thus be ready to hear the inner voice (a meaning associated with high validity), the results showed that guilt increased, reversing the original effect (Lee and Schwarz 2011). Future studies should examine to what extent these variations in meaning can not only have an impact on the use of negative thoughts but also positive thoughts. Furthermore, forthcoming work can benefit from examining whether other products potentially related to deletion (e.g., using a pencil eraser to write thoughts, using disinfecting gel to clean hands) can also be associated with an impact on thought usage (validation) depending on the meaning.

Facilitating Access to Products and One’s Thoughts

Marketers are interested in making it easier for consumers to access products. For example, Topolinski, Zürn, and Schneider (2015) found that when the name of a product, a brand, or a company is easier to pronounce, attitudes are more positive than when the same name was difficult to articulate. Valenzuela, Raghubir, and Mitakakis (2013) showed that products located in an easy to reach location (e.g., in the middle, top of a shelf) received more positive evaluations than when the same products were located in a difficult location (e.g., at the very bottom of the shelf). Elder and Krishna (2012) showed that the orientation of a product toward a participant’s dominant hand was associated with more positive evaluation of the product. There are many reasons why this might occur such as one’s dominant hand serving as a simple positive cue. But there are other possibilities as well. For example, in line with the self-validation hypothesis, we argue that using one’s dominant hand could increase thought confidence (because it is an easier action than using one’s nondominant hand), by polarizing positive evaluations for liked products but also increasing negative evaluations of disliked products.

In one study using this induction conducted by Briñol and Petty (2003), participants were asked as part of an ostensible graphology study to think about and write down their best or worst qualities using their dominant or nondominant hand. Then participants rated the confidence they had in the thoughts they listed, and they reported their self-esteem. Because writing with the nondominant hand is difficult and looks shaky (low confidence), whereas writing with the dominant hand is quite easy and looks good (high confidence), it was expected and found that using the nondominant hand decreased the validity associated with the thoughts people had listed. Consequently, the effect of the best or worst qualities manipulation on stated self-esteem was significantly greater when participants wrote their thoughts with their dominant rather than nondominant hand.

Importantly, we argue that the persuasive effect of ease on thought reliance is meaning dependent. People generally perceive difficulty in retrieving thoughts as something bad, thereby associating difficulty with low validity. In contrast, by default, ease is good and therefore associated with high validity. However, as was the case with feelings of power induced by power postures, if people’s naïve theories regarding the meaning of ease vary (or could be manipulated to vary), then different judgments could arise following the experience of ease. In one study investigating this possibility, Briñol, Petty, and Tormala (2006) asked participants to write their thoughts using an easy to read font (black over white) or a difficult font (pink over yellow). In addition, they also manipulated the perceived meaning of ease versus difficulty. Half of the participants were told that ease in generating thoughts generally reflected thoughts that were low in complexity (low validity) and that intelligent people, who have more complex thoughts, typically experienced more difficulty in generating thoughts than unintelligent people. The remaining participants received the opposite information, suggesting that ease was an indicator of intelligence (high validity). Consistent with expectations, results indicated that the traditional fluency effect emerged only among participants who received the “ease is positive” instructions. That is, among these participants, from those who listed their thoughts in black over white format reported more favorable attitudes than did those participants listing positive arguments in pink over yellow. Among participants receiving the “ease is negative” instructions, the opposite effect emerged; this group reported more favorable attitudes when listing thoughts that were difficult to list rather than easy.

In conclusion, in this first section we have introduced the idea that the glue that ties together the studies on the impact of physical actions on judgment is the meaning that those actions have for the people who engage in them. In particular, we demonstrated that when the meaning of an action is something associated with high validity (as is often the case for ease, agreement, happiness, high power, intelligence, etc.), then the action will increase thought usage. In contrast, if the meaning of the very same physical action is associated with low validity (difficulty, disagreement, sadness, low power, stupidity, etc.), then it will decrease thought usage. Furthermore, we went one step further to show that even these default meanings can be modified for certain people or in certain situations. Thus, we showed that if people came to believe that ease or high power were associated with something bad, the typical effects of these normally positive states would reverse.

The Meaning of Actions: Protecting versus Hiding Objectified Thoughts

Subsequent research has shown that variations in the meaning of the action taken with objectified thoughts are also important in this domain. For example, Kim et al. (2014) had participants first writing positive or negative thoughts about increasing their level of physical exercise. Then, all participants were asked to move what they wrote to a box labeled as a “trash can” (low validity) or as a “safety box” (high validity). Results showed that physically moving objectified thoughts to a trash can led them to be mentally discarded as well, whereas physically moving them to a safety box led them to be relied upon more, even though the physical action was the same in both conditions. Conceptually similar results were obtained when people were told to put their thoughts in their pockets but in some conditions this was described as “out of sight” (low validity; e.g., Li, Wei, and Soman 2010; Sparrow, Liu, and Wegner 2011) and in other conditions it was described as “a safe place” (high validity). That is, there was less thought used in the former than the latter condition.

These results are conceptually consistent with work by Hadi and Valenzuela (2014) showing that performing a positive action (e.g., having people hugging a product with the excuse of testing how easy it was to carry around the laboratory) increased product evaluation and the willingness to pay for that product compared to performing other physical actions with neutral meanings. There is probably no need to protect or to hug the object since merely touching an object can be enough to increase perceived ownership and liking of it (Peck and Shu 2009). Future research should examine whether hugging and touching of disliked products also reduces negativity (as it does for previously liked products) or if it polarizes the initially negative response (as shown by Briñol et al. 2013). We argue that the former (touching operating as a valence cue) is more likely to occur under low thinking conditions, whereas the latter (touching validating thoughts) is more likely to happen under high thinking conditions.

An important assumption in this kind of research on “physical contagion” is that ownership probably increases liking and thought usage due to, at least in part, the self being evaluated positively. Consistent with this view based on psychological balance principles (Heider 1958), recent research has shown that ownership effects are more likely to emerge for people high (vs. low) in self-esteem (Horcajo, Briñol, and Petty 2010). Also consistent with this assumption, Argo, Dahl, and Morales (2008) found that products that were touched by a perceived highly attractive person were better evaluated compared to products touched by a person of average attractiveness.

Thought Endowment: Buying and Selling Objectified Thoughts

Given that thoughts can be objectified and treated as physical objects, an interesting extension is to what extent the physical manifestation of our thoughts can be bought, sold, and traded as if they were commercial products. Drawing from research on the endowment effect, we suggest that people will value their objectified thoughts more if those thoughts are perceived as belonging to themselves rather than someone else. In their seminal research on the endowment effect, Kahneman, Knetsch and Thaler (1990) told participants that they would be participating in a simulated market, and they could trade tokens. Participants were assigned either to be the owners or the buyers of the tokens. Participants were told they should indicate their price for buying or selling the tokens. Results indicated the selling prices were more than twice as high as the buying prices.

Although most research on endowment has examined the possession of physical products (such as lottery tickets or coffee mugs), we argue that endowment principles can be also applied to the possession of less tangible objects (such as intellectual or legal property). In line with this proposition, Gascó et al. (2016) examined whether objectified thoughts could also be treated as physical products to be bought and sold. In one of the studies, students were told that they were going to participate in a role-playing exercise where they would have to play the part of a publicist working on an advertising campaign about fast food. The role of the publicist was to come up with ideas, arguments, and slogans for the campaign. Participants were asked to write down either positive or negative thoughts about fast food on different pieces of paper, and then they were assigned either to play the role of buyers or sellers. Participants in the sell condition were told they could sell their thoughts to another student, but by doing this, they would lose the right to use their thoughts in their own campaigns. On the other hand, participants in the buying condition were told they could buy thoughts from other students in order to use them in their own campaigns. Finally, all participants had to indicate their attitudes toward fast food as the main dependent measure. Participants’ self-esteem was also measured as a potential moderator since previous research has suggested that self-esteem might moderate traditional endowment effects (Horcajo et al. 2010).

In this research, selling thoughts to others was found to increase thought usage for low (vs. high) self-esteem participants. This suggests that the action of selling was construed by those with low (vs. high) self-esteem as providing value for the thoughts, making them more useful (“if others are buying my thoughts they might actually have some value”). That is, for low self-esteem people, seeing others buy their thoughts was associated with high validity, thereby increasing thought usage. In contrast, for low self-esteem people, buying others’ thoughts was not informative of validity regarding their own thoughts, therefore keeping them from using them (the effect by default for those with low self-esteem; Briñol and Petty, forthcoming). Among other things, this work on thought endowment illustrates that people can make different inferences from the very same action, which can have different downstream consequences.

The Role of Physical Action

Another important question for future research concerns the extent to which it might be necessary to physically act for behavior to produce attitude change or whether imagining behavior is sufficient to produce the same effects. That is, would visual illusions suggesting that a person acted, providing false feedback about behavior, and computer-controlled digital representations of the person acting in a virtual environment have the same effects as physical actions (e.g., Krishna, Morrin, and Sayin 2014)? It could be that actual behavior (embodiment) produces the same effects as any other form of priming (e.g., semantic priming of behavioral words, recalling past behavior, imaging a behavior, observing a behavior in others). We think that there are several reasons why effects can be stronger for actual rather than imagined actions. First, it is possible that performing an action or embodying an object allows for more precise associations than mere observation of the same object. Consider as an example the research we covered on wearing reading glasses on validation processes (Belding et al. 2012). Although this research had people physically wear the glasses, it could be that merely seeing reading glasses sitting on a table could have produced equivalent effects. Alternatively, seeing reading glasses might prime multiple constructs including intelligence, the elderly, and so forth whereas wearing the item may disambiguate the meaning associated with that item and more clearly prime intelligence. Furthermore, it is possible that performing an action or embodying an object leads to more complex representations (with more associations of different kinds, including experiential memory) than thinking about the action or looking at the object (Niedenthal et al. 2005).

A second way in which observation and embodiment or enclothing of objects may differ is in the way they influence one’s self-concept. The active-self account of prime-to-behavior effects suggests that primes can change the content of one’s self-concept and linking the prime to the self-concept increases the impact of primes on judgments and behavior (DeMarree, Wheeler, and Petty 2005). Perhaps performing physical actions or wearing an external object such as glasses makes one feel more intelligent compared to seeing the glasses. That is, although wearing the glasses and seeing them might both prime the concept of intelligence, it would be mostly those wearing the glasses who come to view themselves as intelligent. To the extent that people feel more intelligent, they are more likely to act intelligently by processing a persuasive message more carefully. If intelligence is primed but people do not incorporate this trait into the self-concept, the prime is less likely to affect behavior and judgments. This is consistent with recent research showing that perceived agency is an important aspect of embodiment effects (Taylor, Lord, and Bond 2009). This is also consistent with research by Peck and Shu (2009), who found that touching an object can increase the perceived ownership of that object, and with research by Kettle and Häubl (2011) showing that signatures influence consumer behavior by priming the self.

A third way in which observation and embodiment or enclothed cognition might differ from other forms of priming is that performing physical actions could function as a stronger prime. Research on priming suggests that stronger primes show larger effects. For example, participants primed with professor for nine minutes answered significantly more questions correctly on an intelligence test than those primed with professor for only two minutes or a control condition (Dijksterhuis and Van Knippenberg 1998). It is possible that although the time of exposure to the objects was held constant in embodiment research, physical acting or physically wearing the object increases the strength or salience of the prime compared to mere observation/imagination and therefore produces stronger effects.

These three possibilities provide some reasons why actual behavior could produce stronger effects than imagined behavior. In addition to these conceptual reasons, there is some empirical evidence suggesting that actual behavior can produce stronger effects than imagined behavior (e.g., Briñol et al. 2013).

Deliberative Use of Behavior for Self-Change

Another important matter to consider is the question of whether physical actions can be used deliberatively to produce changes in one’s own psychological processes. On the one hand, we know that people can deliberately choose to experience negative emotions such as anger when they think that those emotions can help them to achieve a desired goal (e.g., fighting; Tamir and Ford 2012). We also know that people can deliberatively correct their attitudes from the impact of embodiment if they perceive bodily responses to be an unwanted source of bias (Wegner, Erber, and Zanakos 1993). On the other hand, the effects of most bodily responses (and other incidental inductions, including the retrieval of past memories) are likely to be eliminated when people become aware of their incidental nature (e.g., Schwarz and Clore 1983). Therefore, it is not clear whether people can use their behaviors deliberatively to intentionally influence their evaluations.

We think that many of the embodiment effects described in this article can (and are likely to) operate by influencing evaluation without awareness. The presence of bodily responses capable of changing evaluation and the association between those actions and their meanings can indeed work automatically to affect attitudes. However, sometimes consumers might be aware that their bodily responses might be affecting their evaluations, and they might also be aware of the meaning of those actions. When people perceive their actions to change their attitudes, they can respond differently depending on whether they want that potential impact to occur or not (Wegener and Petty 1997). For example, it might be helpful for some consumers to know that their actions can influence their likes and dislikes. In fact, our bodies can provide us with valuable information in many cases (e.g., elevated heart rate and stomach butterflies when encountering a brand informs us that we like that brand). However, if people who believe that their judgments are somehow biased or influenced by their bodily actions do not want this to occur, they may adjust their judgments in a direction opposite to the expected bias (correction processes). Of course, if people think that the bias is desirable, they can even adjust their judgments to enhance the biasing effect (e.g., McCaslin, Petty, and Wegener 2010).

Thus, future research should examine the question of whether, when, and how awareness of the effects of incidental physical experiences can decrease and increase their impact. Future research should also examine the role of awareness and placebo effects when deliberatively choosing to use consumer products designed to improve body postures (e.g., physical devices that help people to keep an erect back).

What Are the Implications for Replicability on Embodiment?

Our view based on understanding the psychological processes underlying embodiment effects and the moderators of these processes may provide unique insight regarding attempts (successful and unsuccessful) to replicate various embodiment effects (see, e.g., Strack 2016). For example, it is now clear from the research we reviewed from the self-validation perspective, that bodily responses and physical actions can produce different (even opposite) effects on attitudes depending on the circumstances. The presence of opposite effects in the attitude change domain can lead to some confusion (e.g., for a historical review, see Briñol and Petty 2012; for a recent example, see Luttrell, Petty, and Xu 2017). It is confusing that something which seems as simple as nodding one’s head, smiling, wearing a baseball cup, wearing a swimsuit, or pushing one’s chest out can both increase and decrease evaluation in a persuasion paradigm (or it might not work at all in producing any persuasion effects). It can also be challenging when the same body-action (e.g., making an expansive posture, smiling, or putting thoughts in a box) can decrease but also increase elaboration and validation processes, leading to different evaluations as a function of argument quality. However, evaluating these effects within the framework of the ELM used throughout our article helps to explain this complexity. Furthermore, our approach is very consistent with McGuire’s (1983) contextualist framework for social psychology. The attitude change research we have presented in this article indicates that any given phenomenon (e.g., relating to body postures, wearing clothes, moving thoughts to boxes and beyond) can produce multiple effects by operating through multiple processes that operate under specific conditions. The ELM presents a guiding framework to organize and comprehend whether, when, and why embodiment effects are likely to appear, when they are not, and what direction of effect is most likely.