The Sociology of Harry Potter: An Analysis of Adolescent Friendship Networks in the - Harry Potter Series

The Sociology of Harry Potter:

An Analysis of Adolescent Friendship Networks in the

                Harry Potter Series

                   Katie Christie

              The College of Charleston

              Department of Sociology

Tables and Figures

Table 1. Actor Degree Before Dumbledore’s Army…p. 22

Table 2: Reachability Before Dumbledore’s Army…p. 23

Table 3a. E.I. Index- Homophily by Gender Before Dumbledore’s Army… p. 24

Table 3b. E.I. Index- Homophily by Gender After Dumbledore’s Army… p.25

Table 4. E.I. Index- Homophily House Before Dumbledore’s Army… p. 26

Table 5. Closeness Centrlity Before Dumbledore’s Army… p. 27

Table 6a. Betweenness Centrality Before Dumbledore’s Army…p. 28

Table 6b. Betweenness Centrality After Dumbledore’s Army… p.29

Figure 1a. Friendship Ties Before Dumbledore’s Army…p.30

Figure 1b. Friendship Ties After Dumbledore’s Army… p.31

The Sociology of Harry Potter:

           An Analysis of Adolescent Friendship Networks in the Harry Potter Series

                                            Abstract

       Recent studies in sociology of literature have focused on the ways in which readers

construct meaning in their lives based on the stories that they read. This study adopts the idea

that literature can be analyzed sociologically, applying social theories to characters and plots, by

examining stories as social models. Using the internationally acclaimed Harry Potter series as

an example, this study will focus on two networks of adolescent friendship ties within the

fictional school of Hogwarts, analyzing basic social network properties, such as density,

reachability, and centrality. Particular emphasis will be placed on the presence of homophily in

friendship ties at Hogwarts. The character Harry Potter will also be singled out and analyzed in

terms of power and connectedness. The main objective of this paper is to provide a model for

analyzing contemporary fantasy novels geared towards adolescents, utilizing sociological

network methods and procedures in order to put a spin on literary analysis.

Introduction

       In 1999, Harry Potter and the Sorcerer’s Stone was released in the United States. Since

then, the seven-part series has gone on to sell millions of copies worldwide, earning author J.K.

Rowling international fame. Millions of children around the world have become devoted fans to

the series, and the term,“Potter-mania” has been used to refer to the way in which the series has

swept through the world of children’s literature and burst forth into mainstream society. The

character Harry Potter has become a cultural icon, and the record-breaking series has left its

mark on contemporary culture. The Harry Potter series is important not only because of its

commercial and financial success in the publishing, marketing, and entertainment industries, but

also because it has created characters with which contemporary children heavily identify.

       The Harry Potter series is more than a collection of popular children’s books; it can be

viewed as a mirror of contemporary society. As such, the characters and events in the story take

on a new meaning. The characters have special friendships and relationships with one another,

and the majority of the events take place in a school setting. If Rowling’s fictional world is

viewed through a sociological lens, the wizarding school, Hogwarts, where Harry and his friends

attend, can be thought of as its own network environment. The actors in this social network are

the students at Hogwarts, and the relations that link the students are ties of friendship. To give

the world of Harry Potter true sociological significance, it must be viewed as possessing the

characteristics of any other social network. To achieve this, I review the literature that examines

adolescent friendship networks, many within the context of schools.

       To properly examine Hogwarts as a social network, the Harry Potter series must be

analyzable and frameable in a sociological context. Wouter de Nooy (2001: 364) proposed

studying stories as social facts or models and applying social theories to characters and plots in

the same way that those theories would be applied to social institutions. De Nooy argued that

texts and stories offer models for interpersonal relations, and fictional stories serve as models in

everyday life, shaping people's self-perceptions and the way in which they share experiences (de

Nooy 2001: 365). Similarly, Griswold argued that recent advances in the sociology of literature

have focused on readers' construction of meaning, and research has indicated that there’s a

“psychology of reading” involved whereby readers transform the words being read into mental

associations that relate the story to real life experiences (Griswold 1993: 458). Therefore, stories

and novels, such as Harry Potter, can be viewed as fields of research in need of further

exploration.

       In order to put the character, Harry Potter, his friends, and Hogwarts School of

Witchcraft and Wizardry into a sociological context, I first examine literature that relates to

adolescent friendship networks. Unfortunately, much of the research that exists on friendship

networks is disconnected and suffering from inconsistencies in methodological procedures.

Crosnoe (2000) reviewed contemporary research on the topic of adolescent friendship, and

suggested using the concept of life-course theory in order to fuse existing data. Crosnoe focused

on four broad themes found in contemporary literature. Among his themes are friendships as

developmental contexts, the nature and structure of friendship, and the social context of

friendship (Crosnoe 2000: 377). In organizing my literature, I chose to use most of his themes,

and expanded upon certain elements within each category. I also created sub-categories within

the category of ‘Friendships as developmental contexts’ and ‘The social context of friendship.’

It should be noted that the themes identified by Crosnoe are by no means exclusive, as I found

multiple studies in which the research bridged these categories. However, for organizational

purposes, I found Crosnoe’s themes to be extremely useful in putting existing data into themes.

Friendships as Developmental Contexts

       This category deals with literature that focuses on individual analysis; in other words, the

literature discussed uses a micro-level approach in analysis. There is a wide range of literature

dealing with friendships and how individuals learn cooperation, forge new roles, identities, and

shape their views based on these friendships (Crosnoe 2000:378). Berndt (1982) cited three

explanations for the significance of adolescent friendships. He found that biological, social, and

cognitive forces are at work in shaping adolescent friendships (Berndt 1982: 1447-48). The

social environment of adolescents is of particular interest and relevance. In early adolescence,

the adolescent reaches a unique position because he is no longer a child, but also doesn’t yet

have all the responsibilities or freedom that come along with adulthood (Berndt 1982: 1447).

The majority of social interactions in this stage exist between close friends and peers, and

friendships are egalitarian (Berndt 1982: 1447-48). In the fictional setting of Hogwarts, social

interaction occurs at an even higher rate: the children are removed from their homes and are

forced into constant interaction with friends and peers.

       Due to the unique setting at Hogwarts, an environment that serves as both school and

home for the students, socialization and friendship play an even larger role in the daily lives of

the students. It is therefore important to consider the ways in which friendships are formed, and

the role friendships play in the setting of Hogwarts. Berndt (1982, 1995) discussed the role of

intimacy within friendships, and how adolescents build self-esteem by sharing feelings and

information with friends. In a later study, Berndt (1995: 1327) examined the influence of friends

on development as it related to school adjustment and found that adjustment is affected by the

characteristics and quality of friendships.

The Nature and Structure of Adolescent Friendship

       Research in this category can be divided into two categories: the nature of adolescent

friendships, which can be further divided into the elements or characteristics of friendship and

the quality of friendships, and the structure of friendships, which has slightly more complicated

sub-divisions (Crosnoe 2000: 377). Studies have been done in which the focus is on particular

features of adolescent friendship, such as George and Hartmann’s 1996 study of popularity.

Characteristics or elements of adolescent friendships may include age, gender, reciprocity,

identity, similarity, and stability. George and Hartmann (1996) conducted a study where

students were given the opportunity to rate their classmates on how much they were liked by

their peers. The purpose of this research was to examine three popularity groups: children were

defined as popular, average, or unpopular, in order to determine friendship network

characteristics and friendship prevalence (George and Hartmann 1996: 2304). George and

Hartman found that unpopular children were friendless in comparison to other children.

However, they also concluded that prior research done on popularity overestimated

friendlessness in unpopular children (George and Hartmann 1996: 2311). Another interesting

discovery was that approximately 75% of children who were popular had friends who were

approximately the same age (George and Hartmann 1996: 2312). Unpopular children, on the

other hand, were more likely to be friends with children who were different ages (George and

Hartmann 1996: 2312).

There is also a significant amount of research done on the structure of adolescent

friendship. I chose to highlight social networks and friendship groups, as they are most relevant

to my topic. Social network analysis includes information on cliques and isolates. Shrum (1987)

argued that examining liaisons in the social networks of adolescents, rather than focusing solely

on cliques as the hub of peer relations, is important in social network analysis. A liaison can be

defined as an individual belonging to a social system, who serves the purpose of linking outside

members to the system (Shrum 1987: 218). Clique members are defined as participants in a

close-knit group of interacting peers (Shrum 1987: 218). Shrum argued that distinguishing

liaisons from cliques is especially important for contemporary theories of the school as a context

for developing peer relations (Shrum 1987: 219). Social networks can also be viewed as

changing, adaptive processes (Cairns, et al 1995: 1330). Cairns, et al found (1995: 1343) that

fluidity in the strength of friendships and group members is an important feature of social

relationships.

Social Context

       Research with an emphasis on social context examines various aspects of adolescent

friendship as they are embedded in a larger social context. Examples of this include homophily,

race, gender, and physical location/geography. McPherson, et al (2001) defined homophily as

the tendency for people to have more contact with people who are similar to them than to people

who are dissimilar. In the United States, the biggest divide in social networks is between race

and ethnicity (McPherson, et al 2000: 420). The idea of homophily is important because it says

that people are more likely to confide and share with individuals who are similar to them. When

the principle of homophily is applied to adolescents, the literature suggests that teenagers tend to

associate with other teenagers who are similar in terms of behavior or achievement (McPherson,

et al 2000: 428). Sources of homophily include space, family ties, and organizational ties (this

could include schools) (McPherson, et al 2000: 429-434). Shrum, Cheek, and Hunter (1988)

examined racial and gender homophily in children in association with grade level. Their

findings showed that racial homophily increased as grade level increased, although the reverse

was true for gender homophily (Shrum, Cheek, and Hunter 1988: 237). For gender homophily

during elementary school, boys were more likely to be friends with other boys than girls were

with other girls, but this reversed in middle school. In other words, middle school girls were

found to have a significant increase in same-gender preference, leading to the conclusion that

there is more exclusion and intimacy in girl friendships (Shrum, Cheek, and Hunter 1988: 236).

One of the most interesting and significant findings of this study was that children develop an

early awareness of racial identity, and as they age, there is an increasing tendency to develop

homophilous friendships (Shrum, Cheek, and Hunter 1988: 236). The topic of racial homophily,

in particular, is an interesting one to examine in the environment of Hogwarts, because racial

divides are not based on skin color, but on blood type. Racial homophily based on blood type,

along with gender homophily, will be discussed at length later in this study.

Summary of Literature

       Based on the substantial amount of literature available on adolescent friendships, several

themes appear to be consistent. Building on Berndt's idea that the majority of social interactions

in early adolescence are between close friends and peers, and these friendships tend to be

egalitarian in nature, this study will examine whether or not this is true within the setting of

Hogwarts, and how or if this changes over time. In other words, this study will determine the

extent to which homophily plays a role in the friendship ties among students in Hogwarts by

examining the interactions occurring between friends who are similar in terms of race, gender

and organization. For the purpose of this study, race will consist of purebloods, or individuals

belonging to a genealogy that consists of no "muggles," or humans, and non-purebloods, which

constitutes individuals who have mixed parentage. By combining Berndts' ideas with Shrum,

Cheek, and Hunter's findings that young adolescents are more likely to spend time with members

of the same sex and same race, it is expected that this study will confirm the theory that young

adolescents interact primarily with friends who are similar to them, especially in terms of race

and gender. A second objective of this study is to determine how much power and influence

the character, and actor, Harry Potter has in his network of friends. The actor Harry Potter is

also expected to have the most connections in the network. The final and most important

objective of this study is to test social network procedures on the friendship networks present in

Hogwarts, in order to set the stage for further sociological analysis of literature.

       Because Harry Potter is a prominent series in contemporary society, it is important to

recognize the power that the books have in shaping children’s ideas about friendship, love, and

even notions of good and evil. Using friendship ties as the relations linking the students (the

actors) of Hogwarts (the network) to one another, particular emphasis will be placed on

homophily, connectedness, network size, and distance between actors. Although there is a great

deal of literature available on the subject of adolescent friendship, many studies highlight

inconsistencies in similar studies. Also, weaknesses appear in the methodological procedures

used. There is a great deal of research available on cliques and group level analysis, but literature

consisting of liaisons or other linking actors could be useful. Literature consisting of

sociological analyses of children’s books seems to be rare, but the importance of Harry Potter in

youth culture cannot be denied. The aim of this research is to bridge the world of fantasy that

children find so compelling and highlight the similarities that exist between actual adolescent

friendship networks and the social networks in the fictional setting of Hogwarts.

Methods
       The data collection process of this study involved unique procedures. Two networks

were examined, and the actors in the networks were students at Hogwarts School of Witchcraft

and Wizardry. Because the two networks examined did not consist of people or organizations,

but characters in literature, the samples were chosen based on the characters' relevance to the

story. In other words, characters were picked as actors according to the amount of information

that was available in the series about the character, and whether or not they were featured

frequently enough to be considered key actors. To begin collecting data, a pool of potential

actors was devised. A list of actors for the entire network of Hogwarts could not be determined,

because no complete list of characters is provided within the book. Hogwarts consists of seven

grades, grades one through seven, but the number of students within each grade is not specified.

The data collection process, therefore, involved identifying a pool of actors to be examined,

based on their relevance in the series. For the first set of data, or Network A, an actor was

classified as relevant and included if he or she was mentioned throughout the series, beginning

with the first book. A preliminary list of actors was drawn up by examining each character who

was named in the first book during the sorting ceremony at Hogwarts. The sorting ceremony

alphabetically listed the names of many of the characters in Harry's grade. From this pool of

actors, relevance was determined based on which characters went on to be featured regularly in

consecutive books. Any character who did not appear in each story was considered to lack

sufficient information for further analysis and was excluded from the list of actors. In order to

determine which characters were mentioned frequently enough to be included, the online source

Wikipedia was used.       Although Wikipedia is not an academic source, it provided basic

information on each of the characters, including which books featured particular characters,

which houses they belonged to, etc. Wikipedia featured an alphabetical list of all the characters

mentioned in the Harry Potter series, and by searching the list, information could be found in

sufficient detail about each actor.

       The next step in the data collection process was to identify the relational ties between

actors in the Hogwarts network. With few exceptions, the actors who are identified in the series

are mentioned in relation to Harry Potter, and the vast majority of characters with regular

mentions are in the same grade with him. For this reason, all actors chosen were in the same

grade as Harry Potter. The character Harry Potter starts Hogwarts in the first book of the series,

Harry Potter and the Sorcerer's Stone, at age eleven. In the six books that follow, Harry Potter

ages one year, and advances one grade. All students in the series belong to one of four houses,

which is determined in the Sorting Ceremony in the first year at Hogwarts. The four houses are

Gryffindor, Hufflepuff, Ravenclaw, and Slytherin. Rowling identifies twenty-three characters

during the Sorting Ceremony for Harry's grade in Book One, but only sixteen characters play an

active role in the series in the first four books (Rowling 1997). These sixteen characters are the

actors comprising the first data set, and the relations linking these actors are ties of friendship.

An actor was classified as a friend, and a tie was formed when actor A spent time outside

classes, engaged in conversation, shared leisure time, or otherwise interacted in an intimate

setting, with actor B. Being members of the same house or sharing classes together was not

considered sufficient grounds for a friendship tie. It should be noted that this measure of

friendship ties allowed for friendships to exist inside as well as outside houses. However, due to

the structure of Hogwarts, there are considerably more opportunities for members of the same

house, or inter-house members, to interact than members outside of houses, or extra-house

members. Members of the same houses eat lunch together, share a common room, and attend

classes together.

       Two networks were examined in this study. During the preliminary stages of research, it

was observed that friendship patterns change throughout the series, which is to be expected as

the characters age. However, differences in friendship ties became especially pronounced after

the formation of Dumbledore’s Army in Book Five. Dumbledore’s Army is a club that was

formed in Harry Potter and the Order of the Phoenix by Harry Potter, Ron Weaslely, and

Hermione Granger for students wishing to practice using defensive spells. The first network,

Network A, therefore, consisted of actors who played a prominent role throughout the series, and

were friends before the formation of Dumbledore's Army. Network A also represented early

adolescence, as the actors in this network were between the ages of eleven and fifteen in the first

four books. Network B referred to the ties of friendship formed after Dumbledore’s Army. In

Book Five, Rowling chose to introduce six new characters in Harry’s grade. These characters go

on to be featured regularly in the series, so they were included in Network B. The two friendship

networks consisted of symmetric, undirected ties and a binary scale of measurement.             To

measure friendship ties between nodes, the presence or absence of friendship ties was determined

based on the question, "Is node A friends with node B?" A score of 1 was given for a tie of

friendship, and a score of 0 was given where a tie was absent. All ties of friendship were

symmetric and undirected. In other words, if actor A considered B to be a friend, B also

considered A to be a friend.

       The network size for the first data set, friends before Dumbledore's Army, or Network A,

was found to be sixteen by performing a simple count of nodes. The network diameter was

found by running the Ucinet procedure for identifying geodesic distances. Geodesic distances

refer to the shortest possible path from one actor to another, usually resulting in the most

efficient connections between actors. The diameter of the network was the furthest distance

between actors, and in the case of friendship ties before Dumbledore's Army, the number was

two.   The second network, or Network B, was comprised of all of the actors present in the first

network, with the addition of six actors who were not featured until Book 5. The network size of

the friends after the formation of Dumbledore's Army was twenty-two. The diameter of this

network was one. The change in the diameter between the first network and the second can be

attributed to the removal of steps between actors when more friendships were formed due to the

actors' involvement in Dumbledore's Army. These results will be further explored and discussed

in the Analysis and Results section of this study.

Analysis and Results

       In order to analyze the two friendship networks at Hogwarts, basic network properties of

connection and distance were first examined. A total of four measures were used relating to

connection and distance: network size, actor degree, density, and reachability, where network

size and actor degree were considered to be basic demographics, and density and reachability

were measures of connection. To determine the size of Network A and Network B, a simple

count of nodes was performed. The total number of nodes for Network A was sixteen. This

means that a total of sixteen ties were possible, since the network is symmetric. In other words,

by using the equation, k* (k-1) / 2, where k is number of nodes, and plugging in the count of

nodes, 16, the total number of ties was found. It is important to establish a count of ties in order

to determine the networks' density. The total number of nodes for network B was twenty-two,

and the total number of ties, therefore, was twenty-two. The difference in sizes between

networks A and B is important because as the number of nodes increases, so does the number of

potential relationships. This generally means that the level of complexity increases as the size of

a network increases. Figures 1a and 1b show the differences in network structure before the

formation of Dumbledore's Army and after.

       Actor degree can be defined as a count of ties linking a node to other nodes, where the

highest possible value is (k-1). Actor degree is useful in determining how connected an

individual is in relation to the overall social structure. The actors with the highest degree for

Network A (see Table 1a) were Harry Potter, and Ron Weasley, with Crabbe, Malfoy, and Goyle

coming in closely behind. However, if we compare Network A to Network B, we see that the

actor degree changes significantly: actors 1-17, i.e., the actors in Dumbledore’s Army, had a

degree of fifteen, and actors 18-22, the actors not in Dumbledore’s Army, had a degree of five.

These results are not surprising, because with the formation of Dumbledore's Army, friendships

that did not exist initially within the houses of Gryffindor, Hufflepuff, and Ravenclaw, sprang

up, so that every actor in those three houses was connected with every other actor. It therefore

follows that these actors would become more connected with one another.

       Density is the proportion of ties present to the total number of ties. Density is an

important trait to examine, because generally the denser a network is, the greater the

mobilization of resources and spread of information. The density of Network A was found to be

.2250, which means that 22.5% of all possible ties were present. As would be expected, the

density increased for Network B, which had a density of .5844. These results show more than a

30% increase from Network A to Network B.

       The final measure conducted for connection was reachability. Reachability can be

defined as the set of connections or links that exist between actors, even if those actors are not

adjacent or directly connected. In a symmetric network, if actors are not reachable, the network

can become divided and sub-populations can form. A division was encountered in this study: a

major rift existed between the actors belonging to the Slytherin house, and all other actors. This

is true for both Network A and Network B (refer to Figures 1a and 1b). For Network A, the

actors Brown and Patil are not reachable to any other actors, and gaps exist between actors

within the house of Gryffindor. However when Network B is examined, (see Table 2) we see

that all actors are equally reachable, and a natural block formation appears, showing the divisions

between Dumbledore's Army members and non- Dumbledore’s Army members.

       The results from the Reachability measure lead us to the conclusion that the friendship

ties that sprung up due to the formation of Dumbledore's Army led to a cohesion in the

friendship network, further resulting in direct paths being formed between all actors in that

group. Therefore, due to the high level of connection between members of Dumbledore's Army,

we would expect the flow of information and mobilization of resources to be high. Indeed, a

perfect example of this occurred at the end of Book Five. In Book Five, Harry was nearly

attacked on the Hogwarts Express at the end of the school year by Malfoy, Crabbe, and Goyle.

Before they are able to attack him, however, several members of Dumbledore’s Army, including

Hannah Abbott, Susan Bones, Anthony Goldstein, and Terry Boot, who were not friends of

Harry’s prior to that year, jump to the rescue and prevent Harry’s attack (Rowling 2003). This

example illustrates the differences that existed after the formation of Dumbledore’s Army: a

significant increase in camaraderie and a banding together, or higher level of mobilization, could

be observed.

Results

       The first objective of this study was to identify the presence of homophilous friendship

ties by determining if young adolescents interact predominantly with friends who are similar in

terms of race and gender. Organizational association, or association with house-members, was

also examined. The methods used to test this procedure were fairly straight-forward. To test the

hypothesis, external group ties and internal group ties were measured by running the E-I Index

procedure on Ucinet. Actors were coded separately for Gender, Race, and House and were used

as attribute vectors for comparing Network A and Network B.      Race was measured by coding

individuals with a pure-blood status as one, and individuals without a pure-blood status as zero.

E-I Index was used to measure the percentage of ties being sent outward and the percentage

remaining internal. The E-I Index was run for Network A and Network B, to compare the

changes in homophily after the formation of Dumbledore’s Army. For gender, the re-scaled E-I

Index was found to be -.185. The negative value shows a weak tendency for internal ties.

However, when the same procedure was run for Network B, the E-I index was found to be .083,

showing a higher tendency for external ties. Tables 3a and 3b show which actors have the

highest and lowest internal and external ties.

       To further investigate gender homophily, observations were performed by simply

examining the network structure of friendship ties for Network A compared to Network B. The

friends in Network A consisted of a total of six females and ten males. The node Herminone

Granger had a total of three ties, all of which were connected to males. The node Hannah Abbott

had two ties, both of which were connected to boys. Both Milicent Bulstrode and Pansy

Parkinson had four ties, three of which were to boys, and one of which was to a girl. The nodes

Lavender Brown and Parvati Patil stood out: not only do they constitute a sub-division in the

network, but they are tied to only each other, showing them as isolates from the rest of the

network. In the case of Network A, it does not appear that females show a tendency towards

friendship with members of the same sex. However, when the boys in the network were

examined, there was a considerable difference. The actors Dean Thomas and Seamus Finnigan

each had three ties, all of which were directed at males. Ron Weasley had a total of five ties,

four of which were directed at males. The actor with the most ties, Harry Potter, had a total of

seven ties, and six of these seven ties were directed towards males. Although there was a

disproportionate ratio of males to females, it is clear that within Network A, there was a

discrepancy between males' tendency to be-friend other boys, and girls' tendencies to befriend

other females. These findings are therefore partially inconsistent with Shrum, Cheek, and

Hunter's observations that gender homophily is strong among young adolescents: although males

showed a strong tendency for gender homophily, females did not. Shrum, Cheek, and Hunter

found females to be much more likely to associate with same-sex friends beginning in early

adolescence, and the reverse appears to be true for females at Hogwarts.

When ties were measured for house membership, Network A was shown to have an E-I

Index of -.852, representing a strong tendency towards internal ties. The results were even

stronger for Network B, which had an E-I Index of -1. However, this can be attributed to

Slytherin’s complete separation from the other houses (see Table 4). Slytherin was the only

house after the formation of Dumbledore’s Army to not have external ties. Finally, to measure

racial homophily, the E-I Index was run using the blood status attribute vector. The results from

this test showed a re-scaled index of -.037, representing a very weak tendency toward internal

ties for Network A. This means that, within Network A, ties of friendship were more likely to be

to people of the same race (in this case, pure-blood or non-pure-blood), but only slightly. These

results changed significantly for Network B, which showed a strong tendency towards internal

ties. These results show that actors were much more likely to be friends with people of the same

race after the formation of Dumbledore’s Army. This is consistent with Shrum, Cheek, and

Hunter's claims that racial homophily increases with age.

       In order to determine if Harry was the most central and powerful actor in Network A and

Network B, Freeman’s measures of Closeness Centrality and Betweenness Centrality were

calculated. The idea behind closeness centrality is that actors who are closer to others have

greater power. Betweenness centrality is the idea that actors who serve as intermediaries

between other actors have the most power. According to Freeman’s Closeness Centrality

measure, the most powerful actors in Network A were Harry Potter, Ron Weasley, and Ernie

Macmillan (see Table 5). The weakest actors were Parvati Patil and Lavender Brown. However,

when the same measure was performed for Network B, all members of Dumbledore’s Army had

the same amount of closeness centrality, which was substantially more than the non-members.

The change from Network A to Network B can be attributed to the decreased distance between

actors from Network A to Network B. In other words, all actors were at an equal distance from

one another, so no one actor had a greater amount of power.

       Freeman’s Betweenness Centrality measure yielded very interesting results between

Network A and Network B. In Network A, the most powerful actors for betweenness were

Harry Potter, Ernie Macmillan, Justin Finch-Fletchley, and Ron Weasley (see Table 6). All other

actors had the same amount of betweenness power. However, when these results were compared

to Network B, it was found that no actors had betweenness power: all actors had a betweenness

power of 0. The reason for this is that the creation of Dumbledore’s Army removed

intermediaries, neutralizing the power of Harry Potter and the other formerly powerful nodes. It

should also be noted that the actors who were not a part of Dumbledore’s Army had the same

amount of power between actors because they also had no intermediaries: all actors were the

same distance, a distance of 1, from other actors. The results of Freeman’s Centrality measures

show that Harry Potter did have the most power in terms of betweenness and closeness prior to

the formation of Dumbledore’s Army, but that power was voided, or perhaps reallocated into a

measure of power not devised in this study, after the formation of Dumbledore’s Army.

Discussion

       This study analyzed the Harry Potter series through a sociological lens, treating the

characters and social network of friendship ties in Hogwarts as any other social network with

basic properties and characteristics. The study examined friendship ties during Harry’s first four

years at Hogwarts, and compared them to friendship ties after Dumbledore’s Army, which was

formed in Harry’s fifth year. The primary objective of this study was to highlight the similarities

that exist between friendship networks in the series, and recent sociological studies on the same

subject, showing that the Harry Potter series can be analyzed sociologically by identifying basic

network properties and characteristics. An additional objective of this study was to test existing

theories about homophily, specifically, the idea that adolescents are friends with those who are

similar to them in terms of race, gender, and organization. This was accomplished by comparing

changes in friendship patterns before Dumbledore’s Army, when the actors were between the

ages of 11-14, and after, when the actors were 15 and older. A final objective of this study was

to determine if Harry Potter, the focal point and most recognizable character in the series,

wielded the most power and influence within his social network.

       The results of this study showed a decrease in gender homophily between friends over

time, further supporting Shrum, Cheek, and Hunters findings that gender homophily decreases

over time. However, in contrast to their research, my findings showed that males were more

likely to be friends with other males prior to the formation of Dumbledore's Army, whereas

females exhibited many cross-gender ties. Also consistent with Shrum’s studies on adolescent

friendship, I found that racial homophily did increase over time, but this study did not perform an

in-depth analysis on the nature of the relations among those actors. Organizational homophily

was not shown to have a significant change over time, but a strong tendency towards internal ties

was discovered, suggesting that the placement of students into houses played a heavy role in the

friendships those students formed. In the fictional setting of Hogwarts, students are required to

share dormitories and common rooms, attend classes, and eat lunch with members of the same

house, so it is not surprising to find that ties of friendship within houses are stronger than ties of

friendships outside houses.

Conclusions

       An unexpected finding in this study was the discovery that among early adolescents, the

males were more likely to exhibit gender homophily than the females. One of the most notable

exceptions to the theory that gender homophily is strongest in early adolescence is Hermione

Granger. Until the formation of Dumbledore’s Army, Hermione did not have friends of the same

sex in her grade. She spent the majority of her time with Harry and Ron. Although I did not

measure the intensity of friendship ties, or examine the social dynamics between close friends or

“best friends,” this would be an interesting study to conduct. The characters Dean Thomas and

Seamus Finnigan, Lavender Brown and Parvati Patil, Harry, Ron, and Hermione, and Malfoy,

Crabbe, and Goyle, are understood to be best friends within the series, and an analysis of the

similarities between those dyads, or triads, could further support sociological studies pertaining

to the nature of adolescent friendship.

        Another useful study that could be conducted would be to examine the break-down of

gender in the friendship networks at Hogwarts. There is a disproportionate number of males to

females mentioned within the friendship networks studied, and a comparison study could be

drawn up between the Harry Potter series and other works of juvenile fantasy. What social roles

do the female characters occupy in this series in comparison to other stories? What about males?

Although I did prove my hypothesis that Harry Potter held the most power and connection in his

friendship network as a young adolescent, it was interesting to discover that the level of power

dispersed as Harry lost his position as an intermediary. Future studies could place more

emphasis on the power of intermediaries in social networks within literature, perhaps using

Bonacich’s power measure as a basis for determining actors’ power.

The biggest weaknesses in this study related to the methodological procedures. The

procedure used to establish ties of friendship was an entirely subjective measure: an arbitrary

measure was drawn up to link actors as friends, and no distinction was made between different

levels of friendship. In other words, an actor could be tied to his or her best friend and someone

classified as a friend who is not as intimately linked, i.e. does not share secrets with this person,

and no distinction would be made. Because only friends in Harry’s grade were included in the

sample, any friendship ties with older or younger students were ignored. It is possible that

patterns of friendship would be different if all ties of friendship were examined. Additionally,

there were a number of procedures that would have been helpful to use in examining ties of

friendship, such as transitivity, block densities, and point connectivity, but these were not

included. Due to the limited amount of information available about the actors in Network A and

Network B, details from the first four books were often lumped together into a “before”

category, despite the fact that each book experienced a passage of time and featured emotional

changes in characters. The behavioral, emotional, and psychological aspects of the sociological

parameters in this study were not discussed, but this topic would be fertile ground for further

study.

References

Berndt, Thomas J. 1982. "The Features and Effects of Friendship in Early Adolescence." Child
        Development 53:1447-60.
Berndt, Thomas J. and Keunho Keefe. 1995. "Friends' Influence on Adolescents' Adjustment to
        School." Child Development 66: 1312-29.
Cairns, Robert B., Man-Chi Leung, Lisa Buchanan, and Beverly D. Cairns. 1995. "Friendships
        and Social Networks in Childhood and Adolescence: Fluidity, Reliability, and
        Interrelations." Child Development 66: 1330-45.
Crosnoe, Robert. 2000. "Friendships in Childhood and Adolescence: The Life Course and New
        Directions." Social Psychology Quarterly 63: 377-91.
De Nooy, Wouter. 2001. “Stories and Social Structure: A structural Perspective on Literature in
        Society.” The Psychology and Sociology of Literature: 359-75.
George, Thomas P. and Donald P. Hartmann. 1996. "Friendship Networks of Unpopular,
        Average, and Popular Children." Child Development 67: 2301-16.
Griswold,Wendy. 1993. “Recent Moves in the Sociology of Literature.” Annual Review of
        Sociology 19: 455-67.
Hartup, Willard W., Doran C. French, Brett Laursen, Mary Kathleen Johnston, and John R.
        Ogawa. 1993. "Conflict and Friendship Relations in Middle Childhood: Behavior in a
        Closed-Field Situation." Child Development 64: 445-54.
McPherson, Miller, Lynn Smith-Lovin, and James M. Cook. 2001. "Birds of a Feather:
        Homophily in Social Networks." Annual Review of Sociology 27: 415-44.
Rowling, J.K. 1997. Harry Potter and the Sorcerer’s Stone. U.S.A.: Scholastic Books.
Rowling, J.K. 2003. Harry Potter and the Order of the Phoenix. U.S.A.: Scholastic Books.
Shrum, Wesley and Neil H. Cheek, Jr. 1987. "Social Structure During the School Years: Onset
        of the Degrouping Process." American Sociological Review 52: 218-23.
Shrum, Wesley, Neil H. Cheek, Jr., and Saundra MacD. Hunter. 1988. "Friendship in School:
        Gender and Racial Homophily." Sociology of Education 61: 227-39.
http.//en.wikipedia.org/wiki/Harry_Potter, accessed November 15, 2007.
http.//en.wikipedia.org/wiki/Dumbledore’s_Army_Members, accessed November 15, 2007.
http://en.wikipedia.org/wiki/List_of_ characters_in_Harry_Potter_books, first accessed
        November 25, 2007.

Table 1. Actor Degree- Before Dumbledore’s Army

 Actor                  Degree
 Harry Potter           7
 Ron Weasley            5
 Vincent Crabbe         4
 Draco Malfoy           4
 Hannah Abbott          2
 Lavender Brown         1
 Parvati Patil          1

Table 2: Reachability- Before Dumbledore’s Army

              L   S   H   N   P   H   D   R   H   J   E   M   V   G   D   P

Brown         0   0   0   0   1   0   0   0   0   0   0   0   0   0   0   0

Finnigan      0   0   1   1   0   1   1   1   1   1   1   0   0   0   0   0

Granger       0   1   0   1   0   1   1   1   1   1   1   0   0   0   0   0

Longbottom 0      1   1   0   0   1   1   1   1   1   1   0   0   0   0   0

Patil         1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0

Potter        0   1   1   1   0   0   1   1   1   1   1   0   0   0   0   0

Thomas        0   1   1   1   0   1   0   1   1   1   1   0   0   0   0   0

Weasley       0   1   1   1   0   1   1   0   1   1   1   0   0   0   0   0

Abbott        0   1   1   1   0   1   1   1   0   1   1   0   0   0   0   0

Finch-Fletc   0   1   1   1   0   1   1   1   1   0   1   0   0   0   0   0

Macmillan     0   1   1   1   0   1   1   1   1   1   0   0   0   0   0   0

Bulstrode     0   0   0   0   0   0   0   0   0   0   0   0   0   1   1   1

Crabbe          0     0   0   0       0      0   0   0   0   0    0    1    0   1   1   1

Goyle           0     0   0   0       0      0   0   0   0   0    0    1    1   0   1   1

Malfoy          0     0   0   0       0      0   0   0   0   0    0    1    1   1   0   1

Parkinson       0     0   0   0       0      0   0   0   0   0    0    1    1   1   1   0

                                             Table 3a:

E.I. Index- Homophily by Gender- Before Dumbledore’s Army

Actor                             Internal                       External

Lavender Brown                    1                              0

Seamus Finnigan                   3                              0

Hermione Granger                  0                              3

Neville Longbottom                2                              1

Parvati Patil                     1                              0

Harry Potter                      6                              1

Dean Thomas                       3                              0

Millicent Bulstrode               1                              3

Table 3 b.

        E-I Index- Homophily by Gender- After Dumbledore’s Army

Actor                    Internal                External

Lavender Brown           5                       10

Seamus Finnigan          9                       6

Hermione Granger         5                       10

Neville Longbottom       9                       6

Parvati Patil            5                       10

Harry Potter             9                       6

Dean Thomas              9                       6

Ron Weasley              9                       6

Hannah Abbott            5                       10

Susan Bones              5                       10

Justin Finch-Fletchley   9                       6

Ernie Macmillan            9                         6

        Terry Boot                 9                         6

        Michael Corner             9                         6

        Anthony Goldstein          9                         6

        Padma Patil                5                         10

        Millicent Bulstrode        1                         4

        Vincent Crabbe             3                         2

        Gregory Goyle              3                         2

                Table 4- E-I Index- Homophily by House- After Dumbledore’s Army

Actor                              Internal                  External

Lavender Brown                     7                         8

Seamus Finnigan                    7                         8

Hermione Granger                   7                         8

Neville Longbottom                 7                         8

Parvati Patil                      7                         8

Harry Potter                       7                         8

Dean Thomas                        7                         8

Ron Weasley                        7                         8

Hannah Abbott                      3                         12

Susan Bones                        3                         12

Justin Finch-Fletchley             3                         12

Ernie Macmillan                3                       12

Terry Boot                     3                       12

Michael Corner                 3                       12

Anthony Goldstein              3                       12

Padma Patil                    3                       12

Millicent Bulstrode            5                       0

Vincent Crabbe                 5                       0

Gregory Goyle                  5                       0

              Table 5. Closeness Centrality- Before Dumbledore’s Army

      Actor                                Closeness

      Harry Potter                         12.397

      Ron Weasley                          12.097

      Ernie Macmillan                      12

      Justin Finch-Fletchley               12

      Dean Thomas                          11.905

      Hermione Granger                     11.905

      Neville Longbottom                   11.905

      Seamus Finnigan                      11.905

      Hannah Abbott                        11.450

      Vincent Crabbe                       8.333

      Draco Malfoy                         8.333

Milicent Bulstrode                  8.333

Gregory Goyle                       8.333

Pansy Parkinson                     8.333

Lavender Brown                      6.667

Parvati Patil                       6.667

        Table 6a. Betweenness Centality- Before Dumbledore’s Army

Actor                               Betweenness

Harry Potter                        16.190

Ron Weasley                         2.857

Ernie Macmillan                     2.857

Justin Finch-Fletchley              1.905

Dean Thomas                         0

Hermione Granger                    0

Neville Longbottom                  0

Seamus Finnigan                     0

Hannah Abbott                       0

Vincent Crabbe                      0

Draco Malfoy                        0

Milicent Bulstrode                  0

Actor                    Internal               External

Gregory Goyle                       0

Pansy Parkinson                     0

Lavender Brown                      0

Parvati Patil                       0

        Table 6b. Betweenness Centrality- After Dumbledore’s Army

Lavender Brown           0   0

Seamus Finnigan          0   0

Hermione Granger         0   0

Neville Longbottom       0   0

Parvati Patil            0   0

Harry Potter             0   0

Dean Thomas              0   0

Ron Weasley              0   0

Hannah Abbott            0   0

Susan Bones              0   0

Justin Finch-Fletchley   0   0

Ernie Macmillan          0   0

Terry Boot               0   0

Michael Corner           0   0

Anthony Goldstein        0   0

Padma Patil              0   0

Millicent Bulstrode      0   0

Vincent Crabbe           0   0

Gregory Goyle            0   0

Draco Malfoy             0   0

Figure 1a. Friendship Ties Before Dumbledore's

                 Army

Figure 1b. Friendship Ties After Dumbledore's
                    Army