Channeling Identity: A Study of Storytelling in Conversations Between Introverted and Extraverted Friends

Abstract

This narrative study examined the process of personal storytelling between college-age friends who were similarly introverted or extraverted. Participants were 19 introverted and 20 extraverted same-sex pairs (49 percent female) who had been friends for an average of 18 months. Stories emerged spontaneously during 10-minute catch-up conversations. Extraverted friends more often told stories that changed the topic, and more often co-constructed story plots. Introverted friends more often told stories that were embedded in a developing theme, and constructed story plots solo. With regard to content, extraverted friends told stories about romance more so than introverted friends, whose stories more often concerned family/hometown, and older events. The findings suggest that the traits of extraversion and introversion channel the identity-making process.

One cannot grasp the most profound logic of the social world unless one becomes immersed in the specificity of an empirical reality (Bourdieu, 1993, p. 271)

With the advent of narrative psychology (Bruner, 1990), the concept of personality has expanded beyond the domain of traits, such as extraversion, to encompass identity, or the sense that people make of their lives through telling stories (McAdams, 2001). This expansion of the concept of personality to include life stories has enhanced the degree to which personality psychology attends to the whole person (McAdams, 1995), but the expansion has not been easy. A notable difficulty in connecting traits and life stories is that traits tend to be construed as genetic endowments, and life stories as psychosocial constructions (McAdams, 2001). However, it is now widely recognized that traits such as extraversion, while genetically based (e.g., Plomin, Chipuer, & Loehlin, 1990), become developmentally elaborated through psychosocial processes (Caspi, 1998).

The present study builds on the idea that traits help to channel the development of other domains of personality (McLean & Pasupathi, 2006; Winter, John, Stewart, Klohnen, & Duncan, 1998). We investigated this channeling by exploring how people actively construct or “do” personality in everyday life (Cantor, 1990). A performative approach to the development of personality traits has been well-articulated by Caspi (1998), whose model of developmental elaboration views individuals as actively selecting, construing, and evoking situations that are compatible with their predispositions. For example, one way in which traits become developmentally elaborated is through the selection of friends and mates who are similar in personality (Caspi & Herbener, 1990; Kandel, 1978). A performative approach to the development of identity is somewhat similar in that individuals actively select and tell their stories to listeners whom they feel will appreciate and understand their experience, and that in the process of storying experience, identities are actively created in vivo (Bamberg, 2004).

Our construal of extraversion and introversion, although compatible with evidence that these individual differences are genetically and psycho-physiologically based (Eysenck & Eysenck, 1985; Shiner & Caspi, 2003), is grounded in the phenomenological view that introverts and extraverts proffer personal experience differently by virtue of the way they prefer to live in the world, with extraverts constructing a more inclusive, common space, and introverts a more exclusive, individual space (Jung, 1923/1976). Past research suggests that these different orientations are apparent in narratives. In a study of written narratives, introversion was found to be associated with a tendency to make distinctions through the use of exclusive words such as “but” and “except,” and tentative words such as “perhaps” and “maybe,” while extraversion was associated with a higher frequency of inclusive words, such as “and” and “with” (Pennebaker & King, 1999). In a study of TAT stories, extraverts more often imported other characters into stories about a solitary figure than did introverts (Shapiro & Alexander, 1975). And in a study of conversations, extraverted strangers were found to explore more topics and to claim more common ground than did introverted strangers (Thorne, 1987).

Although extraversion and introversion may be developmentally elaborated on many levels of language use, the story has a special status in personality development because stories are a means of preserving, evaluating, and broadcasting experiences that are deemed sufficiently noteworthy to be remembered and told (McAdams, 2001; Thorne, 2000). The present study used the story rather than the utterance as the basic unit of analysis, and focused on the story-telling process in conversations with friends. Whereas stories have been found to be rare in brief conversations between strangers (Thorne, 1987), stories are arguably more likely to emerge with friends because a story can be more easily launched if the teller and listener are privy to each other's lives. To establish a backdrop for the present study, we first consider why storytelling between friends is important for understanding the process of identity development. We then consider how the dynamics of storytelling might differ for introverted and extraverted friends.

The Contours of Stories and Storytelling

Stories center on events that disrupt experience and are deemed worthy of sharing with others (Bruner, 1990; Rimé, Phillipot, Boca, & Mesquita, 1991). The longest story is the autobiography or extended life story, which weaves together experiences across an entire life (McAdams, 2001). The present study is concerned with small stories, the kinds of stories that tend to emerge in everyday conversations (Bamberg, 2004). Although small stories are common in everyday life, and are a joint product of the teller and the listener (e.g. Bamberg, 2004; Ochs & Capps, 2001), the process of everyday storytelling has rarely been studied by personality psychologists (Thorne, 2004).

Sociolinguists have long been fascinated by the contours of the story-telling process, a body of work from which we drew for the present study. Qualitative studies of natural storytelling have found that stories tend to be heralded with time stamps that alert listeners that a story is about to be told (“Yesterday, I…”; Labov & Waletzky, 1967/1997). The backbone of a story is the plot, a series of action clauses that are temporally sequenced (Coates, 2003; Labov & Waletzky, 1967/1997). Ochs and Capps (2001) unpacked the myriad ways in which stories emerge in everyday conversations. They observed that stories can surface voluntarily or be bidden, and can be relevant to an ongoing topic or emerge suddenly, “out of the blue.” Stories can be scattered throughout a conversation or can emerge in clusters, on the heels of prior stories. Stories can be about one's own experience, a shared experience, or something that happened to someone else, and can be unilaterally or collaterally told.

Connecting the Story-telling Process with Introversion and Extraversion

Two aspects of the story-telling process potentially map onto well-established findings with regard to extraversion and introversion. One feature concerns how stories are initiated, either abruptly, or more gradually. The social initiative of extraverts is one of their cardinal qualities, and is a staple in the item content of extraversion-introversion scales, such as the MBTI (Briggs & Myers, 1998), the NEO-PI (Costa & McCrae, 1985), and the Big Five (Benet-Martinez & John, 1998). This social initiative is noticed by friends and strangers, who have been found to describe extraverts as more active, outgoing, and sociable than introverts, who were described as more reserved, quiet, and shy (e.g., McCrae & Costa, 1987; Thorne & Gough, 1991). In laboratory settings, extraverts have been found to excel on tasks that require speed, and introverts on tasks that require focus, vigilance, or reflection (Harkins & Geen, 1975; Matthews, 1992). In a study of conversations between strangers, extraverted dyads were found to initiate a wider array of topics than did introverted dyads, who more often focused on only one or two topics (Thorne, 1987). These findings led us to expect that extraverts would more often tell stories that suddenly changed the topic than would introverts, whose stories would tend to be more anchored in an ongoing theme.

In addition to examining how stories were initiated, we also examined differences in who told the story, that is, whether the story was solo- or co-told. A story can be told by one person, or by more than one person. In a recent experience sampling study, extraverted adults reported engaging in more frequent back and forth exchanges of stories than did introverted adults (McLean & Pasupathi, 2006). Such volleying, in which one person tells a story and then another person tells a story, is a form of unilateral exchange, because each speaker is telling his/her own story. Bilateral storytelling occurs when a story is told by both speakers. A particularly strong form of bilateral storytelling occurs when both speakers construct the plot of the story. Because the plot is the backbone of a story, co-telling the plot of a story represents a notable level of bilateral engagement.

Based on the notion that extraverts inhabit and evoke a more common world, we expected that extraverted friends would more often co-tell story plots than would introverted friends, who would more often construct the story plot solo. In examining solo- versus co-told plots, we also attended to whose experience the story was about. We were particularly interested in stories about past events that had been experienced by only one speaker, since a mutually experienced event is presumably more co-tellable than an individually experienced event. However, for extraverted friends, it is possible that events that were experienced by only one speaker might also be considered fair game for co-construction.

Although we were primarily interested in exploring the dynamics of the storytelling process, we also examined the topic and time frame of the stories. A prior study of conversations between college student strangers found that introverted strangers primarily talked about school, whereas extraverted strangers voyaged beyond this common ground into a host of other topics, especially extracurricular activities, such as hobbies and travel (Thorne, 1987). Because friends share more history than do strangers, we expected that their conversations would be relatively more personal. We also assessed the time frame of the stories, from recent to older events.

Our method was very similar to that of a prior study of strangers, in which same-sex college students engaged in brief, 10-minute, conversations (Thorne, 1987). However, instead of examining the frequency of particular speech acts across the entire conversation, the present study focused on how stories were produced. We compared conversations of extraverted and introverted dyads because dispositionally similar dyads have been found to accentuate predisposed tendencies and to thereby enhance the visibility of discursive patterns in personality (Thorne, 1987).

Method

Participants

Participants were 39 same-sex dyads (49 percent female) who had been friends for at least six months, and were living away from home at a public university in Northern California. Their average age was 19.5 years (SD = .97). Participants were required to be native English speakers, and 90 percent self-identified as European American. To enhance their likelihood of recruiting a friend whom they had known for at least six months, participants were required to have been enrolled for a minimum of three academic quarters. One member of each dyad was recruited on the basis of pre-testing in large psychology courses, and participated in the study to fulfill a course requirement; the friend was compensated $20. There were 19 introverted dyads (11 male, 8 female) and 20 extraverted dyads (9 male, 11 female). ¹

Measures

Extraversion-Introversion

Participants were selected on the basis of scores on a subset of items on the Extraversion-Introversion scale of the Myers-Briggs Type Indicator (MBTI), Form M (Briggs & Myers, 1998). The MBTI was also used in the prior study of strangers (Thorne, 1987). The forced-choice items generally refer to sociability and initiation of conversation in general gatherings of people, versus reserve. Scores on this scale have been found to correlate substantially, r = .72, with Extraversion scores on the NEO-PI (McCrae & Costa, 1989), as well as with other frequently used extraversion scales (Thorne & Gough, 1991), and to map sensibly onto observer ratings collected over periods ranging from 10 minutes to 3 days (Thorne, 1987; Thorne & Gough, 1991). The 10 MBTI items used in the present study correlated highly with the full 21-item MBTI Extraversion-Introversion scale, r (158)= .93, p < .001. The alpha reliability of the 10-item scale was .82. Scale scores ranged from 10 to 20, with higher scores indicative of extraversion.

Friendship survey

Each participant responded to a survey that included a question about the length of their friendship in months. They were also asked to rate the closeness of their friendship (“How close do you feel to this friend, compared to your closest same-sex friend?”) from 1 (not at all) to 5 (very).

Procedures

Pretest

One member of each dyad was part of a pre-test group, averaging 250 students per quarter, recruited between Fall, 1999, and Spring, 2002. Students were administered a survey in large psychology courses; the survey included 10 extraversion 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 10-12. These cut-offs were maintained for subsequent pre-test samples, which showed very similar distributions.

Recruitment

Approximately one week after pre-testing, candidates were telephoned and asked to participate in a 2-hour “study of the friendship process.” They were asked to bring along a same-sex friend whom they had known for at least six months. The candidates could participate in the study to fulfill a course requirement, and their friends were offered $20 for participating. The nature of the study was not revealed until they arrived.

Catch-up conversations

Upon arrival, the dyad was greeted by a same-sex undergraduate research assistant. The friends were seated on couches at a right angle to each other in a comfortable room with children's art on the wall. Participants were told that the purpose of the study was to explore conversations between friends. They were asked to participate in a brief audio-recorded conversation, and to use the 10 minutes to simply catch up and talk about anything whatsoever. The directions were intentionally left vague so as to not to restrict the conversational topics. Consent was obtained to audio-record the conversation, with assurances that anonymity would be preserved. The conversations usually sputtered initially, and then transformed into ordinary and natural sounding conversations, replete with everyday expressions such as the use of common slang and inside jokes. The conversations were terminated by a knock on the door.

Questionnaires

After the conversation, each partner was administered a series of questionnaires and other activities that included a friendship survey and the MBTI. Participants were not informed that extraversion-introversion was the focus of the research; the decision to withhold this information was made when a prior study found that informing the participants of this fact tended to enhance their tendency to stereotype each other (Thorne, 1987).

Selection of dyads

The personality score of the friend was determined after the session, when the MBTI was scored for the same items used on the pre-test. In order to maximize the sample size, we relaxed the cutoffs by one scale point for the friend: scores between 10-13 designated introverts, and scores between 17-20 designated extraverts.

Conversation Length and Speaking Turns

Each conversation was transcribed and independently checked for accuracy. Computerized word counts were obtained to assess the number of words per conversation. Computer counts were also used to determine the total number of speaker turns, an index of how frequently the speakers changed. A speaking turn could be as short as a speaker saying, “Yes”, or a long monologue, but a turn was usually at least a clause or sentence in length.

Coding of Conversations

Coding categories were developed through a grounded theory process (Charmaz, 1983) by exploring a random subset of the transcribed conversations in order to discern how to parse stories, and how to comprehensively capture the ways in which stories were initiated, how plots were told, and what the stories were about. These discussions were used to develop the final iteration of the coding manual. Coders were the authors and four undergraduate research assistants, all of whom were blind to the personality test scores of the participants. Once the coders achieved acceptable reliability, the actual coding began. On the first pass through the transcripts, stories were identified. On subsequent passes, the following features of each story were identified: the time frame of the story, the story topic, who experienced the event in the story, how the story was initiated, and who constructed the plot. All coding categories were dichotomous (presence-absence), and subcategories were defined as mutually exclusive. Reliabilities were randomly sampled throughout the coding process to prevent coder drift. Approximately half of the stories were randomly chosen to be reliability coded, and disagreements were settled by consensus. A detailed coding manual is available upon request (Thorne, Korobov, & Morgan, 2006).

Stories

Because the plot constitutes the backbone of a story, we chose the plot as the unit of analysis for identifying stories within conversations. As defined by Labov and Waletzky (1967/1997), a plot consists of a set of action clauses that are temporally ordered with respect to each other. The action clauses build on each other around some sort of problem or complication that eventually terminates or resolves, signaling the end of the story. Action clauses tend to be interspersed with comments that set the scene or evaluate the action, and need not be contiguous in order to be temporally ordered. Following Coates (2003), an action clause was defined as a simple sentence containing an active verb in the past or historical present tense, for example, “I dashed to the window;” “Mom said ‘Jane dashed to the window’”; “So I'm dashing to the window;” “Did you dash to the window?” Because we did not know if the event in the story had actually occurred, we did not distinguish between real plots and imagined plots. Although a plot can consist of two or three action clauses, for example, “She dashed to the window, sobbed, and waved goodbye,” we found that very short plots tended to be overlooked by coders. In prior work, we found that a plot with at least six action clauses was more reliably detected than was a shorter plot (Korobov & Thorne, 2006). Since longer plots also have the advantage of being more developed, we continued the practice of using the lower bound of six action clauses to identify stories in the present study. Inter-rater agreement for identifying stories was 86 percent.² The following example illustrates how we parsed stories into action clauses, which are underlined. Clause breaks are marked with slashes. Note that some of the action is reported speech of a story character.

You should have come with me and Kate to Kinko's yesterday,/ ‘cause I had to make photo copies of those porn magazines/ for my presentation. / Kate was just like, ‘You're such a wuss/. Why can't you do this by yourself?’/ And I was just like, ‘It's so embarrassing.’/ And there were, a few people there / and we went to the copy machine / and you know just like totally like, a huge naked woman on the cover and everything. / So I'm lifting up the Kinko's thing / and looking around./ Shoving it in, /and this woman kept coming behind me/ and using like the stapler thing/ or… and I'm just like, ‘Hmm hmm.’/ (laughing)

How the story was initiated

Five mutually exclusive ways of launching a story were identified. First, stories could be embedded in an ongoing theme; embedded stories were volunteered and relevant to the ongoing theme of the conversation, but did not come on the heels of a prior story. Second, stories could emerge suddenly, out of the blue; these stories abruptly changed the theme, for example, my story about a car accident pops up while we are discussing a school assignment. Third, stories could be solicited by the partner; these stories were specifically requested by the partner, for example, “What happened with Albert?” (“What's new?” did not count as a request because it did not inquire about a particular event). The last two modes of initiation concerned stories that followed on the heels of a prior story, or second stories. Second stories were defined as occurring within three rounds of speaking turns from a prior story, and as continuing the theme of the prior story. One kind of second story, a second story volley, involved a change in teller, for example, my story about Joe follows your story about Joe. Second story dribbles involved the same teller, for example, I tell one story about Joe and then tell another story about Joe. Independent coders reliably differentiated these five ways of initiating stories (kappa = .82).

Who experienced the event in the story

Stories could be about one's own experience, a shared experience, or a third party's experience. A me story was about an event that was directly and actively experienced by only one speaker. A we story was about an experience that in whole or in part, both speakers directly experienced, for example, a story about our bumpy flight to New York, or a story about my boyfriend flirting with you and my stomping out of the house to argue with him in the garage. A they story was about a third person or persons, and concerned an event in which neither of the speakers had actively participated; such stories were known through eavesdropping, or through another person who had reportedly recounted the story to one of the speakers, who was portrayed as a passive listener. A fourth category, a you story, was very infrequent in these conversations, and concerned stories told about the listener, for example, “You are really secretive around your roommates” (kappa = .90)³. Of these subcategories, our primary interest was in distinguishing “we” stories from “me” stories.

Who constructed the plot

The plot of a story could be constructed by one speaker, or co-constructed. Co-constructed plots were defined as plots in which each speaker contributed at least 25 percent of the action clauses, with a minimum of three such clauses per speaker; this rule was developed to insure that both shorter plots (a minimum of six action clauses) and longer plots would require substantial listener contributions in order to be considered co-constructed. A listener could co-construct a plot by asking a question about action (“So what did he say?”), by contributing new action (“I heard that too, and he said you were crazy”), or by contributing real or imagined dialogue (“So he was probably going, ‘Dude! You are such an idiot!’”). Plots that were not co-constructed were coded as solo-constructed (kappa = .83).

Time setting of the story

Most stories referenced a time period in which the story happened, such as last night, last Wednesday, last year, or fifth grade. We coded each story for whether the original event reportedly had occurred within the past week, longer than the past week but during college-age, or prior to college, that is, between childhood and high school (kappa = .85).

Story topic

Topic categories were developed by sorting the stories according to what the story generally concerned. Seven categories emerged: school (classes, homework, teachers); romance (a sexual or romantic involvement or interest that was clearly non-platonic); peers (platonic relationships with peers); family members/home town; jobs (paid work); basic needs (health, food, money, transportation) and unclassifiable. Trump rules were used to decide cases that had overlapping concerns. Because this age group is generally transitioning from family to romantic attachments (Arnett, 2004), we designated romance as the top trump category, followed by family, and peers. For example, a story about taking one's boyfriend home to meet family was coded as a romance story, and a story about taking peers home to meet family was coded as a family story. The kappa reliability of these seven topic categories was .88.

Results

Data Analysis Strategy

Dyad as the unit of analysis

Because conversations and stories are joint productions, the dyad was used as the unit of analysis for the story measures. We also used the dyad as the unit of analysis for the personality and friendship measures, since dyads were selected due to similarity on extraversion-introversion. Averaging the personality scores and friendship ratings within dyads was justified by the substantial similarity between target (the person initially recruited) and partner (the friend who was brought to the study) not only in terms of Extraversion-Introversion scores, r = .88, but also with regard to ratings of friendship closeness, r = .75, and length of friendship, r = .95. Dyadic indices of these variables were computed by summing the two friends' ratings on the variable and dividing by two.

Metric for story categories

Because each dyad produced different numbers of stories, the metric for the story categories was the percentage of stories per dyad showing a particular subcategory. For example, story initiation had five subcategories. If a dyad produced five stories, with one story in each of these subcategories, this pattern was represented as .20 for each subcategory.

Effects for dyad personality and gender

The effects of dyad personality (introverted or extraverted) and dyad gender were compared with 2 × 2 ANOVAS for all variables except the story categories. Because the subcategories for story initiation, plot construction, whose experience the story was about, time setting, and topic were proportional and therefore interdependent, these five sets of categories were each analyzed with 2 × 2 MANOVAs.

Base Rate of Story Categories

Table 1 shows the base rate of the story categories in the overall sample of 141 stories. With regard to the time frame of the stories, the majority of the stories, n = 80, concerned events that had transpired in the past week; events that had happened before coming to college were least likely to be storied, n = 14. No particular topic dominated the conversations, although the most frequent topics were romance, n = 49, peers, n = 30, school, n = 21, and family/hometown, n = 15. The stories were primarily about experiences that had happened to only one partner (“me” stories), n = 99, rather than both partners (“we” stories), n = 24. With regard to who told the plot of the story, the large majority of the stories were solo- rather than co-constructed, n = 112 versus 29. Finally, the most frequent modes of initiating a story were to embed the story in ongoing dialogue, n = 60; to solicit a story, n = 32, or to volunteer a story suddenly, out of the blue, n = 29. Relatively few stories emerged on the heels of a prior story, although volleys (in which speakers took turns telling a story) were more prevalent than dribbles (in which one speaker told serial stories).

Table 1. Overall Frequency of Story Variables.

Coding category	Overall frequency (N = 141 stories)	Percentage
Time setting of story
Past week	80	57
During college	45	32
Pre-college	14	10
Unclassifiable	2	1
Story topic
Romance	49	35
Peers	30	21
School	21	15
Family, hometown	15	11
Job	9	6
Basic needs (finances, food, shelter)	7	5
Not classifiable	10	7
Who experienced the event in the story
Me	99	70
Us	24	17
They	15	11
You	3	2
How the story was initiated
Embedded in ongoing dialogue	60	43
Sudden, out-of-the-blue	29	21
Solicited by friend	32	23
Second story volley	15	11
Second story dribble	5	4
Who told the story plot
Solo-told	112	79
Co-told	29	21

Gender Differences

Only two significant main effects were found for dyad gender. First, female dyads rated this friendship as closer than did male dyads, M = 4.32, SD = .65 versus M = 3.65, SD =.95, F (1,39) = 6.53, p < .05, partial η² = .16, a gender difference that tends to be found with survey methods (Way & Pahl, 1999). Second, male dyads changed speakers more frequently, M = 82.81, SD = 18.41, than did female dyads, M = 66.50, SD = 19.34, F (1,39) = 9.02, p < .01, partial η² = .20. Only one significant interaction of dyad gender and personality was found, for conversation length. Inexplicably, male introverted friends talked more than female introverted friends: males averaged 1869 words, and females averaged 1696 words, and the reverse was true for extraverted friends: males averaged 2049 words, and females averaged 2263 words, F (1,39) = 4.09, p < .05, partial η² = .10. A recent meta-analysis found that gender differences in turn-taking and conversational length tend to be quite context-dependent (Leaper & Ayres, 2006). Overall, differences between male and female dyads were minimal, and no gender differences were found for the story categories. We now turn to the focus of the study, main effects for personality.

Personality Differences in Friendship Closeness and Length

Introverted and extraverted dyads did not differ with regard to their assessment of the closeness or length of their friendship. With regard to closeness, M = 4.03, SD = .86 for introverted friends, M = 3.93, SD = .91 for extraverted friends, F (1,39) = .49, p = ns, partial η² = .01. With regard to length of friendship, M = 17.92 months, SD = 15.06 for introverted friends, M = 18.50, SD = 16.48 for extraverted friends, F (1,39) = 1.42, p = ns, partial η² = .00. Generally, the friendships were regarded as quite close, averaging 4 on a 5-point scale. The average length of friendship was 18 months, ranging from five months to six years.

Personality Differences in Story Length, Frequency, and Latency

Table 2 presents descriptive statistics for extraverted and introverted dyads with regard to the general contours of the conversations, and the coded story categories. With regard to general contours, the stories as well as the conversations of extraverted dyads were found to be significantly more wordy than those of introverted dyads. Extraverted stories averaged 282 words in length, while introverted stories averaged 170 words in length, p < .001, an average difference of over 100 words. Differences in volume of spoken words produced have been found in prior extraversion research (Dewaele & Furnham, 1999). However, introverted and extraverted dyads did not differ with regard to the number of stories produced; overall, an average of three or four stories were told per dyad (range = 1-8). Introverted and extraverted dyads also did not differ with regard to how long it took for the first story to emerge in the conversations. As shown in Table 2, conversations showed considerable variability on this index; on average, the first story was launched at line 40, which tended to be about 2 minutes into the 10-minute conversation. Across the entire conversation, no personality differences were found for the frequency with which speakers rotated. We next examine differences in story content.

Table 2. Descriptive Statistics for Story-telling by Introverted and Extraverted Dyads.

	Introverted		Extra verted		F (1, 39)	Partial η2
	M	(SD)	M	(SD)
Speaking turns per conversation	70.95	(16.81)	78.23	(22.87)	2.64	.07
Conversation length (words)	1791.37	(287.42)	2166.45	(307.04)	16.30***	.32
Story length (words)	169.97	(63.49)	282.29	(113.33)	12.93***	.27
Latency of first story (line number)	51.47	(46.80)	34.70	(35.15)	1.14	.03
Number of stories per dyad	3.16	(1.43)	3.95	(1.99)	1.54	.04
Time setting of story
Past week %	.57	(.39)	.59	(.33)	.02	.00
During college %	.21	(.28)	.40	(.34)	5.03*	.13
Pre-college %	.21	(.34)	.01	(.04)	5.79*	.14
Story topic
Romance	.16	(.28)	.44	(.31)	9.74 **	.22
Peers	.19	(.21)	.29	(.36)	.72	.02
School	.25	(.28)	.10	(.15)	3.62 †	.09
Family/hometown	.16	(.29)	.04	(.10)	4.03 *	.10
Job	.08	(.17)	.03	(.12)	.93	.03
Basic needs (finances, food, shelter)	.02	(.06)	.05	(.10)	1.14	.03
Not classifiable	.14	(.31)	.05	(.11)	1.01	.03
Who experienced the event in the story
Me %	.75	(.30)	.63	(.29)	1.99	.05
Us %	.12	(.26)	.22	(.22)	1.35	.04
They %	.11	(.17)	.15	(.26)	.65	.02
You %	.02	(.07)	.01	(.04)	.51	.01
Who told the story plot
Co-told (vs. solo-told) %	.08a	(.17)	.35b	(.33)	9.55**	.21
How the story was initiated
Embedded in ongoing dialogue %	.57	(.30)	.35	(.24)	6.43*	.16
Out-of-the-blue %	.11	(.19)	.34	(.33)	6.59*	.16
Solicited by friend %	.21	(.31)	.21	(.26)	.00	.00
Second story volley %	.07	(.15)	.08	(.13)	.00	.00
Second story dribble %	.03	(.09)	.02	(.07)	.02	.00

^†p < .10;

^*p < .05;

^**p < .01;

^***p < .001, two-tailed.

Personality Differences in Story Content

Time setting of the story

As shown in Table 2, for both types of dyads, the majority of the stories in each conversation, M = 58 percent, were about events that had occurred in the past week, a finding that is not surprising given our request for a “catch-up” conversation. A MANOVA found that two time settings differed significantly by dyad type. Extraverted dyads told a significantly greater proportion of stories about events that happened during the college years (but not within the past week) than did introverted dyads, M = 40 percent versus 21 percent, respectively, p < .05. Introverted dyads, in contrast, told a significantly greater proportion of stories about pre-college events than did extraverted dyads, M = 20 percent versus 1 percent, respectively p < .05. Introverted dyads produced 12 stories about pre-college experiences, and extraverted dyads produced 2 such stories. In all cases, the pre-college experiences had occurred before the friendship had formed.

Story topics

As shown in Table 2, extraverted friends more often told stories about romantic relationships, nearly three times as much as did introverted friends, M = 44 percent versus 16 percent, p < .01. Introverted friends did not specialize in a particular topic, although their highest density of stories was about school, M = 25 percent, followed by stories about peers, M =19 percent, romance, M = 16 percent, and family/hometown, M = 16 percent. Relative to extraverted friends, introverted friends told significantly more stories about family/hometown, M = 16 percent versus 4 percent, p < .05, and marginally significantly more stories about school, M = 25 percent versus 10 percent, p < .07.

Whose experience the story was about

Introverted and extraverted dyads did not differ with regard to whether stories were about “me” or “we” experiences. As can be seen in Table 2, for both types of dyads, approximately 70 percent of the stories in each conversation were about one's own experience (me stories), 16 percent were about a shared experience (we stories), and 13 percent were about a third party (they stories). We now turn to the focal analyses, personality differences in the storytelling process.

Personality Differences in the Storytelling Process

How the story was initiated

A MANOVA found that 2 of the 5 modes of story initiation differed significantly by dyad personality. As can be seen in Table 2, introverted dyads specialized in embedding the story in an ongoing topic, so that the story nested within the ongoing dialogue; whereas 57 percent of the introverted stories were initiated in this way, this was true of only 35 percent of the extraverted stories, p < .05. Extraverted dyads, in contrast, more often volunteered stories out of the blue, departing from the theme of the ongoing dialogue; whereas 34 percent of the extraverted stories were offered out of the blue, this was true of only 11 percent of the introverted stories, p < .05. No significant differences for dyad personality were found for story solicitations (requests), which averaged 21 percent per conversation, or for second story telling (volleys or dribbles), which cumulatively accounted for about 10 percent of stories per conversation.

Who constructed the story plot

The large majority of the story plots were constructed individually (overall M = 78 percent), a pattern that did not differ by dyad type. However, co-constructed story plots were significantly more prevalent for extraverted than introverted dyads. As shown in Table 2, 35 percent of the story plots of extraverted dyads were co-constructed, compared with only 8 percent of the story plots of introverted dyads, p < .01. The vast majority of stories told by introverted friends were solo-constructed, M = 92 percent, versus M = 65 percent for extraverted friends.

To co-construct a plot would seem to require knowledge of what the story was about. To explore whether co-constructed plots were primarily stories about a shared experience, we arrayed co-told stories by whose experience the story was about (me, us, they, or you). For introverted dyads, all of the 5 co-told stories were about a shared, we, experience. For extraverted dyads, 15 co-told stories were about we experiences, but 8 co-told stories were about me experiences. Thus, plot co-construction was limited to shared experiences for introverted dyads, but also extended to individually experienced events for extraverted dyads. A closer examination of the method by which extraverts' me stories were co-constructed showed that listeners participated in me stories primarily by repeatedly asking questions about the action: The ratio of action questions to the overall frequency of plot co-constructions (which, in addition to action questions, could be statements about action or dialogue) averaged .72 (SD = .32) for extraverted listeners' co-constructions of me stories (n = 8 cases). The parallel ratio for extraverts' co-constructions of we stories averaged .09 (SD = .11; n = 15 cases), with a similar ratio of .12 (SD = .07) for introverts' co-constructions of we stories (n = 5 cases). Thus, co-constructed we stories primarily consisted of action statements rather than questions, which makes sense because both speakers were privy to aspects of the original event.⁴

Summary of Story Findings

With regard to the storytelling process, introverted and extraverted friends differed in how they initiated stories, and in their engagement in co-telling stories. Specifically, extraverted friends more often initiated stories suddenly or out of the blue, and more often co-told the plot of the story. Introverted stories, in contrast, developed more gradually within the bounds of an ongoing topic, and their story plots were primarily constructed solo. With regard to story content, we found that extraverted friends more often told stories about romance than did introverted dyads, who more often told stories about family/hometown, and, marginally significantly, about school. With regard to time setting, introverted dyads more often told stories about pre-college experiences, which overlapped with their telling more stories about family and hometown.

In order to get closer to the data than numbers permit, we now consider some narrative cases of these different story-telling practices. The cases were chosen because they represented multiple aspects of the findings and afford a more fluid and veridical presentation of the process by which stories got launched and elaborated. Names have been changed to protect identities.

Narrative Examples of Out of the Blue versus Embedded Stories

The following sudden, out-of-the-blue, story, told by extraverted friends, suggests how such stories can legitimately emerge in conversations despite their seeming irrelevance to an ongoing topic. The first speaker, Ann, has just told a story about opting not to hang out with friends last night, and then casts herself and her friend as gossipers. The friend, Sue, then launches a story about a new topic, going home and meeting a cute bassist. Underlined text marked with a slash indicates an action clause, the unit of analysis for a story plot.

• Ann: we're the gossip talkers on this tape, gotta do the gossip

• Sue: cause I went home for like a week./ When you went down to, uh

• Ann: did you have fun?

• Sue: yeah, yeah, I met this really cute bassist/ [proceeds with story]

In the above excerpt, Ann's frame of “we're the gossip talkers” seemed to suggest that it was legitimate to abandon the prior conversational topic, which Sue promptly does by beginning to launch a story about what happened at home last week, when she met a cute bassist.

Out-of-the-blue stories did not necessarily get framed as gossip. For example, Tony and Alan, extraverted friends, were talking about their plan to get a television next year so they can watch the Lakers (a basketball team) right from their own couch. Then suddenly, Tony launched into a story about his problems with his girlfriend Elizabeth. Action clauses, which constitute the plot of the story, are underlined.

• Tony: that will so perfect though next year when we have our own TV. We'll watch the Lakers

• Alan: yeah, pony-keg at ours

• Tony: win a three-peat'-

• Alan: eah

• Tony: we'll watch the three-peat' next year

• Alan: oh my god

• Tony: and we'll be right on our own couch

• Alan: hangin' out. I think that'll be nice [pause]

• Tony: yeah so I just have to find out after this what the deal is with Elizabeth

• Alan: um-hm

• Tony: and just clarify everything with her

• Alan: yeah, see what's up

• Tony: she just walked out last night/

• Alan: really? (Tony continues with story)

The newsworthy nature of the story of what happened with Elizabeth last night seems to be its ticket to legitimacy. Although the story does not seem to be relevant to the plans to get a TV, it conforms to the instruction to engage in a “catch up” conversation.

Some extraverted out-of-the-blue stories emerged during a lull in the conversation, when the prior topic seemed to have played itself out. For example, one such story, told by Dora, occurred in the midst of making a plan to meet at the library that evening:

• Dora: okay and we'll start that tomorrow around 6, at what library? Hoover?

• Mary: I guess yeah at Hoover yeah, then we could go to the Art Center

• Dora: okay you know, who is that girl that Andy was with today?

• Mary: I don't know

• Dora: [story begins]

Instead of offering stories out-of-the-blue, introverted friends primarily offered stories that were embedded in an ongoing topic. Sometimes such stories seemed to emerge in an effort to rectify a misconception on the part of the listener. Here is an example from a conversation between introverted friends Bill and Tom. Bill is explaining that he is missing his class today to participate in the present study, and that Lisa is going to take notes for him. That Lisa is in Bill's class is a surprise to Tom, and Bill hastens to explain that he does not purposely take the same courses as his friends, giving the example of the courses he happens to be taking with another friend, Ernie; then ensues the story about Ernie:

• Bill: um, but Lisa's going to take notes for me because I'm always helping her out with her notes because

• Tom: well, you go to Lisa's classes

• Bill: me and Lisa are in the same class

• Tom: oh

• Bill: we randomly decided we

• Tom: you're both doing Physics or Psychology?

• Bill: Psychology, yeah. Um because uh she occasionally has trouble finishing all the notes before the professor changes the slide or changes the page in the notes. I'll help her finish up her notes later on

• Tom: yeah

• Bill: so she's going to cover for me today, so

• Tom: oh, I see ok yeah

• Bill: I figured coming here and getting two hours

• Tom: I didn't even know you were doing the same classes

• Bill: well, it wasn't on purpose

• Tom: well um Intro, Intro to Psych

• Bill: yeah, Intro to Psych

• Tom: yeah

• Bill: um, yeah, it wasn't even on purpose. It's like me and Ernie. Me and Ernie have chosen one of the same classes/ all three quarters without actually being, you know, without planning it./

• Tom: yeah

• Bill: like he, I'd figured that we weren't going to do it this same, this quarter/ and while he was choosing classes one day/ he's like, all his classes he wanted to do for his third one were full./

• Tom: yeah, yeah

• Bill: he's like ‘how is this planter discovery class?’ / I'm like ‘no, that's my class / wanted one class without you’/

• Tom: yeah

• Bill: and so

• Tom: so, you are doing, all of you at least doing a class with someone else

• Bill: well, not on purpose

• Tom: like with all of your friends

At the end of Bill's story, Tom summarizes the point (“So….all of you….are doing a class with someone else”) and Bill reiterates his position that it's not on purpose. They subsequently proceed to catalog which friend is with which other friend in which class, which gradually becomes a discussion of someone's film class, and the telling of a story about that film class. The effect is a meandering conversational theme within which stories nest, without abruptly changing the topic. The more nested nature of introverted stories, and the more transient nature of extraverted stories have potential implications for identity construction.

Examples of Co-constructed versus Solo-constructed Plots

We now turn to the question of how an extraverted listener managed to co-construct a plot about an event that only her/his friend had experienced. Such co-constructions were sometimes encouraged by inviting the listener to guess what happened, which lured the listener to keep guessing about what happened. The listener was presumed to have sufficient knowledge about the teller's life to make educated guesses, although the guesses were as often wrong as right. The following excerpt from extraverted friends illustrates this dynamic of plot co-construction. The story belongs to Marie, but Donna is an avid participant in guessing the plot. Action clauses contributed by the listener are indicated in brackets as [L].

Marie: um last night, um, (laugh) I am gonna sound like such a little gossiper
Donna: that's alright
Marie: um guess who called me?/
Donna: did Sally call?/	[L]
Marie: no. That would be like
Donna: did Chris?/	[L]
Marie: no
Donna: I'm trying. Terry called?/	[L]
Marie: no
Donna: I'm trying to think of like, people who would, who?
Marie: scary Susie called/
Donna: uhhh (gasp)
Marie: yeah
Donna: oh no!
Marie: yeah okay, she called me once/ and then um, I didn't tell you about, okay, so like, what happened was um, that time I got back from Fresno/ you know like way back in January
Donna: like the first time you got back?/	[L]
Marie: yeah, and I was like all sleeping for two days/ and she called /and I didn't know who it was/ and I didn't call her back/ And like she um
Donna: wait was she the one who left that weird message on your machine,/ like, ‘oh, I wish you were home’, /member, and I thought it was Rhonda?/	[L]
	[L]
	[L]
Marie: Nancy? oh maybe that was her. I don't know. I never found out who that was/but anyways, so she called /um Kim's like, ‘you have a lot of girl Stalkers’/ and I was like, ‘I know/ I don't like 'em’/
Donna: I know
Marie: but she she
Donna: there's something about you
Marie: she um was um she was like ‘why didn't you call me back?’/ da da da da and I was like, ‘well you know um I didn't know / I was like sleeping for two days, I was out of it’/
Donna: wait this is last night that she's asking about this?/	[L]
Marie: no, okay, no this is over winter break
Donna: right right

Guessing about the plot is an intriguing way to involve the listener in the construction of an event that was solely experienced by the teller. These extraverted friends seem be reaching for overlapping knowledge and experience, actively inserting their knowledge into representations of each other's lives.

As a final contrast, here are two stories about the romantic problems of third parties; in each case, only one speaker “knows” the story. However, the plot of the first story is co-constructed by extraverted friends. The plot of the second story is constructed solo by an introvert.

Joe: I talked to Johnny today/
Bob: really? What'd he have to say?/	[L]
Joe : uh he said that his girlfriend's angry with him/because
Bob: is she pulling the camping thing again? /	[L]
Joe: oh no, she's pulling, well not so much that, she's pulling the ‘I don't want you to go home this summer’ /
Bob: oh God, yeah, she's attached
Joe: ‘because I'll never get to see you, Johnny’/
Bob: ‘and there's so many other women Johnny’/	[L]
Joe: ‘that Joe’/
Bob: ‘he's a bad influence on you’ /	[L]
Joe: ‘he's the devil!’/ so I don't know what's, I don't know what's gonna happen with him

Although Bob, the listener, is not privy to this particular event, he appears to know enough about the story characters to manufacture apparently credible imaginary dialogue, which he uses to co-tell the plot of the story. Of the 10 action clauses in the story, Bob, the listener, contributes four--two questions about the action, and two imaginary dialogues.

The last case, from introverted friends, is also a recount of a third party's romantic troubles, but this plot is solo-constructed by Al. Friend Bo contributes only one action clause, the question, “What'd he do?” A single question from the listener about what happened was quite typical in solo-told stories. Although Bo talks quite a bit during Al's story, Bo is mostly contributing evaluations of Al's story.

Al: and like a couple weeks ago, he told me how he made this big mistake with her/
Bo: what'd he do?/	[L]
Al: well like, he-heh-he was like joking around/ though, but sometimes she takes things personally, because that's how girls are. You know how guys always joke around and sometimes a girl hears, then totally takes it the wrong way?
Bo: I haven't had it happen to me, but I've seen it happen so I know what you're talking about
Al: ok, well he said like, ‘yeah yeah you were so easy’/ and everything ‘when I got with you’/
Bo: mhwahahaha
Al: yeah, well he was joking around/ because that's how he is, he jokes around and well, she didn't take it jokingly/
Bo: oooh
Al: so like, but then he admitted he was wrong/ so he called her/ and stuff and said you know, ‘I'm sorry I'm sorry,/ I was wrong,/ that was a stupid mistake,/ the dumbest mistake you could make’ /
Bo: I hate girls that make guys do that though
Al: yeah, well the worst part is ok, even though he apologized /she wouldn't accept it/

Although the action of the story is almost solely advanced by Al, each friend contributes critical evaluations about what is wrong with girls. The story seems to be an occasion for co-constructing their moral positions about girls.

Summary of Narrative Cases

These narrative examples illustrate how stories told by introverted friends developed within the bounds of an ongoing theme that meandered slowly as the speakers took up a setting or a character and nudged it in to a story. We also saw how introverts' story plots were advanced by the storyteller, with little explicit effort to invite the listener to co-construct the plot. In the process, we also saw an interesting, yet to be explored, engagement in the other's story--evaluating its meaning. Cases of extraverted story telling, in contrast, showed appreciably more listener engagement in telling the plot of the story. Sometimes this listener engagement was invited by the teller, and sometimes the engagement was volunteered by the listener; either way, extraverted friends essentially invited and asserted claims on each other's experience. We also saw how extraverted friends cast their sudden stories as news or gossip, a frame that seemed to give these stories a different kind of legitimacy than the more thematically relevant stories that emerged in introverted conversations.

Discussion

When we began to collect conversations for this study, we did not anticipate that stories would be sufficiently plentiful to afford systematic comparisons of how the storytelling process was socially organized. Our expectation that introverted and extraverted friends would show different patterns of participation in storytelling was guided by a general rubric that extraverts tend to invite and assert themselves into others' lives more so than introverts, who are more protective of their personal experience. This rubric appears to have been a serviceable framework for discerning how friends who were similarly extraverted or introverted participated in channeling each other's stories.

Extraverted friends more actively engaged each other in constructing the action of the story, and more abruptly imported stories from the world outside of the conversation. Case analyses illustrated how extraverted friends accomplished this feat: through such mechanisms as peppering the storyteller with questions, or asking the listener to guess what happened, extraverts invited each other to co-tell the story. In order to guess correctly about an experience that had only happened to the teller, the listener had to be up to date on what might be newsworthy to the teller, but even an incorrect guess displayed knowledge of other aspects of the teller's life, thereby exuding common ground. Another way in which extraverted friends' sense of their lives was more overlapping was that the events in their stories predominantly had happened during the college-age years; they rarely offered stories about experiences that had happened before they had become friends. Overall, these findings suggest that extraverted friends were working from an inclusive sense of self that actively and abruptly incorporated the listener and the news of the day.

Introverted friends, in contrast, less often invited each other to participate in constructing the plot of the story, resulting in fewer co-constructed stories. In addition, introverted friends' stories emerged more gradually, within the confines of the ongoing conversational theme. Case analyses showed the meandering way in which introverted friends' conversations slowly condensed into a story that explained a misunderstanding or clarified a point of view. Overall, the process of introverted storytelling suggested a more cautious offering of stories. Notably, these differences did not appear to be driven by introverts generally engaging in less back and forth dialogue, since we did not find personality differences in overall frequency of speaking turns. Rather, these different dynamics suggest different psychological orientations.

With regard to story topics, the finding that extraverted friends told a higher proportion of stories about romance suggests that they may have been more romantically active than introverted participants. Although we did not survey participants' romantic histories or current romantic status, less sociable people have been found to report lower frequencies of falling in love during the first years of college than more sociable people (Asendorpf & Wilpers, 1998). Another possibility is that introverted friends felt that discussion of their romantic relationships was inappropriate for this setting, or perhaps more generally (e.g., Levesque, Steciuk, & Ledley, 2002). Introverts' more frequently telling stories about family and hometown may indicate a greater attachment to their former lives, compared to extraverted friends, all but one of whose stories centered on experiences during the college years. An alternative explanation is that introverts may favor the telling of old or retold stories, which may allow them to stay on comfortable or familiar ground.

Overall for this college student sample, stories about school and family were notably less prevalent than stories about experiences with peers, either romantic or non-romantic. Peer-based concerns have also been found to prevail in college students' stories of self-defining events (McLean & Thorne, 2003), suggesting that everyday and momentous stories center on similar kinds of concerns. This peer focus seems to reflect the kinds of intimacy concerns that are considered crucial in the transition to adulthood (Arnett, 2004; Erikson, 1968). Given the developmental import of peer relations for this age group, future research should go beyond the study of general story topics to a closer examination of how introverted and extraverted friends position themselves vis-à-vis intimacy in telling stories about close relationships (Korobov & Thorne, 2006).

The present findings should be viewed with caution because the findings have no precedent in past research, and the effect sizes were modest. In addition, the laboratory setting may not have reflected how introverts, in particular, tend to tell stories with friends. The instruction to engage in a recorded conversation may have felt particularly unnatural to introverts and thereby accentuated personality differences. Alternatively, the context may have tended to mask differences that would be more likely in everyday contexts. For example, the finding that both extraverted and introverted dyads primarily told me stories rather than we stories may have been an artifact of the instruction to “catch up.” It is also important to note that the sample was limited to college students who were primarily of European-American descent; studies of storytelling in other cultural communities and age groups suggest different practices with regard to how stories are told, and by whom (e.g., Miller, Wiley, Fung, & Liang, 1997). Other important limitations are that we did not incorporate the domain of goals or motives in this study, another key feature of personality (Winter et al., 1998), nor did we examine how stories were evaluated, an important aspect of storytelling and identity making (McAdams, 2001; McLean & Thorne, 2003).

With regard to our conceptualization of introversion and extraversion, is it important to note that the exclusive-inclusive distinction may seem to resonate with other kinds of distinctions that have become common in psychology, such as independent-interdependent selves, and individualism-collectivism. The latter constructs, however, have been developed to explain differences that are attributed to socio-cultural values rather than to predispositions. Arguably, the greater sensitivity of introverts to external stimulation (Eysenck & Eysenck, 1985), and their more cautious approach to social interaction exudes not individualism but rather a more exclusive sense of self that is reluctant to surface and does not easily invite others in, perhaps even friends. Similarly, the social boldness of extraverts exudes not collectivism but rather social assertion, an assumption that a person's stories are representations that are fair game for public engagement.

The notion that introverts exude more of a sense of inhabiting a distinctive, exclusive world has been suggested in past research conducted at the level of the single word or clause (Pennebaker & King, 1999; Thorne, 1987), in surveys that have found that introverts report sharing memories with fewer people, and less comfortably, than extraverts (McLean & Pasupathi, 2006), and in daily experience samples, in which introverts report less often exchanging memories with each other (McLean & Pasupathi, 2006). This confluence across an array of contexts and methods suggests that extraversion and introversion are communicated in many ways. And so, it is important to ask why it is necessary to go to the trouble of parsing conversations for stories, and then detecting how the stories emerge. What is so special about the storytelling process? One reason to study storytelling is that it draws much-needed attention to the reciprocal workings of personality traits in social interaction, since only through listener cooperation can a story be launched and completed (Ochs & Capps, 2001). Also, an array of studies have found that how one tells a story and how the listener responds can impact not only memory for the story, but also physical and emotional well-being (e.g., King & Raspin, 2004; Pasupathi, Stallworth, & Murdoch, 1998; Pennebaker, 1997). At present, very little is known about longer-term outcomes of different kinds of storytelling processes. For example, a more exclusive, introverted, storytelling process might result in relatively fewer memories which are relatively more rehearsed, or possibly stronger links or connections drawn across these memories, or a greater imperviousness to listener feedback.

Overall, the present study represents the kind of process research that can get to the heart of how individuals reciprocally construct their lives in the context of social relationships, a major source of personality continuity (Caspi, 2000). More specifically, the findings have important implications for synthesizing trait and identity approaches to personality in that friends who were similarly extraverted or introverted showed distinctive patterns of telling spontaneous stories of their lived experience. Although extraversion-introversion is only one feature of the self-concept, this shared individual difference was found to channel how friends participated, albeit briefly, in storying their lives. Traits and stories, when studied together, in close relationships, and in vivo, provide a dynamic and vivid view of how traits contribute to the social construction of identity, and how traits emerge in the process of storying one's life.

Figure 1.

How stories were initiated by introverted and extraverted dyads: Mean proportion per dyad of each kind of initiation.