Which of the following theoretical perspectives supports the claim that outgoing personality is a result of reinforcement?

J Pers. Author manuscript; available in PMC 2012 Oct 1.

Published in final edited form as:

PMCID: PMC3094598

NIHMSID: NIHMS265782

Abstract

Close college-age friendships provide differential opportunities for reinforcing dispositional tendencies and fostering accommodation or change. This finding was obtained from a cross-sectional study of 66 pairs of same-sex college-age friends (58% female). Each pair of friends was extreme and either very similar or different with regard to extraversion-introversion. Interviews with each friend were analyzed for references to each other's role in various friendship domains, including the setting of the friendship and position with regard to chatting, disclosing, expressing opinions about peers, and energizing the friendship. Matched friends mutually reinforced each other's similar dispositional tendencies. Friends with contrasting personalities showed patterns of personality accommodation as well as complementary reinforcement. Implications are discussed for embedding reciprocal theories of personality development in close friendships.

Keywords: Extraversion-introversion, reciprocal-interaction, friendship practices, mutual reinforcement, complementarity

One person's experience of another person has a special quality, quite different from his perception of a piece of cheese…. It is marked by a recognition of mutuality…accompanied by an appreciation of the feelings of the other person and some willingness to adjust to them…. It is difficult to interpret any interpersonal proceeding without knowledge of the history of both personalities and a knowledge of their current thoughts and feelings. (Murray, 1951, pp. 438–440)

Observational studies have long recognized that particular personalities produce a distinctive social press or force (Block, 1971; Funder, 1999; Murray, 1938). To date, the social force of personality has usually been observed unilaterally, with less attention to how the personality of the observer interacts with that of the observed. In addition, studies of the bilateral force of personalities have usually focused on strangers rather than people who have a long history of interacting with each other (Sadler & Woody, 2003; Thorne, 1987; Tracey, 2004). While observations of strangers can reveal the instantaneous press of personality, interviews with friends can reveal how they come to live with and potentially adjust to each other's dispositional similarities and differences. Accordingly, this unprecedented study explored how close friends who were extreme and either very similar or very different with regard to extraversion-introversion reportedly engaged in an array of friendship practices.

Mutuality, Complementarity, and Accommodation in Person-Situation Interactions

The dynamic interaction of personality and environments has increasingly gained broad theoretical attention (e.g., Caspi & Roberts, 2001; Fleeson, 2004; Mischel & Shoda, 1998). A commonality among dynamic approaches is adherence to the generally accepted view that personality has both bio-genetic and psychosocial foundations. New, however, is an emphasis on examining trait processes (e.g., how a person manages to enthusiastically chat with strangers) in addition to outcomes (e.g., I like to talk with strangers), and on being more situated, specific, and explanatory regarding what is meant by global dispositions such as extraversion. Dynamic approaches also emphasize that individuals are not necessarily passive agents who respond on call to their environments. For example, individuals have been shown to actively choose and construct environments that reinforce their dispositional tendencies (Caspi, Elder, & Bem, 1987; Emmons, Diener, & Larsen, 1986; Roberts & Robins, 2004). Self-reinforcement theories of personality stability, however, assume a pliable environment that bends to the force of an individual's personality. What if the environment is not so pliable? What if the environment is peopled with personalities that push back?

The customary answer to how personality can be consistently reinforced across social environments is that individuals choose friends and mates whose personalities are similar to their own. A number of studies have found evidence of personality similarity between friends and spouses (e.g., Blaz, 1983; Byrne, 1971; Dishion, Patterson, Stoolmiller, & Skinner, 1991; Klohnen & Luo, 2003). Because similar personalities should enhance the potential for sharing broad likes and dislikes and engaging in similar activities, friends with similar personalities have been theorized to mutually reinforce each other's behavior patterns (Caspi & Roberts, 2001). However, the degree to which people choose friends whose personalities are similar to their own has been found to be modest, suggesting that many friendships survive in spite of or because of personality differences (Hamm, 2000; Umphress, Smith-Crowe, Brief, Dietz, & Watkins, 2007).

Surprisingly, close friends with significant personality differences rarely have been observed together in a systematic manner. This lack of empirical attention may reflect a common assumption that friendships between different personalities are rare and/or can come to no good. However, two theories, Carson's (1969) interpersonal theory and Aron and Aron's (1996) self-expansion theory, challenge this assumption. Interpersonal theory (Carson, 1969) posits that opposites attract for traits related to control; for example, in a complementary fashion dominance invites submission and vice versa. In contrast, for affiliative traits similarities are theorized to mutually attract; for example, warmth invites warmth and coldness invites coldness. Although considerable support has been found for dominant behavior evoking submissive behavior and warmth evoking warmth, these behavioral patterns have not been established with bilateral studies of personality (Sadler & Woody, 2003; Tracey, 2004). Another limitation of interpersonal theory is that many traits are combinations of control and affiliation, with extraversion being a notable example (McCrae & Costa, 1989b). With regard to extraversion, interpersonal theory would predict that extraverts are attracted to both introverts and extraverts. Extraverts should be attracted to introverts because introverts are more likely to submit to and thereby reinforce in a complementary fashion the extravert's social dominance. On the other hand, extraverts should be attracted to other extraverts because extraverts are more likely to mutually reinforce each other's positivity and warmth.

Aron and Aron (1996) proposed a broader understanding of why personality differences can be attractive. They argued that people develop close relationships with dissimilar others in order to enhance or expand the self. According to self-expansion theory, when people experience feeling close with each other they have integrated the “other's resources, perspective, and characteristics into the self” (Aron, Paris, & Aron, 1995, p. 1103). To test the hypothesis that close relationships expand the self, Aron et al. (1995) tracked the romantic status of college students over a period of three months. Every two weeks the students wrote a list of self-descriptive terms and completed a “fallen in love” measure. When students were in love they reported more diverse selves, suggesting that close relationships produce more complex selves. In a similar vein, friends with different personalities might accommodate to each other's distinctiveness to expand their selves. This expansion may not only be cognitive but also based in action, involving accommodation to the roles or activities of each friend. For example, extraverted friends might come to appreciate the stillness of bird watching with their introverted friends, and introverted friends might enjoy being a sidekick on the social adventures of extraverted friends.

Extraversion-Introversion and Friendship Practices

Although extraverts generally report having more friends than introverts (Selfhout et al., 2010), most introverts do report having friends, simply fewer of them. A friendship with an introvert, however, may be qualitatively different than a friendship with an extravert. The nature of these friendship qualities is suggested by a host of prior correlational findings. Extraversion is broadly associated with sociability (McCrae & Costa, 1989b), and is associated more specifically with social dominance, affiliation, adventure, ambition, and energy (Watson & Clark, 1997). More recently, close ties between extraversion, reward sensitivity, and positive affect have challenged the notion that sociability is the key feature of extraversion (e.g., Lucas, Diener, Grob, Suh, & Shao, 2000). However, when Ashton, Paunonen, and Lee (2002) separated social attention from reward sensitivity and positive affect, they found that extraversion was most strongly connected to social attention. They concluded that extraversion is not simply sociability, but more specifically the tendency to both participate in and take pleasure in social interaction (Ashton, et al., 2002).

Because social interaction is so central to friendship, the dimension of extraversion-introversion should be associated with a number of friendship activities or domains. Prior research has articulated several such domains of friendship, including where and what kinds of activities friends engage in, the provision of energy to mobilize the friendship, and how friends participate in conversation or dialog and intimacy (Radmacher & Azmitia, 2006; Sharabany, Gershoni, & Hofman, 1981; Vitaro, Boivin, & Bukowski, 2009). Companionship occurs in specific places or locales. Extraversion is linked to more socially outgoing activities, such as competing in team sports and partying (Argyle & Lu, 1990), while the other pole of the dimension, introversion, is associated with less physically active pursuits that manifest closer to home, such as watching television or reading (Argyle & Lu, 1992; Eysenck, Nias, & Cox, 1982). With regard to energy, the more adventurous spirit of extraverts would seem to mobilize or propel a friendship, whereas the more cautious nature of introverts may help to stabilize the friendship (Cuperman & Ickes, 2009). With regard to conversation and dialog, extraverted individuals have been found to more often take the lead and assert their voice (Cuperman & Ickes, 2009; Rim, 1977), and to be more skilled than introverts at talking to strangers (Thorne, 1987). Introverts, on the other hand, have been found to be more skilled at activities that are potentially advantageous for listening, such as focused concentration (Blumenthal, 2001). With regard to intimacy and self-disclosure, introverts have been found to be more sensitive to issues of privacy (Stone, 1996), and extraverts to be more prone to share personal feelings (Peter, Valkenburn, & Schouten, 2005). The present study extended these findings, which are largely based on unilateral studies of individuals, to explore how both friends' personalities reportedly shaped the roles that each friend took in these domains of friendship.

The Present Study

Friendships during the college years are fertile ground for studying the reciprocal impact of personality because peers play an increasingly important role in the lives of adolescents and emerging adults (Arnett, 2000; Rubin, Bukowski, & Parker, 1998). During the college years, youth tend to be living away from home for the first time, starting new friendships, and exploring new venues of mutuality and intimacy. Personality may be particularly open to social influence during this transitional period. For example, a recent meta-analysis found that personality traits changed more in young adulthood than in any other period of the life course (Roberts, Walton, & Viechtbauer, 2006).

To date, however, close friendships in young adulthood typically have been studied at some remove from the activities in which friends actually engage. Friendships usually have been assessed using global surveys that rate levels of support and conflict (Mendelson & Kay, 2003; Sturaro et al., 2008). Although such global ratings indicate the general flavor of the relationship (e.g., “My friend would help me if I needed it,” “I can get into fights with my friend,” and “I feel happy when I am with my friend;” Bukowski, Hoza, & Boivin, 1994, p. 475) discerning the role that each friend plays in various friendship domains requires methods that are sensitive to what emerges in particular friendships.

To achieve a proximal understanding of the experience of each other's personality in particular domains of friendship, the present study explored the press of personality phenomenologically, in the words of the friends themselves, based on a private interview with each friend. The research built on a prior study in which interviews were used to inform understanding of the press of similar and different personalities in brief conversations between strangers (Thorne, 1987). In those interviews, introverts said that the conversation with the extraverted stranger was a breath of fresh air because they could relax and simply let the extravert talk. Some extraverts, on the other hand, said they did not feel so pressed to say nice things and to refrain from complaining when talking with an introverted versus extraverted stranger.

In designing the present study, we reasoned that friends, compared to strangers, have had much more occasion to observe each other's behavior and to experience each other's characteristic conduct and concerns. Therefore, interviews with each friend were expected to go beyond the immediate interaction to identify the characteristic activities and roles enacted by each friend in the relationship. Interviews with each friend were analyzed quasi-inductively by systematically coding accounts of each friend's characteristic roles in each of six domains. Based on prior studies of friendship, we examined the characteristic roles that each friend played in the domains of Locale (where the friendship generally took place—away from home or close to home), Energy (who energized the friendship and who steadied the friendship), Dialog (who did most of the talking and who did most of the listening), and Intimacy (who usually disclosed personal information and who usually contained personal information). We also studied two additional friendship domains, the general attitudes toward peers or Outsiders, and the primary Content of conversations. The latter domains emerged in the coding process as the informants discussed what each friend usually talked about with regard to their own life (Content) and the lives of their peers (Outsiders).

We developed three hypotheses to anticipate the roles played by each friend in matched (E-E and I-I) and mixed (E-I) dyads. The first hypothesis concerned mutual reinforcement in matched dyads:

(H1) Friends in matched dyads will a) report similar roles, b) and the roles will be consistent with the personality of each individual.

This hypothesis draws from the work of Caspi and Roberts (2001), who suggested that people seek out friends whose personalities are similar to their own because the similarity reinforces their own personality tendencies. Such self-reinforcement was found in a comparison of the conversational styles of mutually extraverted versus mutually introverted strangers (Thorne, 1987), but has yet to be established with friends. In the domain of Locale, we expected that mutually extraverted friends would be more likely to describe their friendship as taking place away from home while mutually introverted friends would more likely describe their friendship as taking place close to home (Argyle & Lu, 1990). In the domain of Energy, we expected mutually extraverted friends to report taking the lead in energizing or propelling the friendship and mutually introverted friends as stabilizing or anchoring the friendship (Cuperman & Ickes, 2009). We expected that in the domain of Dialog, mutually extraverted friends would more often be described as the talkers, while mutually introverted friends would more often be described as the listeners (Rim, 1977). In the realm of Intimacy, we expected mutually extraverted friends to report as disclosers and mutually introverted friends to report as confidants (Levesque, Steciuk, & Ledley, 2002).

The most intriguing and novel question in the present study concerned the roles of friends in mixed dyads, one of whom was extremely introverted and the other extremely extraverted. One viable possibility is that friends in mixed dyads would reinforce each other's dispositional tendencies in a complementary fashion. This hypothesis draws from interpersonal theory (Carson, 1969), which would predict that the control aspect of extraversion would accentuate the extravert's tendency to be socially dominant and the introvert's tendency to be socially submissive. For example, complementarity might be manifest in the domain of Dialog by the extravert reportedly taking the role of talker and the introvert reportedly taking the role of listener, by the extravert serving as relationship energizer and the introvert as relationship stabilizer, and the other dispositional role differences referenced in hypothesis 1. Thus, the second hypothesis predicted complementary reinforcement in mixed dyads:

H(2): Friends in mixed dyads will a) report complementary roles that are b) consistent with the personality of the individual.

The third hypothesis also concerned mixed friendship but countered the second hypothesis. Rather than reinforcing dispositional tendencies, mixed friendships may require stretching to meet the friend “where they live.” This hypothesis also draws on the self-reinforcement theory of Caspi and Roberts (2001). If people seek out friends whose personalities are similar to their own because such similarity is self-reinforcing, friends with different personalities might disrupt each other's typical expression of personality. Such disruption might result in accommodation to each other's dispositional routines, potentially enriching one's skill set, in accordance with self-expansion theory (Aron & Aron, 1996). Also in support of the accommodation hypothesis is that the mutuality of a friendship might be compromised if friends always assume different roles. For example, if one friend typically wants to stay home and the other usually wants to go out, some accommodation would have to be made for the friendship to endure. Because extraverts tend to be more socially dominant than introverts (McCrae & Costa, 1989b) and dominant personalities press for submission on the part of the interaction partner (Carson, 1969), our final hypothesis predicted that if role accommodation in the mixed personality dyads occurred, accommodation would be unilateral: the introvert would adopt the extravert's practices. For example, in mixed dyads, the introverted friend would be more likely to hang out at parties with the extraverted friend, and forgo his/her preference to associate closer to home. Thus, the third hypothesis predicted accommodation on the part of the introverted friend in mixed dyads:

(H3) In mixed dyads, when accommodation occurs, the introverted friend will report accommodating to the dispositional role of the extraverted friend rather than vice versa.

Method

Participants

Participants were 66 pairs of same-sex friends consisting of 19 matched introvert pairs (58% female), 22 matched extravert pairs (45% female), and 25 mixed introvert-extravert pairs (72% female). Participants averaged 19.6 years of age (SD = 0.9) and attended a public university in Northern California. They had been friends for at least six months (62% of the pairs either currently or had formerly lived together; 38% had never lived together). All participants were required to be native English speakers, and 90% self-identified as white or European American. One member of each dyad was recruited on the basis of pre-testing in a large psychology course, for which participation in the study fulfilled a course requirement. The recruit subsequently brought along a friend who was compensated $20.

Procedure

Recruitment and participant selection

One member of each dyad was part of a pre-test group, averaging 250 students per academic quarter, recruited between Fall, 1999, and Spring, 2002. Students were administered a survey in large psychology courses; the survey included 10 extraversion-introversion items and demographic questions. To determine cut-offs for recruiting extraverts and introverts, scale scores in the Fall, 1999 pre-test sample were compiled into a distribution (M = 15.2; SD = 2.9, range = 10–20), and students scoring in the upper and lower quartiles were identified as candidates. The scores for extraverted candidates ranged from 18–20, and introverted candidates ranged from 10 to 12. These cut-offs were maintained for subsequent pre-test samples, which showed very similar distributions. Approximately nine candidates were randomly recruited each academic quarter, representing a small portion of the approximately 125 potential participants available. This small percentage reflected the difficulty of scheduling a two-hour session in which the candidate was available to bring a friend along to the study.

A few days after taking the survey, candidates were telephoned and invited to bring a same-sex friend whom they had known for at least six months to participate in a two-hour “friendship study.” The personality of the friend was identified at the end of the study when the extraversion-introversion scale was re-administered. Cut-offs for the friend were relaxed one scale point (10–13 for introverts, 17–20 for extraverts) to obtain viable sample sizes. Friendship pairs consisting of two high scoring extraverts, two high scoring introverts, or a partner who scored high on extraversion and a partner who scored high on introversion were selected to form the three groups in the present study. A total of 79 dyads were run in the study, 66 of whom met the selection criteria; the remaining 13 dyads consisted of a high scoring introvert who brought along a mid-scoring friend (n = 8 dyads), and a high scoring extravert who brought along a mid-scoring friend (n = 5 dyads). These 13 dyads were not included in the present study.

Conversation

Upon arriving for the study, an undergraduate research assistant who was the same sex as the dyad greeted the friends. The research assistant was blind to the personality scores of the friends. The dyad was directed to a comfortable room decorated with children's art and seated on couches arranged in an L-shape. The dyad was informed that we were interested in understanding friendship dynamics. They were asked to first engage in a brief 10-minute conversation in order to catch up or talk about whatever came to mind. The research assistant promised anonymity, acquired consent for audio-recording the conversation, and then left the room. Conversations were ended when the research assistant knocked on the door. These conversations have been analyzed in prior studies (Thorne, Korobov, & Morgan, 2007; Thorne, Shapiro, Cardilla, Korobov, & Nelson, 2009) and are ancillary to the present study, which focused on interviews with each friend.

Interview

At the end of the conversation, the research assistant interviewed one of the friends in an adjacent private room while the other friend separately completed various questionnaires; the friends then switched rooms and activities. The same research assistant separately interviewed both friends in a dyad. The interview was audio-recorded and typically lasted 30 to 40 minutes.

The interview provided an opportunity for each friend to discuss the history and behavioral patterns of the friendship in increasing depth. Throughout the interviews, the research assistant treated participants as expert informants concerning their behavior and their friendship. The interviewer followed a semi-structured interview protocol. The first part of the interview began with collecting general information about the friendship, “Could you tell me how you got to know each other? When and how did you meet?” Next, the interviewer asked about what the friends did together and what stood out in this friendship compared to their other relationships.

In the second part of the interview the research assistant made use of the friends' catch-up conversation to further explore friendship dynamics. The interviewer described to the participant how the audiotape of the conversation would be played back for both of them to hear. The interviewer or the participant could stop the tape at any time to discuss what was happening. For example, if the friends started to laugh in the conversation, the interviewer might pause the tape and ask the participant to explain the laughter. This procedure allowed participants to elaborate on the meaning of specific communications. When asked to explain something in the conversation participants often reported background stories concerning the relationship that enhanced the interviewer's understanding of the friendship beyond the conversation itself, for example, “We always do that. Like one time, we….” The interviewer would also ask whether conduct referenced in the conversation was common for the friendship, or unique to this particular conversation. At the end of the interview, the participant was asked to add anything about the friendship not already discussed.

Questionnaires

While each friend was being interviewed, the other friend was administered surveys about the friendship, the conversation, and their personality. Personality was assessed to determine the extraversion-introversion score of the friend who had been brought along.

Debriefing

At no point were the participants told the study concerned extraversion or, in general, personality. This was done to avoid priming the participants to think about each other in terms of personality types (see Thorne, 1987). With regard to the intent of the study, participants were only informed that the research centered on friendship dynamics.

Measures

Extraversion-Introversion scale

Because initial recruitment occurred during class time, 10-items from the Extraversion-Introversion (E-I) scale of the Myers-Briggs Type Indicator (MBTI), Form M (Briggs & Myers, 1998) were used to quickly assess the personality of potential participants. The MBTI extraversion scale contains forced-choice items (1 = introverted direction; 2 = extraverted direction) that mainly refer to preferences for either sociability or reserve in general social settings. Scale scores ranged from 10 to 20, with higher scores indicating extraversion. The internal consistency of the 10-item scale was acceptable (∂ = 0.83), and the subset of items correlated highly with the full 21-item MBTI E-I scale, r(158) = .93, p < .001. The full MBTI E-I scale has been found to correlate strongly with other commonly used extraversion scales (see McCrae & Costa, 1989a; Thorne & Gough, 1991). We also administered both the Big Five Inventory (BFI Binet-Martinez & John, 1998) extraversion items and the 10 MBTI extraversion items to college students (74% female, 57% European American); the two scales were found to correlate robustly, r(87) = .79, p < .001.

Friendship ratings

Participants completed a survey that asked them to rate the closeness of this friendship from 1 (not at all) to 5 (extremely), and the length of the friendship in months.

Conversation typicality

Participants were also asked to rate how typical the recorded conversation was in comparison to their ordinary conversations in the friendship from a 1 (not at all) to a 5 (extremely).

Coding Interviews for Friendship Roles

The concept of friendship role refers to how a friend characteristically behaves in a specific domain of the friendship. For example, in the domain of Dialog, one friend might reportedly do more of the talking in the friendship, whereas another friend might reportedly do more of the listening, or both friends might report being the talkers. In developing a set of friendship roles based on what informants said in the interviews about their own typical behavior and that of the friend, we followed a quasi-inductive thematic analysis procedure (Braun & Clarke, 2006) based on an essentially realist position: we took for granted what the informants said. However, because we separately interviewed each friend, we were able to assess whether the friend's reports confirmed each other's views regarding the roles each friend routinely took in the relationship, thereby bolstering construct validity (see Paunonen & O'Neill, 2010). In addition, the fact that characteristic roles were discussed in the context of hearing the playback of the recorded conversation helped to ground the findings in actual behavior. For example, while listening to the recording of the conversation, the interviewer might notice that the friends talked about going to a party. The interviewer could then ask the friend if going to parties together was a common activity for the friends. Participants usually said that the behavior referenced in their conversation was characteristic of the friendship, but sometimes participants would elaborate on how the friendship was different or more complex than the conversation portrayed.

The system for coding friendship roles was developed by two of the authors who were blind to the personality test scores of the participants. The authors repeatedly read a random subset of the interview transcripts with increasing depth and attention to what the friends reported doing together, what they said they did for each other, and why they said they appreciated each other. Numerous discussions and passes through a random subset of the transcripts led to the creation of six domains of friendship practices, each of which consisted of two complementary roles that the friends could characteristically enact in the practice of their friendship. The following definitions of each domain and its dichotomous roles are also summarized in Table 1.

Table 1

Definition and Examples of Roles in each Domain of Friendship

Locale: outgoing/homebody

The Locale domain identified where the friends generally spent their time together. Partners who emphasized engaging in friendship activities away from home were coded as “outgoing.” Partying, playing sports, traveling, or going to concerts exemplified activities in the outgoing locale. Partners who reported spending time with their friend in proximity to where they lived were coded as a “homebody.” Evidence for this category included friendship activities such as watching television, playing computer games, talking, eating, or studying (kappa = .80).

Energy: propeller/anchor

The Energy domain referred to who infused the friendship with activity and excitement and who tended to keep friendship activities on track and stable. Friends who provided the motion and energetically sped up the relationship were coded as “propellers.” Propellers reported leading conversations by introducing new topics, often described in a “rambling” or “random” fashion, and/or by taking the lead in planning the friends' activities. Friends who reported slowing down and/or stabilizing relationship activities were coded as “anchors.” Evidence of slowing down the friendship included reportedly keeping the conversation on a single topic or exhibiting reluctance and reflection when offered opportunities to engage in spontaneous activities (kappa = .78).

Outsiders: buddy/judge

The Outsiders domain referred to the role or stance the friends took toward peers outside of their relationship. Friends who reported taking a generally friendly position towards outsiders were coded as a “buddy.” Evidence for this role included descriptions of spending lots of time with their friend in groups of other people or generally talking about peers in a positive or neutral, rather than primarily critical, manner. Friends who reported generally taking a more negative or critical position toward peers in conversations with the friend were coded as “judges.” Evidence for this role included reports of friends bonding by defining themselves against others or exchanging critical judgments of other people (kappa = .78).

Dialog: sounder/sounding board

The Dialog domain focused on who mainly spoke and who mainly listened in the friendship. Friends who were described as doing a lot of talking in the relationship were coded as “sounders.” Friends who were described as typically taking a listener role and of asking questions were coded as “sounding boards.” If the friends described taking turns equally between listening and talking, both were designated as “sounders” (kappa = .86).

Content: dramatizer/hobbiest

The Content domain referred to what the friends characteristically talked about in their conversations. Friends who reported talking mostly about people in social dramas that were emotionally charged, concerning their own lives or the lives of others, were coded as “dramatizers.” Friends who reported detailed conversations about hobbies (e.g., photography, quilting, or computing) were coded as “hobbiests” (kappa = .82).

Intimacy: spout/container

Lastly, the Intimacy domain pertained to whether friends shared personal and emotional concerns with each other. People who reported comfort with voluntarily discussing their personal and emotional experiences with the friend were coded as “spouts”. We coded friends who more often reported keeping such concerns private and/or being trusted holders of their friend's secrets, as “containers” (kappa = .83).

Coders for friendship roles were the first author and two research assistants who were also blind to the personality scores of the participants; approximately 60% of the transcripts were coded for reliability. All coding categories were dichotomous (present or absent in the transcript), and subcategories were mutually exclusive (e.g, someone who was characterized as a “sounder” could not also be characterized as a “sounding board”). The individual friend was the unit of analysis because friends could report taking similar roles (e.g., both sounders) or different roles (e.g., one sounder, one sounding board). Coders read the two interviews and compared what each friend said about him/herself, the friend, and the relationship. Because friendship roles identified patterns of behavior, the roles were coded for how the friend usually acted in this relationship. Rare disagreements (< 5% of accounts) between the two friends about a friend's role were resolved by adhering to the more specific and descriptive account that seemed to best reflect the domain of practice in question. Lack of reference to a person's role in a particular domain was coded “no mention.” As noted earlier, coder reliability for each domain ranged from a kappa of .78 to .86, indicating very good agreement. A detailed manual of the coding process used to train coders is available from the authors (Nelson & Thorne, 2009).

Results

Background Statistical Analyses

Comparison of personality scores

Introverts in mixed dyads were not significantly different in their introversion scores (M = 11.76, SD = 1.61) than introverts in matched dyads (M = 11.66, SD = 1.49), t(61) = .26, p = .796, d =.06. Similarly, extraverts in mixed dyads were not significantly different in their extraversion scores (M = 18.64, SD = 1.08) than extraverts in matched dyads (M = 18.77, SD = 1.10), t(67) = .48, p = .628, d = .12.

Conversation typicality

Participants generally rated the recorded conversation as very typical of their usual conversations, with typicality ratings averaging four on a five-point scale. One dyad in the mixed group and one dyad in the matched introvert group did not provide typicality ratings. Ratings between friends were not significantly different according to a paired t-test, t(63) = .33, p = .739, so average ratings for both friends were compared across the three types of dyads. No significant differences in the typicality of the conversation were found among matched introverted dyads (M = 3.97, SD = .81), matched extraverted dyads (M = 4.36, SD = .52), and mixed dyads (M = 4.38, SD = .66), F(2, 61) = 2.36, p = .103, h2 = .07. This finding indicated that the recorded conversation that was used to generate discussion during the interview was generally representative of their friendship.

Friendship length

Friendship length also did not differ across matched introverts (M = 17.92 months, SD = 15.06), matched extraverts (M = 17.68 months, SD = 15.92), and mixed dyads (M = 25.68 months, SD = 37.05), F(2,63) = .72, p = .491, h2 = .02. The large majority of friendships averaged about one year in length; a few exceptions lasted from 3 years to, in one case (a mixed dyad), 14 years. Several outliers in mixed dyads accounted for the relatively high magnitude and variability with regard to friendship length.

Friendship closeness

Ratings of friendship closeness across dyads averaged four on a five-point scale, indicating that most of the friends regarded themselves as quite close. One dyad in the mixed group and one dyad in the matched introvert group did not report closeness ratings. Closeness ratings between friends were not significantly different according to a paired t-test, t(63) = 1.38, p = .172. Therefore, both friend's ratings were averaged to create a single closeness rating for each dyad. No significant differences in dyad closeness were found among matched introverts (M = 3.97, SD = .84), matched extraverts (M = 3.93, SD = .89), and mixed dyads (M = 4.25, SD = .72), F(2,61) = 1.02, p = .367, h2 = .03.

Gender

The distribution of gender at the level of the dyad was analyzed with a 2 (gender) × 3 (dyad personality) chi-square analysis. Although mixed dyads contained more women (72%) than did matched extraverted dyads (58%) and matched introverted dyads (45%), the gender differences did not reach significance, LR χ2(2, N = 66) = 4.15, p = .126, with a small effect size, Cramer's V = .130.

Similar or Different Roles Reported by Dyads

The first set of analyses tested whether friends in matched dyads reported similar roles (Hypothesis 1a) and friends in mixed dyads reported complementary roles (Hypothesis 2a), regardless of role content. To test these hypotheses, a 3 (dyad personality) × 2 (friend pair reporting same/different role) chi-square analysis was performed for each friendship domain. Only dyads in which both friends reported performing a particular role in the domain were included in the analysis. The results of the analyses are shown in Table 2, which presents descriptive statistics, chi-square results, and Cramer's V effect sizes.

Table 2

Percentage of Friend Pairs Reporting Similar Versus Different Roles in each Domain, for each Type of Dyad

Confirming hypothesis 1a, friends in matched dyads reported similar roles in almost all domains. Specifically, reports of similar roles in the matched dyads ranged from ~75% in the domain of Dialog to ~95% in the domains of Content and Locale, indicating that complementarity was quite rare for matched dyads. The sole exception to mutual role reinforcement in matched dyads occurred in the Energy domain, where almost two thirds of matched introvert pairs took different roles.

Partially confirming hypothesis 2a, friends in mixed dyads reported mutually different roles more often than did friends in matched dyads; this was found in four of the six domains: Locale, Energy, Outsiders, and Dialog. As can be seen in Table 2, this complementarity effect for mixed dyads was strongest in the domain of Energy, where mixed dyads were even more likely to report different roles than matched introverts (E-I 88%, I-I 62%, E-E 27%). In the domain of Dialog, friends in mixed dyads were also more likely to take different roles than were friends in matched dyads (E-I 56%, I-I 26%, E-E, 27%), a finding that was marginally significant, p = .061. In the domain of Locale, the complementarity effect for mixed dyads (39%) was significantly greater than that for matched dyads (I-I 0%, E-E 6%), although the majority (61%) of mixed dyads still reported similar Locale roles, suggesting a lack of complementarity in this domain. A strong exception to the general trend toward role complementarity in mixed dyads was found in the Content domain, where 81% of mixed dyads reported the same role.

Descriptive Statistics for Roles Reported by Friends in each Type of Dyad

Table 3 shows the percentage of individuals in each type of dyad who reported taking a particular role in each domain. For example, in the domain of Locale, 76% of all individuals in matched introverted dyads reported as homebodies, while 11% reported as outgoing. In mixed dyads, 92% of the introverted friends reported as homebodies and 8% as outgoing, compared to 44% and 52% of the extraverted friends, respectively. This table will be helpful in interpreting the results of the final set of statistical analyses. The percentages for male versus female dyads in Table 3 generally appear to be more similar than many of the percentages for the personality dyads, and are also compared in the following analyses.

Table 3

Percentage of Roles Reported by Individual Personalities in each Type of Dyad, and for Gender

Distribution of Friendship Roles for Individuals Nested within the 3 Types of Dyads

The prior analyses found strong evidence for mutual role reinforcement in matched dyads, and partial evidence for complementary role reinforcement in mixed dyads. The final set of analyses substantively compared the particular kinds of roles that were taken by individuals in each type of dyad (Hypotheses 1b, 2b, and 3). Because the friendship roles were dichotomous (e.g., sounder / sounding board) and paired together in each dyad, a multilevel logistical regression analysis was employed using SAS GLIMMIX. The unit of analysis was the dyad, with reports from individual friends nested together in the dyad. Because friend reports were paired in the analysis, only dyads in which both friends were coded for a domain were included in the analyses. For each friendship domain, dyad personality, dyad gender, and the interaction of personality × gender were tested. A 4-level dyad personality variable (matched introverts, matched extraverts, introvert in mixed dyads, extravert in mixed dyads) was entered as a dummy coded predictor. All analyses reached convergence. No significant gender interactions were found, perhaps due to the small sample size.

Main effects for gender

Results of the GLIMMIX analyses for roles in each domain can be found in Table 4. Main effects for dyad gender, shown at the bottom of Table 4, were significant for the domains of Outsiders and Intimacy. Table 3 shows the direction of these effects. In discussing third parties (Outsiders), women friends were significantly more likely to take the role of “buddy” by being appreciative (51%) than were male friends (34%). In the domain of Intimacy, women friends more often reported as taking the role of disclosers (spout, 78%) while male friends more often reported as nondisclosers (containers, 61%).

Table 4

SAS GLIMMIX Coefficients and Statistical Tests

We next present the GLIMMIX results that compared roles in each domain by the personalities in particular types of dyads. The results are parsed into three sections. We first compare individuals in the two kinds of matched dyads (I-I, E-E). We then compare the introverted and extraverted friends in the mixed (E-I) dyads. Finally, we compare the friends in matched dyads to the friends in mixed dyads.

Comparing matched introverts with matched extraverts

As can be seen in Table 4, matched introverts and matched extraverts reported significantly different roles in the domains of Locale, Energy, Outsiders, and Content, thereby confirming hypothesis 1b. Specifically, as shown in Table 3, matched introverts reported more often as homebodies, anchors, judges, and hobbiests, compared to matched extraverts, who more often reported as outgoing, propellers, buddies, and dramatizers. Exceptions to this trend occurred only for matched introverts. In the domain of Dialog, matched extraverts mutually reported as sounders, as expected, but surprisingly, so did matched introverts. Similarly, in the domain of Intimacy, matched extraverts mutually reported as spouts or disclosers, as expected, but matched introverts' roles varied; approximately half of the matched introverts mutually reported as spouts and about half mutually reported as containers (see Tables 2 and ​3).

Comparing introverts and extraverts in mixed dyads

As can be seen in Table 4, comparisons of introverts and extraverts in mixed dyads found that these friends' roles were significantly different in four domains: Energy, Outsiders, Dialog, and Intimacy. Role distributions in these domains can be seen in Table 3, which shows that in mixed dyads, introverts and extraverts reported, respectively, as anchor vs. propeller (Energy), judge vs. buddy (Outsiders), and sounding board vs. sounder (Dialog). These findings provide partial support for hypothesis 2b, which predicted dispositional role complementarity in mixed dyads. Findings for the domain of Intimacy were also statistically significant, but less clearcut. The significant differences between friends in mixed dyads for the domain of Intimacy were mainly driven by extraverts, who, as shown in Table 3, showed a definite preference for being spouts (80%) rather than containers (20%), while mixed introverts were more evenly split in serving as either spouts (44%) or containers (56%). No significant differences were found between mixed extraverts and mixed introverts in the domains of Locale and Content, where, surprisingly, both friends reported more often as homebodies (Locale) and as dramatizers (Content).

Comparing extraverts in matched and mixed dyads, ditto for introverts

Hypothesis 3 predicted that when accommodation to the other's personality occurred, the accommodation would be on the part of the introvert. To test hypothesis 3, we compared extraverts in matched versus mixed dyads, and likewise for introverts. Partially confirming the hypothesis, evidence for accommodation occurred in two domains for introverts and in one domain for extraverts. As can be seen in Table 4, the two domains in which introverts reported roles differently in matched than mixed dyads were Dialog and Content. In Dialog, introverts were more likely to report as sounding boards in mixed (60%) than matched dyads (18%, from Table 3); however, the fact that introverts did not adopt the extraverts' role (sounder) in mixed dyads indicates that this finding reflected complementarity rather than accommodation. Good evidence of accommodation on the part of introverts was found in the domain of Content; introverts were more likely to report as talking about drama in mixed (44%) than matched dyads (16%, from Table 3). However, extraverts in mixed dyads also showed accommodation in one domain. As can be seen in Table 4, extraverts reported roles differently in mixed versus matched dyads in the domain of Locale, or where the friendship usually took place. Specifically, extraverts in matched dyads rarely reported as homebodies (14%, from Table 3), whereas 52% reported as homebodies in mixed dyads, suggesting an accommodation to the introverts' preferred locale.

Summary of Patterns of Mutual and Complementary Reinforcement, and Accommodation

Because the analyses were complex, the gist of the findings is summarized in Table 5. Hypothesis 1 predicted that matched dyads would show similar roles consistent with their personality. As is shown in Table 5, this hypothesis was very strongly confirmed for matched extraverted friends, who showed mutual reinforcement of roles across all domains. For matched introverts, mutual reinforcement of dispositional roles was found in only half of the domains: the Locale of the friendship (at home), attitudes toward Outsiders (judgmental), and the typical Content of their conversations (hobbies). For the other domains, matched introverts either did not show a distinctive preference for one role over another (in the domain of Energy, some anchored the relationship and some propelled it; in the domain of Intimacy, some were spouts or disclosers whereas others were containers), or seemingly acted similar to matched extraverts (in the domain of Dialog, matched introverts mutually reported as sounders rather than sounding boards).

Table 5

Overall Pattern in Each Domain as a Reciprocal Function of Personality

Hypothesis 2 predicted complementary reinforcement of dispositional roles in mixed dyads, and was supported in three of the six domains. Specifically, more often than not in mixed dyads, the introvert served as the anchor and the extravert as the propeller (Energy), the introvert was judgmental of peers and the extravert was appreciative (Outsiders), and the introvert served as sounding board and the extravert as sounder (Dialog).

Hypothesis 3 predicted that in mixed dyads, introverts would be more likely than extraverts to accommodate to the friend's role. This hypothesis was not supported. As can be seen in Table 5, mixed introverts accommodated to mixed extraverts in one domain, and mixed extraverts accommodated to mixed introverts in one domain. Specifically, mixed introverts mostly talked about social drama with extraverts (Content), whereas they mostly talked about hobbies with other introverts. Unexpectedly, extraverts accommodated to the introvert in mixed dyads by staying close to home, the general preference of the introvert.

Accounts of Accommodation and Complementarity by Friends in Mixed Dyads

A close look at the interviews suggested why introverts more often served as sounding boards with extraverts than with other introverts. Pairing up with an extravert provided a social space where a preference to listen did not hinder but rather helped the flow of friendship. Some introverted friends said that they valued opportunities to listen to their friends talking about social dramas, although others found the drama a bit excessive or exaggerated. For an extravert, being friends with an introvert tended to afford more opportunity to talk without competition. Access to a good listener provided extraverted friends free reign to talk at length, but it also afforded occasions to appreciate the listening skills of their introverted friends.

The interviews also suggested why some extraverts were willing to yield to the introvert's preference to stay home rather than go out and socialize. By staying home, extraverts could talk at length to their introverted friends, who provided an interested and supportive ear. As to why introverts yielded to extraverts' conversational topics, social dramas seemed to be more generally accessible than were the hobbies that mutually introverted friends discussed with each other. These hobbies, which included quilting, writing stories, and computer programming, tended to require somewhat specific or technical knowledge. Although introverts tended to be less inclined or skilled at talking about social events, they could draw on their personal experiences with others to engage in such talk with their extraverted friends, perhaps affording the honing of both established and novel skills.

Regardless of the friend's personality, introverts generally (though with considerable variation) were found to be less disclosing and more judgmental of outside peers than extraverts, who predominantly were disclosers and more appreciative of their peers. In the interviews, personality differences in disclosure tended to be a source of contention in that extraverts desired more revelations from their introverted friends but tended to be reluctant to cede the floor. Differences in attitudes toward peers were less often discussed as a bone of contention in the interviews than were differences in disclosure. Although discussions of social drama were prevalent in the conversations of mixed dyads, the extraverts' positive affect appeared to immunize them against seeming unduly critical of other people. Hearing their introverted friends criticize others and engaging in such criticism themselves did not impede extraverts' tendency to simultaneously report affection for the targets of the criticism. For introverts, on the other hand, criticizing others appeared to undermine their construction of positive feelings for the targets of the criticism. Introverts' tendency to criticize outsiders was perhaps tied to their feeling of separation from the larger social world.

Discussion

The phenomenon of mutual personality reinforcement in matched dyads received the strongest confirmation in the present study. Specifically, matched extraverted friends reported similar roles consistent with their personalities across all six friendship domains, and similar findings for matched introverted friends emerged in half of the domains. In mixed dyads, partial support was found for complementary reinforcement of personality, with half of the domains showing clear evidence of role complementarity. Specifically, in mixed dyads, the introvert was judgmental of peers and the extravert appreciative of peers, the introvert served to anchor the friendship while the extravert served as propeller, and the introvert served as sounding board (listener) while the extravert served as sounder (talker). However, the findings did not support the prediction that when accommodation occurred in mixed dyads, introverts would accommodate to extraverts rather than vice versa. Instead, the introvert accommodated to the extravert's preferred conversational content (social drama), and the extravert accommodated to the introverts' preferred locale (staying close to home).

An obvious limitation of the present study is that the findings are cross-sectional. A round robin-type study would be necessary to definitively demonstrate that extraverts change the location of the friendship when with introverted rather than with extraverted friends, that introverts are more silent with extraverted than with introverted friends, and that introverts talk more about social dramas with extraverted friends and about hobbies with introverted friends. Nonetheless, the findings suggest that such accommodations are possible and suggest that the roles people play in friendships can shift depending on the press of each other's personality. More generally, the findings lend support to the idea that close friendships provide distinctive opportunities for reinforcing dispositional tendencies and for fostering personality (or at least behavioral) accommodation or change.

By focusing on interviews about the nature of the friendship, the study revealed the phenomenological meaning of the reinforcements and accommodations that develop between friends who are similar or different in personality. Although an introverted friendship may seem paradoxical, such friendships were reportedly as close as the other friendships. However, pairs of introverted friends practiced friendship in ways that afforded a standoffishness from the larger world. Specifically, mutually introverted friends tended to stick closer to home, to keep sensitive matters private, to take a critical view of other peers, and to bond around shared hobbies more than sharing personal feelings. Pairs of extraverted friends, on the other hand, congregated away from home, exchanged their own and others' social dramas, and took a more appreciative stance toward other peers. Whereas introverted friends reinforced each other's tendencies to be socially exclusive, extraverted friends appeared to reinforce each other's tendencies to be broadly socially inclusive, to create, pursue, and appreciate social rewards. Notably, a similar pattern was found in the storytelling that emerged in the conversations of these mutually introverted and mutually extraverted friends (Thorne, Korobov, & Morgan, 2007).

The most unprecedented findings concerned the friendship practices of mixed dyads. Relative to matched dyads, friends in mixed dyads reported a more complex distribution of reinforcement, sometimes taking complementary roles and sometimes mutual or similar roles. We were particularly interested in instances of mutual reinforcement that occurred through accommodation. Friendships between an introvert and an extravert tended to take place at home (the introvert's preferred turf) rather than out in the world, but their conversations tended to concern social dramas, the extravert's turf. These similar roles may be necessary because a friendship requires a consensus about where to congregate and what to talk about. Such accommodations, however, suggest that the introverted partner was strongly wedded to home, winning over the extraverts' tendencies to go out, and that the extraverted partner was particularly wedded to talking about social dramas, winning over the introverts' tendencies to talk about the details of his/her personal hobbies.

The findings show partial support for the self-reinforcement theory of personality (Caspi & Roberts, 2001) and for interpersonal theory (Carson, 1969). Self-reinforcement theory maintains that similar personalities reinforce each other's dispositional tendencies, a pattern that was indeed found for most of the domains in matched dyads. In addition, self-reinforcement theory implies that dissimilar personalities disrupt or challenge each other's dispositional tendencies, and patterns of dispositional disruption did emerge in two domains for mixed dyads: the extraverts more often reported staying home with introverted friends than with extraverted friends, and introverts reported talking about social dramas with extraverted friends more often than with introverted friends. However, we also found support for interpersonal theory, which posits that dispositional opposites can reinforce each other's tendencies in a complementary fashion. Complementary reinforcement occurred most strongly in the domains of Energy and Dialog, where introverts identified as relationship anchors and sounding boards and extraverts as propellers and sounders. Theoretically, these particular accentuations reflect the control or assertiveness facet of extraversion (Carson, 1969).

The differential support for interpersonal theory and self-reinforcement theory highlights the value of taking a proximal approach to the particular dynamics that friends engage in on a routine basis. Similarity and difference did not uniformly affect the expression of personality in interpersonal behavior, but rather selectively patterned particular domains of practice. The approach exemplified in this study holds much promise in elucidating how personality is potentially reinforced and challenged to change through repeated everyday interactions in close relationships.

Lastly, there are three important limitations with regard to our methodology. First, as previously noted, the patterns with regard to personality reinforcement and accommodation are cross-sectional rather than longitudinal. A longitudinal analysis would be particularly valuable in understanding whether friendships between very different personalities have an enduring impact, perhaps expanding a person's flexibility and efficacy (Aron & Aron, 1996). For example, do introverts with a close extraverted friend in high school find it easier to make new friends when they go away to college? How do the dynamics of introvert-extravert friendships change across the life course, particularly as social networks narrow in later life (Carstensen, 1995)? Second, although the interviewers were blind to the personality scores of the participants, the interviewers themselves had distinctive personalities that presumably impacted participants' responses. We tried to randomize and control these impacts by employing a large number of interviewers over the three years of data collection, by matching the interviewer's gender with that of the participant, and by having individual interviewers conduct numerous interviews across all three types of dyads. Consequently, we have no reason to suspect that personality interactions between interviewers and participants systematically skewed the results. Third, the molar coding of the interviews doubtless concealed more subtle variations in friendship practices. For example, although the role of “sounder” prevailed in mutually introverted and mutually extraverted friendships, a companion study of the participants' conversations found that introverted friendships showed more orderly turn-taking during storytelling than did extraverted friendships (Thorne, Korobov, & Morgan, 2007). The use of an inductive coding system allowed us to discover what practices were phenomenologically most relevant to extraversion-introversion in this sample of friendships, but the molar coding of the interviews doubtless concealed some of the dynamics that were observed in the conversations between the friends.

In a widely cited review of advances in personality research during the last decade, Caspi, Roberts, and Shiner (2005) called for studies of the micro-interactional processes that contribute to personality development in the context of close relationships. The present study is one answer to this call, and is notably one of the first studies to systematically investigate friends with extreme personality differences. In so doing, the study represents a stark departure from unilateral studies of personality development, and suggests that birds of both same and different feathers are potentially potent forces in the development of personality.

Acknowledgments

Funding for this study was provided by NIH training grant T32 HD46423 and grants from Consulting Psychologist Press and the Social Sciences Division of the University of California, Santa Cruz. We especially thank Lisa Rucker and Mary Hayes for helping to code the interviews, Per Gjerde for consulting on early drafts, the diligent research assistants involved in data collection, and the participants.

References

• Argyle M, Lu L. Happiness and social skills. Personality and Individual Differences. 1990;11:1255–1261. doi: 10.1016/0191-8869(90)90152-H. [Google Scholar]
• Argyle M, Lu L. New directions in the psychology of leisure. The New Psychologist. 1992;1:3–11. doi: 10.1007/s10902-005-8628-3. [Google Scholar]
• Arnett JJ. Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist. 2000;55:469–480. doi: 10.1037/0003-066X.55.5.469. [PubMed] [Google Scholar]
• Aron A, Aron EN. Self and self-expansion in relationships. In: Fletcher GJO, Fitness J, editors. Knowledge structures in close relationships: A social psychological Approach. Lawrence Erlbaum Associates; New Jersey: 1996. pp. 325–344. [Google Scholar]
• Aron A, Paris M, Aron EN. Falling in love: Prospective studies of self-concept change. Journal of Personality and Social Psychology. 1995;69:1102–1112. doi: 10.1037/0022-3514.69.6.1102. [Google Scholar]
• Ashton MC, Lee K, Paunonen SV. What is the central feature of extraversion?: Social attention versus reward sensitivity. Journal of Personality and Social Psychology. 2002;83:245–252. doi: 10.1037/0022-3514.83.1.245. [PubMed] [Google Scholar]
• Blaz M. Perceived extraversion in a best friend. Perceptual and Motor Skills. 1983;57:891–893. [Google Scholar]
• Block J. Lives through time. Bancroft Books; Berkeley, CA: 1971. [Google Scholar]
• Blumenthal TD. Extraversion, attention, and startle response reactivity. Personality and Individual Differences. 2001;30:495–503. doi: 10.1016/S0191-8869(00)00153-7. [Google Scholar]
• Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3:77–101. doi: 10.1191/1478088706qp063oa. [Google Scholar]
• Briggs KC, Myers IB. The Myers-Briggs Type Indicator, Form M. Consulting Psychologists Press; Palo Alto, CA: 1998. [Google Scholar]
• Bukowski WM, Hoza B, Boivin M. Measuring friendship quality during pre- and early adolescence: The development and psychometric properties of the Friendship Qualities Scale. Journal of Social and Personal Relationships. 1994;11:471–484. doi: 10.1177/0265407594113011. [Google Scholar]
• Byrne D. The attraction paradigm. Academic Press; New York: 1971. [Google Scholar]
• Carson R. Interaction concepts of personality. Aldine Publishing Co; Oxford: 1969. [Google Scholar]
• Carstensen LL. Evidence for a life-span theory of socioemotional selectivity. Current Directions in Psychological Science. 1995;4:151–156. doi: 10.1111/1467-8721.ep11512261. [Google Scholar]
• Caspi A, Elder GH, Bem DJ. Moving against the world: Life-course patterns of explosive children. Developmental Psychology. 1987;23:308–313. doi: 10.1037/0012-1649.23.2.308. [Google Scholar]
• Caspi A, Roberts BW. Personality development across the life course: The argument for change and continuity. Psychological Inquiry. 2001;12:49–66. doi:10.1207/S15327965PLI1202_01. [Google Scholar]
• Caspi A, Roberts BW, Shiner RL. Personality development: Stability and change. Annual Review of Psychology. 2005;56:453–484. doi: 10.1146/annurev.psych.55.090902.141913. [PubMed] [Google Scholar]
• Cuperman R, Ickes W. Big Five predictors of behavior and perceptions of initial dyadic interactions: Personality similarity helps extraverts and introverts, but hurts “disagreeables” Journal of Personality and Social Psychology. 2009;97:667–684. doi: 10.1037/a0015741. [PubMed] [Google Scholar]
• Dishion TJ, Patterson GR, Stoolmiller M, Skinner ML. Family, school, and behavioral antecedent to early adolescent involvement with antisocial peers. Developmental Psychology. 1991;27:172–180. doi: 10.1037/0012-1649.27.1.172. [Google Scholar]
• Emmons RA, Diener E, Larsen RJ. Choice and avoidance of everyday situations and affect congruence: Two models of reciprocal interactionism. Journal of Personality and Social Psychology. 1986;51:815–826. doi: 10.1037/0022-3514.51.4.815. [Google Scholar]
• Eysenck HJ, Nias DK, Cox DN. Sport and personality. Advances in Behavioral Research and Therapy. 1982;4:1–56. doi: 10.1016/0146-6402(82)90004-2. [Google Scholar]
• Fleeson W. Moving personality beyond the person-situation debate: The challenge and the opportunity of within-person variability. Current Directions in Psychological Science. 2004;13:83–87. doi: 10.1111/j.0963-7214.2004.00280.x. [Google Scholar]
• Funder DC. Personality judgment: A realistic approach to person perception. Academic Press; San Diego: 1999. [Google Scholar]
• Hamm J. Do birds of a feather flock together? The variable bases for African American, Asian American, and European American adolescents' selection of similar friends. Developmental Psychology. 2000;36:209–219. doi: 10.1037/0012-1649.36.2.209. [PubMed] [Google Scholar]
• Klohnen EC, Luo S. Interpersonal attraction and personality: What is attractive—Self similarity, ideal similarity, complementarity, or attachment security? Journal of Personality and Social Psychology. 2003;85:706–722. doi: 10.1037/0022-3514.85.4.709. [PubMed] [Google Scholar]
• Levesque MJ, Steciuk M, Ledley C. Self-disclosure patterns among well-acquainted individuals: Disclosers, confidants, and unique relationships. Social Behavior and Personality. 2002;30:579–592. doi: 10.2224/sbp.2002.30.6.579. [Google Scholar]
• Lucas RE, Diener E, Grob A, Suh EM, Shao L. Cross-cultural evidence for the fundamental features of extraversion. Journal of Personality and Social Psychology. 2000;79:452–468. doi: 10.1037/0022-3514.79.3.452. [PubMed] [Google Scholar]
• McCrae RR, Costa PT., Jr. Reinterpreting the Myers-Briggs Type Indicator from the perspective of the five-factor model of personality. Journal of Personality. 1989a;57:17–40. doi: 10.1111/j.1467-6494.1989.tb00759.x. [PubMed] [Google Scholar]
• McCrae RR, Costa PT., Jr. The structure of interpersonal traits: Wiggins's circumplex and the five-factor model. Journal of Personality and Social Psychology. 1989b;56:586–595. doi: 10.1037/0022-3514.56.4.586. [PubMed] [Google Scholar]
• Mendelson MJ, Kay A. Positive feelings in friendship: Does imbalance in the relationship matter? Journal of Social and Personal Relationships. 2003;20:101–116. doi: 10.1177/0265407503020001190. [Google Scholar]
• Mischel W, Shoda Y. Reconciling processing dynamics and personality dispositions. Annual Review of Psychology. 1998;49:229–258. [PubMed] [Google Scholar]
• Murray HA. Explorations in personality. Oxford University Press; New York: 1938. [Google Scholar]
• Murray HA. Toward a classification of interaction. In: Parsons T, Shils EA, editors. Toward a general theory of action. Harper Torchbooks; New York: 1951. pp. 434–464. [Google Scholar]
• Nelson PA, Thorne A. Manual for coding friendship practices in interviews. Department of Psychology, University of California; Santa Cruz: 2009. Unpublished manuscript. [Google Scholar]
• Paunonen SV, O'Neill TA. Self-reports, peer ratings, and construct validity. European Journal of Personality, Special Issue: Construct validity in self and observer reports of personality. 2010;24:189–206. doi: 10.1002/per.751. [Google Scholar]
• Peter J, Valkenburg P, Schouten A. Developing a model of adolescent friendship formation on the internet. Cyber Psychology and Behavior. 2005;8:423–430. doi: 10.1089/cpb.2005.8.423. [PubMed] [Google Scholar]
• Radmacher K, Azmitia M. Are there gendered pathways to intimacy in early adolescents' and emerging adults' friendships? Journal of Adolescent Research. 2006;21:415–448. doi: 10.1177/0743558406287402. [Google Scholar]
• Rim Y. Personality variables and interruptions in small discussions. European Journal of Social Psychology. 1977;7:247–251. doi: 10.1002/ejsp.2420070210. [Google Scholar]
• Roberts BW, Robins RW. Person-environment fit and its implications for personality development: A longitudinal study. Journal of Personality. 2004;72:89–110. doi: 10.1111/j.0022-3506.2004.00257.x. [PubMed] [Google Scholar]
• Roberts BW, Walton KE, Viechtbauer W. Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin. 2006;132:1–25. doi: 10.1037/0033-2909.132.1.1. [PubMed] [Google Scholar]
• Rubin KH, Bukowski W, Parker JG. Peer interactions, relationships, and groups. In: Damon W, Eisenberg N, editors. Handbook of child psychology: Social, emotional, and personality development. Vol. 3. John Wiley & Sons Inc.; Hoboken, NJ: 1998. pp. 619–700. [Google Scholar]
• Sadler P, Woody E. Is who you are who you're talking to? Interpersonal style and complementarity in mixed-sex interactions. Journal of Personality and Social Psychology. 2003;84:80–96. doi: 10.1037/0022-3514.84.1.80. [PubMed] [Google Scholar]
• Selfhout M, Burk W, Brainie S, Denissen J, van Aken M, Meeus W. Emerging late adolescent friendship networks and Big Five personality traitis: A social network approach. Journal of Personality. 2010;78:509–538. [PubMed] [Google Scholar]
• Sharabany R, Gershoni R, Hofman JE. Girlfriend, boyfriend: Age and sex differences in intimate friendship. Developmental Psychology. 1981;17:800–808. doi: 10.1037/0012-1649.17.6.800. [Google Scholar]
• Stone DL. Relationship between introversion/extraversion, values regarding control over information, and perceptions of invasion of privacy. Perceptual and Motor Skills. 1986;62:371–376. [Google Scholar]
• Sturaro C, Dennissen JJA, van Aken MAG, Asendorpf JB. Person-environment transactions during emerging adulthood: The interplay between personality characteristics and social relationships. European Psychologist. 2008;13:1–11. doi: 10.1027/1016-9040.13.1.1. [Google Scholar]
• Thorne A. The press of personality: A study of conversations between introverts and extraverts. Journal of Personality and Social Psychology. 1987;53:718–726. doi: 10.1037/0022-3514.53.4.718. [Google Scholar]
• Thorne A, Gough HG. Portraits of type: An MBTI research compendium. Consulting Psychologists Press; Palo Alto, CA: 1991. [Google Scholar]
• Thorne A, Korobov N, Morgan EM. Channeling identity: A study of storytelling in conversations of introverted and extraverted friends. Journal of Research in Personality. 2007;41:1008–1031. doi: 10.1016/j.jrp.2006.12.001. [PMC free article] [PubMed] [Google Scholar]
• Thorne A, Shapiro L, Cardilla K, Korobov N, Nelson PA. Caught in the act: How extraverted and introverted friends communally cope with being recorded. Journal of Research in Personality. 2009;43:634–642. doi: 10.1016/j.jrp.2009.03.014. [PMC free article] [PubMed] [Google Scholar]
• Tracey TJG. Levels of interpersonal complementarity: A simplex representation. Personality and Social Psychology Bulletin. 2004;30:1211–1225. doi: 10.1177/0146167204264075. [PubMed] [Google Scholar]
• Umphress EE, Smith-Crowe K, Brief AP, Dietz J, Watkins MB. When birds of a feather flock together and when they do not: Status composition, social dominance orientation, and organizational attractiveness. Journal of Applied Psychology. 2007;92:396–409. doi: 10.1037/0021-9010.92.2.396. [PubMed] [Google Scholar]
• Vitaro F, Boivin M, Bukowski W. The role of friendship in child and adolescent psychosocial development. In: Rubin KH, Bukowski WM, Laursen B, editors. Handbook of peer interactions, relationships, and groups. Social, emotional, and personality development in context. Guilford Press; New York, NY: 2009. pp. 568–585. [Google Scholar]
• Watson D, Clark LA. Extraversion and its positive emotional core. In: Hogan R, Johnson JA, editors. Handbook of personality psychology. Academic Press; San Diego, CA: 1997. pp. 767–793. [Google Scholar]