A situational touch: How touch affects people's decision behavior

Abstract

While the majority of previous findings have shown that interpersonal touch positively affects human interactions (e.g., cooperation), it is unclear whether touch truly is only positively rather than negatively received, and which situations might influence this differential effect. The present research argued and demonstrated that in a competitive situation touch negatively affects cooperative interactions between the persons involved. Specifically, we showed that in a competitive rather than a supportive environment interpersonal touch on the shoulder will reduce helping behavior towards the person invoking the touch.

The first author is a doctoral fellow of the Agency for Innovation by Science and Technology in Flanders (IWT). This work was supported by STRT1/10/013TBA from the Research Fund of the KULeuven.

Keywords:

• Touch
• Social context
• Nonverbal behavior
• Decision making
• Cooperation

The haptic senses are one of the most essential sources for humans’ physical and mental wellbeing (Weiss, 1986, 1990). Moreover, touch is one of the most important adaptations of human evolution and provides us with essential information even before we are born (Barnett, 1972). The significance that touch has in daily life, however, is contrasted with the attention it has received in behavioral and cognitive research (e.g., Gallace & Spence, 2010). In human development the sense of touch is considered essential in fostering infants’ physical and mental growth (Frank, 1957). A simple touch like maternal tickling, for example, can end a baby's crying and make it laugh instead (Nwokah & Fogel, 1993; Trevarthen, 1979). Even as young as 6 months, children use the sense of touch to acquire information from their environment (Meltzoff & Borton, 1979), and a few hours after labor mothers can already identify their babies by stroking the back of their hand (Kaitz, Lapidota, Bonner, & Eidelman, 1992). Touch is also an integral part of social learning as it communicates distinct emotions (Hertenstein, Holmes, McCullough, & Keltner, 2009; Hertenstein, Keltner, App, Bulleit, & Jaskolka, 2006). Although touch can invoke both positive and negative emotions and perceptions (Floyd, 2000; Gutek, Morasch, & Cohen, 1983; Major & Heslin, 1982; Struckman-Johnson & Struckman-Johnson, 1993), a “common” touch (i.e., excluding sexual harassment and physical aggression) has generally been described as a positive event for consequent behavior (e.g., Gallace & Spence, 2010; Kurzban, 2001). In the present research we argue that touch need not always invoke positive behavior, but can also invoke aversion, yet this is a function of the situation in which touch occurs.

Even though little research has focused on the influence of touch, past studies have consistently reported that interpersonal touch has a positive impact on the subsequent behavioral interactions between the persons involved. Touch enhances prosocial behavior towards the person invoking the touch and fosters people's willingness to agree with a request (e.g., Kleinke, 1977; Patterson, Powell, & Lenihan, 1986). For example, consumers in a supermarket complied more often to both a taste and a purchase request if they were touched by the demonstrator (Hornik, 1992), and students whose teacher touched them on the forearm during the request were more likely to demonstrate the solution to a given exercise on the blackboard (Guéguen, 2004).

Moreover, even in the absence of a request, touch seems to have a positive effect on the receiver's behavior. For example, people who were previously touched on the arm by a stranger were more likely to spontaneously help this person when he or she dropped some diskettes (Guéguen & Fisher-Lokou, 2003), and women felt more positive and evaluated a library clerk more favorably if their hand was briefly touched by this person when returning a book (Fisher, Rytting, & Heslin, 1976). Further supporting the positive impact of touch, research revealed that in certain situations (e.g., for schizophrenic patients) self-induced tickling even provides the receiver with the same benefits as other-induced touch (Blakemore, Smith, Steel, Johnstone, & Frith, 2000). Also, touch could be shown to have a positive influence on people's decision making, as it increased people's cooperative choices in a social dilemma game (Kurzban, 2001). Furthermore, early season touch between players of the same team in NBA competition even predicted an improvement in performance for both individual players and their team later in the season through fostering cooperation between the teammates involved (Kraus, Huang, & Keltner, 2010). However, one exception to these findings is a recent study showing that interpersonal touch between Polish men decreased compliance and helping behavior as a result of homophobic attitudes (Dolinski, 2010). Taken together, we can conclude that previous research indicated that “social touch can have a number of positive effects on people's behavior, but one might question whether this is true in all interpersonal contexts …” (Gallace & Spence, 2010, p. 250).

While the existing research has reported on the positive consequences of interpersonal touch on people's behavior, findings based on people's perceptions of interpersonal touch displayed in photographs indicate that touching others may not always be welcomed (e.g., Floyd, 2000; Gutek et al., 1983). For example, in addition to positive evaluations, touch can also be perceived as dominant, putting the other person in a submissive position (Major & Heslin, 1982). Although these findings seem to contradict each other at first, the differences might be explained by the social situation in which the touch occurs. That is, contrary to touch displayed in photographs, a real-life touch always takes place in a specific situation, possibly affecting the way the touch is received. Furthermore, situations provide a person with information and often have a profound impact on people's subsequent behavior (Rusbult & Van Lange, 2003). Here we argue that, in certain circumstances in which touch may indeed be perceived as negative (i.e., dominant) by the person being touched, touch will have negative implications for people's decisions. That is, although several studies demonstrated that displaying authority (e.g., Cialdini, 2001) and acting in a dominant way (Tiedens & Fragale, 2003) positively affect others’ compliance, it may well be that there are situations in which such dominant behavior will be less accepted. Indeed, Mazur and Cataldo (1989) revealed that dominant individuals (with an initial equal status) tend to elicit perceptions of dominance in others, resulting in decreased cooperative behavior. Furthermore, Lewis and Fry (1977) argued and demonstrated that in a competitive setting people are more likely to use distributive bargaining tactics (i.e., the use of threats and persuasive arguments in order to impose unilateral concessions) if the other person is perceived as acting in a dominant way. Therefore, if touch is perceived as dominant by the person being touched (see Major & Heslin, 1982), this might influence the cooperative interaction in a negative way in a competitive situation.

In Experiment 1 we examined the effect of touch on participants’ decision behavior towards the person touching them in a situation in which touch might not be welcomed (i.e., a competitive situation). Will touching others improve or discourage cooperative behavior in this specific situation? As touch is expected to communicate dominance (Major & Heslin, 1982), and given that situational cues will guide people's subsequent behavior in interactions (Lewis & Fry, 1977; Mulder & Nelissen, 2010), we expected that participants would be less accepting of touch in a competitive situation and thus would be less helpful when touched by another person. In Experiment 2 we further explored the contradiction of research reporting both positive (e.g., Kurzban, 2001) and negative implications (Experiment 1) of interpersonal touch. We argue that the social context is essential for interpreting the social significance of touch. That is, a supportive rather than a competitive context shapes positive interpersonal interaction (e.g., Tedesco & Fromme, 1974). As a result we expect that in a supportive setting people can be positively “touched” by another person (generating cooperative behavior). Yet a competitive context may trigger being “touched at a sore spot” or “touching a raw nerve” and consequently generate non-cooperative behavior. We therefore expected that in a competitive rather than a supportive environment, interpersonal touch will negatively affect cooperative behavior.

Experiment 1

Participants and design

In exchange for course credit 74 undergraduate psychology students (63.8% women), with a mean age of 19.16 years (ranging from 17 to 28 years old) participated individually in this study. They were told that the study examined the effect of the presence of others on performance. Participants were randomly assigned to either the touch or control condition.

Procedure

Prior to the experiment one male and one female student were trained to take the confederate role. During these training sessions the confederates practiced administering three consecutive gentle pats on the shoulder of a participant seated next to them and received feedback on their behavior from the experimenter. We opted for three consecutive pats on the shoulder, rather than one (which has been used in previous research), as for our sample this type of touch is a common type in social interactions, and specifically three pats on the shoulder is more common, and more often used than one single pat. We repeated the training until we judged both the touch to be natural and the other (non)verbal communication of the confederates to be identical between the touch and the no touch condition.

Upon arrival participants were brought to a first room where they received an individual task (a maze). After completing this task participants were asked to follow the experimenter to a second room. Here they were seated next to a fellow (confederate) student and had to complete a second maze. Both participants and the confederate were told that they received the same maze and that they were not allowed to talk to each other during this task. They were also instructed to work individually and to solve this maze as quickly as possible once the starting signal was given. These instructions made the situation inherently competitive. In order to minimize the effect of differences in perceived social status, we ensured that the confederate solved the maze a few seconds slower than each participant. Participants were then asked to move to another room for the last part of the experiment. Participants either left the room without further nonverbal interaction (control condition) or received three consecutive gentle pats on the shoulder by the confederate before leaving the room (touch condition). Both in the touch condition and the control condition the confederate wished the participants good luck with the rest of the experiment and accompanied this statement with a gentle smile. We opted for this procedure in order to minimize possible awkwardness of the touch (as three taps on the shoulder without further verbal statements might be awkward for the receiving person).

As previous research (Major & Heslin, 1982) showed that touch communicates perceptions of dominance, in order to exclude potential fear of retaliation the last part consisted of a dictator game (Hoffman, McCabe, & Smith, 1996) to measure helping behavior. Participants learned they had to divide 11 credits between themselves and the other student in the negotiation. Each credit gave them a chance of winning a movie ticket (so credits were valuable). As a final task, to examine whether participants were aware of being touched or not and whether they experienced the other person as unpleasant (in order to control for unwanted side effects of our touch manipulation), participants were asked to respond to the questions “Has there been any kind of physical contact between yourself and the other participant?” and “Do you find the other person unpleasant?” (on a 5-point scale ranging from 1 = not at all to 5 = very strong).

Results and discussion

Given that previous research (see for example Dolinski, 2010) revealed that the gender of the person invoking the touch as well as the gender of the receiving person sometimes affects people's reactions to interpersonal touch, we included both gender variables (and all possible interaction terms) in our analyses. There were no significant main effects, nor interaction effects, of gender on people's donation behavior. Therefore these will not be reported here. An ANOVA on the amount of credits that participants donated showed a main effect for touch, F(1, 72) = 7.86, p < .01, η²= .10. As predicted, participants in the touch condition donated fewer credits to the other person (M = 3.70; SD = 1.76) than those who were not touched (M = 4.70; SD = 1.27). Furthermore, an ANOVA on participants’ answers to the question of whether interpersonal physical contact was present or absent showed a significant main effect for touch, F(1, 72) = 52.69, p < .001, η²= .42. Participants who were touched on the shoulder by another person stated more often that physical contact was present (M = 2.65; SD = 1.23) than participants who were not touched on the shoulder (M = 1.11; SD = 0.39). Given that participants in the touch condition (or at least some of them) were aware of the fact that they were touched, we explored whether this awareness affected people's donation behavior. In line with previous research (Guéguen, 2002) however, being aware of the touch did not affect people's (donation) behavior. Finally, an ANOVA on participants’ perceptions of unpleasantness of the confederate revealed no main effect for touch, F(1, 72) = 0.51, p = .48, η²= .01. Participants in the touch condition did not find the other person more unpleasant (M = 1.11; SD = 0.39) than participants who were not touched (M = 1.19; SD = 0.57). Furthermore, we also asked participants in the touch condition to what extent they experienced the touch as unpleasant. The average score for this question was 2.71, which seems to indicate that participants experienced the touch as rather neutral. Hence the latter two findings indeed suggest that the difference in donation behavior between the touch condition and the control condition is not due to an (unintentional) awkwardness of the touch.

While the majority of previous research has indicated that interpersonal touch positively affects people's helping behavior and compliance, our findings showed that touch also negatively affects people's decisions and more specifically helping behavior. In Experiment 1 participants donated fewer credits to another person if this person touched them on the shoulder than when touch was absent. In Experiment 2 we examined when touch is perceived as negative or positive.

Experiment 2

Participants and design

A total of 54 undergraduate students (64.8% women), with a mean age of 21.28 years (ranging from 18 to 26 years old), participated in this study. Participants were randomly assigned to one out of four conditions of a 2 (interpersonal touch: touch or control) × 2 (frame: competitive or supportive) between-participants design.

Procedure

At the start of the experiment participants were told that they would participate in two supposedly unrelated experiments. In the “first” experiment participants were introduced to a fellow (confederate) student and were randomly assigned to the frame conditions. In the supportive frame condition participants were seated next to their fellow student (a female confederate who was trained following the procedure described in Experiment 1) and were told that, in order to have a chance of winning a movie ticket, they together had to score higher than a predetermined score in the task (a puzzle). In the competitive frame condition participants were seated at a different table than their fellow student and were told that in order to have a chance to win a movie ticket they had to perform better than the other person in the task (the same puzzle that was used in the supportive frame condition). After finishing the puzzle the fellow student was asked to move to another room for the “second” experiment and either patted participants three times on the shoulder before leaving the room (touch condition) or left the room without further interaction (control condition). As in Experiment 1 the confederate wished the participants good luck with the rest of the experiment and accompanied this statement with a gentle smile. In what was said to be the “second” experiment participants again took part in a dictator game (as a measure of cooperative behavior). They learned that they had to divide 20 credits between themselves and their fellow student. As in Experiment 1 each credit gave participants a chance of winning another movie ticket.

Results and discussion

A 2 (interpersonal touch) ×2 (frame) ANOVA on the amount of credits donated revealed a main effect of frame, F(1, 50) = 38.59, p < .001, η²= .36. Participants in the competitive frame condition (M = 8.14; SD = 2.15) donated fewer credits to the other person than participants in the supportive frame condition (M = 9.76; SD = 0.79). There was also a main effect of interpersonal touch, F(1, 50) = 13.58, p = .001, η²= .13. Participants donated fewer credits to the other person when they received three consecutive pats on the shoulder (M = 8.88; SD = 1.94) than when interpersonal touch was absent (M = 9.33; SD = 1.40). Finally, the ANOVA revealed a significant Frame × Interpersonal Touch interaction effect on the amount of credits that participants donated, F(1, 50) = 26.62, p < .001, η²= .25. An ANOVA revealed that, in the competitive frame condition, participants donated fewer credits to the other person if they were touched on the shoulder (M = 6.14; SD = 1.46) than when interpersonal touch was absent (M = 9.14; SD = 1.70), F(1, 19) = 15.78, p = .001, η²= .45. In contrast, in the supportive frame condition, participants in the touch condition donated slightly more credits to the other person (M = 10.00; SD = 0.01) than those who were not touched on the shoulder (M = 9.50; SD = 1.10), F(1, 31) = 3.55, p < .07, η²= .10.

In sum, in Experiment 2 we argued and showed that the influence of interpersonal touch on the shoulder on people's helping behavior is dependent on the context in which human interactions take place. That is, when participants interacted in a competitive context, interpersonal touch decreased participants’ helping behavior towards the person invoking the touch. In a supportive context (although the effect was only marginally significant), interpersonal touch on the shoulder slightly increased participants’ helping behavior.

General Discussion

Even though previous research largely focused on the positive outcomes of interpersonal touch with respect to the other person's consequent behavior (e.g., Fisher et al., 1976; Kraus et al., 2010), some studies do suggest that touch can communicate negative as well as positive emotions (e.g., Floyd, 2000; Gutek et al., 1983; Major & Heslin, 1982; Struckman-Johnson & Struckman-Johnson, 1993). However, these latter findings are often based on participants’ perceptions of interpersonal touch displayed in photographs—which does not take into account all the aspects of a real-life dynamic touch (Floyd, 1999)—and, more importantly, do not address people's behavior in response to interpersonal touch. Here we add to the existing literature on interpersonal touch, as we show that depending on the social context touch can have negative in addition to positive implications for the ongoing relationship.

Depending on the specific situation, touch can be perceived as uncomfortable and uneasy. More specifically, we argued and demonstrated that in certain situations (i.e., a competitive situation) touch indeed has negative implications for people's cooperative behavior (Experiment 1). Finally, in Experiment 2 we showed that touch may have positive or negative implications for people's helping behavior depending on the context in which people were touched. Interpersonal touch on the shoulder (slightly) increased the receiver's helping behavior if participants interacted in a supportive context, while it decreased helping behavior in a competitive context.

At first glance our findings might seem to contradict previous research by Tiedens and Fragale (2003), which indicated that people tend to take a reciprocal submissive role when they are confronted with a dominant-acting person (e.g., a person invoking a touch). However, a closer look at the series of studies by Tiedens and Fragale (2003) reveals that the task participants had to complete (and during which they were exposed to the dominant behavior) was clearly cooperative in its nature. That is, participants were told that they had to identify the picture projected on the wall behind their head (which they were unable to see themselves) out of a series of pictures, based solely on the description provided to them by a fellow student (a confederate). The procedure used by Tiedens and Fragale (2003) shows several similarities with the procedure that we used to create a supportive situation (see Experiment 2) as participants had to work together in order to achieve their goal. As we showed that in such a supportive situation interpersonal touch (i.e., acting in a dominant way) leads to increased donation behavior (which can be seen as acting in a submissive way), we believe that our results do not contradict but rather build upon the findings reported by Tiedens and Fragale (2003).

Limitations and suggestions for future research

In addition to the contributions described above, our studies have some limitations that need to be acknowledged in order to interpret the findings reported here correctly, and which should be addressed in future research. First of all, we were unable to reveal the process underlying our results and thus explain why participants donated—for example in terms of felt emotions or because they perceived the other person as acting in a dominant way (as we argued). Consequently we cannot completely rule out alternative explanations. One such alternative explanation for our findings can be found in the fact that the confederate completed the task a bit more slowly than the participants. As a result, touch might have been perceived as strategic (i.e., the confederate is trying to enhance his/her profit by being nice to the participant), with lower donations indicating a punishment for this strategic behavior. However, the confederate completed the task almost immediately after the participants and the experimenter did not communicate which person was the winner of the task. As a result we feel confident that this is not a likely explanation for our findings. Nevertheless, future research should explore the process underlying our findings, for example by having participants interact in all conditions while the interpretation of touch is varied. This will allow us to confirm why participants altered their donation behavior depending on the social context in which the touch occurred.

Also, even though we showed that the social situation is essential in shaping how interpersonal touch is received, other factors might subsequently affect this relationship. One such element concerns the quality of the relationship between the persons involved. In the two studies presented above, participants interacted with a person with whom they were not familiar prior to the experiment. Not surprisingly, touch may be interpreted differently depending on the relationship between the person invoking the touch and the person receiving it. For example, previous research has shown that, contrary to participants looking at interpersonal touch between strangers, people who had to imagine that the touch occurred between themselves and their most preferred opposite-sex partner rarely reported perceptions of dominance and control. Instead, perceptions of love and warmth were most typical (Pisano, Wall, & Foster, 1986). In contrast, high-status persons in particular often engage in the act of touching others on the shoulder as a way to communicate their dominant position (Hall, 1996; Henley, 1973), which may result in an increased feeling of uneasiness. Moreover, another person's status indeed affects others’ cooperative behavior (De Kwaadsteniet & van Dijk, 2010). As perceptions of and feelings towards interpersonal touch seem to be affected by relational information, it is likely that people will also behave differently depending on their relationship with the person touching them. Therefore future research should examine whether the relationship between the persons involved affects the way people behave in response to interpersonal touch in a competitive and supportive situation, and other social situations.

In addition to the quality of the relationship between the persons involved, people's behavior in response to interpersonal touch in a certain situation may also differ depending on the specific body part at which the touch is directed. In daily life, for example, people with a high status more often use touch on the shoulder or the arm, while lower-status individuals initiate more formal touches and handshakes (Hall, 1996). Moreover, Burgoon and Walther (1990) revealed that people show different expectations and evaluations towards different kinds of touch (e.g., a handshake, an arm around the shoulder) between strangers. As we did not vary the kind of touch in the present research, we cannot completely rule out that the negative influence of touch is caused by the awkwardness of this specific kind of touch. However, given that our findings indicate a tendency for people to react positively to our touch manipulation in a cooperative context, and that participants in the touch condition experienced the touch as rather neutral (Study 1; M = 2.71; on a 5-point scale), we feel confident that our findings are not due to an unwanted awkwardness of our touch manipulation. Still, although our findings revealed that interpersonal touch on the shoulder is interpreted differently depending on the situation in which it occurs, this might be different for other kinds of touch. Interpersonal touch on the knee, for example, might be perceived as too invasive of one's personal space (Burgoon & Jones, 1976) and consequently affect cooperation in a negative way regardless of the situation in which it occurs. As we focused on interpersonal touch on the shoulder and did not vary the touch in our experiments, we cannot exclude the possibility that other kinds of touch might invoke slightly different behavior (depending on the situation in which these touches occur). We therefore suggest that future research should extend our findings with regard to interpersonal touch directed at other areas of the human body.

Another interesting finding of the present studies concerns the results with regard to the manipulation check used in Experiment 1. As reported above, participants in the touch condition reported more often that interpersonal touch was present than participants in the no touch condition. Although we asked this question as a manipulation check, the obtained results also indicate that participants in the touch condition were clearly aware that the other person had touched them. This finding is especially interesting given the fact that in daily life touch can also be used in very subtle ways, which might result in people not necessarily being aware of the touch they received. Moreover, a majority of previous research—often providing evidence for the positive consequences of interpersonal touch—has used more subtle touch manipulations (e.g., Fisher et al., 1976; Guéguen & Fisher-Lokou, 2003) than the one used in our studies. Surprisingly, only one study (Guéguen, 2002)—at least to our knowledge—has examined whether identical touch elicits different behavioral reactions depending on the receiving person being either aware or unaware of the touch. Although being aware of the touch had no impact on the donation behavior in our study, it may be interesting for future research to further explore whether this affects the way touch is received and reacted upon.

In Experiment 1 we argued that interpersonal touch negatively affects people's donation behavior in a competitive situation. However, even though our findings provided support for the hypothesized effect, we did not explicitly measure whether participants indeed perceived the situation as competitive and thus we cannot completely rule out alternative explanations. As such, even though Experiment 2 provided supplementary evidence for the negative effect of interpersonal touch on people's helping behavior in a competitive situation, future research is needed to measure people's perceptions in specific situations and explore whether these perceptions are affected by the kind of interpersonal touch that occurred (i.e., people's perceptions of a competitive situation might alter depending on the number of pats they receive or on which specific body part the touch occurs).

A final limitation of our studies concerns the potential effects of the gender of both the confederate and the person receiving the touch. Although we were unable to retrieve any gender effects in a competitive situation (Experiment 1), it could well be that the gender of both the person invoking the touch and the person receiving the touch affects people's donation behavior in a supportive situation. As we did not vary the gender of the confederate in our second experiment, we were unable to examine whether situational effects of interpersonal touch are independent of people's gender. Therefore we suggest that future research should examine whether the gender of the person invoking the touch and the gender of the receiving person affect the impact of interpersonal touch in different social situations.

Conclusion

In general the present research was able to show that the haptic senses, and more specifically touch on the shoulder, have a significant impact on people's behavior. Our findings show that the simple act of touching others during social interaction can have profound consequences for subsequent interactions, and that the context of the interaction moderates these consequent social interactions. Despite what some people might think, touching someone else may thus not always have desirable social consequences. A simple tap on the shoulder, even with the best intent, will do nothing but harm when used in the wrong place at the wrong time.

Notes

The first author is a doctoral fellow of the Agency for Innovation by Science and Technology in Flanders (IWT). This work was supported by STRT1/10/013TBA from the Research Fund of the KULeuven.