ArticlePDF Available

A Comparison of Social Network Mapping and Personal Network Visualization

Authors:

Abstract

This article presents an analysis of personal network visualization based on systematic evaluations of alter pairs compared to freestyle drawings respondents made of their personal network. In most cases, personal network visualization provided important details that are different from respondents' perceptions. Several case studies are discussed that highlight the additional data provided when using personal network visualization.
http://fmx.sagepub.com
Field Methods
DOI: 10.1177/1525822X06298592
2007; 19; 145 Field Methods
Christopher McCarty, José Luis Molina, Claudia Aguilar and Laura Rota
Visualization
A Comparison of Social Network Mapping and Personal Network
http://fmx.sagepub.com/cgi/content/abstract/19/2/145
The online version of this article can be found at:
Published by:
http://www.sagepublications.com
can be found at:Field Methods Additional services and information for
http://fmx.sagepub.com/cgi/alerts Email Alerts:
http://fmx.sagepub.com/subscriptions Subscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://fmx.sagepub.com/cgi/content/refs/19/2/145
SAGE Journals Online and HighWire Press platforms):
(this article cites 15 articles hosted on the Citations
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
A Comparison of Social Network
Mapping and Personal Network
Visualization
CHRISTOPHER MCCARTY
University of Florida
JOSÉ LUIS MOLINA
CLAUDIA AGUILAR
LAURA ROTA
Universitat Autonoma de Barcelona
This article presents an analysis of personal network visualization based on systematic
evaluations of alter pairs compared to freestyle drawings respondents made of their per-
sonal network. In most cases, personal network visualization provided important details
that are different from respondents’ perceptions. Several case studies are discussed that
highlight the additional data provided when using personal network visualization.
Keywords: personal networks; visualization; cognition
In this article, we focus on the visual representation of people a respon-
dent knows—their personal network. Personal networks are a type of ego-
centric network; they consist of the set of family, friends, and acquaintances
surrounding a focal person. In social network analysis, personal networks
are contrasted with whole (sociocentric) networks in which the focus is on
the pattern of interactions within a focal group.
Visual representations of personal networks are not used much in social
science research but are common in counseling psychology and social
work. Genograms are techniques used by mental health therapists to cap-
ture the relationships, both past and present, surrounding a client (DeMaria,
Weeks, and Hof 1999; McGoldrick, Gerson, and Shellenberger 1999). By
representing men, women, and children as circles and triangles, as in a kin-
ship diagram used by anthropologists, therapists attempt to understand the
social environment that may have contributed to the conditions that led
them to seek help. The visual representation helps both therapists and
145
This research was supported by the National Science Foundation, Award No. BCS-0417429.
Field Methods, Vol. 19, No. 2, May 2007 145–162
DOI: 10.1177/1525822X06298592
© 2007 Sage Publications
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
clients understand the social environment that may be contributing to or
hindering mental health. Genograms tend to focus on close, mostly family
relationships and typically represent the social environment chronologi-
cally, including relatives who are both living and dead.
The hierarchical mapping technique (Antonucci 1986; Ajrouch, Antonucci,
and Janevic 2001) uses three concentric circles to represent the personal net-
work of the respondent. In the middle of the circle is the word “YOU.
Respondents are asked to put the first name of people they know closely in the
innermost circle and those they know but are less close to in the outer circle.
The resulting map gives the researcher some sense of the size of the network
and the distribution of their network based on closeness.
Carrasco, Hogan, Wellman, and Miller (forthcoming) used a similar
approach with four concentric circles. In their study, respondents free listed
alters first, then placed them on the network map—those closest to the
respondent in the inner circles and those less close in the outer circles.
Respondents were asked to place those who knew each other nearby and
finally to draw circles around groups of respondents. This method adds
network structural features to the hierarchical mapping technique.
Perhaps the most straightforward technique for acquiring an image of a
personal network is to ask respondents to draw them freestyle, with little
instruction as to how they are represented. For example, respondents can be
told to represent people with dots and groups of people with circles. Unlike
genograms and hierarchical mapping, both of which start with some struc-
tural constraints, freestyle drawing captures the variability in the way
respondents represent their network. Despite its simplicity, there are few
examples of this approach in the literature. Coates (1985) used this tech-
nique in the study of the personal networks of black adolescents.
Another approach is to elicit names of network members from the respon-
dent, then ask him or her to evaluate the relationship between each individual
pair of alters. We call this personal network visualization. This method differs
from the others because the respondent is being asked to evaluate a set of
binary relationships that are then built into a representation of personal net-
work structure, as opposed to the respondent being asked to try to summarize
all relationships into a structure from memory. McCarty (1992, 2002) used this
approach in a study of structure within personal networks. Mitchell (1994)
used this approach for a small sample of homeless women in Manchester,
England. More recently, Widmer and La Farga (2000) used visualizations of
personal networks to study the variability in the structure of families.
In this article, we contrast the two techniques that are the most different.
At one end of the spectrum are freestyle drawings that allow respondents to
represent their network however they like. At the other end are personal
146 FIELD METHODS
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
network visualizations based on the systematic evaluation of the relations
between all pairs of network members. We illustrate these differences using
case studies showing the types of conclusions an interviewer might make
about the respondent’s personal network using each technique. In the next
section, we outline the principles of personal network data collection for
those who are unfamiliar with the process.
COLLECTING PERSONAL NETWORK DATA SYSTEMATICALLY
The study of personal networks typically involves, at a minimum,
acquiring a list of a person’s network members (alters). In studies of social
support, for example, people are asked to name some number of alters
(three, five, ten) on whom they rely for advice or material help (Burt 1984;
Wellman and Wortley 1990). Respondents may be asked to think of five
people they talk to about important matters, or three people they talk to
about health-care decisions. In studies of support that involve weak ties
(acquaintances—rather than relatives, close friends, or coworkers) respon-
dents may be asked to list up to sixty people they know (McCarty 2002).
The method for sampling respondents varies greatly depending on the
study. A balance must be achieved between the number of respondents, the
number of alters they will be asked about, the amount of information
elicited about each alter, and the mode of data collection. Some network
studies have only a handful of respondents whereas others have thousands.
Most analyses of personal network data summarize the composition of
the network as a set of variables that become attributes of the respondent
(Fischer 1982; Schweizer, Schnegg, and Berzborn 1998; Hampton and
Wellman 1999). Along with the age, education, and income level of a
respondent, the researcher may have the average age of their alters, the
average strength of their ties with alters, the proportion of their network that
is family or coworkers, or the proportion of their network from whom
people say they can borrow money or get a ride to the doctor (Campbell and
Lee 1991; McCarty et al. 1997). These measures may, in turn, be used as
independent variables to predict other variables. They may also be used as
dependent variables or predicted by typical demographic variables or vari-
ables more specific to the topic of interest.
Some personal network researchers also try to measure structure within
each respondent’s network (McCarty 2002). To do this, the researcher must
ask respondents to report not only on their relationship with each alter but
also on the relationships of all pairs of alters. The number of ties grows geo-
metrically as alters are added (see Figure 1). For a network of ten alters, a
McCarty et al. / NETWORK MAPPING AND VISUALIZATION 147
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
respondent must report on forty-five ties. For a network of fifty alters, they
must report on 1,225 ties. There are conceptual and empirical issues sur-
rounding the application of structural measures to personal network data
(McCarty and Wutich 2005), specifically, whether to include or exclude ego
for a given measure.
The collection and analysis of personal network structural data has been
difficult in the past, given the absence of software devoted to that purpose.
Recently, a program called EgoNet was developed that is designed specifi-
cally for the collection, analysis, and visualization of personal network data.
EgoNet consists of four modules—questions asked of the respondent
about themselves, questions used to generate the names of network alters,
questions asked of the respondent about those alters, and questions asked of
the respondent about the existence of relations between alters. It is designed
as a questionnaire authoring language that allows researchers to tailor the
interview to their specific research interests. The program displays a visu-
alization of the respondent’s personal network. As an illustration, consider
the example in Figures 2 and 3. Figure 2 represents an adjacency matrix for
148 FIELD METHODS
0
1000
2000
3000
4000
5000
6000
1 112131415161718191
Alter Pair Evaluations
Number of Alters
FIGURE 1
Respondent Burden when Collecting Structural Data
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
a ten-alter personal network. In most cases, we ask the respondent to describe
the relations between their alters as symmetric, meaning both alters agree to
the level of knowing. In cases where the question being asked is simply the
presence or absence of a tie, respondents are quite consistent in their assess-
ments. With a ten-alter symmetric network, the respondent must evaluate
forty-five unique pairs of alters. This matrix can then be used to calculate sev-
eral measures, and to visually represent the ties in the network, as depicted in
Figure 3. The visualization in Figure 3 used a spring-embedding algorithm to
place the nodes relative to each other (see Freeman 2000).
Several attributes of this visualization immediately suggest questions we
might ask the respondent. First, we can see that George is a central figure
in this network and that he connects two groups of people. Depending on
the research topic, we may first want to know who George is and how he
relates to the respondent. Next, we may want to know what the two groups
are and why George is or is not connected to all of them. Disconnections
within a group may be a sign of the potential formation of factions. Finally,
although Cindy is part of the personal network, she is disconnected from
McCarty et al. / NETWORK MAPPING AND VISUALIZATION 149
Don Janice Debra Steve Randy Alex Cindy Alvaro Lisa George
Don 11000 100 01
Janice 1 1 0 0 0 0 0 0 0 1
Debra 0 0 1 0 0 1 0 0 1 1
Steve 0 0 0 1 1 0 0 1 0 1
Randy 0 0 0 1 1 0 0 1 0 1
Alex 1 0 1 0 0 1 0 0 1 1
Cindy 0 0 0 0 0 0 0 0 0 0
Alvaro 0 0 0 1 1 0 0 1 0 0
Lisa 0 0 1 0 0 1 0 0 1 0
George 1 1 1 1 1 1 0 0 0 1
FIGURE 2
Adjacency Matrix for Ten-Alter Personal Network
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
everyone else. Is she a broker to another set of ties that may be potentially
beneficial to the respondent? Or does she represent a social pressure that
could impact the respondent negatively?
METHOD
To test the utility of personal network visualizations against that of freestyle
drawings of networks, we conducted nineteen interviews in Barcelona, Spain
(seven men and twelve women, with a mean age of 33). Respondents were
selected from different ethnic groups to maximize the differences dis-
covered through the qualitative interview (ten from Spain, three from
Ghana, two from Serbia, one from Senegal, one from Croatia, and two from
Bosnia).
Respondents were first interviewed without the aid of the personal net-
work visualization. A paper instrument was used to collect the names of
exactly forty-five free-listed alters. The criteria for including an alter were:
“You know them and they know you, by sight or by name. You have had
150 FIELD METHODS
Debra
Cindy
Lisa
Alex
Don
Janice
George
Steve
Alvaro
Randy
FIGURE 3
Visualization of Ten-Alter Personal Network
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
some contact with them in the past two years (i.e. phone, face-to-face,
e-mail, mail) and you could contact them again if necessary.
Respondents then provided data on each alter, including how they knew
the alter, the language of communication, the intensity of the relation, fre-
quency of contact, method of communication, occupation of the alter, and
whether or not they consider them a foreigner. Respondents were then asked
to draw a representation of their network using the following instructions:
We would like you to draw your social circles. The more people in each cir-
cle, the larger it will be. Circles that are farther apart mean that they are less
socially connected. Circles can also overlap. You can also specify single indi-
viduals. When you are done drawing, you can put a label or name in each cir-
cle to tell us what it is.
Following this task, respondents were interviewed about the way that their
network impacts their ethnic identity. The interview was recorded. Each of
the three interviewers then entered the alter names provided by their respec-
tive respondents into EgoNet. Interviewers only entered the names for the
forty-five alters. The interviewers then arranged a second session with each
respondent, no later than one week after the original interview, having them
complete the 990 alter-tie evaluations. The tie evaluation question was
whether the two alters would talk independently of the respondent. Following
the fourth module, the program displayed a visualization of the personal net-
work of the respondent based on the adjacency matrix from the tie evalua-
tions. The interviewer, assisted by the personal network visualization, then
asked the respondents the same questions they did during the first interview.
The analysis consisted of the qualitative assessment of the benefit to the
respondent and the interviewer of having the visualization as a cue. This
method isolated the benefit specific to the visualization.
RESULTS OF PERSONAL NETWORK
VISUALIZATION INTERVIEWS
Most of the respondents were surprised by the personal network visual-
ization. The process of evaluating the 990 alter pairs gave no indication of
the structure of the graph. Indeed, the task of making so many evaluations
makes it virtually impossible to fake. After the alter pair evaluations, which
on average lasted about 20 minutes, respondents were interested in the
product of their efforts. Most respondents were excited about the opportu-
nity to describe the visualization and talk about their network.
McCarty et al. / NETWORK MAPPING AND VISUALIZATION 151
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
With two exceptions, respondents verified that the personal network
visualization made sense to them, given what they knew about the social
environment around them. They were all able to identify groups of people
and people in structurally important roles. In some cases, respondents made
errors in some of the 990 alter-pair evaluations and recognized those in the
personal network visualization.
Those respondents who did not recognize groupings are suspected of not
entering the alter-pair evaluations correctly. For example, one respondent
coded nearly all the potential pairs as ties, resulting in a visualization with
one large grouping of all forty-five alters. Although this is not impossible,
the respondent’s reaction indicated that it did not match her conception of
her network.
Comparisons between the initial drawing and the personal network
visualization are more revealing. In some cases, the visualization closely
resembles their initial drawings. Figure 4 shows the network of Elia, a 30-
year-old woman from Barcelona. Elia’s drawing and her network visual-
ization share many things in common. The drawing shows a large group of
people at the center who are a mixture of family and friends. This is also
depicted in the network visualization. In both, Elia separates maternal and
paternal family and shows a strong connection with the central family and
friend group.
In both graphs, Elia shows groups from a town in the Netherlands, the
Halle, as well as groups in Florence and Berlin. However she shows a con-
nection between the Halle group and the Florence and Berlin groups that
152 FIELD METHODS
FIGURE 4
Drawing and Network Visualization of Elia, a 30-Year-Old
Woman from Barcelona
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
does not exist in the visualization. She also shows a connection between the
Florence group and her family that is not shown in the visualization. The
connection between the Halle group and Berlin in the social circles graph
is mainly due to one respondent who is from Berlin, but lives in Halle.
When asked about that connection, the respondent said that the nature of
the tie question (would the two alters talk independently of the respondent)
negated that tie. In this case, the social circles represents links based on cat-
egory association (being from Berlin), whereas the personal network visu-
alization represents actual communication.
Figure 5 shows the drawing and the network visualization of Marta, a
26 year old. Comparison of the drawing and the visualization show that the
categories are very much the same. Marta is aware that she has a large fam-
ily group that overlaps with her cousins’ friends group. She also depicted a
group of friends from school who are connected to her family. Marta is an
Erasmus student, a European program that sends students from one
European country to study in another. She maintains a group of friends
from the Erasmus program who are depicted in both the drawing and the
visualization. Similarly, she depicts a group of friends from a small town
called Calella in both pictures.
There are, however, some distinct differences. The main difference is the
ability to identify key people in the visualization. The drawing depicts only
broad categories of people, and, in only one case, a connection between
those categories. The visualization shows bridging between categories from
McCarty et al. / NETWORK MAPPING AND VISUALIZATION 153
FIGURE 5
Drawing and Network Visualization of Marta, a Catalan, 26 Years Old
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
key alters. In this interview concerning ethnic identity, the role of bridging
ties can be critical.
For example, the visualization shows four key ties between the school
group and Marta’s family. There is also a bridging tie between the Erasmus
group and her family. This raises two types of questions. Why do these
people serve a bridging role? Why do the others within the group not serve
one? The utility of the visualization for the interview would be made
stronger by overlaying characteristics of the alters, such as sex, ethnic group,
and age. Figure 6 compares the drawing and visualization of Milanka, a
Serbian migrant in Barcelona. In this case, the two representations of her
network are not quite so similar. She depicts her family group at the center,
with a mixture of neighbors, extended family, friends, and work connected
directly to it. The only indirect connection she shows is acquaintances from
work through friends from work.
The personal network visualization tells a different story. Although the
right side of the visualization does bear some similarity to the drawing,
showing family work and neighbors, we see that the work group is actually
connected through the neighbors by one key tie. On the left, we also see that
Milanka has a set of family ties in Serbia that are connected to the family
group on the left by one key tie. In the upper left, we see a group of five
alters from a former job who were not represented in the drawing at all.
Finally, in the far upper left is a single alter (an isolate), her hairdresser,
who is not tied to anyone.
The ability to identify isolates is another key advantage of the personal
network visualization. Isolates represent interesting subject matter for dis-
cussion. In the case of migrants, an isolate may represent attempts to reach
out to other groups. Discussion about their success or failure in doing so
may be facilitated by discussion about the reason an alter is an isolate.
Figure 7 shows the network of Edin, a Bosnian migrant. The drawing
depicts a well-organized network, with several groups nested within each
other. We would conclude from this drawing that Edin’s maternal family
and his university friends were connected and at the center of his network.
The personal network visualization on the right shows no connection
between his family on the lower right and his university friends. Edin verified
that this was the case. He does have a set of friends who live in the same town
in Spain as his family, but they are connected by only one person. Information
about this friend and why he is so key is useful for understanding how Edin
integrates into Spanish society. We also see from the personal network visu-
alization that Edin has an adoptive family in Spain, something that is not
depicted in the drawing. The long list of isolates in the upper right represent
clients from his work. The fact that he chose six of his forty-five alter choices
154 FIELD METHODS
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
155
FIGURE 6
Drawing and Network Visualization of Milanka, a 34-Year-Old Serbian Migrant
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
for clients who did not know each other could be significant. The interviewer
should explore whether Edin considers these clients to be potential bridges to
new alters.
Finally, we show the personal network visualization and the drawing from
Regina, a 33-year-old woman from Ghana (Figure 8). Regina obtained a
Masters in Spanish translation at the University of Valladolid in Valladolid, a
city in the north of Spain. She’s kept some friends from this stage in her life,
but the transnational community of students whom she met in Vallodolid are
now in other countries. Some of them are represented as isolates in the upper
left corner of the visualization. The two main groups of Regina’s network are
her family in Ghana and the people from Ghana who also live in Vic (a town
in Barcelona). She met those people in Vic for the first time, with the excep-
tion of Sam, her husband. Sam is the dot connecting both worlds—the fam-
ily and the transnational community. It is interesting to note the support role
of the Evangelist church in Vic for migrants from Ghana.
DISCUSSION
In seventeen of the nineteen cases, both the interviewer and the respondent
recognized lines of questioning that were available to them using the personal
156 FIELD METHODS
FIGURE 7
Drawing and Network Visualization of Edin, a 26-Year-Old Man from Croatia
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
157
FIGURE 8
Drawing and Network Visualization of Regina, a 33-Year-Old Woman from Ghana
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
158 FIELD METHODS
network visualization as a cue that were not available without it. Memories
and information are not stored randomly, but arranged in hierarchical sets that
are quickly accessible. The same is true for names of people. When partici-
pants free list names, they tend to cluster or mention successively persons
from the same social context (e.g., family, work, school, church, etc.; Brewer
and Yang 1994). This suggests that memories of people are, to some degree,
organized by social structure.
When we ask respondents to describe their personal network (i.e., the
social environment in which each respondent lives), we are asking them to
describe the structure of their memory, using those labels that they have avail-
able to them. This is not a task that most people ever do, and is thus affected
by their ability to report the structure of their memory accurately. Even if
respondents can report the structure of their memories, there is reason to
believe that this system of storing information differs significantly across
respondents and that the labels used between people are radically different,
and, in many cases, ad hoc. Assuming people can report accurately and store
information in the same way, we see no reason to believe that they store names
based on an accurate representation of social structure, although there are
suspicions that social structure is the basis for the organization of people in
memory (Brewer and Garrett 2001; Brewer et al. 2005). Thus, the drawings
that respondents made in the first interview represent their attempt to apply
those labels they have available to describe this structure.
The fourth module of the interview—the alter-tie evaluation—is, on the
face of it, a much easier set of questions to answer. But it is more than that.
The perception that two people are connected in some way does not require
the respondent to abstract the nature of the relationship and fit it into a cat-
egory that then must be labeled. Relations that cross conceptual categories,
such as people we both work with and socialize with, can be evaluated by
one criterion: whether they are socially connected, given the definition of a
connection. The respondent can consider the multiplex relationships they
have with their alters so that they accurately answer the question. The result
is a picture of the structure that shows groupings, when they exist, and
bridges between groups. It also shows people who are isolated and groups
who are isolated and allows for the exploration of those relations.
Another reason that the personal network visualizations may differ from
the categories generated by respondents is that respondent-generated cate-
gories tend to follow a certain dimension, such as types of behavior the
respondent engages in with the alter (e.g., family, work, church), whereas
social structures do not necessarily conform to one domain. In small-world
studies in which the object is to try to get a message to randomly selected tar-
gets, it is not unusual for respondents to use alters that fit a behavioral domain
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
McCarty et al. / NETWORK MAPPING AND VISUALIZATION 159
for local targets and a locational domain for those far away. This can be
observed in the personal network visualizations in which clusters of respon-
dents are identified by a mixture of behavioral and locational domains.
As can be seen from the analysis above, it is also quite common for the
personal network visualizations to make distinctions between categories.
For example, it is typical for respondents to talk about family relations. Yet,
personal network visualizations often show clustering within families,
based on kinship, location, or interaction. It is not uncommon to see groups
within families who have no ties whatsoever. Such distinctions cannot be
realized without the systematic evaluation of alter pair ties.
One potential disadvantage of the personal network visualization, at
least the one used here, is that it does not accommodate overlap. From the
drawings, we see that some respondents indicate comembership of people
within more than one group. This is shown by overlapping circles. This ver-
sion of the visualization software does not accommodate overlap. Alters are
only shown in one position.
This method is not a panacea. When the subject matter does not concern
a topic that is fundamentally interpersonal, the personal network visualiza-
tion would help very little, if at all. For example, if the subject matter of an
interview were specifically about knowledge or technique (such as the
names of medicinal plants or how to construct a hut), it is doubtful that the
tedium of collecting personal network structural data would be worthwhile.
On the other hand, if the research topic is fundamentally interpersonal
(such as influences on migration choices or political opinion), then the
personal network visualization provides a perspective on these topics that
cannot be gained otherwise. For example, in the case of the interviews
above concerning ethnic identity, the personal network visualizations show
how some respondents compartmentalize alters of different ethnicities. This
is most evident in Figure 8, where Regina maintains ties to a group of
Catalans and several groups of Ghanaians. With the personal network visu-
alization, the interviewer could question Regina specifically about people
within those groups and how they interact, rather than hypothetical rela-
tions between abstract categories of relations.
A new version of EgoNet has been released that provides some additional
capability that will no doubt be useful (available at www.mdlogix.com). Figure
9 shows the visualization, using this new software, of a second-generation
Gambian woman living in a Catalan city. The circles with numbers by them are
hierarchical clusters calculated by the program. As we can see, in the previous
examples the circles do not fully match the naturally occurring groups that the
informants defined from the visualization. However, they do provide a standard
and objective way for interviewers to talk to respondents about the groupings
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
160 FIELD METHODS
in their personal network. The numbers make it easy to record these interviews
and indicate to which group the respondent is referring. Having an objective
way of identifying groupings increases the reliability of the qualitative inter-
pretation of the visualization.
We can also see that the software has allowed us to represent attributes of
alters. In this case, color is used to represent skin color (Black = Black, Light
grey = White, Dark grey = Brown), size represents how close the respondent
feels to each alter (larger nodes are closer), and shape represents whether or
not the alter smokes (Circles = Nonsmokers, Squares = Smokers). The nodes
are also labeled by the country in which the alter was born. Using these data
gives a much more detailed view of the social context of the respondent. In
this case, the interviewer can easily see how alter attributes are distributed
through the personal network and if the respondent compartmentalizes alters
based on that. For example, most of the isolates the respondent lists are smok-
ers. The respondent lives with her Muslim family. When questioned about the
number of isolates and the fact that they smoke, she revealed that she also
smokes and stays with these friends on overnight trips where she can smoke
FIGURE 9
EgoNet Visualization of the Personal Network of a 22-Year-Old
Gambian Woman
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
McCarty et al. / NETWORK MAPPING AND VISUALIZATION 161
freely and go to parties. The ability to overlay alter attributes over structure is
quite powerful.
Future modifications of the method should include ways to reduce the
respondent burden, such as presenting alter-pair evaluations in a way that is
easier for respondents to input, either visually or in groupings. More work
must be done on perfecting name generators to elicit the names of alters in an
unbiased way, or to elicit names so that the bias is known and manageable.
NOTE
1. McCarty (2002) found that respondents recoding a set of alter-pairs were 97% accurate
in recoding ±1 point.
REFERENCES
Ajrouch, K. K., T. C. Antonucci, and M. R. Janevic. 2001. Social networks among Blacks and
Whites: The interaction between race and age. Journals of Gerontology Series B—
Psychological Sciences and Social Sciences 56 (2): S11218.
Antonucci, T. C. 1986. Social support networks: Hierarchical mapping technique. Generations
10 (4): 10–12.
Brewer, D. D., and S. B. Garrett. 2001. Evaluation of interviewing techniques to enhance recall of
sexual and drug injection partners. Sexually Transmitted Diseases 28 (11): 666–77.
Brewer D. D., J. J. Potterat, S. Q. Muth, P. Z. Malone, P. Montoya, D. L. Green, H. L. Rogers,
and P. Cox. 2005. Randomized trial of supplementary interviewing techniques to enhance
recall of sexual partners in contact interviews. Sexually Transmitted Diseases 32 (3): 189–93.
Brewer, D. D., and B. L. Yang. 1994. Patterns in the recall of persons in a religious commu-
nity. Social Networks 16 (4): 347–79.
Burt, R. S. 1984. Network items and the General Social Survey. Social Networks 6 (4): 293–340.
Campbell, K. E., and B. A. Lee. 1991. Name generators in surveys of personal neighbor net-
works. Social Networks 13 (3): 203–21.
Carrasco, J.-A., B. Hogan, B. Wellman, and E. J. Miller. Forthcoming. Collecting social net-
work data to study social activity-travel behavior: An egocentric approach. Journal of the
Transportation Research Board.
Coates, D. L. 1985. Adolescent social record, your social map, social network record [SNR].
In Handbook of tests and measurements for black populations, 1996, 2 vols., ed. R. L.
Jones, 269–83. Hampton, VA: Cobb & Henry Publishers.
DeMaria, R., G. Weeks, and L. Hof. 1999. Focused Genograms: Intergenerational assessment
of individuals, couples, and families. New York: Brunner-Routledge.
Fischer, C. S. 1982. To dwell among friends: Personal networks in town and city. Chicago:
University of Chicago Press.
Freeman, L. 2000. Visualizing social networks. Journal of Social Structure 1:1.
Hampton, K. N., and B. Wellman. 1999. Netville online and offline—Observing and surveying a
wired suburb. American Behavioral Scientist 43 (3): 475–92.
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
162 FIELD METHODS
McCarty, C. 1992. Perceived clique definition in ego-centered networks. Unpublished doctoral
dissertation, Department of Anthropology, University of Florida.
———. 2002. Measuring structure in personal networks. Journal of Social Structure 3:1.
McCarty, C., H. R. Bernard, P. D. Killworth, E. Johnsen, and G. Shelley. 1997. Eliciting rep-
resentative samples of personal networks. Social Networks 19 (4): 303–23.
McCarty, C., and A. Wutich. 2005. Conceptual and empirical arguments for including or
excluding ego from structural analyses of personal networks. Forthcoming. Connections.
McGoldrick, M., R. Gerson, and S. Shellenberger. 1999. Genograms: Assessment and inter-
vention. New York: Norton.
Mitchell, C. 1994. Situational analysis and social networks. Connections 17 (1): 16–22.
Schweizer, T., M. Schnegg, and S. Berzborn. 1998. Personal networks and social support in a
multiethnic community of southern California. Social Networks 20 (1): 1–21.
Wellman, B., and S. Wortley. 1990. Different strokes from different folks: Community ties and
social support. American Journal of Sociology 96 (3): 558–88.
Widmer, E. D., and L. A. La Farga. 2000. Family networks: A sociometric method to study
relationships in families. Field Methods 12 (2): 108–28.
CHRISTOPHER MCCARTY is currently survey director of the Bureau of Economic and
Business Research at the University of Florida; he gained his PhD at the same university
in the Department of Anthropology. Among his areas of research is the development of
new methods and tools for studying personal networks in a transcultural framework and
authoring the software EgoNet. Some recent publications are (with P. D. Killworth, H. R.
Bernard, E. C. Johnsen, and G. A. Shelley) “Comparing Two Methods for Estimating
Network Size” (Human Organization, 2000) and (with P. D. Killworth, H. R. Bernard,
E. C. Johnsen, J. Domini, and G. A. Shelley “Two Interpretations of Reports of Knowledge
of Subpopulation Sizes” (Social Networks, 2003).
JOSÉ LUIS MOLINA is full professor in the Department of Social and Cultural
Anthropology, at the Universitat Autonoma de Barcelona (UAB) and director of the
Personal Networks Lab, UAB. His areas of interest are in economic anthropology and
social networks, especially the change among ethnic groups moving to other countries.
Two recent publications are “El estudio de las redes personales: Contribuciones,
métodos y perspectivas” (Empiria, 2005) and, with S. Borgatti, “Toward Ethical
Guidelines for Network Research in Organizations” (Social Networks, 2005).
CLAUDIA AGUILAR is a graduate student in the anthropology program at the
Universitat Autonoma de Barcelona (UAB). She is conducting her research in Sarajevo
(Bosnia) about changes in personal network and ethnic identifications. She has recently
published “Visualización de redes personales en Sarajevo” (REDES-Revista hispana
para al Análisis de Redes Sociales, 2005).
LAURA ROTA is a graduate student in the anthropology program at the Universitat
Autonoma de Barcelona (UAB). She is conducting her research in Senegal about changes
in personal network and ethnic identifications. A recent publication is, with J. L. Molina,
C. McCarty, and C. Aguilar, “La estructura social de la memoria” (in Interacción, Redes
Sociales y Ciencias Cognitivas, 2006).
distribution.
© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized
at PENNSYLVANIA STATE UNIV on February 6, 2008 http://fmx.sagepub.comDownloaded from
... Such freestyle renderings are thought to be superior for revealing respondents' cognitive representations of their local social environments (e.g. Maya-Jariego and Cachia, 2019;McCarty et al., 2007b). All sociograms visualize the entire network and depict all alters simultaneously. ...
... It is important to ask whether all study participants share these sensibilities . Qualitative investigations and respondent debriefings suggest that some respondents do not embrace these metaphors, as is evident (e. g.) when freestyle drawings and structured sociograms are compared (McCarty et al., 2007b). Some lower-status respondents in the Hollstein et al. (2020) study expressed ambivalence about the concentric circle template, preferring a funnel representation in which they were positioned as outsiders looking in. ...
Article
This article reviews recent methodological research that bears on the collection of egocentric network data. It begins with background on setting egocentric network boundaries and principal types of instruments that obtain information about such networks. It then discusses innovations in data collection and studies of data quality. The bulk of these address questions about “name generator” instruments that obtain information about the alters and relationships in a subject's network. Among topics receiving substantial attention in recent research are mitigation of respondent burden, interviewer effects, survey mode, and the performance of name generators in longitudinal studies. Potentially fruitful innovations supplement conventional question-and-answer surveys with visual elements that promise to better engage respondents and reduce the demands that name generator-based data collection poses. We close by highlighting both accomplishments of this body of research and some open issues.
... This technique asks participants to draw a series of circles, the centermost representing them (i.e., the bullseye), and each surrounding circle representing a different tier of their social network (see Figure 1). Extant work has shown that HMT can help explain the types of attachments that people share with their network members (Rowe & Carnelley, 2005), explore discrepancies between perceived and actual network qualities (McCarty et al., 2007), and help differentiate which network members are trusted with end-of-life interactions (Canaway et al., 2019). Both social and clinical sciences have employed HMT to better understand the nature of human interaction; however, interpersonal communication has yet to adopt HMT as a method for understanding the specific people that comprise a social network. ...
Article
Full-text available
Social networks remain a relatively understudied topic in the field of interpersonal communication. As such, the hierarchical mapping technique (HMT) can help shed light on the levels of closeness that people share with their network members. For interpersonal, organizational, and family communication researchers, this gap is tremendously important, as uncovering trends in relationship closeness is pivotal to understanding the individuals/roles/relationships that bear the most weight in a person’s life. The present study utilized HMT to explore which type(s) of relationships people believe are closest to them, as well as which people are mildly close and also rather distant. Results demonstrated that most people consider their parents, siblings, children, best friends, and roommates as first-tier network members. Common second-tier relationships included cousins, aunts/uncles, friends, and non-biological siblings. Common third-tier relationships included patrons, acquaintances, neighbors, and idols. Moreover, there were statistical similarities across a variety of network relationships, indicating that there is substantial variance in the sorts of relationships that people typically consider close, interpersonal, and distant. Results are discussed in terms of both theoretical and methodological significance.
... For instance, a graphical representation of data may reveal potential clusters or outliers that require closer analysis (Mellinger and Hanson 2017). Likewise, visual mappings of networks may uncover relationships that may not be easily identified or described in text alone, and instead provide another means by which to understand relationships among various data points (e.g., McCarty, Molina, et al. 2007). Visualisation techniques can also be used to present research findings, synthesising large amounts of data into accessible visual infographics or creating interactive dashboards. ...
Article
As corpus-based translation studies continues to expand, researchers have employed data analytic techniques from neighbouring disciplines, such as corpus linguistics, to explore a wider variety of research questions. The field has evolved from early frequency-based approaches to corpus-based translation studies to now include more advanced statistical analyses to understand the complex web of variables encapsulated by the translation process. Big data analytic techniques that originated in data analytics and related quantitative fields could be usefully applied to research questions in translation and interpreting studies. To assess their applicability, this article first outlines what distinguishes big data from general corpora in translation and interpreting studies, identifying how data volume, variety, and velocity are applicable properties to be considered in corpus-based translation and interpreting studies research. Then, the article presents three types of big data analysis techniques, namely crosslingual and multilingual data analysis, sentiment analysis, and visual analysis. These analyses are presented in conjunction with potential research areas that would benefit from these complementary analytical approaches. The article concludes with a discussion of the implications of big data analytics in corpus translation studies, while charting the trajectory of a more quantitative, corpus-based approach to translation studies.
... Many studies also ask ego questions about the relationships between the alters, for example, to what extent the alters are in contact with each other. This is not absolutely necessary if certain statistical measures or questions are not considered essential for one's own question (McCarty et al., 2016). In contrast to the whole network analysis, where the contact persons are specified by some kind of border, the interviewee is free to name them. ...
Chapter
Full-text available
Thanks to improvements in living standards and health behavior as well as medical progress since the second half of the twentieth century, old age has become a life phase in its own right. This phase usually begins by the transition from working life to retirement (Kohli, 2000). Both the chance of reaching retirement and the life expectancy after retirement have increased significantly (Eisenmenger & Emmerling, 2011). The post-work phase spans several decades for many people now. In addition, people who retire are considerably healthier and more independent than their peers of earlier birth cohorts (Crimmins, 2004). The expansion of this phase of life has been accompanied by a differentiation of older people in terms of health and independence: healthy and active people experience this phase, as do people in need of help and care. This fact is considered by distinguishing between old and very old people (Baltes, 2007). Characteristics of old age are absence of non-compensable health restrictions, self-determination of various activities (e.g., traveling, hobbies, voluntary work), and strong social integration. Overall, the demands of old age can be coped well in this phase. Very old age is characterized by an increase in physical and cognitive losses and diseases, and a decrease in the abilities and possibilities of compensating for deficits (Baltes, 1997; Baltes & Smith, 2003).
Article
Full-text available
Our study contributes a contextual perspective on entrepreneurs’ networking, shifting focus from individual-level network structure and networking activities toward understanding networking as a multilevel process involving individual and contextual mechanisms. Through a multiple case study of entrepreneurs in Silicon Valley and Berlin, we introduce networking temperature as novel concept that captures context-bound templates for how entrepreneurs should network, ranging from colder to warmer. As core implications, networking temperature enables a contextualized understanding of tie quality, to explain why networking takes different forms in different contexts, and why entrepreneurs gain more cumulative advantage from their existing relationships in warmer than colder contexts.
Chapter
This edited volume demonstrates the potential of mixed-methods designs for the research of social networks and the utilization of social networks for other research. Mixing methods applies to the combination and integration of qualitative and quantitative methods. In social network research, mixing methods also applies to the combination of structural and actor-oriented approaches. The volume provides readers with methodological concepts to guide mixed-methods network studies with precise research designs and methods to investigate social networks of various sorts. Each chapter describes the research design used and discusses the strengths of the methods for that particular field and for specific outcomes.
Chapter
This edited volume demonstrates the potential of mixed-methods designs for the research of social networks and the utilization of social networks for other research. Mixing methods applies to the combination and integration of qualitative and quantitative methods. In social network research, mixing methods also applies to the combination of structural and actor-oriented approaches. The volume provides readers with methodological concepts to guide mixed-methods network studies with precise research designs and methods to investigate social networks of various sorts. Each chapter describes the research design used and discusses the strengths of the methods for that particular field and for specific outcomes.
Chapter
This edited volume demonstrates the potential of mixed-methods designs for the research of social networks and the utilization of social networks for other research. Mixing methods applies to the combination and integration of qualitative and quantitative methods. In social network research, mixing methods also applies to the combination of structural and actor-oriented approaches. The volume provides readers with methodological concepts to guide mixed-methods network studies with precise research designs and methods to investigate social networks of various sorts. Each chapter describes the research design used and discusses the strengths of the methods for that particular field and for specific outcomes.
Chapter
This edited volume demonstrates the potential of mixed-methods designs for the research of social networks and the utilization of social networks for other research. Mixing methods applies to the combination and integration of qualitative and quantitative methods. In social network research, mixing methods also applies to the combination of structural and actor-oriented approaches. The volume provides readers with methodological concepts to guide mixed-methods network studies with precise research designs and methods to investigate social networks of various sorts. Each chapter describes the research design used and discusses the strengths of the methods for that particular field and for specific outcomes.
Chapter
This edited volume demonstrates the potential of mixed-methods designs for the research of social networks and the utilization of social networks for other research. Mixing methods applies to the combination and integration of qualitative and quantitative methods. In social network research, mixing methods also applies to the combination of structural and actor-oriented approaches. The volume provides readers with methodological concepts to guide mixed-methods network studies with precise research designs and methods to investigate social networks of various sorts. Each chapter describes the research design used and discusses the strengths of the methods for that particular field and for specific outcomes.
Article
Full-text available
The structural properties of personal networks are potentially fruitful variables for explaining differences in attitudes, behaviors and conditions across individuals. When researchers apply structural measures to personal network data, they must decide whether to include or exclude ego from the adjacency matrix. This research note discusses several conceptual and empirical issues that should be considered in making that decision. Most personal network research over the past forty years has focussed on network composition (summaries of alter attributes) rather than network structure (analysis of the pattern of ties between alters). While compositional analyses yield vital information about the network and how it impacts the respondent, structural properties of networks offer a unique perspective and are a worthwhile pursuit (McCarty, 2002). As researchers create new software that makes it easier to construct studies that collect alter-to-alter tie evaluations from respondents (the basis for structural analyses of personal networks), structural features will add to the set of network features that are used to explain respondent attitudes, conditions, and behaviours. There is, of course, an established tradition of constructing egocentric networks within sociocentric networks. For example, Burt’s notion of structural holes is a concept derived from looking at the
Article
Full-text available
Since the mid-1850s, scholars have debated how technological innovation would affect community. The debate continues as the Internet makes it increasingly possible for people to socialize, shop, work, learn, and participate in leisure activities all from within their home. Will the movement of these previously public activities into the private realm lead to reduced participation in public activities? What will be the fate of community and social relations as a result of the growth of computer-mediated communication? Netville is a suburban Toronto development equipped with a high-speed network as part of its design. The clustering of homes within this area allowed us to study the social networks, civic involvement, Internet use, and attitudes of residents. This article explores the research approach of the Netville project and describes its main sources of data collection: surveys collected using computer-assisted interviewing and ethnographic fieldwork.
Article
Full-text available
Community ties with friends and relatives are a principal means by which people and households get supportive resources. Quantitative and qualitative data from the second East York study are used to evaluate six potential explanations of why different types of ties provide different kinds of supportive resources: tie strength, contact, group processes, kinship,network members' characteristics, and similarities and dissimilarities between network members in such characteristics. Most relatioships provide specialized support. The kinds of support provided are related more to characteristics of the relationship than to characteristics of the network members themselves. Strong ties provide emotional aid, small services, and companionship. Parents and adult children exchange financial aid, emotional aid, large services, and small services. Physically accessible ties provide services. Women provide emotional aid. Friends, neighbors, and siblings make up about half of all supportive relationships. The ensemble o...
Article
Full-text available
Presents a hierarchical mapping technique using a diagram of concentric circles to represent the degree of closeness of members of the social support network. This method does not assume that particular family or social relationships automatically ensure network membership. The method provides a relatively unbiased account of social support. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Chapter
Full-text available
With today‘s ubiquity and popularity of social network applications, the ability to analyze and understand large networks in an efficient manner becomes critically important. However, as networks become larger and more complex, reasoning about social dynamics via simple statistics is not a feasible option. To overcome these limitations, we can rely on visual metaphors. Visualization nowadays is no longer a passive process that produces images from a set of numbers. Recent years have witnessed a convergence of social network analytics and visualization, coupled with interaction, that is changing the way analysts understand and characterize social networks. In this chapter, we discuss the main goal of visualization and how different metaphors are aimed towards elucidating different aspects of social networks, such as structure and semantics. We also describe a number of methods where analytics and visualization are interwoven towards providing a better comprehension of social structure and dynamics. KeywordInformation visualization-visual analytics-semantic filtering-centrality
Article
Full-text available
Family researchers have made extensive use of household survey data that predefine the boundaries of the family system and focus on a small number of family dyads. The complexity of contemporary family contexts that have emerged as a result of divorce and remarriage might not be well served by those techniques. This article shows that social network methods can improve existing methods used to study families.
Article
To investigate the consequences of name generators for network data, we compare characteristics of egocentric networks from Wellman's East York survey, Fischer's Northern California Communities Study, the General Social Survey, and our study of networks in 81 Nashville, Tennessee neighborhoods. Network size, age and education heterogeneity, and average tie characteristics were most strongly affected by the name generator used. Network composition, and racial and sexual heterogeneity, were more invariant across different kinds of name generators.
Article
This is an argument for obtaining network data in the General Social Survey (GSS). The proposal requires a discussion of how and why at least minimal network data ought to be obtained in a probability sample survey of attitudes and behaviors.I begin with general concerns; briefly describing the proposal, available experience with the proposed items in large probability samples, how the proposed items are different from existing GSS items, kinds of variables that the proposed items would generate, and kinds of research questions that could be addressed if the proposed items were included in the GSS.I then address comparatively focused questions likely to arise in deliberations over the proposal; explaining how much interview time the proposed items are expected to require, why one rather than multiple name generators are proposed, why recording five alters is proposed, why intimacy is proposed as the name generator criterion content, why a short form is proposed for obtaining formal data, how priorities among name interpreter attribute items were established, how the proposed items elicit data on the strength and content of relationships, and how the proposed data might be coded for easy access by GSS users.