The gendered transition to college: The role of

culture in ego-network evolution

Danielle Kane

Duke University Thompson Writing Program, Box 90025, Durham, NC 27708, USA

Available online 8 July 2011

Abstract

This article argues that gender norms and students’ entering network structure affect their transition to

college. More specifically, just as Bourdieu (1977) posited that the education system privileges the culture of

the dominant classes, I argue that elite universities favor those students who enter with sparse, diverse

networks – the network capital of the adult upper class – and that their presence creates incentives that lead

all students to adopt this structure. I predict that cultural mandates that encourage women to cultivate and

manage ties actively will foster a more satisfying social transition than for men, who rely more on the very

dense networks that the elite university environment undermines. I find that after one year at an elite university,

students’ networks are sparser and more diverse. Interviews reveal that men from dense networks experience a

particularly difficult social transition to the university. Because gendered cultural norms contribute so greatly to

tie formation in this sample, I conclude that culture plays a key role in network evolution.

# 2011 Elsevier B.V. All rights reserved.

1. Introduction

College influences a variety of stratification outcomes (Bowen and Bok, 1998; Kingston et al.,

2003), but the mechanisms of influence are poorly understood because sociologists have focused

on single points in time – enrollment and graduation – while neglecting the lived experience of

college (Stevens et al., 2008).1 This gap in higher education research can obscure the

stratification processes that take place during college. For instance, much of the female advantage

in college completion comes from gender-specific behaviors of young adults while in college

www.elsevier.com/locate/poetic

Available online at www.sciencedirect.com

Poetics 39 (2011) 266–289

E-mail addresses: Danielle.kane@duke.edu , dckane@gmail.com.1 In the United States, colleges provide undergraduate education; that is, students earn bachelor’s degrees. Typically

universities comprise a number of colleges and also offer graduate education; therefore universities are often larger than

free-standing colleges. Although the US has thousands of colleges and universities, elite schools wield a disproportionate

influence in higher education issues.

0304-422X/$ – see front matter # 2011 Elsevier B.V. All rights reserved.

doi:10.1016/j.poetic.2011.05.003

(Buchmann and DiPrete, 2006, p. 535, emphasis mine). This suggests that the gender gap in

college graduation rates is attributable to men’s lagging college completion, which in turn seems

specific to their – as yet unexplored – college experience.

Therefore, a better understanding of the role of college in stratification will require more insight

into the college experience. The large-scale quantitative datasets typically used to investigate

higher education give little insight into this experience. However, a departure point may come from

another tradition of higher education research, which suggests that forming social ties is so

important that we might conceive of the transition to higher education as network evolution, an area

that has generated an increasing amount of interest among network sociologists (Clydesdale, 2007;

Doreian and Stokman, 1997; Stevens et al., 2008).2 The longstanding emphasis on social ties in

higher education research suggests that a network approach would be a promising way to explore

differences in college experience.

Because sociologists have established that gender differences in college behaviors relate to

different degree completion rates – a key factor in stratification – this article examines gender

differences in the transition to higher education. This example also provides an opportunity to refine

our understanding of network change (or evolution) and the role that culture plays in this change.

Sociologists have only just begun to consider how culture shapes social network patterns. While this

literature has focused on the role of tastes, I propose that two other forms of culture can shape

network formation, institutionally valued network structures and a culturally informed ‘‘network

know-how,’’ the name I give to the spectrum of skills that allow one to adapt to a new network

environment.

1.1. The transition to college and tie formation

Tie formation is fundamental to the study of network change. Most researchers examine

network evolution through mathematical models and simulations, emphasizing technical

considerations (e.g., Doreian, 2006; Huisman and Steglich, 2008; Moody et al., 2005; Robins

and Pattison, 2001; Willer, 2007). This work often aims to identify how global networks

evolve and is usually less concerned with the effects of ego-network evolution. Ego-networks

refer to ties around specific agents (in this case, students), while global networks refer to

‘‘the pattern of connections in the network as a whole’’ (say, for an entire university)

(Scott, 2000, p. 72).

However, a qualitative, empirical approach to ego-network change has the potential to inform

sociological research in two ways. First, qualitative work could inform the rational-actor theories

that drive many of the models of network evolution. For instance, Doreian (2006) acknowledges

that his models’ assumption that all actors calculate costs and benefits in the same way is overly

restrictive. He notes that different cost-benefit regimes could be included (Doreian, 2006);

empirical work that examines how individuals think about tie formation could inform these

modeling decisions.

D. Kane / Poetics 39 (2011) 266–289 267

2 Doreian and Stokman (1997, p. vii) introduced their volume on the topic by noting that network evolution is a

fundamental problem whose fruitful solution is necessary for realizing the revolutionary potential of network analysis in

the social sciences. McPherson et al. (2001, p. 438) conclude their 2001 Annual Review article on social networks by

arguing for attention to ‘‘the ways in which networks evolve over time through cumulative processes of tie creation and

dissolution. . .’’ Social Networks devoted a special issue to the topic in 1997. Similarly, Marsden (1988) has advocated

determining how homogeneity patterns arise and why they vary in magnitude across groups.

Second, some important social issues are specific to ego-network change. For instance, Sageman

(2004) demonstrated how moving from a rural network to a university network increased young

Egyptian men’s susceptibility to joining terrorist groups. Their social mobility ‘‘had distanced them

from their original families or friends and, when they were socially isolated in a big university city,

they found a family substitute in the [terrorist group]’’ (Sageman, 2004, p. 133). Implicit in this

analysis is the idea that the challenge to these men arose from their leaving their home network; if

instead their home network had evolved rather than their personal networks, they would not have

faced this transformation alone and presumably there would have been a different outcome. In fact,

many subfields of sociology examine at least some problems (or opportunities) that are specific to

this kind of network change. The fact that numerous social issues involve the evolution of personal

rather than global networks means that theorizing ego-network change has the potential to inform a

variety of sociological concerns. (The rest of this article deals exclusively with ego-networks,

which, for ease of reading, I will refer to simply as ‘‘networks.’’)

2. Theorizing network change: the intertwining of networks and culture

Network change is undertheorized, complicating efforts to make predictions about the

transition to college. However, some research indicates that culture may play an important role in

the creation of new networks. This emerging area has focused on the importance of shared

cultural tastes and worldviews for tie formation (DiMaggio, 1985; Lizardo, 2006; Vaisey and

Lizardo, 2009). A key insight of this research is that cultural preferences or interests (such as

liking or disliking a certain type of music) contribute to network formation and maintenance.

The case of college transition provides the opportunity to contribute to the exciting area of

cultural influences on social networks by suggesting other forms of culture that may come into

play when people are faced with exogenous shocks to their networks. In particular, I propose that

two forms of culture apart from shared interests shape tie formation: ‘‘network know-how’’ and

institutional norms. In the remainder of this section I discuss ‘‘network know-how,’’ particularly

as it relates to gender norms, and in the following section I will discuss the role of institutional

norms in shaping network change.

2.1. Network know-how and gender

I use the term ‘‘network know-how’’ to refer to the idea that people acquire meta-knowledge

about ties themselves—that is, how to form relationships, how to stay in them, and what are

reasonable expectations of friends. Network know-how includes the interpersonal skills

surrounding the give-and-take of interaction, the skills of ‘‘peer group entry’’ and conflict

management, as well as the ability to self-disclose, and the ability to provide and elicit emotional

support—all key factors for gaining and maintaining friendships (Samter, 2003). Overall, we

might think of network know-how as shorthand for discerning the rules of interaction.

As an example of these rules in an empirical context, in his ethnography of a working-class,

Italian-American community in Boston, Gans (1982) depicted what is perhaps the prototype of a

dense network in a modern American context. (‘‘Density’’ refers to how interconnected a

person’s network is—that is, to what extent his or her ties know one another.) This group had

clear if unspoken rules about interaction. For instance, Gans described how ‘‘any attempt by an

insecure person to build himself up in the group at others’ expense is considered out of place. It is

politely ignored in his presence, and harshly criticized when out of earshot’’ (Gans, 1982, p. 82).

More generally, those who do not follow the West End rules of action ‘‘are rejected’’ (Gans, 1982,

D. Kane / Poetics 39 (2011) 266–289268

p. 93). The group consensus that allows a common response, such as gossip or ostracism, is

testament to the presence of rules for behavior, which in turn reflects a shared culture for how to

think about network ties and interaction.

Network know-how, itself a cultural form, is in turn shaped by a variety of other cultural factors.

For the purposes of this article, the most relevant factors are the culturally based gender norms that

may influence network transition. For instance, some scholars argue that men’s social interactions

are part of ‘‘doing masculinity,’’ a key component of which is suppressing any feminine traits

(Connell, 1995; Doyle, 1995; Kimmel, 1996; West and Zimmerman, 1986). As a result, although

men desire as much expressivity and intimacy in friendship as do women (Cordova et al., 2005;

Patrick and Beckenbach, 2009; Reid and Fine, 1992; Reisman, 1990), they are constrained by

gender norms that require them to avoid the intimacy and expressivity that are considered to be the

hallmark of women’s relationships (Connell, 1995; Doyle, 1995; Kimmel, 1996; Migliaccio, 2009;

Patrick and Beckenbach, 2009). According to masculinities research, ‘‘masculinity is coerced and

policed relentlessly by other guys’’ (Kimmel, 2008, p. 51). As a result, masculinity is a test, and men

always need to prove themselves (Kimmel, 2008). If masculinity is ‘‘always up for grabs’’

(Kimmel, 2008, p. 51), then social interactions serve as a forum for displaying masculinity. For

instance, male-bonding rituals in boarding schools ‘‘involved enactments of macho bravado, such

as hot-dog-eating competitions and midnight dodgeball’’ (Gaztambide-Fernandez, 2009, p. 187).

Male status is won through ‘‘being cool and demonstrating power’’ (Chase, 2008, p. 92).

Psychologists (e.g., Pollack, 1998) suggest that boys are taught from an early age to suppress their

emotions and avoid showing vulnerability, with the result that boys feel effeminate not only for

expressing their emotions, but also just for feeling them (Kimmel, 2008, p. 53).

Yet for both women and men, emotional support seems to produce more satisfying

relationships; in a sample of undergraduates, because of less intimacy and more competitiveness

in their relationships, men were less happy than women with their friendships (Singleton and

Vacca, 2007).3 More generally, scholars argue that competition in friendships create male

inhibitions about intimacy and self-disclosure, traits that are strong predictors of friendship

satisfaction for both men and women (Jones, 1991). Taken together this research demonstrates

that although men value relationships as much as women, they are somewhat hampered in their

ability to form satisfying friendships because of norms that denigrate the kind of intimacy that is

typically a perquisite of these sorts of relationships.

2.2. Theorizing network change in higher education

I have argued that network know-how influences success at tie formation and that gender

shapes network know-how. I now contextualize these processes in higher education. In particular,

D. Kane / Poetics 39 (2011) 266–289 269

3 There is a perception that women seek out small, interconnected (i.e., dense) groups of friends, but the literature has

not borne this out (Burt, 1998). Using General Social Survey data, Marsden (1987) found no gender differences in

density; nor does Moore (1990) using the same data. Similarly, Burt (1998) found no tendency among female executives

to build dense networks and a significant tendency to build larger networks (which typically have lower density). More

recently, Lewis et al. (2008) actually found that women had less dense networks than men of Facebook friends. The

impression that women have denser networks may derive from earlier findings that women had networks more dominated

by kin (e.g., Marsden, 1987); however, Moore (1990) found that even differences in proportion of kin disappear (or are

significantly reduced) when variables related to family, age, and employment are controlled. Most relevant to this study,

Chase (2008, p. 40) found at an elite boarding school that men and women equally desired both autonomy as well as

bonds and connections.

I argue that gender differences in tie formation are part of a larger stratification that takes place at

many universities. More specifically, I suggest that elite university culture promote norms that

reflect the diverse, sparse (not-dense) network structure closest to that of the adult upper classes

(Goldstein and Warren, 2000; Marsden, 1987). This happens in three ways: first, these

universities feature a high concentration of students with an achievement ethos that prioritizes

individual accomplishment over social ties; the high concentration of these students transforms

this ethos into a norm of the environment, altering the relationship expectations and possibilities

for all students. At the same time, elite universities lack the mechanisms employed by other kinds

of colleges that would facilitate dense networks (or ‘‘community’’ to use the language of those

schools). Finally, diversity is a clear institutional priority, which entails celebrating difference

rather than emphasizing the sameness that is often the basis of dense-network solidarity. The

concentration of achievement-oriented students, the emphasis on diversity, and the lack of

density-promoting mechanisms create in the elite university environment a sparser (i.e., less

dense) network structure than what most students will have experienced previously.

Students who enter with relatively sparse networks or those with the network know-how to

adapt to this kind of network structure should be best positioned to master tie formation, a key

concern of entering freshmen (Clydesdale, 2007), thus securing an easier transition to college.

Students without this network know-how or the ability to acquire it quickly will likely encounter

difficulties in forming new ties. Therefore, we would expect that those who enter without this

structure would be at risk for a difficult transition.

2.2.1. The reproduction of network advantage

The structure of students’ social networks influences college outcomes. In particular, while the

benefits of high-density networks for social support are well documented (e.g., Marsden, 1987),

low-density networks appear to convey more benefits in college (Derryberry and Thomas, 2000;

Kilgannon and Dary, 1992; Pascarella and Terenzini, 2005; Thomas, 2000). In fact, participation

in the best-known college-level dense network – the fraternity (an all-male social organization

that is often economically and racially homogeneous) – is associated with lower achievement

(Arum and Roksa, 2011; Sax, 2008).

For instance, in their review of three decades of research on higher education, Pascarella and

Terenzini (2005) underscore the numerous studies that point to the positive moral and cognitive

benefits for students with diverse, low-density networks. Similarly, low-density networks are

associated with greater exposure to new ideas, values, and experiences, as well as more

‘‘ethically oriented’’ behavior (Derryberry and Thomas, 2000; Kilgannon and Dary, 1992). In

addition, in one of the unusual studies of higher education that uses calculated network measures,

Thomas (2000) found that students who are heavily invested in a single peer group (i.e., high

density) attain a lower GPA and have less social and academic integration. Taken together, these

studies point to the seemingly intrinsic value of sparse (low-density), diverse networks for

intellectual and moral development.

However, it can hardly be coincidental that such a beneficial network structure is also the one

most commonly held by the upper social classes (Campbell et al., 1986; Fischer, 1982; Goldstein

and Warren, 2000) and reflects the cultural priorities of universities themselves. More

specifically, elite universities privilege diversity and do not privilege density. Elite universities

clearly favor diversity and have the resources to make this priority a reality; a diverse student

body has become a point of pride and competition among these institutions (Stevens, 2007).

Diversity is a distinguishing feature of the elite university environment, which is dominated by

choice for potential friends (at least in manifest availability); people from an astonishing array of

D. Kane / Poetics 39 (2011) 266–289270

racial, national, and religious backgrounds are available for interaction. This campus-wide

diversity is reflected in networks: as school diversity increases, friendship diversity also increases

(Fischer, 2008).

Similarly, as elite high schools (e.g., boarding schools, prep schools) mimic those higher

education institutions their students aspire to attend, they go to great lengths to diversify their

faculty and student body, sometimes sacrificing valued school hierarchies to accomplish this goal

(Chase, 2008; Khan, 2011). Khan (2011) describes how St. Paul’s School, a prestigious New

England boarding school, eliminated faculty rank-based seating assignments in chapel when the

administration realized that all of the nonwhite faculty (who had been hired recently) would be

seated in the back. This change did not reflect a more general move away from hierarchy; once a

sufficient number of nonwhite faculty gained seniority and could move forward in their seating,

St. Paul’s reinstituted the seating hierarchy (Khan, 2011, p. 72).

The departure and eventual return to hierarchical seating reveals some of the necessary trade-

offs between diversity and density. Temporarily suspended to accommodate diversity,

hierarchical seating maintains status groups that play an important role in ensuring that

everyone has a role in the community (Gaztambide-Fernandez, 2009; Khan, 2011). Gaztambide-

Fernandez (2009) similarly describes the seating generated by students themselves in an elite

high school as they chose hierarchical arrangements, ranging from the ‘‘popular/attractive

student’’ section to the much more nuanced ‘‘second-second tier.’’ These distinctions can only be

made in a social environment with high consensus—that is, where everyone knows one another

and has a reputation: a dense network. Elite universities typically do not have such nuanced

seating at any event; indeed, there are few shared rituals that provide such an opportunity.

In addition to diversity, an ethos of individual achievement leaves little room for students to

invest in time-consuming dense networks. A counselor at an elite boarding school observed that

the pressures toward accomplishment mean that ‘‘prep schools facilitate giving up childhood’’

(Chase, 2008, p. 27). According to this counselor ‘‘every moment is committed until about

10 p.m.,’’ and only then do these students engage in ‘‘their own kid culture’’ (i.e., a group-

oriented social life) (Chase, 2008, p. 26).

Their packed schedules reveal their priorities; Chase (2008, p. 43) observes that ‘‘even though

both boys and girls publicly focus on Other, they covertly value self more, with its consequences

of academic, athletic, and career and financial success.’’ Writing about ‘‘overachieving’’ high

school students, Robbins (2006) profiles a student whose mother seldom allows him to spend any

time outside of school with friends; instead he must devote his time to studying and

extracurricular activities to gain admission to Harvard. Another student complains that because

of his workload, he ‘‘can’t do the things [he] used to do, like hang out with friends or go to

parties’’ (Robbins, 2006, p. 263). Robbins (2006, p. 262) observes that ‘‘many high school

students today are like robots, programmed to go mechanically through the motions of academia

in a desperate attempt to craft a competitive college resume.’’

While Robbins (2006) does not detail what she means by ‘‘many,’’ even if this number is

small, these students can disproportionately influence the higher education system by setting a

standard for others. While it may be a minority of students who adopt this singular focus on

individual achievement, because of their over-representation at elite universities, their behavior

nonetheless becomes the measure for many others. For instance, high-achieving students are

sometimes known as ‘‘curve busters’’ because their high scores can influence how exams are

graded.

Although students value social ties, their commitment to professional advancement places a

premium on individual achievement, which can mean making social ties a lower priority. As an

D. Kane / Poetics 39 (2011) 266–289 271

example of the disincentives for dense networks, imagine a student who tells her professor that

she cannot take the final exam because she stayed up the previous night comforting a friend

whose parents recently divorced, or a student who cannot complete a homework assignment

because he spent the previous evening with a group of friends. These explanations are generally

not accepted by faculty, who expect students to manage their personal lives in a way that allows

them to complete their work on time; using this explanation, a professor voices the university’s

priority of individual achievement. There are academic penalties for not conforming to the

university’s priority; if faculty do not accept these reasons from a student, his or her grade is at

risk. If the student exhibits a pattern of prioritizing social ties – a priority necessary for a dense

network – his or her academic credential may be on the line.

Upper-class students are more likely to be initiated early into the achievement ethos that can

compete with social bonds, making the transition to college less of a rupture. In fact, students may

have a particular investment in this ethos, as ethnographies of elite high schools reveal that their

belief that their hard work entitles them to elite membership, obscuring the multitude of

structural factors that helped gain them admission in the first place (Gaztambide-Fernandez,

2009; Khan, 2011).

Finally, while diversity and an ethos of individual achievement can serve to undermine dense

networks, elite universities lack the mechanisms that foster high density at other kinds of schools.

Community – the layperson’s term for dense networks – is by definition a collective good, and as

such is subject to free riding (Iannacone, 1992). As a result, many collectives, including colleges

that assign a relatively higher priority to community, institute mechanisms to ensure high

density—mechanisms that elite universities do not implement. For instance, at some military and

religious colleges, the value of community is realized through compulsory participation in a

dense network focused on a shared culture.4 Reflecting their diversity priority, elite universities

embrace multiculturalism, a concept that can be at odds with the idea of a shared culture.

Perhaps most significantly, at some religious colleges, monitoring of participation in school

rituals reminds students that the community (i.e., a dense network) takes priority over individual

preferences. For instance, students at Lubbock Christian University (LCU) are required to attend

chapel services four days during the week, and they must document their participation by

scanning their student ID card as they leave the chapel (Lubbock Christian University, 2010-

2011). Pages 25 through 27 of the student manual detail both exemptions and the disciplinary

actions that correspond to the number of chapel absences, culminating in a withdrawal of

institutional funding and prohibition of pledging campus clubs and participation in other campus

activities. Because LCU places a high value on community, administrators have put into place

mechanisms that ensure individuals’ full participation. These mechanisms are generally lacking

at an elite university, at least at a university-wide level. (Greek organizations and sports teams are

an exception to this rule and arguably draw some of their power over students from the lack of

highly ritualized alternatives.)

For most students, this kind of compulsory participation in a dense network has more in

common with high school than with college life. In typical high schools, students may spend

entire days together with the same group of friends, not to mention evenings for sport practices.

By contrast, although most elite college students have roommates, few students share the same

D. Kane / Poetics 39 (2011) 266–289272

4 For instance, David Lipsky’s (2003) discussion of the military college West Point reveals a high degree of monitoring

of individual behavior; and ‘‘the terms of success. . .are belonging and not belonging. The official word for expulsion is

separation’’ (Lipsky, 2003, p. 35).

roster of classes and activities that solidify the group bonds (positive or negative) experienced

more commonly in high school (Nathan, 2005). Therefore, the high school graduate has

experienced a ritually intense, relatively homogeneous social life (Stearns et al., 2009). This kind

of dense-network social life, while not necessarily pleasant, is often stable and predictable. It is

also what will feel most familiar to many students.

Most elite university students need to give up this familiarity in the transition to college. The

sparse, diverse network favored by the university is most often found among the adult upper class

(Collins, 1994; Goldstein and Warren, 2003). In other words, while upper-class students may

have an advantage in exposure to this network structure, students themselves are likely to have

denser networks than their parents (Chase, 2008; Robbins, 2006). Therefore, the transition to the

university network environment has the potential to be stressful for all incoming students. The

question for stratification research, then, is who is most at risk for a difficult transition?

In order to address this question, I make two hypotheses based on the disjuncture between high

school and college network environments and cultural-based gender norms related to network

know-how.

H1. After one year at an elite university, students’ networks will evolve to mimic the university’s

network environment; that is, they will become sparser and more diverse.

H2. As compared to men from sparse networks and all women, men from dense networks will

experience the most difficult transition.

3. Data and methods

3.1. Total sample

This article reports on data from a larger study of social network and attitude change and

includes two waves of a survey from an initial sample of 512 incoming students at an elite

university, which I refer to as ‘‘Ivy.’’ The sample responding one year later dropped to 291 for a

57 percent retention rate. I conducted an analysis of sample attrition (reported in Appendix A) to

determine whether background, demographic, or network characteristics were associated with

non-participation in the second wave. I found that older students and Asian students were less

likely to complete a Time 2 survey (5 percent less and 46 percent less, respectively). I also found

that students with higher religious diversity at Time 1 were more likely to respond at Time 2—

about 8 percent more likely for a 0.1 unit increase in religious diversity. (See below for a

discussion of religious diversity.) This difference is therefore of a small magnitude; moreover, it

is unrelated to density so is unlikely to affect the findings of the article. Overall, there were few

patterns of sample attrition based on background, demographic, or network variables, and those

that emerge do not seem to bear on the conclusions I draw.

Beyond these measured variables, I can imagine two major paths toward non-participation in

the second wave. First, many students might simply have been too busy to reply, particularly in

their second year of school when students might feel pressure to ‘‘jump in’’ to their workload and

have already committed themselves to a full roster of extra-curricular activities.

Perhaps a more serious problem could be that those who experienced the worst adjustment to

Ivy would be reluctant to participate in a study of their experience, especially if, as I argue later,

some students experience a difficult transition as a personal failure. If this were the case, then this

study would be less likely to have the most poorly adjusted students and hence minimize the

impact of a poor transition. In terms of the theoretical framing, those students with the least

D. Kane / Poetics 39 (2011) 266–289 273

network know-how may not be included in this study. In that situation, this study would reveal

illustrative cases of poor transition (of which there are several examples) but would not capture

the full magnitude of the problem. In that scenario, this article would serve as a departure point

for a larger investigation.

Students completed the first wave of the survey within two weeks of their arrival. The second

wave was conducted as a web survey one year after the initial data collection. (Apart from the

web format, the surveys were the same.) For the sample completing both waves of the survey,

47% were female; about half (52%) of respondents were American citizens, while 26% came

from Asia, 17% came from Europe, and 5% came from other countries. About half (49%) of this

sample would be characterized as entering Ivy with high density; that is, the density of their

networks was greater than the mean for the sample (see Table 1, columns 1 and 2). The sample

was relatively affluent, as measured by father’s education: fathers’ mean education level was a

college degree. For American respondents who completed both waves of the survey, 53% were

white, 26% were Asian, 7% were Hispanic, 2% were African American, and 12% identified as

other.

3.2. E-surveys and interviews

The article also reports on e-survey data from a subset of 108 of these students and interview

data from a further subset of 43 of these students, chosen to maximize variation in terms of

gender, race, nationality, and network type.5 Undergraduates were recruited at a campus

dormitory that housed 25% of the freshman class. Sixty percent of incoming freshmen in the

targeted dormitory completed the survey.

I used two instruments to capture the experience of transition as it was happening, a mid-year

‘‘e-survey,’’ and interviews conducted with a further subset of these students. E-surveys asked

students how their Ivy friends compared with friends from home; how they met their friends;

what campus activities they were involved in; and what changes they had noticed in themselves

D. Kane / Poetics 39 (2011) 266–289274

Table 1

Descriptive statistics.

T1–T2 total sample (N = 291) E-survey (N = 122) Interviews (N = 43)

Mean SD Mean SD Mean SD

Female 0.47 0.50 0.42 0.49 0.50 0.51

Parent education 3.49 0.75 3.44 0.75 3.34 0.85

Origin: US 0.52 0.50 0.34 0.47 0.69 0.47

Origin: Asia 0.26 0.44 0.27 0.45 0.03 0.16

Origin: Europe 0.17 0.38 0.34 0.48 0.26 0.44

Origin: Other 0.05 0.21 0.05 0.22 0.03 0.16

Pr. high density T1 0.49 0.50 0.52 0.50 0.46 0.51

Notes: Parent education ranges from 1 = ‘‘high school diploma’’ to 4 = ‘‘some graduate education.’’ ‘‘Pr. high density T1’’

refers to the proportion of respondents with a density greater than the mean (0.68) at time 1.

5 As an inducement to participate, for completion of the first wave of the survey, each student in the total sample

received $5 and a chance to win $1000 in a raffle. For completion of the second wave, each student received $10 and a

chance to win one of two $500 prizes in a raffle. Among the subsections of students, e-survey respondents received $5 for

their participation, and students received $10 for completing an interview.

since coming to the university. The nested nature of the sampling (in which e-survey respondents

were selected from a larger pool of survey respondents) maximized diversity in responses along

the lines of gender, country of origin, and entering density. Students were contacted one at a time,

and when a student refused, every attempt was made to find another student matching the

refusee’s profile. Overall, 149 students were contacted, and 108 responded, for a 72% response

rate. Forty-two percent of respondents were women. Thirty-four percent were Americans,

twenty-seven percent were from Asia, and thirty-four percent were from Europe. Fifty-two

percent were high-density (see Table 1, columns 3 and 4).

From this pool of 108 students, I also conducted 43 semi-structured interviews lasting from an

hour to an hour and a half in the spring of the participants’ first year. E-survey respondents were

categorized by gender and entering network structure, and interviewees were selected to fill out

each of the four gender-network combinations (that is, by gender and high or low density). Within

these constraints, I also aimed for an interview sample that would reflect the variety of world

regions represented in the study. When a student did not reply to an interview request, two

additional emails were sent requesting an interview at the location of his or her choice. If there

was still no response, I moved to the next name on the list. In these interviews students were asked

about their experiences thus far at the university, their biggest challenge in adjusting to university

life, what they looked for in friends, and their experiences of their current and past friendship

networks (in addition to any topics that students initiated). One student requested not to be tape-

recorded; all other interviews were taped and transcribed. Interview and e-survey data were

manually coded and analyzed. Half of the interviewees were female. Americans were

overrepresented in the interview sample (69 percent); the next biggest group were Europeans (26

percent). Forty-six percent entered with high-density networks (see Table 1, columns 5 and 6).

My interpretive strategy was aimed at understanding how students understood and evaluated

their experience of network change, particularly as these experiences and evaluations related to

gender and density. I therefore coded descriptions of their lives at Ivy (and pre-Ivy) in terms of

explicit evaluations, emotion words, and affect. Finally, I analyzed positive and negative

evaluations by density and gender and within these subgroups looked for recurring themes.

Among men, for instance, the theme of ‘‘competition’’ emerged more often than in interviews

with women, and this theme was almost always associated with a negative assessment.

3.3. International students in this sample

As distinct from most research on higher education, this study included international students.

International students are becoming an increasingly important part of the higher education

landscape, and including data on their transition will help us better understand this understudied

segment of high education.

In the interview data, I therefore looked for themes by world region (Europe, Asia, and United

States, the categories which most respondents fit into). However, I found that students’ responses

about the transition were more patterned by gender than by international status. This finding is

consistent with a pair of research findings: on one hand, international students valued the same

friendship skills as Americans (Xu and Burleson, 2001), but both Eastern and Western societies

offer women more opportunities than men for developing some friendship-promoting skills

(Burleson, 2003).

Taken together, these findings suggest that men from different countries might respond more

similarly to the transition to Ivy than would men and women from the same country. For instance,

in my interviews American men and international men were equally negative about Ivy—even

D. Kane / Poetics 39 (2011) 266–289 275

though they used different reference points.6 By contrast, international women were similar to

American women in their enthusiasm. As an example, a Taiwanese student said she felt

uncomfortable in her courses because there was pressure to speak aloud to the class (something

she was unaccustomed to) and because she sometimes found Americans intimidating in their

straightforwardness. But she was as enthusiastic about the ‘‘freedom’’ she experienced at Ivy as

American women were, and she was not eager to return home. Her remarks were more similar to

the American women who complained of the surveillance by friends and families at home than to

the Asian males who missed the sense of belonging and cohesiveness in their home networks.

This is not to say that there were no differences in perceptions by international status but that

evaluation of the transition was more related to gender. For instance, when international students

made comments about the US (directly or indirectly), these too fell along gender lines, with

women having more positive, optimistic reports of living in the US. The commonalities found by

gender across countries are consistent with the entrenched gender norms found to affect

education in multi-country studies (Charles and Bradley, 2002).

While international students’ perceptions seem to fall along gender lines in a way similar to

Americans’, international students are sometimes more conscious of and hence are more articulate

about the transition. Being double outsiders to American higher education (i.e., as freshmen and as

foreigners), they emerge in some ways as amateur social scientists, and from the standpoint of being

both inside and outside the system, they sometimes capture the dilemmas of transition in a

particularly evocative way. Therefore, international students are important to study within their own

right, because they are an important part of the undergraduate population and yet are understudied,

and because of the particular light they shed on the transition to college more generally.

3.4. Network instrument

Apart from gender, the most important analytical category in this article is network density.

The network survey in this study asked the respondent to list up to six people (‘‘alters’’) and to

complete a profile for each name listed, including the alter’s gender, race, religion, nationality,

and source of tie (‘‘From where do you know this person?’’). The respondent then completed a

network matrix to indicate whether each alter knew the other alters listed (see Appendix B).

Network density is typically defined as the proportion of possible ties that are actually present

(Scott, 2000, p. 32).7 (The calculation of network measures can be found in Appendix C.) This

definition says nothing about the kinds of ties that are measured. Network research thus far has

emphasized close ties, perhaps at the expense of other types of network ties that are salient for

respondents. The General Social Survey (GSS) Network Module, for instance, asked respondents

to list people with whom they discussed important matters. It is acknowledged that this prompt

elicits ‘‘reasonably strong ties, with prominent representation of kin among those cited’’

(Marsden, 1987, p. 123). In other words, this type of network prompt limits respondents to

reflecting on only portions of their social lives.

D. Kane / Poetics 39 (2011) 266–289276

6 For instance, while some Asian men attributed Ivy’s competitiveness to American culture, a Midwestern student

attributed the same characteristic to East Coast culture, and a student from the South attributed it to Northern culture.

Significantly, almost all of these students’ interactions were confined to the Ivy campus.7 Density ranges from 0 (no ties, or ‘‘alters,’’ know one another) to 1 (all alters know one another). Note that the link

between alters and the focal person (‘‘ego’’) are not counted; density is a measure of interconnection within a network

rather than the number of ties the focal person has. Low-density networks are often referred to as ‘‘sparse’’—that is, a

relatively low proportion of alters in the network know one another. Sparse, therefore, is the opposite of dense.

The current study follows the path of other work that seeks to gain a broader picture of

respondents’ networks rather than focusing exclusively on individuals with whom one was close

enough to discuss important matters (e.g., Burt, 1998; Lewis et al., 2008; Wimmer and Lewis,

2010). To that end, each respondent was asked to list up to two people in three specific categories:

(1) people with whom they studied or worked; (2) people with whom they spent leisure time; and

(3) people with whom they discussed important matters. Students were describing their networks

at that moment in time, so the first wave, administered during freshman orientation and before

classes began, included ties from their home network. The respondent was then asked whether

each of these people knew the other people the respondent listed. (The paper survey used in Wave

1 can be found in Appendix B.) The same network instrument was used at both points in time; the

second wave, however, was conducted on the web rather than on paper.

4. Results and discussion

4.1. Network change at Ivy

I hypothesized that being in the sparse, diverse network environment of the elite university

would facilitate the creation of sparse, diverse networks. Table 2 shows the average score on five

network characteristics at the time of students’ arrival at the university and one year later.

As predicted, a decrease in density and an increase in three measures of diversity indicate that

students’social networks became sparser and more diverse. Average density decreased by.12; racial

diversity and world region diversity both increased by .16. Average religious diversity increased by

0.13. Gender diversity remained the same, with only a 0.02 change between Time 1 and Time 2.

Curiously, student network size seems to shrink (from 5.01 to 4.65). Although I made no specific

hypotheses about network size, the decrease in network size runs contrary to expectations about

college social life. The difference, which is relatively small in magnitude, could be due to the fact

that Time 1 measures students’ well-established social networks; at Time 2 students are still in the

process of network formation. Overall, in the transition between high school and college, students’

networks seemed to evolve in response to the incentive structure of the university.

Table 3 shows network change by gender. The decrease in density and the increase in diversity

are extremely close for men and women. For instance, men’s average density dropped by .10

while women’s dropped by .13; men’s religious diversity increased by .16 while women’s

increased by .10, and both women and men experience a .17 increase in racial diversity. The

biggest difference for men and women was that men’s network size shrunk by .46 (about one third

D. Kane / Poetics 39 (2011) 266–289 277

Table 2

Network change in the total sample.

N Time 1 Time 2 Sig. diff.

Mean SD Mean SD

Density 267 0.68 0.27 0.56 0.26 ***

Ethnic heterogeneity 275 0.19 0.26 0.35 0.29 ***

Gender heterogeneity 280 0.70 0.33 0.72 0.34

Regional heterogeneity 265 0.17 0.26 0.33 0.32 ***

Religious heterogeneity 250 0.40 0.31 0.53 0.27 ***

Network size 291 5.01 1.32 4.65 1.41 ***

Notes: ***p � 0.001. Significance based on two-tailed matched-pair t-tests.

of a standard deviation) while women’s shrunk by only half as much: .24. However, even for

network size, there is no significant difference between men and women at Time 1 or 2. Overall,

we see no significant gender differences in the amount of network change. How men and women

respond to this network change, on the other hand, diverges considerably.

4.2. Gender differences in the transition experience

Interview findings reveal that the combination of past network experience and gendered

cultural mandates produced a particularly difficult transition for men from dense networks.8

Their difficulties seem to have stemmed from trying to reproduce dense networks at Ivy. This

strategy had two problems: first, the high level of commitment required to sustain these groups

was often not shared by Ivy students, who had other priorities. Moreover, this investment of time

did not always pay off: by relying on a group for ties, they sometimes found that they did not

actually like group members.

4.2.1. Women as ‘‘tie entrepreneurs’’

While past network experience alone might lead us to predict that dense-network women

would experience similar trials, instead women of both dense and sparse network structures were

more likely to act as ‘‘tie entrepreneurs,’’ actively cultivating friendships. Developing ties was

not necessarily easier for women, but they were more likely to persist at initiating friendship than

were men. They sometimes did this by reframing undesirable qualities in potential ties. For

instance, in the e-survey, Maral (Time 1 network density = .13),9 a Turkish student, began with

D. Kane / Poetics 39 (2011) 266–289278

Table 3

Network differences by gender (total sample).

Male Female

N Mean SD N Mean SD

Panel A: Time 1

Density 143 0.65 0.28 131 0.71 0.26

Ethnic heterogeneity 147 0.18 0.27 132 0.20 0.26

Gender heterogeneity 149 0.71 0.34 134 0.69 0.32

Regional heterogeneity 144 0.17 0.27 125 0.18 0.27

Religious heterogeneity 134 0.38 0.31 119 0.42 0.31

Network size 154 5.01 1.40 136 5.00 1.22

Panel B: Time 2

Density 150 0.55 0.28 132 0.58 0.24

Ethnic heterogeneity 152 0.35 0.30 133 0.37 0.30

Gender heterogeneity 152 0.71 0.36 134 0.72 0.32

Regional heterogeneity 152 0.35 0.32 132 0.32 0.32

Religious heterogeneity 152 0.54 0.27 133 0.52 0.28

Network size 154 4.56 1.47 136 4.76 1.35

Notes: None of the differences are statistically significant based on two-tailed t-tests.

8 A reviewer suggested that gender differences in the transition experience could actually be driven by social class

differences; however, I found no significant association between gender and social class (i.e., father’s level of education).9 All density scores in this section of the article refer to Time 1 density. Henceforward I will refer to this simply as

‘‘density.’’

the complaint that ‘‘at Ivy relationships are very superficial.’’ As distinct from men’s reactions to the

perceived shortcomings of Ivy, however, Maral believed that ultimately Ivy students were well

meaning. After remarking on students’ superficiality in an interview, she qualified her explanation

to say, ‘‘People are afraid to talk to each other. I realized that they want to talk but just because it is

not a social norm they avoid doing it. But once you start talking to them, they are nice.’’

Here, Maral displays a high level of network know-how in identifying the ‘‘rules’’ of

friendship in the new situation and using them to understand others’ behavior. In this case, Maral

goes beyond her immediate reaction (i.e., that people are willing to have only superficial

friendships) to give more of a meta-explanation that addresses the new social context she finds

herself in (i.e., people are locked into social norms that limit their ability to achieve the deeper

relationships they actually seek).

Maral also reframed the problem to give herself a greater sense of control and ultimately to

achieve the desired end. Later in an interview she described how she became closer to her current

friends:

I try to teach them something from my culture. For example that giving is more important

than taking. I show them how I can get satisfaction by doing or giving to them. . .I help them

with [their] papers. First I noticed that one of them was just using me. My roommate told me

that she definitely thinks I am stupid, and she [the other friend] is just using me. But I was

patient, and she realized that I am a real friend, and [now] we are very close to each other.

In this description, Maral encountered the same selfishness that men from dense networks

describe in their responses, as we will see shortly. Maral reacts differently, however, and reframes

the selfish friend as a child in need of moral guidance. Maral then offers herself as an example to the

other woman of how friends ought to treat one another. In this way, Maral assumes a position of

power rather than one of victimhood and accomplishes the desired end of having a closer

relationship. If the relationship was hard-won, then it is only that much more of an accomplishment.

Although Maral was an international student, her heavy investment in single ties was typical

of both foreign and American women. Probably because of the emotion work they expended,

many women actively cultivated only a small number of intimate ties in which they then invested

themselves heavily. Ellen, an American, was fairly typical of many women respondents when she

wrote on her e-survey, ‘‘I have two best friends here that I spend all of my time with.’’

Women across network types were more likely to pursue this strategy of tie formation of

cultivating individual ties and investing heavily in them. A measure of the intensity of these

relationships is revealed in the recurring metaphor of a love relationship that appeared in

women’s but not men’s descriptions of their friendships. In describing an old friend of hers, Edith

(density = 1.0) commented in an interview,

We’ve changed a lot since we’ve known each other. . .but we’re still really good friends

now, and when I talk to her we still can understand each other, or at least try to understand

each other. . .just the fact that we can grow and become different but still be able to relate to

each other, and I mean, to me, that’s the mark of a real friend. . .I guess like marriage – you

know, the fact that you can grow together.

Similarly, a French student (density = .73) remarked, ‘‘Ah, the real friends. . .it’s like love, you

know? From the first moment you meet them, you say okay, this person was my friend, I really

like this person. . .They are the ones I call my real friends.’’

Women from a variety of network structures spoke of the heavy investment they made in

individual friendships, often invoking familial or romantic metaphors. In fact, women in this

D. Kane / Poetics 39 (2011) 266–289 279

sample seem to apply many of the standards of choosing romantic partners to choosing friends.

Perhaps as a result, women of both network structures described success stories in gaining friends.

4.2.2. Men’s friendships

In contrast to women’s consciously cultivated individual ties, men from dense networks were

more likely to join a group and expect friendship to ‘‘come. . .in the normal course of peer group

relationships’’ (Gans, 1982, p. 92). Consistent with some research on masculinity (Kimmel, 2008;

Messner, 1990), large groups, especially sports teams, provided an important basis for friendship

and intimacy. (Schultz and Breiger (2010) identify sports-talk as a form of ‘‘weak culture’’ that can

help individuals from a variety of backgrounds bond.) For instance, John (density = 1.0) noted:

Because I play basketball I spend most time with the other basketball players. In high

school it was pretty much the same. My closest friends were on the basketball

team. . .[Here] I see them pretty much every day, and even when we don’t have practice one

day we usually hang out a lot.

Although friends might come from group activities, establishing friendships could not be a

goal as it is for many women, who actively cultivated ties. For instance, Gavin (density = .60),

put it this way:

I think if you’re enjoying a sport, you go and do it knowing that there’s gonna be a group of

kids there, you know, so that you get the friends from the sports. If you try and sort of

reverse it. . .like I go after the sport, and I end up with the sport and the friends. . .I don’t

think if I go in looking for friends, I necessarily come out with sport and friends, you know,

I might just come out with friends.

As distinct from some of the comments made by women, many of these men did not see

cultivating friendships as an end in itself, as exemplified by Gavin’s comment that without

focusing on a sport, he might ‘‘just’’ get friends.

Perhaps as a result of relying on group activities to accumulate friends, men from dense

networks were less satisfied than other groups with their ties in college. A shortcoming of the

‘‘group strategy’’ is demonstrated by men who could not play sports in college; their nostalgia

about the dense network of friends they had in high school reveals their disappointment at failing

to establish such groups at Ivy. Aaron (density = 1.0) explained:

I miss the group. And I think that’s shown by the fact that as soon as I go home I fall back

into the group. Like, I was home over the weekend, and I called each one of my close

friends again. We wanted to all get together. . .yeah, I miss that. . .And so here, I’ve been

here for nine months, and I haven’t found like even five or six people that I’m that close

with, so I’ve got like one or two. . .yeah, I miss that about home.10

D. Kane / Poetics 39 (2011) 266–289280

10 Although this article does not include data from interviews with graduate students because they made the transition to

higher education prior to this study, it is worth noting that comments from graduate men from dense networks were

consistent with undergraduate men of the same network structure. Bill, a male American graduate student (density = .80)

talked about his disappointment when he learned that, as a graduate student, he was ineligible to play on a varsity team;

moreover, intramural sports were poorly organized, and no one had returned his calls. He explained, ‘‘I feel just like the

physical outlets, like sports, are really intertwined with my social network, and I feel like that has definitely been

something that has hurt my ability to make friends here. . .because there’s so much camaraderie that comes out of playing

sports. I just feel like that’s missing. . .I definitely feel like that’s been a big factor in not getting as close with people. . .’’

Some men were so anxious to recreate the dense networks of high school that they would

actually stay in groups with people they did not like. Earlier in the interview, I had asked Alistair,

a student from Hong Kong (density = 1.0), what sort of people he found attractive and what sort

of people he considered repellent. He responded that he particularly disliked ‘‘fake’’ people,

‘‘people who want to be popular, for example, and not be themselves. . .people who need an

audience, not friends. . .’’ Later, when I asked whether he would be friends with his Ivy friends if

he had met them at home, he sighed and said,

One of my Hong Kong friends is actually quite a fake person. [He laughs.] I wouldn’t be

friends with him, but um, but it’s like. . .um. . .[pause] I don’t know, it’s just that. . .because

you’re from Hong Kong, therefore, you know, we chill together sometimes. I mean, it just

becomes a reason; like, it wouldn’t be a reason when we were back in Hong Kong. I would just

say, ‘‘okay, I’m not going to be your friend because, you know, you’re fake.’’ It wasn’t an issue,

like being Hong Kongese in Hong Kong. It’s very interesting – I never thought about this

before.

Alistair seeks a safety zone with a big group of similar others—even if he does not

particularly like them. This dynamic is consistent with Kimmel’s (2008) observation that

even when young men disapprove of friends’ behavior, they still tend to remain loyal to the

group. Kimmel was writing about more extreme behavior – like carousing or even rape – but

Alistair’s comment demonstrates how some young men commit themselves to their group

even at some cost to themselves—and without thinking about it. In contrast to Maral, who

contemplated her friendships enough to develop a theory of Ivy’s rules of interaction,

Alistair’s lack of reflection about the dynamics in his group of friends betrayed a much lower

level of network know-how.

This lack of know-how seems to have had a high cost. Alistair was the most distraught student I

interviewed. For most of the interview his affect was very flat, but occasionally he seemed on the

verge of tears. While he was the only student who came close to crying, his comment may provide

some insight into a struggle other men in this sample faced. When talking about his loneliness, he

said,

It’s not Ivy’s fault; it’s totally my fault. . .The environment is fine. . .and it’s just, like, me

who’s closing myself, and there are abundant opportunities for basically whatever things

you want to do. . .And it’s just me who’s, like, saying no, I’m not going to do that. . .Ivy is

such a dynamic place, it’s not boring like [the British town where he attended boarding

school] – we had to turn to the friendships there. . .but Ivy is – Ivy and the world – is so

dynamic, so unconstrained, that probably wouldn’t happen.

Alistair clearly blames himself for his difficult transition to Ivy, and it is possible that some of

the disappointment in Ivy voiced by men in this sample also reflected a sense of personal failure.

By contrast, Maral had used her network know-how to transform initial difficulties in tie

formation into a source of personal accomplishment. More generally, when we compare the

comments of these men to those of some of the women in this sample, we see a bitter irony: men

are trapped by cultural mandates that denigrate the kind of explicit cultivation of ties that women

may use effectively to ease their adjustment to college. Stated in another way, men may blame

themselves for a failure to gain friends but perceive themselves as powerless to do anything

about it.

D. Kane / Poetics 39 (2011) 266–289 281

4.2.3. ‘‘Goal-oriented and somehow selfish:’’ the impact of academic competitiveness on

dense-network men’s tie formation

Some men from dense networks found that they could not rely on groups for acquiring

satisfying friendships, but Ivy’s focus on academic ambition created what they perceived to be a

hostile environment for forming friends on their own. While students of both sexes from dense

networks remarked on the greater academic ability of their new peers, men were much more

likely to be anxious about this difference and concerned about Ivy’s ‘‘competitiveness.’’ (Eight

men as compared to two women raised this concern in interviews.)

For these men, the celebration of individual achievement that underlay the school’s

competitive culture undercut the very ties they sought to forge. On the e-survey, an American

undergraduate (density = 1.0) complained, ‘‘A lot of people are not accessible outside of projects

and school activities. At home these people would have been at your fingertips.’’ Similarly, an

Indian student (density = 1.0) remarked that at home his friends would play together even when

they had work to do, but here there was only work. In comparing Ivy friends to friends from

home, it was overwhelmingly men from dense networks who described their friends at Ivy as

more ambitious—and more selfish. (Of the ten students who raised this issue, nine were men.)

For dense-network men, academic competitiveness could elide into a more general distrust of

Ivy peers. For instance, one Eastern European (density = 1.0) explained why he spent most of his

time with his roommate, whom he knew from home:

About the other people I met here I do not spend much time with them. They are different

because most of them are goal-oriented and somehow selfish. . .I do not try to communicate

with new people. I am not sure whom I should trust as a friend. . .

Here, ‘‘goal-oriented’’ and ‘‘selfish’’ go hand-in-hand: both indicate a greater priority placed

on the individual rather than – perhaps even at the expense of – relations with others.

Perhaps most disturbing, the pressures induced by academic competitiveness provoked in a

few men from dense networks the fear of losing their own sense of self. As one undergraduate put

it on the e-survey, ‘‘It’s hard not to get swept away in the busy lifestyle engaged in by everyone

around you.’’ Similarly, a European-American man (density = 1.0) made implicit yet powerful

statements about Ivy’s culture:

I feel like I’m a meaner person here, actually. I don’t know, I feel like I was a lot nicer

in high school, but here I feel like if I’m nice people are gonna take advantage of me,

so I, I don’t know what the word is, I feel more aggressive towards people here. I don’t

like that.

Each of these men described feelings of mistrust and threat at Ivy, and the mistrust seemed to

derive from the inability to rely on others, a trait that is usually associated with dense networks.11

D. Kane / Poetics 39 (2011) 266–289282

11 Again, comments from graduate students with dense networks echoed those of undergraduates, sometimes dramati-

cally. For instance, an American (density = 1.0) commented that ‘‘I feel a lot more uptight and a lot more defensive [at

Ivy]. . .I just haven’t been able to let anybody too close.’’ Part of his reaction may be due to the response of other students

in his program when his car was broken into: ‘‘I got here in August and by November my car was broken into twice, and I

mean, I felt like every time I needed something, and I needed to turn to somebody to ask for something, and I figured it

was pretty run of the mill, they acted like it was the biggest thing they’d ever had to do for somebody. And it was

incredible to me because it was not something that I was used to. . .it just feels less friendly here, and that’s something I’ve

just adapted to. . .I think I’ve grown pretty bitter, to be honest with you.’’

The Eastern European student dealt with Ivy by simply withdrawing to a social comfort zone of

students from his home country. The American men did not or could not pursue a similar

strategy, and they described the cost of engaging with their peers in terms of their becoming

‘‘meaner,’’ ‘‘aggressive,’’ and ‘‘bitter.’’ Taken together, these men’s reactions bring to mind

McLean’s (2007, p. 229) contention that networking is a game that remakes the players

participating in it.

In contrast to the embittered reports of men from dense networks, men from sparse networks

described a much easier transition. Men from sparse networks, like women across network

structures, were more likely to talk about the opportunities at Ivy. An American undergraduate

male from a sparse network (density = .13) wrote on an e-survey that ‘‘it’s cool getting to know

people from different backgrounds.’’ Another American undergraduate from a sparse network

(density = .53) said,

. . .at Ivy my friends are much more diverse, and have brought out many different interests

in me. I would never have gone to art houses or museums with my friends on a weekend at

home but at Ivy I do.

As distinct from the dense-network men in this study, the comparison to the home network is

flattering to Ivy students.

5. Conclusion

In the transition to college, men from dense networks emerged as doubly disadvantaged:

first by their network structure, which made the transition especially jarring, and then by

gender norms that cut across network structure and left men ill-equipped to form friendships

outside of a big group. Women in this sample followed a gendered cultural mandate that

encourages them to actively cultivate and manage their ties. Men from dense networks also

conformed to a gendered cultural mandate, one that denigrates men’s explicit cultivation of

ties and instead encourages reliance on group participation, even if they do not particularly

like the group. This strategy often did not work in Ivy’s high-pressure, individualist-oriented

environment, and men from dense networks were more likely to feel unsatisfied with their

ties.

Ivy’s emphasis on individual achievement also fostered a competitive environment that made

many of these men suspicious of their peers, further undermining their ability to cultivate the kind

of friendships that might serve as a buffer to the difficult transition. The same gender norm that

discourages men’s conscious cultivation of friendships implicitly discourages self-reflection

about their social situations, leading some men to blame themselves or feel helpless to improve

their situation.

These findings suggest a new avenue of inquiry for men’s lagging completion of a college

education (Buchmann and DiPrete, 2006), one that emphasizes culture and networks. More

specifically, future analyses might move beyond social class analyses of the gender gap in college

completion to consider social network factors. The difficulty in adjustment that men from dense

networks experience could create problems that ultimately lead to some of them to leave college

before graduating. Moreover, given that dense networks are more often found among the

working-class and people of color (who are also disproportionately likely to drop out of college

(Leonhardt, 2005), this study suggests that exploring the role of social networks could be a useful

direction in educational stratification research.

D. Kane / Poetics 39 (2011) 266–289 283

These findings also extend Bourdieuian theory by broadening our conception of culture and

including social networks in an explanation of the reproduction of advantage in the education

system. Bourdieu (1977) famously argued that the education system reproduces the social class

structure by favoring the culture of the upper classes. This article sought to demonstrate that an

education system may also favor or undermine particular network structures—structures that,

like cultural capital, are rooted in the class system. In addition, while Bourdieu’s assertion

that particular cultural tastes matter for success in the education system (and society at

large) has come under criticism (e.g., Erickson, 1996), this article demonstrates the

importance of another form of culture – namely something I termed ‘‘network know-how’’ –

for success in the transition to higher education. This more process-oriented form of culture

(in contrast to the more static ‘‘taste’’ form of culture that is typically studied) also suggests

new directions for the work currently being done connecting culture to social network

formation and maintenance formation (DiMaggio, 1985; Lizardo, 2006; Vaisey and Lizardo,

2009).

The current study is limited by the relatively small sample size and the research site, an elite

university. To understand better the role of networks in stratification will require sampling a

broader array of schools and students. In addition, to protect students’ privacy, the network

instrument used in this study did not require that subjects list the full names of ties in their

network; as a result, the findings mask possible multiplexity—that is, the extent to which certain

ties might perform multiple roles in the student’s network. Future work might consider the

relationship between multiplexity and network change.

Researchers have hypothesized that men and women should experience university attendance

differently (Sax, 2008), but no one has seriously considered how the experience of university life

would vary with students’ network structure. The research here reveals that network structure and

cultural norms about gender intersect to produce very different transitions to university life. In this

sense, this work can be considered part of a research trajectory initiated by Emirbayer and

Goodwin’s (1994) call to integrate culture into network theory (e.g., Bearman, 1993; Cardon and

Granjon, 2005; Erickson, 1996; Fuchs, 2001; Gould, 1995; Kane, 2004; McAdam, 2003; Mische,

2003; Passy, 2003; Smilde, 2007). To the extent that institutional priorities – reflecting the network

structure of the university environment – shaped students’ network change, this study also builds on

Uzzi and Spiro’s (2005) call for more work integrating our understanding of global and ego-

networks.

In this work, culture and networks interpenetrate in the production of gender differences

in the transition to an elite university. Past network experience can shape reactions to

new environments, and cultural mandates can celebrate or denigrate the kind of emotion

work that leads to new tie formation. Hence, culture seems to play a key role in network

evolution.

Acknowledgements

This study was funded by a US National Science Foundation Dissertation Improvement Grant

and two summer research grants from the University of Pennsylvania Department of Sociology. I

thank Randall Collins, Ezra Zuckerman, Grace Kao, Shawn Bauldry, Karin Velez, and Ben

Albers.

D. Kane / Poetics 39 (2011) 266–289284

Appendix A. Retention of sample respondents

Appendix B. Network instrument

D. Kane / Poetics 39 (2011) 266–289 285

Table A1

Logit model predicting completing time 2 survey; N = 512.

Est SE Sig.

Female �0.28 0.20

Age �0.05 0.02 *

Parent education 0.06 0.12

Urban 0.18 0.23

Origin: Asia �0.62 0.29 *

Origin: Europe �0.39 0.30

Origin: Other �0.69 0.43

Academic program 0.05 0.20

T1 network size 0.03 0.08

T1 network density �0.24 0.39

T1 ethnic heterogeneity �0.33 0.45

T1 gender heterogeneity 0.10 0.29

T1 regional heterogeneity 0.43 0.42

T1 religious heterogeneity 0.77 0.36 *

_constant 1.19 0.90

Notes: *p � 0.05. Based on 10 multiply imputed datasets. Reported estimates are log-odds.

Table B1

Part III. Please answer the following questions for people you know in your home country. For instance, for Persons 1 and

2 think of two people you study or work with. Write their first name or initials in the first blank and then fill in your answers

in the blanks that follow. If there is a person who you share more than one of these activities with, you need only write their

name or initials the second time you list them.

Persons 1 and 2 refer to

people with whom

you study or work.

Persons 3 and 4 refer

to whom you spend

leisure time.

Persons 5 and 6 refer to

people with whom you

discuss important matters.

Person 1 Person 2 Person 3 Person 4 Person 5 Person 6

First Name/Initials

Gender

Race

Nationality

Religion

From where do you know this

person? (e.g., went to school

with, friend of the family)

For how long have you known

this person?

Would this person agree or

disagree with the following:

‘‘When jobs are scarce, men

should have more right to a

job than women.’’

Appendix C. Measures of network variables

C.1. Network density

Network density measures the proportion of possible ties in a respondent’s network that are

actually present.

density ¼ actual ties

possible ties

Like Marsden (1987) and distinct from Fischer (1982), I include networks of size 2, which

Marsden found to increase average density.

C.2. Network diversity

Because all information in the alters’ ‘‘profiles’’ were nominal characteristics, diversity was

measured using the index of qualitative variation (Agresti and Agresti, 1977, p. 208), following

Marsden’s (1987) example. The IQV provides an intuitive metric for measuring diversity among

qualitative variables. A standardized version of the diversity index is:

I ¼ 1�Sp2

1� 1=k:

Network diversity was calculated for race/ethnicity, gender, world region, and religion.

Ethnicity and race of alters were classified into five categories of descent: Asian, African,

European, Hispanic, and Other. Descriptions of the religious preferences of alters were grouped

into six categories: Agnostic/Atheist, Catholic, Protestant, Islamic, Jewish, and Other. The

citizenship of alters was grouped into four world regions: Asia, Europe, US, and ‘‘other.’’

D. Kane / Poetics 39 (2011) 266–289286

Table B1 (Continued )

Persons 1 and 2 refer to

people with whom

you study or work.

Persons 3 and 4 refer

to whom you spend

leisure time.

Persons 5 and 6 refer to

people with whom you

discuss important matters.

Person 1 Person 2 Person 3 Person 4 Person 5 Person 6

How many friends do you and

this person have in common:

most, some, a few, none

Does s/he know Person 1? X

Does s/he know Person 2? X

Does s/he know Person 3? X

Does s/he know Person 4? X

Does s/he know Person 5? X

Does s/he know Person 6? X

First-wave network survey. (The second wave was collected with a web survey and used the same questions.) The

table at the bottom (starting with ‘‘Does s/he know Person 1?’’) creates a ‘‘mirror image’’ on either side of the Xs; in

other words, the same data appear on either side of the diagonal. This allowed for a check on reliability for calculating

density.

References

Agresti, A., Agresti, B.F., 1977. Statistical analysis of qualitative variation. In: Schuessler, K.F. (Ed.), Sociological

Methodology 1978. Jossey-Bass, San Francisco, CA, pp. 204–237.

Arum, R., Roksa, J., 2011. Academically Adrift. University of Chicago, Chicago.

Bearman, P., 1993. Relations into Rhetorics: Local Elite Structure in Norfolk, England 1540–1640. Rutgers University

Press, New Brunswick, NJ.

Bourdieu, P., 1977. Cultural reproduction and social reproduction. In: Karabel, J., Halsey, A.H. (Eds.), Power and

Ideology in Education. Oxford University Press, New York, pp. 487–510.

Bowen, W., Bok, D., 1998. The Shape of the River: Long-Term Consequences of Considering Race in College and

University. Princeton University Press, Princeton, NJ.

Buchmann, C., DiPrete, T., 2006. The growing female advantage in college completion: the role of family background and

academic achievement. American Sociological Review 71, 515–541.

Burleson, B.R., 2003. Emotional support skill. In: Greene, J.O., Burleson, B.R. (Eds.), Handbook of Communication and

Social Interaction Skills. Lawrence Erlbaum, Mahwah, NJ, pp. 551–594.

Burt, R., 1998. The gender of social capital. Rationality and Society 10, 5–46.

Campbell, K.E., Marsden, P.V., Hurlbert, J.S., 1986. Social resources and socioeconomic status. Social Networks 8, 97–

115.

Cardon, D., Granjon, F., 2005. Social networks and cultural practices: a case study of young avid screen users in France.

Social Networks 27, 301–315.

Charles, M., Bradley, K., 2002. Equal but separate?. a cross-national study of sex segregation in higher education.

American Sociological Review 67, 573–599.

Chase, R., 2008. Perfectly Prep: Gender Extremes at a New England Prep School. Oxford University Press, New York.

Clydesdale, T., 2007. The First Year Out: Understanding American Teens after High School. University of Chicago Press,

Chicago.

Collins, R., 1994. Four Sociological Traditions. Oxford University Press, New York.

Connell, R.W., 1995. Masculinities. University of California Press, Berkeley, CA.

Cordova, J., Gee, C., Warren, L., 2005. Emotional skillfulness in marriage. Journal of Social and Clinical Psychology 24,

218–235.

Derryberry, W.P., Toma, S.J., 2005. The friendship effect. About Campus 5, 13–19.

DiMaggio, P., 1985. Classification in art. American Sociological Review 52, 440–455.

Doreian, P., Stokman, F.N. (Eds.), 1997. Evolution of Social Networks. Gordon and Breach Publishers, Amsterdam.

Doreian, P., 2006. Actor network utilities and network evolution. Social Networks 28, 137–164.

Doyle, J., 1995. The Male Experience, second edition. W.C. Brown, Dubuque, IA.

Emirbayer, M., Goodwin, J., 1994. Network analysis, culture, and the problem of agency. American Journal of Sociology

99, 1411–1454.

Erickson, B., 1996. Culture, class, and connections. American Journal of Sociology 102, 217–251.

Fischer, C., 1982. To Dwell among Friends: Personal Networks in Town and City. University of Chicago Press, Chicago.

Fischer, M.J., 2008. Does campus diversity promote friendship diversity? a look at interracial friendships in college.

Social Science Quarterly 89, 631–655.

Fuchs, S., 2001. Against Essentialism: A Theory of Culture and Society. Harvard University Press, Cambridge, MA.

Gans, H., 1982. The Urban Villagers. The Free Press, New York.

Gaztambide-Fernandez, R., 2009. The Best of the Best: Becoming Elite at An American Boarding School. Harvard

University Press, Cambridge, MA.

Goldstein, J.R., Warren, J.R., 2000. Socioeconomic reach and heterogeneity in the extended family: contours and

consequences. Social Science Research 29, 382–404.

Gould, R., 1995. Insurgent Identities: Class, Community, and Protest in Paris from 1848 to the Commune. University of

Chicago Press, Chicago.

Huisman, M., Steglich, C., 2008. Treatment of non-response in longitudinal network studies. Social Networks 30, 297–

308.

Iannacone, L.R., 1992. Sacrifice and stigma: reducing free-riding in cults, communes, and other collectives. Journal of

Political Economy 100, 271–291.

Jones, D.C., 1991. Friendship satisfaction and gender: an examination of sex differences in contributors to friendship

satisfaction. Journal of Social and Personal Relationships 8, 167–185.

Kane, D., 2004. A network approach to the puzzle of women’s cultural participation. Poetics 32, 105–112.

D. Kane / Poetics 39 (2011) 266–289 287

Khan, S., 2011. Privilege: The Making of an Adolescent Elite at St. Paul’s School. Princeton University Press, Princeton,

NJ.

Kilgannon, S.M., Dary, E.T., 1992. A longitudinal study about the identity and moral development of Greek students.

Journal of College Student Development 33, 253–259.

Kimmel, M., 1996. Manhood in America. The Free Press, New York.

Kimmel, M., 2008. Guyland: The Perilous World Where Boys Become Men. Harper, New York.

Kingston, P.W., Hubbard, R., Lapp, B., Schroeder, P., Wilson, J., 2003. Why education matters. Sociology of Education

76, 53–70.

Leonhardt, D., 2005. The College Dropout Boom. New York Times May 24.

Lewis, K., Kaufman, J., Gonzalez, M., Wimmer, A., Christakis, N., 2008. Taste, ties, and time: a new social network

dataset using Facebook.com. Social Networks 30, 330–342.

Lipsky, D., 2003. Absolutely American: Four Years at West Point. Random House, New York.

Lizardo, O., 2006. How cultural tastes shape personal networks. American Sociological Review 71, 778–807.

Lubbock Christian University, 2010–2011. Student Handbook. www.lcu.edu/resources/student-handbook.html (accessed

21.05.11).

Marsden, P.V., 1987. Core discussion networks of Americans. American Sociological Review 52, 122–131.

Marsden, P., 1988. Homogeneity in confiding relations. Social Networks 10, 57–76.

McAdam, D., 2003. Beyond structural analysis: toward a more dynamic understanding of social movements. In: Diani,

M., McAdam, D. (Eds.), Social Movements and Networks. Oxford University Press, New York, pp. 281–298.

McPherson, M., Smith-Lovin, L., Cook, J., 2001. Birds of a feather: homophily in social networks. Annual Review of

Sociology 27, 415–444.

McLean, P.D., 2007. The Art of the Network: Strategic Interaction and Patronage in Renaissance Florence. Duke

University Press, Durham, NC.

Messner, M., 1990. Boyhood, organized sports, and the construction of masculinities. Journal of Contemporary

Ethnography 8, 416–444.

Migliaccio, T., 2009. Men’s friendships: performances of masculinity. Journal of Men’s Studies 17, 226–241.

Mische, A., 2003. Cross-talk in movements: reconceiving the culture-network link. In: Diani, M., McAdam, D. (Eds.),

Social Movements and Networks. Oxford University Press, New York, pp. 258–280.

Moody, J., McFarland, D., Bender-deMoll, S., 2005. Dynamic network visualization. American Journal of Sociology 110,

1206–1241.

Moore, G., 1990. Determinants of men’s and women’s personal networks. American Sociological Review 55, 726–735.

Nathan, R., 2005. My Freshman Year: What a Professor Learned by Becoming a Student. Penguin, New York.

Pascarella, E.T., Terenzini, P.T., 2005. How College Affects Students: Three Decades of Research. Jossey-Bass, San

Francisco, CA.

Passy, F., 2003. Social networks matter: but how? In: Diani, M., McAdam, D. (Eds.), Social Movements and Networks.

Oxford University Press, New York, pp. 21–48.

Patrick, S., Beckenbach, J., 2009. Male perceptions of intimacy. The Journal of Men’s Studies 17, 47–56.

Pollack, W., 1998. Real Boys: Rescuing Our Sons from the Myths of Boyhood. Henry Holt, New York.

Reid, H., Fine, G., 1992. Self-disclosure in men’s friendships: variations associated with intimate relations. In: Nardi, P.

(Ed.), Men’s Friendships. Sage, Newbury Park, CA, pp. 132–152.

Reisman, J., 1990. Intimacy in same-sex friendships. Sex Roles 23, 65–83.

Robins, G., Pattison, P., 2001. Random graph models for temporal processes in social networks. Journal of Mathematical

Sociology 25, 5–41.

Robbins, A., 2006. Overachievers: The Secret Life of Driven Kids. Hyperion, New York.

Sageman, M., 2004. Understanding Terror Networks. University of Pennsylvania Press, Philadelphia.

Samter, W., 2003. Friendship interaction skills across the lifespan. In: Greene, J.O., Burleson, B.R. (Eds.), Handbook of

Communication and Social Interaction Skills. Routledge, London, pp. 637–684.

Sax, L.J., 2008. The Gender Gap in College: Maximizing the Developmental Potential of Women and Men. Jossey-Bass,

New York.

Schultz, J., Breiger, R.L., 2010. The strength of weak culture. Poetics 38, 610–624.

Scott, J., 2000. Social Network Analysis. Sage, Thousand Oaks, CA.

Singleton, R.A., Vacca, J., 2007. Interpersonal competition in friendship. Sex Roles 57, 617–627.

Stearns, E., Buchmann, C., Bonneau, K., 2009. Interracial friendships in the transition to college: do birds of a feather

flock together once they leave the nest? Sociology of Education 82, 173–195.

Smilde, D., 2007. Reason to Believe: Cultural Agency in Latin American Evangelicalism. University of California Press,

Berkeley, CA.

D. Kane / Poetics 39 (2011) 266–289288

Stevens, M.L., 2007. Creating a Class: College Admissions and the Education of Elites. Harvard University Press,

Cambridge, MA.

Stevens, M.L., Armstrong, E.A., Arum, R., 2008. Sieve, incubator, temple, hub: empirical advances in the sociology of

higher education. Annual Review of Sociology 34, 127–151.

Thomas, S.L., 2000. The ties that bind: a social network approach to understanding student persistence. Journal of Higher

Education 71, 591–615.

Uzzi, B., Spiro, J., 2005. Collaboration and creativity: the small world problem. American Journal of Sociology 111, 447–

504.

Vaisey, S., Lizardo, O., 2009. Can cultural worldviews influence network composition? Social Forces 88, 1595–1618.

West, C., Zimmerman, D., 1986. Doing gender. Gender & Society 1, 125–151.

Willer, R., 2007. The role of metanetworks in network evolution. Journal of Mathematical Sociology 31, 110–119.

Wimmer, A., Lewis, K., 2010. Beyond and below racial homophily: ERG models of a friendship network documented on

Facebook. American Journal of Sociology 116, 583–642.

Xu, Y., Burleson, B.R., 2001. Effects of sex, culture, and support type on perceptions of spousal social support: an

assessment of the ‘‘support gap’’ hypothesis in early marriage. Human Communication Research 27, 535–566.

Danielle Kane is a post-doctoral fellow at the Duke University Thompson Writing Program. Her work focuses on the

sociology of culture as well as comparative-historical sociology. She is currently working on a project that examines the

impact of clan involvement and colonial experience on state formation and women’s rights in Central Asia.

D. Kane / Poetics 39 (2011) 266–289 289