Monday, August 29, 2011

Can the important people in your life help you transcend group differences?

Our significant others (even our X's) influence our new relationships (source)
When I first started graduate school I was enthralled by the idea that much of how we relate to other people in our social lives is driven by our past relationships with important/close others. This theory, which my graduate advisor Serena Chen called the relational self suggests that who we are and how we think about the self is fundamentally shaped by our relationships with others. These influential relationships (they must be influential, stable, long-term relationships, though not necessarily positive/well-adjusted ones) create working models-- like scripts in a play that actors follow-- for how to interact with other people. Thus, whenever you are in a context that reminds you of a significant other you should apply these working models.

This can happen pretty much anywhere. Let's say you meet someone who looks like your ex-girlfriend (big X from now on). Your interaction with this person will be based--at least in a small part-- on how you'd expect an interaction with your X to go. So for example, if you expect your X to freak out about you being five minutes late to pick her up at the airport, you'd expect a person who resembles the X to do the same. In addition, if you tend to feel down or blue around your X, again, you should feel a slight dip in your own mood when around this new person.

What I found intriguing about the idea of the relational self was that at any point, if a person reminds you of an influential and important person in your life, you should feel toward that person, just like how you feel toward your important person. This includes even when the resembling other is someone from a different social category. Thus, if you met someone who was a political conservative, but who also happened to resemble a person who you really like, respect, and admire, then you should like, respect, and admire the political conservative you just met, even if you normally hold deep-seated prejudiced attitudes towards conservatives.


This is essentially what I did along with a couple of enterprising honors students (Vicky Lee and Laura Straus) at UC Berkeley and Professor Serena Chen, my advisor. Together, we conducted two separate studies wherein we had undergraduates come into the lab and give detailed descriptions of a significant person from their lives that they like, admire, and have a significant relationship with. We then brought these unsuspecting participants back to a second experiment weeks later. In this experiment, we lied to participants, explaining that they were about to have an interaction with another university student. We then gave them a detailed description of the person before their meeting. This description was designed either to have the same traits as the significant other or not. Additionally, the new person was either described as someone from the same social group as the participant (same ethnicity or same political affiliation) or a different social group. We then measured participants liking impressions for their partner.

What we found was striking! When the new person did not resemble the significant other, participants unilaterally expressed preferences for people from their social category (e.g., liberals liked liberals, whites liked whites). This preference is consistent with decades of research in social psychology: people tend to prefer people who are similar to they are.

What was particularly surprising was that those participants meeting a new person who resembled their significant other actually felt very positively toward that person even when they were from a different social group. That is, liberals tended to dislike their conservative counterparts unless the conservative reminded them of a significant person whom they liked, admired, and respected.

The most striking finding concerned what participants did when they were asked to move rooms and meet the partner. Participants were brought into a separate room where ostensibly the backpack and waterbottle of the new person were placed next to a chair. Participants were instructed to bring their own chair into the room and to place it so that they were seated directly across from their partner. Participants meeting a person who resembled their own significant person sat closer, even when the person was from a different social category.

It is in this fashion that the unique scripts and working models that we develop with our positive significant-other relationships can transcend our underlying stereotypic beliefs toward members of other social categories. Broadly, this research speaks to the powerful influence that positive relational models can have on all our future social interactions with people from a variety of different walks of life.

Is it surprising to you that we use what we know about significant others to transcend group differences? Let us know in the comments.

Kraus, M., Chen, S., Lee, V., & Straus, L. (2010). Transference occurs across group boundaries Journal of Experimental Social Psychology, 46 (6), 1067-1073 DOI: 10.1016/j.jesp.2010.05.003


  1. Not surprising at all. Observation, experience, and imitation are the most fundamental of human learning techniques. It's how babies learn from their parents. It is still how we learn to do many things, especially when it comes to social interaction.

    As far as familiarity transcending other group differences, I would guess that the connection to a significant other is much stronger than any connection made due to race, religion, or political leanings. As a communal species, we reach out enthusiastically for those kinds of connections.

  2. Hey David, thanks for the comment! You're absolutely right that a significant other connection should, in theory, be stronger than any other connection a person has to a social group, and Susan Andersen of NYU has argued this point throughout her research career. Basically, we spend way more time playing out our models for relationships than anything else, and that is why these models are so easily activated.

    Thanks a lot for reading!