“We Are Only as Strong as We Are United, as Weak as We Are Divided” a Dynamic Analysis of the Peer Support Networks in the Harry Potter Books ()
1. Introduction
Since in June 1997 the first Harry Potter book was issued, the popularity of the series has steadily increased. According to the New York Times [1], more than 11 million copies of the final book were sold during the first 24 hours in the US, UK, and India alone. For the very few who are not familiar with the story of Harry Potter: Harry is an orphan living with his abusive uncle, aunt and cousin. At age 11, he is recruited by Albus Dumbledore as a wizard for the Hogwarts School of Witchcraft and Wizardry to learn magic. At Hogwarts, he finally finds himself at home, making friends (especially Ronald Weasley and Hermione Granger) and finding people who truly care for him. Throughout his education at Hogwarts, Harry learns more about his family history and his status as “the Chosen one” to defeat the evil wizard Voldemort. With the help of his friends, he faces his archenemy several times throughout the books, until a final confrontation in book 7.
As can be deducted from the short summary, one of the main topics within the series is peer support, which is also a very well established concept in psychology. Within psychology, social support has been defined in many ways [2,3]. One of the broadest definitions on social support is the one of Tardy [2]. Tardy defines social support as a concept with five dimensions: direction, disposition, description or evaluation, content and network. Direction refers to whether social support is being given or received. Disposition has two dimensions: availability (what support someone has access to, or perceived social support), and enacted (what support someone has used, or received social support). Description or evaluation points to whether an evaluation of the support or a description of the support was given. There are four different types of content in Tardy’s model of social support: emotional, instrumental, informational, and/or appraisal. Emotional support refers to support in the form of trust, love, and empathy. Instrumental support consists of resources such as money and time. Informational support refers to provided information or advice on a certain topic. Appraisal support is evaluative feedback to other individuals. The last dimension in this model of social support is network, or the source(s) or the member(s) of an individual’s support network.
So far, most researchers have focused on perceived social support, as perceived, and not enacted social support, is the most consistent and compelling indicator of health benefits [4,5]. Nonetheless, empirical associations between both dimensions of support are quite modest [6]. However, Hobfoll [7] argues that this is an artifact of the manner in which the two variables are measured. Enacted support is often assessed by means of self-report questionnaires with respect to a delimited retrospective period, and often in relation to a particular stressful experience, whereas self-report measures of perceived support allow the respondent to generalize over a wide array of social interactions over long spans of time. Furthermore, it has been argued that the modest associations between enacted social support and mental health outcomes may be in part because it is supportive interactions of which the subject is unaware (i.e., invisible support) that have the most substantial benefits for mental health [8]. In this mind, self-report questionnaires might not be the best way to assess enacted support. Observations of interactions might be a better option to assess enacted support, focusing only on the descriptive dimension of social support in Tardy’s model [9,10].
Furthermore, contemporary research examines social support entirely from the perspective of the support recipient, relating perceived social support to gender [11] and personality traits [12] of the support recipient. However, for each recipient of social support, there are one or more support providers, with their own characteristics. Furthermore, social support is likely to be reciprocal [8]. Research on social support with a reciprocal or even more general network perspective (see Tardy’s model) is necessary [10]. However, research on social support using these perspectives is still very limited [8].
Finally, research on social support in adolescence has often been cross-sectional. Nonetheless, longitudinal network studies are necessary to study developmental changes in peer support during adolescence [11]. However, so far findings on developmental changes have been contradictory. While most researchers indicate an increase of perceived peer support during adolescence [10,13], others reported that it remains stable [11].
Taking the former knowledge gaps into account, this study aims to add to the literature by focusing on enacted support during adolescence by means of a longitudinal study. The use of observations was recommended in this regard. However, observing students’ interactions with their peers in a longitudinal design is a very expensive and time-consuming enterprise.
An answer to this problem might be found in literature, which brings us back to the world of Harry Potter. Researchers in sociology of literature have long argued that texts and stories offer models for interpersonal relations [14]. Therefore, it is proposed to study stories as social facts and to apply social theories to characters and plots in the same way that those theories would be applied to real-life situations. Moreover, we can address gaps of knowledge on certain phenomena that are difficult to assess in real-life situations. The Harry Potter series seems to be particularly lenient for this purpose. Part of the popularity of the series is explained by the fact that these books sketch a context that is very familiar to Western children, touching popular themes and deep psychological issues [15,16]. Furthermore, every book narrates the adventures of Harry Potter during one school year at Hogwarts, following Harry and his fellow students from age 11 until age 18. Also, students within Hogwarts are assigned to one of the four houses, according to their personality traits. For example, Gryffindors are brave, while Hufflepuffs main characteristic is loyalty. Ravenclaws are known for their intellect and Slytherins are described as selfish and with a large interest into the Dark Arts.
Hence, based on the Harry Potter series, we would like to generate a better understanding of the development of peer support during adolescence. Therefore, we will focus on enacted support between Harry and his fellow students throughout their stay at Hogwarts. The direction of the enacted peer support will be assessed, by means of a network perspective. This study will mainly focus on the description of several types of peer support. Three general research questions will be addressed.
The first general question addresses the structural changes of peer support during adolescence, i.e., 1) What does the structure of enacted peer support look like during adolescence? More specifically, two questions will be assessed: a) Is peer support reciprocal during adolescence? and b) Is there transitivity in the enactments of peer support (e.g., I support person A and person B, so person A will also support person B)? Based on former research [17], we assume that individuals who have increasingly intimate relationships with each other, also support each other more, and that this support is reciprocal. Consequently, we assume a high rate of reciprocity in peer support during adolescence. Finally, as transitivity is often observed in friendship relations in adolescence, we also assume a positive effect of transitivity on peer support.
A second general question addresses the personal factors that might play a role in the development of peer support, i.e., 2) Which personal factors affect the development of peer support? More specifically, three research questions will be addressed: d) Are girls more involved in peer support networks than boys are? e) Does peer support increase as students grow older? and f) Is there a difference in the enactment of peer support between students with different personality traits? Based on former cross-sectional research on perceived peer support [11,13], we hypothesize that girls will be more involved in peer support networks, as they report more perceived peer support. Based on most former studies on perceived peer support [10,13], we assume that peer support increases as children age. As Rowling never mentions students repeating a grade, we can equal age to school year. Hence, we might assume a positive effect of school year on peer support. Furthermore, based on former research relating social support to certain personality traits [12], we expect that students with certain personality traits will be more involved in social support networks than others. Based on the description of the characteristics of the different houses, we might expect that Gryffindors are most supportive, followed by Hufflepuffs, Ravenclaws, and finally Slytherins.
A last general research question focuses on the presence of homophily in the development of peer support during adolescence, i.e., Does homophily affect the development of peer support? Or more specifically; Do students support students who share the same characteristics, i.e., same gender, same age, and same personality traits, more often than they support students who are dissimilar from them? Because adolescents are more likely to engage in social relationships with people who are similar to them (i.e., homophily hypothesis; [18]), they might also be more likely to give support to and receive support from these same peers who they share certain characteristics with, such as gender, age, and personality traits. Support for this hypothesis has long been available in adult women, who have been found to be more likely to provide support, and to both seek and secure support primarily from other women [19]. Therefore, we might expect that students will more often support others who share these same characteristics with them.
2. Method
2.1. Material
The Harry Potter eBooks were used as basic documents for this study [20-26]. The eBooks were coded in NVivo 9 [21], a computerized qualitative analysis program. Each author coded four books. Book 7 was coded by both authors to assess the inter-rater reliability. Interrater reliability was substantial with Kappa = 0.79 [22].
2.2. Participants
As this study focuses on the development of peer support during adolescence, only students were included in the dataset. The study was limited to the first six books as book 7 mainly concerns issues outside school grounds and peer interactions are limited. In total, 80 students were mentioned in the series. However, since these students were all in different years, we needed to indicate for each book (representing a new school year), which characters were at Hogwarts at the time. For sixteen students, no detailed information regarding the exact time they entered and/or left Hogwarts was provided. Consequently, these sixteen students were deleted from the dataset. Therefore, analyses were based on a total amount of 64 students, 41 boys and 23 girls. Twenty-five students were allocated to the Gryffindor house, 11 students were from Hufflepuff, 13 students belonged to the Ravenclaw house and 15 students were member of the Slytherin house. A large proportion of the students (40.63%), including Harry Potter, attended Hogwarts as of 1991. The remaining 38 students entered Hogwarts between 1986 and 1995 (M = 1990.94; SD = 2.09). A detailed description of the students included in the analyses is depicted in Table 1.
2.3. Peer Support
Contact between the 64 Hogwarts students was coded as peer support when one of the four types of peer support, described in Tardy’s model, were found: 1) Student A supports student B emotionally, e.g., in Book 1: Harry, Ron and Hermione assure Neville that he is definitely a Gryffindor when he doubts he is not brave enough to be part of the house; 2) Student A gives students B instrumental help; e.g., in Book 1: Fred and George Weasley help Harry Potter to get his trunk into the compartment of the Hogwarts Express; 3) Student A gives student B certain information to help student B, e.g., in Book 1: Hermione Granger helps Harry Potter with his homework and; 4) Student A praises student B, e.g., in book 5: Terry Boot praises Hermione Granger, for doing a Protean Charm, which is advanced magic.
Furthermore, two extra conditions regarding the context in which peer support appeared needed to be fulfilled as well. First, contact between students was only coded if the peer support was offered voluntarily. Second, only interactions occurring between two living characters, attending Hogwarts at the same moment, were coded as peer support. Consequently, when dead characters reappeared in the books, interactions between these dead characters and living students were not coded. One example for such reappearance is Cedric Diggory’s return at the end of book 4, when Cedric asks Harry to return his dead body to his parents. Furthermore, interactions with former or future Hogwarts students at a certain point in time were not included. For example, although Harry and Ginny met before Ginny attended Hogwarts, peer support relations between both characters were only coded when both students attended Hogwarts together.
2.4. Analysis
Stochastic actor oriented models (SOAM; [23]) were conducted by means of RSiena to analyze the emergence of support networks in the Harry Potter books. The basic idea of SOAMs is that actors choose their network connections depending on an evaluation of their own position in the network and to obtain a satisfying network
Table 1. Characters and their attributes.
configuration. A continuous time process is assumed. Therefore, a continuous Markov process taking also the non-observable time periods between the observations into account is used. During this process, actors can withdraw existing ties or create new ties at random moments. However, at any time point the actor can only change one tie [24,25]. The actors’ preferences for a satisfying network configuration can be expressed by the evaluation function, which can consists of different effects. Three effect categories can be distinguished: 1) Structural effects (e.g., reciprocity) representing network characteristic and their impact on the network formation process; 2) Actor attribute effects which relate to individual characteristics; and 3) Dyadic attribute effects which connect actors among each other [25,26]. In our analysis we included five different forms of effects; three structural effects 1)-3), one attribute effect 4), and one dyadic attribute effect 5).
1) The outdegree (density) effect is defined by the number of outgoing ties [26]. This effect must always be included [26] and is expected to be negative since it controls for the low density in social networks [27].
2) The reciprocity effect is defined by the number of reciprocated ties [26] and is assumed to be positive which means that actors aim to have reciprocated ties.
3) The transitive ties effect is defined by the number of actors to whom a person is directly as well as indirectly tied [26].
4) The covariate ego effect is defined by actors outdegree weighted by his covariate value [26]. This effect captures the impact of individual attributes like gender, school year or house membership on the network formation activity. High parameter values indicate that the covariate increases the likelihood of tie formation.
5) The same covariate effect is defined by the number of ties of one actor to all other actors who have exactly the same value on the covariate [26]. This effect assesses if similar individuals, sharing gender, school year and the same house, are more attracted to each other.
Some extra specifications were needed to suit our data. First, the network composition changed over time since students left and entered Hogwarts at different time points. Consequently, composition change models as suggested by Huisman and Snijders [28] were implemented. This method estimates the network composition of the very time point avoiding the underestimation of the effects. Second, the presence of time heterogeneity, meaning that effects are not constant over time, was to be tested.
3. Results
To get a first overview on the network dynamics, we look at the graphical outline of the students’ network at Hogwarts, depicted in Figures 1-3. The position of each actor within the graph is determined by the layout of the aggregated graph [29].
Figure 1 presents the complete network with all 64 students for the first book. In this figure, the overrepresentation of Gryffindor students is obvious. They constitute the center with the main characters Harry Potter, Hermione Granger, and Ron Weasley. Around the Gryffindor students gather most of the Ravenclaws and Hufflepuffs. All Slytherins and some less known Ravenclaws and Hufflepuffs form a circle around this center. Most actors of the circle keep unconnected throughout all books. The only exception is the Slytherin clique around Draco Malfoy. Figures 2 and 3 focus on the details of the Harry Potter and Slytherin network and show the dynamic network formation process over time.