benlehrburger/college-text-corpus · Datasets at Hugging Face

text stringlengths 0 2.12k
varies with different social network structures. In particular, a “compartmentalized” social
network structure was identified as the most academically supportive. The results of this study
uphold social network analysis as an appropriate method for use in researching digital social
networks and help inform college students about which types of digital and in-person social
network structures can benefit them academically.

Introduction

Our success depends on the integrity of our social networks. For college students,
academic success can be helped or hindered by the quality and structure of the relationships that
a student has. The extent to which a student is able to seek out linear algebra tutoring, history
paper peer reviews, or neuroscience study sessions is determined by who they know and how
they know them.
Literature on the relationship between social networks and academic performance is vast.
Sociologists use social network analysis (SNA) to understand interactions among networks and
their participants, and SNA has been applied in the classroom on many occasions (Serrat, 2017).
Most studies suggest that social network depth and breadth correlate with academic success. One
experiment conducted with 226 undergraduate university students found that socially isolated
students had significantly lower examination grades than socially integrated students and were
more likely to drop out of university (Stadtfeld et al., 2019). Another study on 538
undergraduates reported that the quality of a student’s social network is the most predictive
feature of their academic success, above individual characteristics like personality and class
attendance (Kassarnig et al., 2018).
But exactly how much does academic performance depend on social networks and what
features of social networks is academic performance contingent on? Past research on grade point
average (GPA) and social network characteristics helps clarify the relationship between
academics and social connectivity. One research team observed that variance in social network
size correlated with a half point of GPA; students with the most friends scored about a half point
of GPA higher than students without friends (Khalil et al., 2019). Another team focusing
specifically on the classroom environment found that social network centrality – how connected a student is in the classroom network – explained 47 percent of the variance in a student’s GPA
for any given class (Williams et al., 2019). So, the more people that a student knows and the
better they know them, the higher their GPA will be.
The structure of social networks can differ independent of their size or centrality.
Sociologist Janice McCabe offers a taxonomy for classifying social network structure based on
her research of university undergraduates (McCabe, 2016). She identifies three distinctive types:
tight-knitters, compartmentalizers, and samplers. Tight-knitters have one densely woven
friendship network in which nearly all of their friends are friends with one another. The level of
academic support and motivation that tight-knitters receive is directly determined by the skills
and motivation of their ingroup. Compartmentalizers’ friends form two to four clusters, where
friends know each other within clusters but rarely across them. Compartmentalizers often
succeed academically by seeking support from different clusters depending on their needs.
Samplers make a friend or two from a variety of places, but the friends remain unconnected to
each other. The academic success of samplers is determined by their intrinsic skill and
motivation given that they lack significant support in their network (McCabe, 2016).

The social networks that McCabe and most SNA models focus on are in-person, face-to-
face networks. In the digital age, however, students’ social networks can also exist online.

According to a study of social media use among college-age students, 98 percent of US adults
aged 18 to 24 use social media in a typical month (Perrin, 2020). Social media allows students to
stay connected with their offline network or expand their networks to include new online
relationships. Often, when it is undesirable or impossible for a student’s social network to be
realized in person, they will turn to social media. Research on the reliance of college students on
social media found that about two-thirds of students agreed that social media helps them feel
more connected to their peers and that this connection aids their academic success (Creighton et
al., 2013). Academically-supportive online social networks do not look all too different from
those developed offline. SNA performed on social media platforms Twitter and Facebook

revealed that online communities have very similar structural characteristics to offline face-to-
face networks (Dunbar et al., 2015). Thus, if we see McCabe’s tight-knitters,

compartmentalizers, and samplers in offline networks, then we can expect that structural
taxonomy to carry over to the online world.
The structure of offline social networks is an indicator of academic performance because
who a student knows and how they know them determines how much academic support they will
receive. But what about online social networks? If students are becoming more reliant on their
social media networks for building relationships and receiving academic support, does the
structure of a student’s online network influence academic performance too?
Fieldwork hints that the answer to this question is a resounding “yes.” Past studies
suggest that the structure of offline social networks correlates with academic performance,
students are increasingly reliant on digital social networks for academic support, and that offline
and online social networks share similar structures. Given these premises, we can derive the
following hypotheses:

H1: if online social networks have structures that parallel those of McCabe’s
model, then the number of relationships that students have on social media will
vary significantly across different social network structures

H2: if the structure of social networks influences academic performance, then
students’ GPA will vary significantly across different social network structures

Experimental results supporting these hypotheses would inform future initiatives in SNA
and how college students build their social networks. The primary implication is expanding the
reach of McCabe’s social network taxonomy. Currently, McCabe’s model has only been applied
to in-person social networks. Evidence of varying academic performance and number of online
relationships across McCabe’s social network structures, however, would corroborate the
model’s application to online social networks. Beyond McCabe’s model, supportive empirical
results would also encourage the use of SNA in modeling social networks irrespective of their
medium. Whether it be online or offline, an affirmation of these hypotheses would uphold that
SNA transcends the medium in which people realize their social networks. Lastly, results
upholding these hypotheses would illuminate the importance of social network structure for
college students. If social network structure varies with GPA, then students ought to give more
consideration to how they form friendships. The influence of social network structure would be
compounded even further if we find that offline social network structure can be realized online
too. It is possible that particular social network structures offer more academic support than
others and some students would want to opt for structures that maximize their academic
resources.

Method

varies with different social network structures. In particular, a “compartmentalized” social

network structure was identified as the most academically supportive. The results of this study

uphold social network analysis as an appropriate method for use in researching digital social

networks and help inform college students about which types of digital and in-person social

network structures can benefit them academically.

Introduction

Our success depends on the integrity of our social networks. For college students,

academic success can be helped or hindered by the quality and structure of the relationships that

a student has. The extent to which a student is able to seek out linear algebra tutoring, history

paper peer reviews, or neuroscience study sessions is determined by who they know and how

they know them.

Literature on the relationship between social networks and academic performance is vast.

Sociologists use social network analysis (SNA) to understand interactions among networks and

their participants, and SNA has been applied in the classroom on many occasions (Serrat, 2017).

Most studies suggest that social network depth and breadth correlate with academic success. One

experiment conducted with 226 undergraduate university students found that socially isolated

students had significantly lower examination grades than socially integrated students and were

more likely to drop out of university (Stadtfeld et al., 2019). Another study on 538

undergraduates reported that the quality of a student’s social network is the most predictive

feature of their academic success, above individual characteristics like personality and class

attendance (Kassarnig et al., 2018).

But exactly how much does academic performance depend on social networks and what

features of social networks is academic performance contingent on? Past research on grade point

average (GPA) and social network characteristics helps clarify the relationship between

academics and social connectivity. One research team observed that variance in social network

size correlated with a half point of GPA; students with the most friends scored about a half point

of GPA higher than students without friends (Khalil et al., 2019). Another team focusing

specifically on the classroom environment found that social network centrality – how connected a student is in the classroom network – explained 47 percent of the variance in a student’s GPA

for any given class (Williams et al., 2019). So, the more people that a student knows and the

better they know them, the higher their GPA will be.

The structure of social networks can differ independent of their size or centrality.

Sociologist Janice McCabe offers a taxonomy for classifying social network structure based on

her research of university undergraduates (McCabe, 2016). She identifies three distinctive types:

tight-knitters, compartmentalizers, and samplers. Tight-knitters have one densely woven

friendship network in which nearly all of their friends are friends with one another. The level of

academic support and motivation that tight-knitters receive is directly determined by the skills

and motivation of their ingroup. Compartmentalizers’ friends form two to four clusters, where

friends know each other within clusters but rarely across them. Compartmentalizers often

succeed academically by seeking support from different clusters depending on their needs.

Samplers make a friend or two from a variety of places, but the friends remain unconnected to

each other. The academic success of samplers is determined by their intrinsic skill and

motivation given that they lack significant support in their network (McCabe, 2016).

The social networks that McCabe and most SNA models focus on are in-person, face-to-

face networks. In the digital age, however, students’ social networks can also exist online.

According to a study of social media use among college-age students, 98 percent of US adults

aged 18 to 24 use social media in a typical month (Perrin, 2020). Social media allows students to

stay connected with their offline network or expand their networks to include new online

relationships. Often, when it is undesirable or impossible for a student’s social network to be

realized in person, they will turn to social media. Research on the reliance of college students on

social media found that about two-thirds of students agreed that social media helps them feel

more connected to their peers and that this connection aids their academic success (Creighton et

al., 2013). Academically-supportive online social networks do not look all too different from

those developed offline. SNA performed on social media platforms Twitter and Facebook

revealed that online communities have very similar structural characteristics to offline face-to-

face networks (Dunbar et al., 2015). Thus, if we see McCabe’s tight-knitters,

compartmentalizers, and samplers in offline networks, then we can expect that structural

taxonomy to carry over to the online world.

The structure of offline social networks is an indicator of academic performance because

who a student knows and how they know them determines how much academic support they will

receive. But what about online social networks? If students are becoming more reliant on their

social media networks for building relationships and receiving academic support, does the

structure of a student’s online network influence academic performance too?

Fieldwork hints that the answer to this question is a resounding “yes.” Past studies

suggest that the structure of offline social networks correlates with academic performance,

students are increasingly reliant on digital social networks for academic support, and that offline

and online social networks share similar structures. Given these premises, we can derive the

following hypotheses:

H1: if online social networks have structures that parallel those of McCabe’s

model, then the number of relationships that students have on social media will

vary significantly across different social network structures

H2: if the structure of social networks influences academic performance, then

students’ GPA will vary significantly across different social network structures

Experimental results supporting these hypotheses would inform future initiatives in SNA

and how college students build their social networks. The primary implication is expanding the

reach of McCabe’s social network taxonomy. Currently, McCabe’s model has only been applied

to in-person social networks. Evidence of varying academic performance and number of online

relationships across McCabe’s social network structures, however, would corroborate the

model’s application to online social networks. Beyond McCabe’s model, supportive empirical

results would also encourage the use of SNA in modeling social networks irrespective of their

medium. Whether it be online or offline, an affirmation of these hypotheses would uphold that

SNA transcends the medium in which people realize their social networks. Lastly, results

upholding these hypotheses would illuminate the importance of social network structure for

college students. If social network structure varies with GPA, then students ought to give more

consideration to how they form friendships. The influence of social network structure would be

compounded even further if we find that offline social network structure can be realized online

too. It is possible that particular social network structures offer more academic support than

others and some students would want to opt for structures that maximize their academic

resources.

Method