Notes on Connecting in College — Chapter 1: Friendship Networks

Where to Draw the Line? Measuring Friendship

  • Definition of friendship is variable; participants define and justify who counts as a friend.

  • Key features distinguishing friends from nonfriends include trust, closeness, a shared connection, time spent together, and mutual understanding; friendships involve honesty, similar interests, shared experiences, and duration.

  • In this study, participants defined friendships openly and discussed how they drew the line between friends and acquaintances.

  • Average number of named friends: 18; range from 3 to 60; most students named 6–25 friends.

  • Demographic data were collected for each friend and whether each pair of friends knew each other; this enabled analysis of network density, betweenness centrality, and modularity.

  • Composition of friendships on MU campus: on average, 53% met on campus (MU), 12% from home (and also MU), 26% from home, 9% from other settings.

  • Meeting pattern on campus increases with time on campus: 1st year ~40% on campus; 2nd year ~46%; 3rd year+ ~57%.

  • Proximity and homophily influence who becomes friends (dorms, classes, organizations, shared spaces; race/ethnicity also influences patterns).

  • Most close/frequent contact with friends occurs face-to-face; with recently met or home friends, electronic communication (Facebook, IM, etc.) plays a larger role.

  • Definitions of friendship vary by race, class, and gender; trust and closeness are the most common criteria across groups.

Network Types: Dense Structures in Friendship Networks

  • Social network analysis uses sociograms (nodes = people; edges = ties) and measures like density, betweenness centrality, and modularity to describe structure.

  • Density: proportion of actual ties to possible ties among an ego’s friends.

  • Three network types emerged as thirds of the density distribution:

    • Tight-knitters: highly dense networks (density 0.67D1.00.67 \le D \le 1.0); friends form one cohesive group; average ego-network size smaller.

    • Compartmentalizers: middle-density networks (0.34D0.640.34 \le D \le 0.64); friends organized into 2–4 clusters.

    • Samplers: low-density networks (0.08D0.320.08 \le D \le 0.32); friends are sparsely connected; more one-on-one ties.

  • Density, betweenness centrality, and modularity together differentiate network types.

  • Density range observed: from 0.08 to 1.00; average around 0.56.

  • Betweenness centrality (normalized 0–1): measures how central the ego is in connecting others.

    • Tight-knitters: average centrality 0.040.04

    • Compartmentalizers: 0.250.25

    • Samplers: 0.620.62

  • Modularity (range 0.5 to 1-0.5\text{ to }1; positive means clearer sub-communities):

    • Tight-knitters: 0.020.02

    • Compartmentalizers: 0.190.19

    • Samplers: 0.250.25

  • Implications: betweenness and modularity differentiate network types; density distinguishes between types but does not separate compartmentalizers from samplers.

  • Network type correlates with demographics and outcomes (described below).

Network Types by Demographics

  • Size differences:

    • Tight-knitters: smaller networks (~13 friends)

    • Compartmentalizers and Samplers: larger networks (~20 friends)

  • Race and class associations:

    • Most white students are compartmentalizers; most black and Latino students are tight-knitters.

    • Samplers are a smaller share across groups (~20%).

  • Other characteristics:

    • Samplers more likely to live on campus; less likely to be in a romantic relationship.

    • Tight-knitters more likely to work; however, employed samplers work longer hours than tight-knitters.

  • Academic and success indicators:

    • Tight-knitters tend to have lower ACT scores, GPAs, and graduation rates than other types.

    • Racial composition interacts with network type in shaping experiences and coping strategies.

  • Summary: network type is a meaningful predictor of academic and social outcomes, with race and class shaping which network type students form.

Gender: Similarities More Than Differences

  • Overall, few gender differences in friendship structure; more similarities than differences.

  • Women’s networks: slightly fewer friends overall but similar density and proportions of MU/home connections.

  • Women’s friendships: a higher share of female friends (about 72% of women’s friendships) vs men (about 59% male friends).

  • Fraternity status influences male networks: affiliated men have more same-gender ties (avg ~69% male friends) than unaffiliated men (~54%).

  • Some men form close friendships with women; others maintain strong same-gender ties.

  • Conclusion: both genders form deep and meaningful friendships; gender differences in friendship do not explain large gender gaps in academic outcomes.

Race and Ethnicity

  • Racial composition of friendships varies markedly:

    • White students: ~84% of friends are same-race.

    • Black students: ~76%

    • Latino students: ~61%

  • Diversity of networks:

    • Some students have fully homophilous networks; some have diverse networks.

    • Whites more likely to have diverse friendships than in some other groups, while students of color often report more race-based isolation but also rely on same-race friends for support.

  • Network type and race:

    • White students are more often compartmentalizers; black/Latino students more often tight-knitters.

    • Samplers exist across groups but are less common among whites.

  • Outcomes and support:

    • Networks of color, especially tight-knitters, provide social support in navigating race-based isolation.

    • Network type matters more for academic/social success than racial composition alone.

  • Definitions of friendship differ by race:

    • White students often describe friends as fun or cool people.

    • Students of color emphasize closeness and deeper connection.

Social Class

  • Definitions of friendship vary by class:

    • Lower-class students describe more multi-level closeness and purposeful friendships.

    • Upper-class students emphasize socializing and having fun.

  • Network composition by class:

    • Lower-class students have fewer friends from home.

    • Working-class status correlates with similar network sizes but different home-family composition.

  • Intersections:

    • Race/class interaction shapes friendship size and the way closeness is defined; lower-class students with people of color tend to report different patterns than privileged groups.

  • Overall takeaway: class influences how students experience college and define friendship, but network type often interacts with race to shape support and outcomes.

Intersections of Gender, Race, and Class

  • Complex patterns emerge when all three dimensions are considered.

  • Examples:

    • Latino and Black men: around 21 friends on average; Black women around 12; Latinas around 17.

    • Black/Latino students from lower-class backgrounds: up to ~24 friends; higher than some peers.

  • Bottom line: intersections shape both the size and the meaning of friendship networks and their role in academic/social life.

Other Relationships: Professors, Significant Others, and Parents

  • Professors:

    • Generally not named as friends; a few serve as mentors and provide long-term guidance (graduate school, role modeling).

    • Relationships with professors arise more from chance encounters than systematic mentoring.

  • Significant others (romantic partners):

    • Common and increasingly present as students progress; can provide instrumental, emotional, and intellectual support.

    • Benefits are mixed: partners can help academically but can also distract; women more often help partners with academics than vice versa.

  • Family (especially parents):

    • Family involvement is common; they provide financial support, encouragement, and study guidance.

    • First-generation students rely more on extended family (siblings, aunts/uncles, grandparents) for academic support.

    • For many students, family involvement fades relative to friends, who remain central in daily life.

  • Overall: friends are central to the college experience, while professors, significant others, and family play important but often more variable roles.

Conclusion

  • Friendships are a central, context-dependent element of college life, reflecting broader structural inequities by race, gender, and class.

  • Network type (dense vs. compartmentalized vs. sparse) more strongly relates to academic/social success than raw network size or composition alone.

  • Context (residential setting, campus structure, proximity) shapes how friendships form and function.

  • Implications: recognizing the social role of friendships can inform how educators and administrators support student success; future work should explore how friendship networks evolve after college.

extDensity=extnumberoftiespresentnumber of possible ties;0Density1.ext{Density} = \frac{ ext{number of ties present}}{\text{number of possible ties}}; \quad 0 \le \text{Density} \le 1.
extBetweennesscentrality(normalized)[0,1],higher means more central in connecting others.ext{Betweenness centrality (normalized)} \in [0,1], \quad \text{higher means more central in connecting others}.
extModularity[0.5,1],positive values indicate community structure beyond chance.ext{Modularity} \in [-0.5, 1], \quad \text{positive values indicate community structure beyond chance}.