New Drctns CHLD Adlscnt - 2019 - Hopp - Creative Thus Connected The Power of Sociometric Creativity On Friendship 4

Hopp, M. D. S., Zhang, Z. S., Hinch, L., O’Reilly, C., & Ziegler, A. (2019).
Creative, thus con-

nected: The power of sociometric creativity on friendship formation in gifted adolescents—
a longitudinal network analysis of gifted students. In R. F. Subotnik, S. G. Assouline, P.
Olszewski-Kubilius, H. Stoeger, & A. Ziegler (Eds.), The Future of Research in Talent Devel-
opment: Promising Trends, Evidence, and Implications of Innovative Scholarship for Policy and
Practice. New Directions for Child and Adolescent Development, 168, 47–73.
4
Creative, Thus Connected: The Power of
Sociometric Creativity on Friendship
Formation in Gifted Adolescents—A
Longitudinal Network Analysis of Gifted
Students
Manuel D. S. Hopp , Zhitian Skylor Zhang, Leeanne Hinch,
Colm O’Reilly, Albert Ziegler
Abstract
The current study’s purpose is to explore the influence of peer-perceived creativ-
ity (sociometric creativity) on the short-term development of friendships during
a summer program for high ability students. Specifically, the two main objectives
of our study are: (1) How did students’ friendships network and sociometric cre-
ativity network evolve in the summer program? (2) How did sociometric creativ-
ity influence the friendship formation? The longitudinal study was conducted at
the beginning, middle and the end of a 3-week long program for gifted students
in Ireland. The sample consisted of Irish gifted students (N = 702, aged 13–18
years, 52% female, over thirty-one classes). Overall, our longitudinal multilevel
and multigroup social network analysis shows that gifted adolescents formed
reciprocated friendship ties and cohesive peer group structures in the investi-
gated period; similar age and the same gender predicted friendship formation.
Regarding the sociometric creativity, they tended to nominate a similar age and
same gender student as very creative. Moreover, the sociometric creativity posi-
tively influenced adolescents’ friendship networks on a dyadic level, indicating
NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT, no. 168, November 2019 © 2019 The Authors. New Directions for Child
and Adolescent Development published by Wiley Periodicals, Inc.
Published online in Wiley Online Library (wileyonlinelibrary.com). • DOI: 10.1002/cad.20324 47
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and
reproduction in any medium, provided the original work is properly cited.
48 THE FUTURE OF RESEARCH IN TALENT DEVELOPMENT
that adolescents select friends based on their perception of the other student’s
creativity. Further results, explanations, and implications are discussed. © 2019
The Authors. New Directions for Child and Adolescent Development pub-
lished by Wiley Periodicals, Inc.
A dolescence is a crucial period to develop creative potential, in both

cognitive aspects (Kleibeuker, De Dreu, & Crone, 2016; Lassig,
2013; Stevenson, Kleibeuker, de Dreu, & Crone, 2014) and noncog-
nitive aspects (Barbot, 2018; Barbot & Heuser, 2017; Rothenberg, 1990;
Schaefer, 1973). During this crucial period, adolescents experience the
paradox between building a sense of uniqueness and seeking acceptance
from important others, especially from peers. Previous research suggests
that being creative can have both positive and negative connotations for
peer relationships (Gebart-Eaglemont & Foddy, 1994; Lau & Li, 1996; Lau,
Li, & Chu, 2004; McKay, Grygiel, & Karwowski, 2017). It can be positive
because creative adolescents are often innovative problem solvers, interest-
ing and stimulating (Lau & Li, 1996). Yet, it can be negative because they
are also potential trouble makers, disturbing and nonconforming (Lau & Li,
1996). Thus, some creativity researchers have distinguished between vari-
ous expressions of creative types such as “briefcase” and “subordinate,” as
opposed to “wild bohemian” and “rebellious” (Dawson, 1997; Karwowski,
2017). Whether being creative has a positive or negative influence on their
peer relationships may in turn play a fostering or hindering role in the
further development of their creative potential. In the case of a positive
influence, the resulting peer relationships are helpful and encouraging for
creative adolescents to build their creative identity and maintain motiva-
tion. In the case of a negative influence, the resulting peer relationships
(such as social isolation) can be discouraging for creative adolescents,
potentially leading them to conform to social norms in order to be accepted.
Consequently, as initially raised by Torrence (1961), many researchers
and educators are especially concerned of the social adaptation issue of
the creative students (Callahan & Missett, 2011; Plucker, Beghetto, &
Dow, 2004).
Does Being Creative Have a Positive or Negative Impact on Peer

Relationships?
Previous research of the link between creativity and peer relationships is
inconsistent. Some studies concluded that creativity—either measured as
creative performance or sociometric creativity (i.e., the perceived creativity
from others)—is associated with popularity (Lau & Li, 1996; McKay et al.,
2017), other studies have shown that creative individuals are often disliked
(Lau et al., 2004), or neglected (Gebart-Eaglemont & Foddy, 1994). The
varied results and the corresponding conclusions may partially be attributed
NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

CREATIVE, THUS CONNECTED: THE POWER OF SOCIOMETRIC CREATIVITY 49
to several limitations of previous research. Many of these studies have

primarily focused on the peer-perceived social status, such as “average,”
“popular,” “neglected,” “rejected,” or “controversial” (Gebart-Eaglemont &
Foddy, 1994; Lau & Li, 1996) rather than on natural setting social network
status (McKay et al., 2017). They have all been conducted at a single point
of time instead of a longitudinal period, making it hard to understand the
very nature of an association between being creative and peer relationships.
Furthermore, such studies do not allow to indicate causal direction and
may for students of different intellectual levels exacerbate consequences of
a halo effect (i.e., adolescents may perceive smart students in the classroom
as creative) (Winner, 2000).
In order to investigate the influence of sociometric creativity on the
formation of peer relationship, the present research avoids the aforemen-
tioned methodological limitations by (1) assessing the real-status social net-
work status of high sociometric creative adolescents using (2) a longitudinal
study design; (3) in a sample of comparable intellectual ability students, that
is, gifted adolescents. Specifically, the research involves three main research
questions:
(Q1) What are the fundamental features of gifted adolescents’ friendship

network formation?
(Q2) What are the fundamental features of sociometric creativity network
formation?
(Q3) How does students’ sociometric creativity influence their friendship
formation?
(Q1) What Are the Fundamental Features of Gifted

Adolescents’ Friendship Network Formation?
The first research question aims at examining the characteristics and
dynamics of gifted adolescents’ friendship network evolution in the sum-
mer school context within 3 weeks. We hypothesize that several common
features of social network formation, namely reciprocity, triadic closure and
similarity effects determine the social network evolution.
Reciprocity is a defining characteristic of friendships at any age
(Laursen & Hartup, 2002). In the summer school context, most students are
initially mutual strangers. While facing new friend choices, in general, peo-
ple prefer to be friends with those who want to be friends with them too (van
Duijn, Zeggelink, Huisman, Stokman, & Wasseur, 2003). Although not all
friendship dyads are reciprocated (Vaquera & Kao, 2008), the probability of
choosing someone else as a friend is larger if, all things being equal, one is
chosen by the other individual as well (van Duijn et al., 2003). Considerable
studies have documented the presence of reciprocity in social network stud-
ies among adolescents (Erdley & Day, 2016; Vaquera & Kao, 2008). Hence,

in the present study, we expect a significant positive reciprocity effect as the

social network among adolescents evolves (H1a).
Another common feature of social network formation is triadic closure.
People in a social network tend to be connected in a transitive way, whereby
“the friend of a friend is my friend” (Rapoport, 1953). Two students who
have a mutual friend are more likely to become friends than those who do
not share a mutual friend (Bianconi, Darst, Iacovacci, & Fortunato, 2014).
Triadic closure is often observed in the school setting, where activities are
geographically confined and face-to-face interactions prevail (D. R. Schae-
fer, Light, Fabes, Hanish, & Martin, 2010). Since the summer school pro-
gram in the present study has been organized in a face-to-face fashion with
abundant collaborative activities, we expect that students will form transi-
tive relationships, shown as the positive triadic closure effect (H1b).
Moreover, a vital force driving peer group formation is the similar-
ity effect. Research indicated that peers tend to choose friends who have
social and demographic characteristics that are similar to their own, namely
homophily (McPherson, Smith-Lovin, & Cook, 2001). Thus, we expect
that gifted students also tend to choose friends who share similar attributes
such as gender, ethnicity and age (de Klepper, Sleebos, van de Bunt, &
Agneessens, 2010; McPherson et al., 2001). Since the participants from the
present study are from the same country, we hypothesize the similarity effect
will emerge on the two key demographical characteristics: age (H1c) and
gender (H1d).
To summarize, we expect that during the 3-week summer school
period, the gifted adolescents’ social network evolution shows common
social network features: the friendship relation tends to be reciprocated;
a transitive peer network will be formed shown as triadic closure; students
tend to choose similar age and gender peers as friends.
Hypothesis 1: Gifted adolescents’ friendship network will show a signif-
icant positive effect of reciprocity (H1a), triadic closure (H1b) and similarity
on age (H1c) and gender (H1d).
(Q2) What Are the Fundamental Features of Sociometric

Creativity Network Formation?
Sociometric creativity is abbreviated from the sociometric rating of peer
perceived creativity in the present study. For instance, when student A rates
student B as highly creative, a tie from student A to student B is assumed. All
these ties and students in the course compose the sociometric creativity net-
work. The second research question aims to depict the fundamental features
of this sociometric creativity network. Specifically, whether the sociometric
creativity network also shows the fundamental social network features—
reciprocity, triadic closure and similarity effect on age and gender.
In the context of sociometric creativity network, reciprocity refers
to the tendency that both students rate each other as highly creative.

Transitivity refers to the tendency that two students who rate each other as
highly creative also share a mutual third student rated by them as highly cre-
ative. Sociometric creativity, by nature, is laypeople’s rating of each other’s
creativity level based on their implicit theories (Runco & Johnson, 2002;
Sternberg, 1985). This judgement of another person’s high creativity does
not stem from a vacuum. Instead, it might derive from the recognition of
shared daily creativity in different domains and activities such as inter-
ests and hobbies, or personal characteristics that are implicitly regarded
as highly creative such as open-minded, humorous and so on (Runco &
Johnson, 2002). In this way, it is conceivable some students might mutually
recognize each other as highly creative, therefore show patterns of recipro-
cal ties in the sociometric creativity network (H2a). Based on the devel-
opment of reciprocal ties, two students, who mutually regard one another
as highly creative, are more likely to regard a third person, who probably
shares mutual interests and personality traits, as highly creative, compared
to another student that they have not rated as highly creative. Hence, a
triadic closure might emerge (H2b).
Furthermore, another tendency that underlies participants’ rating of
high sociometric creativity might be the similarity effect. Also based on
the implicit theories of creative individuals, previous studies suggest that
students tend to regard a very creative student as similar to themselves,
such as male students tend to regards the very creative student as male;
whereas a female student regards the very creative student as female
(Hopp, Händel, Stoeger, Vialle, & Ziegler, 2016; Zhang, Hopp, Vialle, &
Ziegler, in press), namely the “same group preference.” Thus, in the present
study, we expect that students might also tend to nominate peers of sim-
ilar attributes as themselves as highly creative regarding age (H2c) and
gender (H2d).
Therefore, drawing the implicit theories of creativity, we assume that
several network features, including reciprocity, triadic closure and similar-
ity effect on age and gender will emerge as gifted adolescents’ sociometric
creativity network evolves.
Hypothesis 2: Gifted adolescents’ sociometric network will show some
of the common network features including reciprocity (H2a), triadic closure
(H2b) and similarity on age (H2c) and gender (H2d).
(Q3) How Does Students’ Sociometric Creativity Influence

Their Friendship Formation?
Previous research casts a contradictory picture on the effect of peer per-
ceived creativity on the friendship formation, both positive (Lau & Li, 1996;
McKay et al., 2017) and negative (Gebart-Eaglemont & Foddy, 1994; Lau
et al., 2004) impacts were reported. Given the dearth of research examin-
ing the influence of students’ sociometric creativity on friendship forma-
tion using longitudinal social network analysis, we leave this hypothesis

exploratory. Specifically, we will explore the effect of sociometric creativity

on friendship formation at a dyadic level (i.e., if student A thinks of student
B as very creative, whether student A tends to befriend student B because
of that) and a group level (i.e., whether having a high sociometric creativity
leads to a higher popularity in the peer group) respectively.
In addition, beside aforementioned sociometric creativity network and
some critical effects of the friendship network itself (reciprocity, transitivity
and similarity effect) that shape the friendship formation, another potential
social network effect, namely Matthew effect of popularity (“popular stu-
dent tends to be more popular overtime”) (Merton, 1968; Ripley, Snijders,
& Preciado, 2018)—is not expected but explored.
To answer above raised research questions and test related hypothe-
ses, the data analysis and results will be presented accordingly in three
parts: the characteristics and dynamics of the gifted adolescents’ friend-
ship network; the characteristics and dynamics of the sociometric creativ-
ity network; and the impact of sociometric creativity on the friendship
network.
Method
Sample. The total sample consisted of gifted students participating at
the Summer School Programme of the Centre for Talented Youth (CTY),
Ireland (N = 702, 52% female, aged 13–18 years, Mage = 15.5 years, SDage
= 1.32 years). CTY Ireland is a program for gifted students at Dublin City
University. The students participated for 3 weeks in a course of their choice,
ranging from behavioral psychology, over film studies to game theory with
a total of eighteen courses. Due to our later analyses, we only included stu-
dents that participated in courses with a high participation rate (>70%) in
our survey. Thus, our final sample consists of n = 268 students (38% of
the total sample, 57% female, Mage = 15.8 years, SDage = 1.32 years), who
participated in thirteen different courses. Each course had between twenty-
one and twenty-six students (M = 24.5, SD = 1.3). This subsample did
not significantly differ with regard to gender (t(260) = 1.47, p = .14, d =
.08); however, these students were approximately 3 months older (t(267)
= −2.67, p = .008, d = .19) than the total sample.
Procedure. Each course was taught by a teacher and a teaching
assistant. The teaching assistant handed out the questionnaires and
collected them at three time points: At the beginning of the summer camp,
after approximately 10 days and at the end (approximately 3 weeks after
the beginning).
Measures and Covariates. The students received a questionnaire in
which they were asked to provide first demographic data, then sociometric
creativity, and finally friendship relations measures for each individual stu-
dent. The sociometric creativity and friendship relationship part was intro-
duced with the following preface.

Dear student,
in this part, you will find questions about your opinions. The answers that you
provide will only be available to the researchers. Your teachers, parents or class-
mates will not see the results. Your name or the names of your classmates will not
be available to the researchers. So please be honest in your answer and complete
the questions as best as you can.
Your teacher’s assistant will now explain to you, how it will work. Please pay
attention.
Friendship Network. First, an example item was presented, where the

mechanism of the rating was explained. For the measurement, each student
had a 30-line list, where each line stood for a different student. The order of
the students was determined by the list of names that the teaching assistant
provided. The student at the top of the name list was the first student to be
rated. The teaching assistant guided the class through the process.
Each student in a course was asked to classify his or her classmates
into three types: friend; acquaintance; and, casual classmate. For that, a
questionnaire was distributed, at the beginning, during and at the end of
the summer camp with the following introductory text:
There are three types of students in your classroom. 1. Somebody who you are
familiar with and would call a friend. 2. Somebody who you know by name and
would not call a friend but an acquaintance. 3. Somebody with whom you are not
familiar at all and would call them a casual classmate. Please select the matching
type for this classmate.
If a student regarded another student as category 1 (“friend”), then the

dyadic variable was coded with “1” in the dataset, which indicates the exis-
tence of a friendship tie. Otherwise, that is, categories 2 and 3 (“acquain-
tance” and “casual classmate”) were coded with “0” as an indication of
absence of a friendship tie. From this we constructed three friendship adja-
cency matrices for each individual course, one at the beginning, one during
and one at the end of the summer camp, which encode all (non)existing
friendship nominations between all students of a given course.
Sociometric Creativity. Every student in a course had to rate his or her
classmates’ creativity on a questionnaire, once at the beginning, once during
and once at the end of the summer camp. For that, the following introduc-
tory text was presented:
Everyone can be creative in many different fields. A person can be highly creative
in one field but less creative in another field. Here we are interested in your class-
mates’ creativity in this course. What do you think, how creative is this classmate
here in this course?

At the first measurement point, we also added the following note, “Hint:
Please use the answer, which comes first into your mind.” The students could
choose one of three options: “not creative,” “somehow creative,” and “very
creative.” If student A described student B as “very creative,” the dyadic vari-
able was set to “1,” otherwise to “0.” Similarly to the friendship network, we
constructed for each course three sociometric creativity adjacency matrices,
one at the beginning, one during and one at the end of the summer camp,
which encode all (non)existing creativity nominations between all students
of a given course.
Covariates.
Age. At beginning of the summer camp, we asked students for their
birth date (month and year). From that, we calculated the age at the time
of testing.
Gender. At beginning of the summer camp, we asked students for their
gender. We coded “1” as male, and “2” as female.
Plan of Analysis. We used longitudinal social network analysis to
examine the characteristics and dynamics of the gifted adolescents’ friend-
ship network; the characteristics and dynamics of the sociometric creativ-
ity network; and the influence of sociometric creativity on the friendship
network over the course of the summer camp. Using the friendship net-
work and sociometric creativity questionnaire data, we were able to create
in total six adjacency matrices for each course, three (measurement points
T1, T2, and T3) for the friendship network and three for sociometric cre-
ativity (measurement points T1, T2, and T3). Adjacency matrices code the
dyadic relations between all students in a course. They are square matrices,
with the dimension n × n, where n stands for the number of students in
a course. Each single entry represents the dyadic relation between student
i and student j. In our case, these dyadic relationships stand for “socio-
metric creative” and “friend.” If a student nominates another student as a
friend, then this is called a (directed) tie. For example, if a course consists of
twenty-three students, the friendship-adjacency matrix to measuring point
T1 consists of 23 × 23 entries. Both the rows and the columns correspond
to the individual students (in the same order). If student 1 has only two
friends, student 5 and student 9, then the first row gets a “1” at the places 5
and 9, which marks the status “friend relationship.” All other entries of the
first row are correspondingly “0.” This coding results also in “asymmetrical”
relations, so that student 1 can indicate that student 5 is his or her friend,
but not vice versa. The totality of the ties and the nodes (i.e., the students)
can also be represented graphically, which provides a quick overview over
the network. Figure 4.1 shows two such graphs. If student A is a friend
of student B, an arrow points (i.e., directed tie) from node A to node B. If
student B is a friend of student A, a directed tie points from B to A.
We first look at the general number of nominations a random student
receives on average, then at a measure for the general connectivity within
the individual courses. One measure of the latter is density. It describes

Figure 4.1. An exemplary sociometric creativity network (left) and

friendship network (right) at measurement point T1 of one specific
course. Each arrow represents a nomination, each disk a student. For
illustrative reasons, the size of a disk is directly proportional to the
number of nomination that student received (i.e., the smaller the size,
the less nominations this student received).
the proportion of relationships observed in the network to the theoretically

maximum number of possible relationships (thus ranging between the val-
ues “0” and “1”).
For the statistical analysis of longitudinal social network data we used
RSiena (Ripley et al., 2018). RSiena incorporates a dynamic network model
family: the Stochastic Actor-Oriented Model (SAOM), which means it mod-
els change from the perspective of the actors (nodes, i.e., students). It
assumes that the change in the friendship network (as well in the sociomet-
ric creativity network) is driven by the actors (i.e., students), who create,
maintain and terminate ties (friendship relationship or sociometric creativ-
ity rating) with other actors. Those decisions (conscious or subconscious)
of actors leading to the aforementioned changes are supposed to be influ-
enced by three categories: (1) the structure of the network itself and (2)
the characteristics and behaviors of the focal actor (ego) who is making
a decision, and (3) those of other actors in the network (alters) (Ripley
et al., 2018). In the simulation process, all students in the course develop
special “preferences,” such as the structural effect that friendships are often
reciprocal, that is, if student A calls student B a friend, then student B
calls student A a friend. These “preferences” are presented in the sense of
log odds ratios (i.e., comparing the probabilities between possible network
states when sending a new tie, for more detail please refer to the RSiena
manual by Ripley et al. (2018). In addition, several networks (in the case
at hand: friendship and sociometric creativity) can influence each other.
For instance, the impression that someone is creative could have a positive

effect on the person’s tendency to become friends. An overview including

a short description of the used preferences (later only called effects) in our
analysis can be found in Table 4.1.
Discrete-time steps are used to approximate the continuous changes
over the course of observation. In each time step, a random student has the
possibility to change his or her friendship or sociometric creativity tie by
considering the current and possible induced status of his or her relation-
ships using the before mentioned “preferences.” For further information on
the model’s functionality and possibilities, see Snijders, van de Bunt, and
Steglich (2010).
In our main analysis, as we obtained data from thirteen different
courses, we conducted single network evolution dynamics estimation for
each course and aggregated the results with a meta-analysis. For that, we
used a random effects model with the restricted maximum-likelihood esti-
mator to combine the models of the courses, as suggested by Ripley et al.
(2018). For assessing heterogeneity, we used Cochran’s Q-test (Cochran,
1950) and calculated tau. The Q-test tests whether the variability in the
observed effect sizes is larger than the sampling variability. A significant Q-
test indicates a heterogeneous true effect. The other computed parameter
tau estimates the standard deviation of the distribution of true (normally
distributed) effect sizes.
For all our analyses, we used R v3.5.1 (Microsoft & R Core Team, 2018)
with the following additional packages: psych v1.8.12 (Revelle, 2018),
igraph v1.2.2 (Csardi & Nepusz, 2006), RSiena v1.2-16 (Ripley et al.,
2018), metafor v2.0-0 (Viechtbauer, 2010), and ggplot2 v3.1.0 (Wickham,
2009).
Results
Descriptives.
Sociometric Creativity and Friendship Networks. Both the mean number
of sociometric creativity (received) nominations and the mean number of
(received) friendship nominations increased steadily over the course of the
three measurement points (Table 4.2). While the average number of creativ-
ity nominations stagnates at around six nominations, the number of friend-
ship nominations increases throughout the summer camp. The two types
of nomination correlate strongly with each other over respective measure-
ment points. Strong correlations can be observed between two consecutive
measurement points within the same nomination type.
The sociometric creativity and friendship networks of thirteen courses
increasingly connected over the duration of the summer camp, indicated by
increasing network densities (see Table 4.3). During the time of the summer
camp, the density of sociometric creativity increases slightly from .22 to .26;
by comparison, the density of friendship networks doubles from .15 to .30.

Table 4.1. Short Description of all Included Effects in the RSiena

Models
Effects Short Description
RSiena Specific Effects
Rate In the simulation of RSiena, each student receives several
decision points in which he or she can either create,
maintain or resolve one of his or her friendship
relationships or sociometric creativity ratings. This rate
parameter gives information about the number of
changes regarding his or her ties.
Outdegree This effect is important to control for the density in the
network. As (Ripley et al., 2018) write in the manual for
RSiena, the specific value of the density parameter is not
very important but should always be included. However,
it should not be interpreted by itself alone since it
correlates with many other effects (Ripley et al., 2018).
Structural Effects
Reciprocity Reflects tendencies to return nominations. A positive value
means that the students tend to form reciprocal ties
with each other, that is, if student A nominates student
B, then student B tends to nominate student A too.
Triadic closure Reflects tendencies of an individual to befriend more
(GWESP) likely the friends of his or her friends. GWESP stands
for geometrically weighted edgewise shared partners
and is a measurement for triadic closure (Hunter, 2007;
Snijders, Pattison, Robins, & Handcock, 2006). If this
value is positive, then there are tendencies for triadic
closure, which means if student A has a tie with student
B and student B shares a tie with student C, then
student A tends to connect with student C and therefore
closes the triad.
Indegree—popularity Reflects tendencies to dispersion in in-degrees, that is,
those who are popular (i.e., high indegree) tend to
gather more nominations because of their popularity.
This can be called the Matthew Effect of popularity
(Merton, 1968; Ripley et al., 2018).
Indegree—popularity The same as the indegree—popularity effect with gender
× gender as a weight. This enables to see, if males or females are
more prone for the Matthew Effect.
Outdegree—popularity Mainly for model-fit reasons. Reflects the tendency to be
nominated often, if one nominates many fellow
students.
Outdegree—activity Reflects tendencies to dispersion in out-degrees of the
students, thus can be called the Matthew Effect of
self-perceived popularity. Usually included for model fit
purposes.
Covariates
(Age / Gender) Examines whether students tend to nominate other
similar / same students based on a similar age, or the same gender
respectively.
(Continued)

Table 4.1. Continued

Effects Short Description
(Age / Gender) ego A positive value means that older / female students are
more likely to nominate others than younger / male
students.
(Age / Gender) alter A positive value means that older / female students are
more likely to be nominated than younger / male
students.
Influence of sociometric
creativity on
friendship
Sociometric creativity A significant positive (negative) effect would mean, if
student A rates student B as highly creative he or she
would nominate student A more (less) likely as a friend.
Reciprocity with Mainly for model fit. If student A thinks student B is very
creativity creative, does student B nominate student A as his or
her friend?
Indegree of creativity Are students with a high sociometric creativity level (i.e.,
on popularity high in-degree in sociometric creativity network) more
popular (i.e., high in-degree in friendship network)
students?
Indegree of creativity Are students with a high sociometric creativity level (i.e.,
on activity high in-degree in sociometric creativity network)
nominating more students as friends (i.e., high
out-degree in friendship network)?
Note. The upper part of the table corresponds to effects regarding the changing friendship net-
work or the changing sociometric creativity network, and the bottom part the interaction between
sociometric creativity and friendship formation. Each part is separated by a dotted line.
Table 4.2. Means, Standard Deviations, and Pearson Correlations of

the Network Measurements (n = 268 Students)
Variable M SD 1 2 3 4 5
1. Sociometric creativity T1 5.17 4.42
2. Sociometric creativity T2 5.96 4.29 .67
3. Sociometric creativity T3 6.07 4.42 .61 .81
4. Friend nominations T1 3.50 2.92 .58 .51 .42
5. Friend nominations T2 6.06 4.12 .42 .60 .48 .75
6. Friend nominations T3 7.20 4.48 .38 .55 .56 .67 .85
Note. M and SD are used to represent mean and standard deviation, respectively. All displayed
correlations were significant on a p < .001 level.
Friendship Network Evolution. For the development of the friend-

ship nominations, a multigroup and multilevel longitudinal network anal-
ysis was carried out. Table 4.4 displays the meta-analysis results over thir-
teen converged courses, and Figure 4.2 the corresponding forest plot. In the
friendship networks, reciprocity and triadic closure effects can be observed,
β reci = 1.41, 95% CI [1.15, 1.68], p < .001; β triadic = 1.65, 95% CI [1.42,

Table 4.3. Means, Standard Deviations, and Minima and Maxima of

the Network Densities (n = 13 Courses)
Variable M SD Min Max
1. Sociometric creativity T1 .22 .06 .12 .31
4. Friend nominations T1 .15 .06 .05 .28
Note. M, SD, Min, and Max are used to represent mean, standard deviation, minimum and maximum
(in the same order) of the network densities.
Table 4.4. Multigroup RSiena Model Results of the Friendship

Network Evolution
Effect Est. SE 95% CI p tau Q pQ
Rate 1 8.61 0.66 [7.31, 9.92] < .001∗∗∗ 1.61 23.08 .027
Rate 2 5.94 0.46 [5.03, 6.84] < .001∗∗∗ 1.22 31.95 .001
Out-degree −1.15 0.64 [−2.40, 0.10] .072† 1.84 35.97 < .001
Reciprocity 1.41 0.14 [1.15, 1.68] < .001∗∗∗ 0.34 26.00 .011
Triadic closure 1.65 0.12 [1.42, 1.88] < .001∗∗∗ 0.00 15.72 .204
Age ego −0.01 0.04 [−0.08, 0.11] .698 0.01 11.30 .504
Age alter 0.01 0.05 [−0.08, 0.06] .791 0.09 17.21 .142
Similar age 0.49 0.11 [0.28, 0.70] < .001∗∗∗ 0.00 14.84 .250
Gender ego 0.03 0.13 [−0.15, 0.30] .807 0.18 15.27 .227
Gender alter 0.07 0.11 [−0.22, 0.28] .523 0.23 16.94 .152
Same gender 0.29 0.07 [0.15, 0.42] < .001∗∗∗ 0.12 16.93 .152
Indegree-popularity −0.38 0.24 [−0.86, 0.10] .117 0.64 28.54 .005
Indegree-popularity 0.09 0.07 [−0.04, 0.22] .191 0.00 8.66 .731
× gender
Outdegree- −0.24 0.11 [−0.45, −0.03] .024∗ 0.18 12.44 .411
popularity
Outdegree-activity 0.02 0.10 [−0.17, 0.20] .853 0.25 27.56 .006
Note. n = 13 courses, all convergence t ratios < .05, all overall maximum convergence ratio < .16;
† p <.10, ∗p < .05, ∗∗p < .01, ∗∗∗p < .001 (each two-sided). Note that all covariates are internally
centered.
1.88], p < .001 respectively. The positive reciprocity effect means that if
student A nominates student B as a friend, then student B tends to nominate
student A as a friend. The positive triadic closure effect means that if student
A nominates student B as a friend and student B nominates student C, then
student A tends to nominate student C as a friend, too. Thus, we accept
both hypotheses H1a and H1b.
The positive “similar age” (β = 0.49, 95% CI [0.28, 0.70], p < .001)
and “same gender” (β = 0.29, 95% CI [0.15, 0.42], p < .001) effects indicate
that students tend to prefer peers of a similar age and same gender as friends.

Figure 4.2. Forest plot of the effect estimates of the friendship

network evolution. Black filled circles correspond to the effect
estimates and the length of the line corresponds to the 95%
confidence interval.
Thus, we accept both hypotheses H1c and H1d. The significant Cochran’s
Q-test values of the reciprocity effect (Q = 26.00, pQ = .011) suggest that
there is strong evidence against the null hypothesis that the variability of
that effect is due to random sampling variation.
We did not find a Matthew effect regarding popularity. Specifically,
we found an insignificant negative value of the related effect “indegree-
popularity,” β = −0.38, 95% CI [−0.86, 0.10], p = .12. This indicates
that popular students do not gain more friends because of their popularity
over time. Moreover, we also examined whether or not gender moderates
the Matthew effect. Based on our coding (male = “1,” female = “2”), the
result revealed no significant gender difference as to the Matthew effect
(“indegree-popularity × gender,” β = 0.09, 95% CI [−0.04, 0.22], p = .19).
In this regard, there is no evidence that a potential Matthew effect shapes
the friendship formation.

Table 4.5. Multigroup RSiena Model Results of Sociometric Creativity

Network Evolution
∗∗∗
Rate 1 11.70 0.99 [9.77, 13.63] < .001 2.58 38.09 < .001
Rate 2 8.09 0.39 [7.32, 8.85] < .001∗∗∗ 0.60 13.90 .307
Out-degree −1.50 0.14 [−1.77, −1.23] < .001∗∗∗ 0.27 18.64 .098
Reciprocity 0.17 0.05 [0.07, 0.27] .001∗∗∗ 0.00 5.59 .935
Triadic closure 0.80 0.10 [0.61, 0.99] < .001∗∗∗ 0.10 14.54 .268
Age ego 0.00 0.03 [−0.06, 0.05] .894 0.04 13.26 .350
Age alter 0.08 0.02 [0.03, 0.13] .001∗∗ 0.07 25.98 .011
Similar age 0.19 0.08 [0.03, 0.34] .020∗ 0.05 12.36 .418
Gender ego 0.06 0.08 [−0.08, 0.17] .470 0.21 27.24 .007
Gender alter 0.05 0.06 [−0.10, 0.21] .463 0.15 20.86 .052
Same gender 0.14 0.04 [0.06, 0.23] .001∗∗∗ 0.00 6.17 .908
× gender
∗p < .05, ∗∗p < .01, ∗∗∗p < .001 (each two-sided). Note that all covariates are internally centered.
Sociometric Creativity Network Evolution. Alongside the evolution

of the friendship network, the evolution of the sociometric creativity net-
works was also simulated. Table 4.5 displays the results of the model, and
Figure 4.3 the corresponding forest plot. The development of creativity
networks is similar to that of the friendship network. The reciprocity and
triadic closure effects are also significant, though weaker than in the friend-
ship network, especially the reciprocity effect (β reci = 0.17, 95% CI [0.07,
0.27], p = .001; β triadic = 0.80, 95% CI [0.61, 0.99], p < .001, respectively).
Based on the evidence we accept both hypotheses H2a and H2b.
Moreover, students of a similar age or with the same gender are more
likely to be described as very creative (“similar age,” β = 0.19, 95% CI [0.03,
0.34], p = .020; “same gender,” β . = 0.14, 95% CI [0.06, 0.23], p < .001
respectively). Thus, we accept both hypotheses H2c and H2d. However,
these effects are again less pronounced than in the friendship networks. Fur-
thermore, at a group level, older students are more likely to be nominated as
“very creative” than younger students by all potential peers, as indicated by
the positive significant “age alter” effect (β = 0.08, 95% CI [0.03, 0.13], p
= .001). None of the significant effects shows signs of increased variability
between different courses.
The result of the Matthew effect regarding sociometric creativity was
insignificant (“indegree—popularity,” β = −0.03, 95% CI [−0.22, 0.17],
p = .78). Moreover, taking gender into consideration, the result pro-
vided no evidence of gender moderation with respect to the Matthew
effect (“indegree-popularity × gender,” β = −0.11, 95% CI [−0.25, 0.03],
p = .13). Hence, there is no evidence that the Matthew effect did occur
while the sociometric creativity network evolved.

Figure 4.3. Forest plot of the effect estimates of the sociometric

creativity network evolution. Black filled circles correspond to the
effect estimates and the length of the line corresponds to the 95%
confidence interval.
Influence of Sociometric Creativity on Friendship Network Evolu-

tion. For the main research question on how sociometric creativity influ-
ences the friendship formation, the interaction between the two networks
was included in the RSiena model. The results are displayed in Table 4.6,
and the corresponding forest plot in Figure 4.4. On a dyadic level, sociomet-
ric creativity has a positive influence on the friendship formation (“socio-
metric creativity,” β = 0.28, 95% CI [0.02, 0.55], p = .035), that is, if student
A thinks of student B as very creative, then student A tends to befriend
student B because of that.
On the group level, as shown in the descriptive correlation in Table 4.2,
there were positive correlations between sociometric creativity and pop-
ularity (i.e., received friendship nominations) at T1, T2 and T3 (on a
r = .6 level). However, the crucial question is whether this positive asso-
ciation also indicates a causal relationship on a group level. Indeed, our
analyses did not support indications for a causal effect, that is, having a high
sociometric creativity does not lead to a higher popularity (i.e., a higher

Table 4.6. Multigroup RSiena Model Results of the Influence of

Sociometric Creativity on the Friendship Network Evolution
∗
Sociometric creativity 0.28 0.13 [0.02, 0.55] .035 0.11 12.19 .430
Reciprocity with 0.45 0.13 [0.19, 0.71] .001∗∗∗ 0.00 5.51 .939
creativity
Indegree of creativity 0.03 0.07 [−0.11, 0.16] .698 0.00 2.06 .999
on popularity
Indegree of creativity −0.45 0.08 [−0.62, −0.29] < .001∗∗∗ 0.00 6.72 .876
on activity
∗ p < .05, ∗∗∗ p < .001 (each two-sided).
Figure 4.4. Forest plot of the effect estimates of the influence of

sociometric creativity network evolution on friendship network
evolution. Black filled circles correspond to the effect estimates and
the length of the line corresponds to the 95% confidence interval.
friendship in-degree; “indegree of creativity on popularity,” β = 0.03, 95%

CI [−0.11, 0.16], p = .70).
The number of (given) friendship nominations of students who had
higher sociomeric creativity at T1 slowed down from T1 to T2; and T2 to
T3 (“indegree of creativity on activity”; β = −0.45, 95% CI [−0.62, −0.29],
p < .001), compared with the students with lower sociometric creativity.
From that, we can conclude that the higher the sociometric creativity status
of a student, the less friends he or she nominates because of that. This effect
was included mainly for model fit reasons.
For model fit purposes, the included reciprocity effect between socio-
metric creativity and friendship was positive and significant (“reciprocity
with creativity”; β = 0.45, 95% CI [0.19, 0.71], p = .001), meaning if
student A thinks that student B is very creative, then student B tends to

nominate student A as a friend because of that. There was no heterogeneity

regarding the effects between the courses.
Discussion
Creative expressions take often place in dynamic social environments
(Beghetto, 2019). The reported empirical study is the first that depicts the
dynamic coevolving of sociometric creativity and peer relationship with
a sample of gifted adolescents based on a longitudinal design. It captured
several fundamental features of the dynamic evolution of the friendship
network and sociometric creativity network over the course of the summer
camp. Specifically, gifted adolescents formed reciprocated friendship ties
and cohesive peer group structures in the investigated period. Similar
age and the same-gender predicted friendship formation. Moreover, gifted
adolescents nominated a similar age and same gender student as very
creative. Last and most importantly, the perceived very creative ones were
more likely to be befriended on a dyadic level in peers, thus demonstrating
the positive influence of sociometric creativity in shaping adolescents’
friendship networks.
Friendship Network Evolution. The students formed friendships in
line with common social network features including reciprocity and tri-
adic closure, as shown in Table 4.4. Similar to previous studies (Ellwardt,
Steglich, & Wittek, 2012; Kiuru, Burk, Laursen, Salmela-Aro, & Nurmi,
2010; Knecht, Burk, Weesie, & Steglich, 2011), gifted adolescents formed
reciprocated friendship ties and cohesive peer group structures even within
a short period of 3 weeks.
Moreover, students nominated friends of similar age and the same gen-
der, which is in line with previous findings demonstrating that adolescents’
friendships were formed or maintained based on similarity such as gen-
der and ethnicity (Rambaran, Dijkstra, Munniksma, & Cillessen, 2015).
In addition, students’ age and gender had not been associated with how
many friends’ nominations they gave or received. This finding complements
former studies, where such a trend had been sometimes found (e.g., Burk,
Kerr, & Stattin, 2015), but sometimes not (e.g., Mercken, Snijders, Steglich,
Vartiainen, & de Vries, 2010; Rambaran et al., 2015).
We did not detect a Matthew effect (Merton, 1968; Perc, 2014) regard-
ing popularity. This might be due to the fact that a classroom is a relatively
closed social network setting, in which students are easier to reach an upper
limit of maximum friendship ties (van Duijn et al., 2003). In contrast, in
more open social network settings such as in social media, the Matthew
effect that the popular ones get more popular is more often detected (Perc,
2014).
In summary, in line with our hypothesis 1, the reciprocity (H1a),
transitivity (H1b), similar age (H1c) and same gender (H1d) were potent

predictors of friendship ties for gifted students’ friendship formation within

a period of 3 weeks.
Sociometric Creativity Network Evolution. Comparable to friend-
ship network, reciprocity and triadic closure were prominent features of
sociometric creativity, although the effects are smaller (.17 vs. 1.41, and .80
vs. 1.65 respectively, compare Tables 4.4 and 4.5). This suggests that ado-
lescents might mutually recognize shared interests (such as art or music) or
traits (such as openness and humor) related to creativity, which yields sig-
nificant mutual and triadic sociometric creativity “hits” among peers (Ram-
baran et al., 2015).
In addition, students’ gender does not influence how many nomina-
tions as “very creative” they give or receive. Younger and older students
also gave a similar amount of “very creative” nominatives. Interestingly,
at a group level, older students are more likely to be nominated as “very
creative” than younger students by all potential peers. Many factors may
contribute to this. For instance, older students seem more mature, and
therefore may appear more knowledgeable, which is an indicator of being
creative (Baer, 2015).
At a dyadic level, the similarity tendency in the friendship network also
applies to sociometric creativity regarding age and gender, that is, students
nominated other students with similar age and the same gender as very
creative. This “same group preference” result is consistent with Hopp et al.
(2016) and Zhang et al. (in press), in which adolescents imagine and expect
the creative peer as the same gender as themselves.
Taken together, as expected in hypothesis 2, these findings suggest reci-
procity (H2a) and triadic closure (H2b) were prominent features of socio-
metric creativity; gifted adolescents tend to nominate a similar age (H2c)
and same gender (H2d) student as very creative.
Influence of Sociometric Creativity on Friendship Network Evo-
lution. Our result showed evidence of a positive influence of sociometric
creativity in shaping adolescents’ friendship networks on a dyadic level.
Adolescents seem to select friends based on the perception of the other
student’s creativity, that is, if student A thinks that student B is very cre-
ative then student A tends to nominate student B as a friend (or keeps
him or her as a friend). In addition, on a group level, although students’
high sociometric creativity is positively associated with popularity from the
beginning to the end of the summer school, the more restrict testing of a
causality-indicating relationship was not supported. This means that there
is no indication that high sociometric creativity increases a student’s pop-
ularity over time. A possible explanation might be that there are one or
more underlying factors that lead students to regard some students as both
creative and popular, which could not be detected in the present study.
The finding that on a dyadic level students befriend each other based
on their perception of high creativity might relieve the concern from pre-
vious research that creative adolescents tend to be socially isolated in the

peer group: either neglected or even disliked (Gebart-Eaglemont & Foddy,

1994; Lau et al., 2004). Contrary to these results, we have found that ado-
lescents who are perceived as very creative tend to be befriended at a dyadic
level, which is in line with research findings that creative adolescents hold
higher social status (Lau & Li, 1996; McKay et al., 2017). These inconsistent
results with previous studies might be due to numerous research method-
ological reasons, which we have explained in the introductory section: the
real-setting social network status versus peer perceived social status; the
longitudinal study versus the study conducted at one point in time; con-
trolled versus uncontrolled halo effect.
Real Social Network Versus Peer Perceived Status. A common way to
investigate the peer relationship from the previous research is to use the peer
perceived status as “average,” “popular,” “neglected,” “rejected,” or “con-
troversial” (Gebart-Eaglemont & Foddy, 1994; Lau & Li, 1996). Although
this approach yields the general judgement of whether creative adolescents
are popular in the peer group, it assesses the peer relationship in an indirect
way. Take a pop singer as an example. He or she differs between perceived
popularity and real-life friendship, that is, the singer might have many fans
(high perceived popularity), whereas few friends that he or she can con-
tact in daily life (friendship tie). The real-life friendship tie, rather than
peer-perceived popularity, effects people socially and emotionally. From this
perspective, the real-setting social network status might be a more proper
proxy of peer relationship for creative adolescents for judging whether they
are popular or isolated.
Longitudinal Study Versus Static Time Point. Almost all the previous
studies have been conducted at a static time point, which depicts the asso-
ciation relation between creativity and popularity, rather than a casual rela-
tion. For example, the finding of Lau and Li (1996) that popular children
were judged as the most creative. Merely from the association, it is hard to
tell whether peers regarded some students as popular, and therefore judge
them as intelligent and creative, or the other way round. The conclusion
“creative children may have easier social development” (Lau & Li, 1996),
therefore, seems not convincing, because it may also be the case that chil-
dren with better social relationships are regarded as more creative, thus
reflecting a halo effect. The results from the present longitudinal research
allows for a more reliable causal indication that high sociometric creativity
influences friendship ties—not the other way around, because we observed
the friendship formation of creative adolescents from the very beginning of
the summer program—however, causal interpretations should be supported
by further results from future studies (Ripley et al., 2018).
Halo Effect. Another potential aspect which may exaggerate the halo
effect is that most studies didn’t control intellectual abilities of the sam-
ple. Indeed, only one Chinese study used only high ability students as par-
ticipants (Lau et al., 2004). Considering the overlap between lay people’s
implicit theories regarding intelligence and creativity (Hopp et al., 2016;

Sternberg, 1985), smarter students might be perceived as more creative

and popular. The present study conducted with only gifted students in the
school context, aimed to reduce the halo effect. Although it was not pos-
sible to completely exclude the halo effect, the gifted adolescents group
is relatively more homogenous than the students from normal classrooms.
However, whether the halo effect is really reduced, considering these gifted
students are different from each other, needs more studies to verify. There-
fore, the conclusion—adolescents with high sociometric creativity obtain
friends—might be more robust, compared to previous studies.
In addition to the three aspects mentioned above, the society’s zeit-
geist might also play a role in producing inconsistent results. For example,
most studies related to creative adolescents and their peer status were con-
ducted in the second half of the twentieth century and the first decade of
twenty-first century. Since creativity is nowadays considered as one of the
most crucial twenty-first century skills, society might gradually be shift-
ing toward a more open and tolerant environment for creative individuals.
Thus, past societies might not have emphasized the value of creativity or
innovation like today. Under this situation, students might have previously
held a more negative image of creative individuals. In contrast, nowadays
people’s perception and attitudes toward creative people might be aligned
with the zeitgeist of society, which influences the conclusions and results
of creativity research.
Based on peer relationship research, actors in the social network form
relationships based on the selection processes, in which they structure—
consciously or unconsciously—their networks based on other actors’
attributes (Rambaran et al., 2015). In our study, we found that the attribute
of ‘high sociometric creativity’ plays a positive role in adolescents’ friend-
ship formation. This finding does not necessarily conflict with Yamamoto’s
(1964) conclusion that creativity, per se, does not play a decisive role in
the choice of best friends. Indeed, sociometric creativity should not be con-
fused with creativity. Specifically, in the context of the present study—the
Irish gifted students in a summer school program—the perceived creative
ones obtain more friends. Unlike previous studies, we have corroborated
this conclusion with a longitudinal approach; with an intellectually gifted
sample. This result suggests that gifted adolescents might have an advantage
of gaining social impact by virtue of their sociometric creativity.
Limitations and Future Research. Several limitations of the present
study need to be mentioned. First, this study was based on a sample of gifted
adolescents from Irish schools. It is an open question if these findings can
be generalized to another sample, culture or social context. Considering
the sociometric creativity is based on laypeople’s implicit theory of creative
individuals, its nature (i.e., what is creativity and who is creative) differs
across samples with distinct characteristics such as profession (Spiel & Von
Korff, 1998) and country (Spiel & Von Korff, 1998; Tang, Werner, & Hofre-
iter, 2018). The impact of sociometric creativity on friendship formation,

therefore, might also disparate in different contexts considering domains

and fields (Csikszentmihalyi, 1988; Sawyer, 2006). Therefore, the individ-
ual and contextual aspects should be taken into consideration while gener-
alizing the findings of this study.
Second, the conclusion of the present study is based on the quantity,
not the quality of the friendship ties. Specifically, the social network analysis
was conducted by categorizing friendship ties as dichotomous, that is, hav-
ing a friendship or having no friendship. The quality of the friendship ties,
such as positive or negative ties (raging from endorsement to opposition),
or the friendship relations with different strengths (raging from a friendly
person to a best friend relationship) was not allowed to be examined with
RSiena (at the time of the current study). It is possible that a more nuanced
examination of friendship ties quality might yield different results. In the
future, it would be worthwhile to examine the impact of high sociometric
creativity on both the quantity and quality of adolescents’ social relation-
ships.
Third, given the participants in the present study were grouped in the
course based on their choice of course content, this might lead to some
bias in the distribution of adolescents’ characteristics or interests related
to the course subjects. Although we conducted a meta-analysis to combine
the results from multiple courses, we did not investigate possible effects of
course level such as categorizing the subjects as art or science. One reason is
that many courses are ambiguous to be categorized as art or science (such
as “Behavior psychology” or “Law”); another reason is that it demands a
bigger course group size than in the present study in order to achieve a
meaningful comparison. However, still, despite the potential subject level
sampling bias, a preliminary overview of the results in various courses
points to the direction of homogeneous rather than heterogeneity among
different courses concerning investigated research questions (as indicated
by the predominantly nonsignificant Q statistics in the meta-analysis; see
Tables 4.4, 4.5, and 4.6). Future studies may consider a multilevel approach
and investigate the role of course characteristics and domain differences
regarding the impact of sociometric creativity on friendship formation.
Finally, although it is beyond the scope of the present study, future
studies may dig deeper into the reasons why adolescents nominate some
students as very creative, thus exploring the mechanism underlying socio-
metric creativity. Specifically, whether some students are perceived as very
creative because they are “really” creative, according to the relatively objec-
tive creativity measurements such as performance on creativity tests and
creative behavior checklists (Kaufman, Plucker, & Baer, 2008), or because
of something else such as personality or even appearance? Such explo-
rations, then, would give further insights into improved strategies to gain
sociometric creativity and social impact.
Implications. This research made use of a longitudinal design, which
starts at the point of time when friendships were newly formed. This

provided a more ecologically valid representation of adolescent friendships,

compared with the majority of previous studies, which had limited them-
selves to a static time point, or used already existed or stable friendships.
The longitudinal design enabled us in particular to examine the dynamic
friendship formation, the characteristics of sociometric creativity, as well as
the interaction between each other in a natural setting.
With regard to the issue if perceived creativity is detrimental with
regard to friendship formation, the present data seems, in light of the
methodological characteristics of the study, especially valuable. They sup-
port clearly the hypothesis that sociometric creativity plays a positive
role in friendship formation. This result has, in turn, several practical
implications. Peer relationship studies have identified many characteris-
tics as influencing factors of friendship formation. Our study suggests
that being perceived as very creative might also be a potential way to
“win friends and influence people.” For gifted students themselves, they
can “dare” to be creative because they might be more likely to befriend
with. Many specific consequences for education and in particular gifted
education follow from this insight. For example, practitioners who orga-
nize summer schools for gifted students, it might be promising to include
creativity-centered activities. Considering that adolescents are more likely
to adjust their behavior to the behaviors of high-status peers (Rambaran
et al., 2015), it might be possible to set role models of creative ado-
lescents in the peer group. In this way, positive behaviors and attitudes
toward being creative may spread via influences process in friendship
networks.
Conclusion
Taken together, our findings demonstrate that high sociometric creativity
yields higher probability to be befriended in gifted adolescents on a dyadic
level. Creativity does not necessarily sacrifice friendships. Indeed, gifted
adolescents who are perceived as creative seem to be more attractive as
friends for their peers. Therefore, educators might find it useful to encour-
age gifted students to dare to be creative.
References
Baer, J. (2015). The importance of domain-specific expertise in creativity. Roeper Review,
37(3), 165–178. https://doi.org/10.1080/02783193.2015.1047480
Barbot, B. (2018). Creativity and self-esteem in adolescence: A study of their domain-
specific, multivariate relationships. Journal of Creative Behavior. Advance online pub-
lication. https://doi.org/10.1002/jocb.365
Barbot, B., & Heuser, B. (2017). Creativity and identity formation in adolescence: A
developmental perspective. In M. Karwowski & J. C. Kaufman (Eds.), The creative
self: Effect of beliefs, self-efficacy, mindset, and identity (pp. 87–98). San Diego, CA:
Elsevier Academic Press. https://doi.org/10.1016/B978-0-12-809790-8.00005-4

Beghetto, R. A. (2019). Creativity in classrooms. In J. C. Kaufman & R. J. Sternberg

(Eds.), Cambridge handbook of creativity(2nd ed., pp. 587–606). New York: Cambridge
University Press.
Bianconi, G., Darst, R. K., Iacovacci, J., & Fortunato, S. (2014). Triadic closure as a
basic generating mechanism of communities in complex networks. Physical Review
E: Statistical, Nonlinear, and Soft Matter Physics, 90(4), 042806. https://doi.org/10.
1103/PhysRevE.90.042806
Burk, W. J., Kerr, M., & Stattin, H. (2015). The co-evolution of early adolescent friend-
ship networks, school involvement, and delinquent behaviors. Revue Française de
Sociologie, 49(3), 499. https://doi.org/10.3917/rfs.493.0499
Callahan, C. M., & Missett, T. C. (2011). Creativity in adolescence. In Encyclopedia
of adolescence (pp. 115–123). Amsterdam, Netherlands: Elsevier. https://doi.org/10.
1016/B978-0-12-373951-3.00009-0
Cochran, W. G. (1950). The comparison of percentages in matched samples. Biometrika,
37(3/4), 256. https://doi.org/10.2307/2332378
Csardi, G., & Nepusz, T. (2006). The igraph software package for complex network
research. InterJournal, Complex Systems, 1–9. Retrieved from http://igraph.org
Csikszentmihalyi, M. (1988). Society, culture, and person: A systems view of creativity.
In R. J. Sternberg (Ed.), The nature of creativity: Contemporary psychological perspec-
tives (pp. 325–339). New York: Cambridge University Press.
Erdley, C. A. & Day, H. J. (2016). Friendship in childhood and adolescence. In M. Hoj-
jat & A. Moyer (Eds.), The Psychology of Friendship (Vol. 6). Oxford, UK: Oxford
University Press. https://doi.org/10.1093/acprof:oSo/9780190222024.001.0001
Dawson, V. L. (1997). In search of the wild bohemian: Challenges in the iden-
tification of the creatively gifted. Roeper Review, 19(3), 148–152. https://doi.org/
10.1080/02783199709553811
de Klepper, M., Sleebos, E., van de Bunt, G., & Agneessens, F. (2010). Simi-
larity in friendship networks: Selection or influence? The effect of constrain-
ing contexts and non-visible individual attributes. Social Networks, 32(1), 82–90.
https://doi.org/10.1016/j.socnet.2009.06.003
Ellwardt, L., Steglich, C., & Wittek, R. (2012). The co-evolution of gossip and
friendship in workplace social networks. Social Networks, 34(4), 623–633. https://
doi.org/10.1016/j.socnet.2012.07.002
Gebart-Eaglemont, J. E., & Foddy, M. (1994). Creative potential and the socio-
metric status of children. Creativity Research Journal, 7(1), 47–57. https://doi.org/
10.1080/10400419409534508
Hopp, M. D. S., Händel, M., Stoeger, H., Vialle, W., & Ziegler, A. (2016). A cross-
national study of implicit theories of a creative person. Education Sciences, 6(4), 1–14.
https://doi.org/10.3390/educsci6040038
Hunter, D. R. (2007). Curved exponential family models for social networks. Social
Networks, 29(2), 216–230. https://doi.org/10.1016/j.socnet.2006.08.005
Karwowski, M. (2017). Subordinated and rebellious creativity at school. In R. Beghetto
& B. Sriraman (Eds.), Creative contradictions in education. Creativity theory and
action in education (pp. 89–113). Cham, Switzerland: Springer. https://doi.org/
10.1007/978-3-319-21924-0_6
Kaufman, J. C., Plucker, J. A., & Baer, J. (2008). Essentials of creativity assessment. In
A. S. Kaufman & N. L. Kaufman (Series Eds.), Essentials of Psychological Assessment
series. Hoboken, NJ: Wiley.
Kiuru, N., Burk, W. J., Laursen, B., Salmela-Aro, K., & Nurmi, J.-E. (2010). Pres-
sure to drink but not to smoke: Disentangling selection and socialization in adoles-
cent peer networks and peer groups. Journal of Adolescence, 33(6), 801–812. https://
doi.org/10.1016/j.adolescence.2010.07.006

Kleibeuker, S. W., De Dreu, C. K. W., & Crone, E. A. (2016). Creativity development

in adolescence: Insight from behavior, brain, and training studies. In B. Barbot (Ed.),
Perspectives on Creativity Development. New Directions for Child and Adolescent Devel-
opment, 151, 73–84. https://doi.org/10.1002/cad.20148
Knecht, A. B., Burk, W. J., Weesie, J., & Steglich, C. (2011). Friendship and alcohol
use in early adolescence: A multilevel social network approach. Journal of Research on
Adolescence, 21(2), 475–487. https://doi.org/10.1111/j.1532-7795.2010.00685.x
Lassig, C. J. (2013). Approaches to creativity: How adolescents engage in the cre-
ative process. Thinking Skills and Creativity, 10, 3–12. https://doi.org/10.1016/
j.tsc.2013.05.002
Lau, S., Li, C. S., & Chu, D. (2004). Perceived creativity: Its relationship to social sta-
tus and self-concept among Chinese high ability children. Creativity Research Journal,
16(1), 59–67. https://doi.org/10.1207/s15326934crj1601_6
Lau, S., & Li, W. L. (1996). Peer status and perceived creativity: Are popular children
viewed by peers and teachers as creative? Creativity Research Journal, 9(4), 347–352.
https://doi.org/10.1207/s15326934crj0904_6
Laursen, B., & Hartup, W. W. (2002). The origins of reciprocity and social exchange
in friendships. In W. G. Graziano & B. Laursen (Eds.), Social Exchange in
Development. New Directions for Child and Adolescent Development, 95, 27–40.
https://doi.org/10.1002/cd.35
McKay, A. S., Grygiel, P., & Karwowski, M. (2017). Connected to create: A social net-
work analysis of friendship ties and creativity. Psychology of Aesthetics, Creativity, and
the Arts, 11(3), 284–294. https://doi.org/10.1037/aca0000117
McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather:
Homophily in social networks. Annual Review of Sociology, 27(1), 415–444.
https://doi.org/10.1146/annurev.soc.27.1.415
Mercken, L., Snijders, T. A. B., Steglich, C., Vartiainen, E., & de Vries, H. (2010). Dynam-
ics of adolescent friendship networks and smoking behavior. Social Networks, 32(1),
72–81. https://doi.org/10.1016/j.socnet.2009.02.005
Merton, R. K. (1968). The Matthew effect in science: The reward and com-
munication systems of science are considered. Science, 159(3810), 56–63.
https://doi.org/10.1126/science.159.3810.56
Microsoft & R Core Team. (2018). Microsoft R Open. Redmond, WA: Microsoft. Retrieved
from https://mran.microsoft.com/
Perc, M. (2014). The Matthew effect in empirical data. Journal of The Royal Society Inter-
face, 11(98), 20140378. https://doi.org/10.1098/rsif.2014.0378
Plucker, J. A., Beghetto, R. A., & Dow, G. T. (2004). Why isn’t creativity more
important to educational psychologists? potentials, pitfalls, and future direc-
tions in creativity research. Educational Psychologist, 39(2), 83–96. https://doi.org/
10.1207/s15326985ep3902_1
Rambaran, J. A., Dijkstra, J. K., Munniksma, A., & Cillessen, A. H. N. (2015).
The development of adolescents’ friendships and antipathies: A longitudinal
multivariate network test of balance theory. Social Networks, 43, 162–176.
Rapoport, A. (1953). Spread of information through a population with socio-structural
bias: I. Assumption of transitivity. Bulletin of Mathematical Biophysics, 15(4), 523–533.
https://doi.org/10.1007/BF02476440
Revelle, W. (2018). psych: Procedures for psychological and personality research
[R package]. Evanston, IL: Northwestern University. Retrieved from
https://CRAN.R-project.org/package=psychVersion=1.8.12
Ripley, R. M., Snijders, T. A. B., & Preciado, P. (2018). Manual for RSiena. Retrieved from
http://www.stats.ox.ac.uk/∼snijders/siena/RSiena_Manual.pdf

Rothenberg, A. (1990). Creativity in adolescence. Psychiatric Clinics of North America,

13(3), 415–434.
Runco, M. A., & Johnson, D. J. (2002). Parents’ and teachers’ implicit theories of chil-
dren’s creativity: A cross-cultural perspective. Creativity Research Journal, 14(3–4),
427–438. https://doi.org/10.1207/S15326934CRJ1434_12
Sawyer, R. K. (2006). Explaining creativity: The science of human innovation. Oxford, UK:
Oxford University Press.
Schaefer, C. E. (1973). A five-year follow-up study of the self-concept of creative
adolescents. Journal of Genetic Psychology, 123(1), 163–170. https://doi.org/10.1080/
00221325.1973.10533200
Schaefer, D. R., Light, J. M., Fabes, R. A., Hanish, L. D., & Martin, C. L. (2010). Funda-
mental principles of network formation among preschool children. Social Networks,
32(1), 61–71. https://doi.org/10.1016/j.socnet.2009.04.003
Snijders, T. A. B., Pattison, P. E., Robins, G. L., & Handcock, M. S. (2006). New specifica-
tions for exponential random graph models. Sociological Methodology, 36(1), 99–153.
https://doi.org/10.1111/j.1467-9531.2006.00176.x
Snijders, T. A. B., van de Bunt, G. G., & Steglich, C. E. G. (2010). Introduction to
stochastic actor-based models for network dynamics. Social Networks, 32(1), 44–60.
Spiel, C., & Von Korff, C. (1998). Implicit theories of creativity: The conceptions of
politicians, scientists, artists and school teachers. High Ability Studies, 9(1), 43–58.
https://doi.org/10.1080/1359813980090104
Sternberg, R. J. (1985). Implicit theories of intelligence, creativity, and wisdom. Jour-
nal of Personality and Social Psychology, 49(3), 607–627. https://doi.org/10.1037/
0022-3514.49.3.607
Stevenson, C. E., Kleibeuker, S. W., de Dreu, C. K. W., & Crone, E. A. (2014). Training
creative cognition: Adolescence as a flexible period for improving creativity. Frontiers
in Human Neuroscience, 8, 1–16. https://doi.org/10.3389/fnhum.2014.00827
Tang, M., Werner, C. H., & Hofreiter, S. (2018). Creativity alone does not make a star:
Social attributes of the nomination of creative icons: Results of a trend study in Ger-
many. Frontiers in Psychology, 9, 1–13. https://doi.org/10.3389/fpsyg.2018.01944
Torrence, E. P. (1961). Problems of highly creative children. Gifted Child Quarterly, 5(2),
31–34. https://doi.org/10.1177/001698626100500201
van Duijn, M. A. J., Zeggelink, E. P. H., Huisman, M., Stokman, F. N., & Wasseur, F. W.
(2003). Evolution of sociology freshmen into a friendship network. Journal of Math-
ematical Sociology, 27(2–3), 153–191. https://doi.org/10.1080/00222500305889
Vaquera, E., & Kao, G. (2008). Do you like me as much as I like you? Friendship reci-
procity and its effects on school outcomes among adolescents. Social Science Research,
37(1), 55–72. https://doi.org/10.1016/j.ssresearch.2006.11.002
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Jour-
nal of Statistical Software, 36(3). https://doi.org/10.18637/jss.v036.i03
Wickham, H. (2009). ggplot2. New York: Springer. https://doi.org/10.1007/978-0-387-
98141-3
Winner, E. (2000). The origins and ends of giftedness. American Psychologist, 55(1),
159–169. https://doi.org/10.1037/0003-066X.55.1.159
Yamamoto, K. (1964). Creativity and sociometric choice among adolescents. Journal of
Social Psychology, 64(2), 249–261. https://doi.org/10.1080/00224545.1964.9919564
Zhang, Z., Hopp, M. D. S., Vialle, W., & Ziegler, A. (in press). Social perceptions of a
creative person: Stereotypes and prejudice of a creative student among German ado-
lescents. Creativity Research Journal,

MANUEL HOPP is a research associate at the chair of Educational Psychology and

Research on Excellence at the University of Erlangen-Nuremberg. His research
focuses on social networks, complex systems and their application in educa-
tional psychology.
ZHITIAN ZHANG is a research associate at the chair of Educational Psychology

and Research on Excellence at the University of Erlangen-Nuremberg. Her areas
of research include creativity, talented development and cultural studies. She
develops courses, workshops and seminars on creativity from both research and
practical perspectives.
LEEANNE HINCH, PhD, is the academic coordinator for the Centre for Talented
Youth, Ireland. She holds a PhD in Science Education from Dublin City Uni-
versity. Her research interests include inquiry based science education, gifted
education, and professional development for teachers.
COLM O’REILLY, PhD is the Director of CTY Ireland at Dublin City University
and secretary for the European Council for High Ability. His research interests
include programmes for gifted students and their social and emotional needs.
ALBERT ZIEGLER, PhD, is the chair professor of educational psychology and

research on excellence at the University of Erlangen-Nuremberg, Germany. He
is the founding director of the State-Wide Counselling and Research Centre for
the Gifted. He has published approximately 350 books, chapters, and articles in
the fields of talent development and educational psychology. Presently, he serves
as the secretary general of the International Research Association for Talent
Development and Excellence (IRATDE) and the Editor-in-Chief of High Ability
Studies.

New Drctns CHLD Adlscnt - 2019 - Hopp - Creative Thus Connected The Power of Sociometric Creativity On Friendship 4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

New Drctns CHLD Adlscnt - 2019 - Hopp - Creative Thus Connected The Power of Sociometric Creativity On Friendship 4

Uploaded by

Copyright:

Available Formats

Hopp, M. D. S., Zhang, Z. S., Hinch, L., O’Reilly, C., & Ziegler, A. (2019).

Creative, thus con-

A dolescence is a crucial period to develop creative potential, in both

Does Being Creative Have a Positive or Negative Impact on Peer

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

to several limitations of previous research. Many of these studies have

(Q1) What are the fundamental features of gifted adolescents’ friendship

(Q1) What Are the Fundamental Features of Gifted

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

in the present study, we expect a significant positive reciprocity effect as the

(Q2) What Are the Fundamental Features of Sociometric

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

(Q3) How Does Students’ Sociometric Creativity Influence

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

exploratory. Specifically, we will explore the effect of sociometric creativity

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Friendship Network. First, an example item was presented, where the

If a student regarded another student as category 1 (“friend”), then the

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Figure 4.1. An exemplary sociometric creativity network (left) and

the proportion of relationships observed in the network to the theoretically

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

effect on the person’s tendency to become friends. An overview including

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Table 4.1. Short Description of all Included Effects in the RSiena

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Table 4.1. Continued

Table 4.2. Means, Standard Deviations, and Pearson Correlations of

Friendship Network Evolution. For the development of the friend-

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Table 4.3. Means, Standard Deviations, and Minima and Maxima of

Table 4.4. Multigroup RSiena Model Results of the Friendship

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Figure 4.2. Forest plot of the effect estimates of the friendship

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Table 4.5. Multigroup RSiena Model Results of Sociometric Creativity

Sociometric Creativity Network Evolution. Alongside the evolution

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Figure 4.3. Forest plot of the effect estimates of the sociometric

Influence of Sociometric Creativity on Friendship Network Evolu-

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Table 4.6. Multigroup RSiena Model Results of the Influence of

Figure 4.4. Forest plot of the effect estimates of the influence of

friendship in-degree; “indegree of creativity on popularity,” β = 0.03, 95%

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

nominate student A as a friend because of that. There was no heterogeneity

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

predictors of friendship ties for gifted students’ friendship formation within

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

peer group: either neglected or even disliked (Gebart-Eaglemont & Foddy,

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Sternberg, 1985), smarter students might be perceived as more creative

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

therefore, might also disparate in different contexts considering domains

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

provided a more ecologically valid representation of adolescent friendships,

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Beghetto, R. A. (2019). Creativity in classrooms. In J. C. Kaufman & R. J. Sternberg

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Kleibeuker, S. W., De Dreu, C. K. W., & Crone, E. A. (2016). Creativity development

NEW DIRECTIONS FOR CHILD AND ADOLESCENT DEVELOPMENT • DOI: 10.1002/cad

Rothenberg, A. (1990). Creativity in adolescence. Psychiatric Clinics of North America,