Persuasive agents and the occurrence of reactance as a result of restricting communication

M.A.J. Roubroeks, C.J.H. Midden, J.R.C. Ham

Research output: Contribution to conferenceOtherAcademic

1 Downloads (Pure)


Nowadays, technologies are being developed to persuade people. Examples of such technologies are embodied conversational agents, robots, or even the personal computer. When people are the subject of persuasion, they are directed to perform a specific behavior. For example, they can be directed to stop smoking, start exercising or increase energy conservation. In response to such messages people can conform and act on the desired behavior. However, people can also experience psychological reactance and consequently perform the opposite of the intended behavior. But when does this happen and why? And more interestingly, does psychological reactance also occurs when interacting with an agent, rather than a human being? According to the theory of psychological reactance (Brehm, 1966), psychological reactance is defined by; "…a state of arousal that occurs when a person experiences a perceived threat to freedom, which motivates to restore that threatened freedom." A model is proposed that consists of four elements; an experienced intrusion to autonomy, a perceived freedom, psychological reactance, and a desire to restore the threatened freedom. First, people perceive that they have the freedom to choose. Then, people are being persuaded to perform a specific behavior, which is experienced as an intrusion to their autonomy. This results in a state of arousal, which is called psychological reactance. This will eventually lead participants to feel the desire to restore that threatened freedom. It is hypothesized, as in line with previous studies, that more psychological reactance will be experienced by a highthreatening message compared with a low-threatening message, or a no-threatening message. Furthermore, it is expected that as a result of this experience of psychological reactance, people will report higher intentions to restore the threatened freedom.In addition, we were interested in the social nature of psychological reactance. The authors of this article suggest that psychological reactance is a social phenomenon that is the outcome of an interaction between human beings. We further suggest that this can also be true when interacting with a social agent. Evidence for this hypothesis can be found from The Media Equation (Reeves & Nass, 2002), in which it is stated that people react to computers as if they were reacting to other people. Furthermore, according to the Social Agency Theory, the more social cues that are available in an interaction, the more social the interaction becomes, and consequently leads people to try to better understand the relationship with the other actor (Mayer, Sobko, & Mautone, 2003). We hypothesize that this also holds for psychological reactance; the more social cues, the more people experience psychological reactance. Lastly, it is hypothesized that the most psychological reactance will occur when a high-threatening message is combined with a high social agent. These hypotheses were tested by using a 3 (Threat: No Threat vs. Low Threat vs. High Threat) x 3 (Social Agency: Text vs. Still picture vs. Brief moving film clip) between-subjects experimental design. Eighty-nine participants participated in an online study, in which they had to read an advisory text about energy conservation with the washing machine. After reading this text, participants were requested to perform a though-listing task and answer some questionnaires. Participants were randomly assigned to a no-threatening message, a low-threatening message or a high-threatening message. This advice was either delivered solely by text, by a text that was accompanied by a still picture of a robotic agent, or a text accompanied by a brief film clip of the same robotic agent. Results show that more psychological reactance was reported by a high-threatening message compared to a low-threatening message or a no-threatening message. Furthermore, it was found that the more social cues the agent possessed, participants experienced the more psychological reactance. Unfortunately, we did not find any support for an interaction between the amount of threat and the level of social agency leading to the strongest reactance when high threat would co-occur with high social agency.
Original languageEnglish
Publication statusPublished - 2009
Eventconference; 2nd International Conference on Human-Robot Personal Relationships; 2009-06-12; 2009-06-12 -
Duration: 12 Jun 200912 Jun 2009


Conferenceconference; 2nd International Conference on Human-Robot Personal Relationships; 2009-06-12; 2009-06-12
Other2nd International Conference on Human-Robot Personal Relationships


Dive into the research topics of 'Persuasive agents and the occurrence of reactance as a result of restricting communication'. Together they form a unique fingerprint.

Cite this