Network Research Experiments
Media
Part of Network Research Experiments
- Title
- Network Research Experiments
- extracted text
-
Network Research Experiments
ALLEN L. LINTON II and BETSY SINCLAIR
Abstract
This essay attempts to lay the foundation of modern social networks research with
a jolt toward innovative ways to create data or finding ways to access newly available data to address meaningful political questions. We focus on outlining potential
new resources for data, discuss the emergent theoretical arguments involving political networks, and present some current empirical estimates for the magnitude of
the effects of political networks. With the rise of social media and new technology,
ordinary citizens socialize online with old friends from elementary school, siblings
across the country, and local neighbors. While these relationships have long been
part of the social fabric of ordinary life, the ability to observe these exchanges directly
and on a daily basis is new, for both researchers and citizens. Records of our social
interactions have the potential to transform our academic understanding of the relationship between communication among family, friends, and coworkers and how we
become informed about politics and act politically. Whether the relationship occurs
on or offline, the social element of the relationship can be incredibly vital in understanding the way individuals react and interact with their political environments.
Processing and understanding these interactions, however, can be difficult without
knowing where to look for new information, what patterns to look for, and how to
interpret data in the context of other findings on the effects of social and political
networks. We conclude by considering the new and exciting directions this research
may take in the future.
INTRODUCTION
With the rise of social media and new technology, ordinary citizens socialize
online with old friends from elementary school, siblings across the country,
and local neighbors. While these relationships have long been part of the
social fabric of ordinary life, the ability to observe these exchanges directly
and on a daily basis is new, for both researchers and citizens. Records
of our social interactions have the potential to transform our academic
understanding of the relationship between communication among family,
friends, and coworkers and how we become informed about politics and
act politically. Whether the relationship occurs through an “offline” meeting
in a neighborhood or a Skype conversation across the world, the social
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
element of the relationship can be incredibly vital in understanding the way
individuals react and interact with their political environments. Processing
and understanding these interactions, however, can be difficult without
knowing where to look for new information, what patterns to look for, and
how to interpret data in the context of other findings on the effects of social
and political networks. This brief essay attempts to lay the foundation of
modern social networks research with a jolt toward innovative ways to
create data or finding ways to access newly available data to address meaningful political questions. We focus on outlining potential new resources
for data, discuss the emergent theoretical arguments involving political
networks, and present some current empirical estimates for the magnitude
of the effects of political networks. We conclude by considering the new and
exciting directions this research may take in the future.
FOUNDATIONAL RESEARCH
With the growth and now ubiquitous nature of online social media, the quantity and depth of social ties between individuals is recorded. This data has
enormous potential to yield insights into the relationships between an individual’s social networks and political choices but is as of yet largely inaccessible to the academic community. Here we describe the potential of this
data and some of the ways in which researchers can leverage these resources
to focus on questions involving social networks and political behaviors and
choices.
NEW DATA RESOURCES
Facebook, one of the most popular social networking sites, allows users the
ability to create online profiles about themselves and share these profiles
within a user-controlled network of other people called “friends.” Facebook
user data can provide researchers with a wealth of information germane to
discussions of political networks and political information. First, one can
find information on “friends” to see their relationship with a user in the
online world and in the offline world. Traditionally, the dichotomy between
on and offline social spheres suggested a barrier between the people and
quality of interactions shared in these spaces. Facebook data provides insight
on the type of relationship with the user (school friends, coworkers, family
members, etc.) and its depth with respect to shared communication. It is
possible to reasonably measure how often people interact with others online
by counting the frequency a person shares stories, comments, or “likes”
different types of communication. It is possible to perform a similar measure
for offline interaction using instances where users are tagged in pictures
Network Research Experiments
3
together. Where self-reporting of certain social networking questions left
doubt about the veracity of responses, records of varying types of connections can help researchers map ties to each other and accurately predict the
strength of those ties.
A second and related strength of online social networking sites is the
insight into the type of political communications shared between users.
Tracking Facebook posts (status updates and information shared on a
user’s wall) and responses to posts can reveal the political heterogeneity of
friendship networks and sources of political information that are identified
as credible. We can now figure out not only where people are getting their
political information but what stories appeal to users from these sources
and what subpopulations respond to different types of political information.
Therefore, we can identify a liberal person who maintains a homogeneous
liberal network and quantify differences between that user’s friends who
respond to security versus environmental politics. Beyond the complexity
of Facebook, Twitter has emerged as an avenue that researches can utilize
to track political information. Twitter defines itself as “a real-time information network that connects you to the latest stories, ideas, opinions and
news about what you find interesting.” Researchers can recreate networks
between users to determine the salient political stories and establish their
transmission through a network at low costs. Potentially, sites such as
Twitter can be a preeminent source for disentangling political information
consumption and constructing clustering patterns based on partisanship,
issue salience, and node centrality within a network.
Structured websites such as Facebook and Twitter should not overshadow
the development of more open-source information-based networking sites
such as Reddit. Users collectively vote on discussions and stories, allowing
people to control the popularity and salience of conversation on the site.
Similar to other social networking sites, Reddit is not restricted to politics
but user’s collective political discourse can reveal several important findings
for scholars. First, users on Reddit are not bonded by friendship; therefore,
“popular” political topics emerge without the same filters worrisome in
mediated relationships. An unmediated community is more susceptible to
cross-partisan interaction and emerging topics may truly reflect the general
population as compared to just friendship networks. Second and consistent
with the theme throughout this section, researchers can identify the role
of different types of media sources on public discourse. While users can
generate topics independent of a traditional news source, many users begin
conversations by linking to a traditional news source or popular blog. These
open source sites allow scholars to access data on the types of information
that resonate with broader audiences by utilizing the voting system for
story popularity. Although not as popular as Facebook or Twitter, Reddit is
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
quickly emerging as a destination for sharing and conversing about political
information.
Regardless of the platform, online social networks should be utilized to
further construct and disaggregate the commotion innate in social network
studies. Networking sites such as Facebook, Twitter, Reddit, and YouTube
remove barriers to mapping and researching social networks by allowing
access to quantifiable connections between users. These companies now
possess the data necessary to answer a myriad of political questions.
Establishing partnerships between these companies and academics would
allow us to understand who is central to information diffusion by seeing
how many people interact with a type of communication, for example, and
other types of questions as well. Are there different levels of social pressure
when viewing how users respond to interactions from family (immediate
vs distant) and friends (close vs far)? What is the relationship between
network structure and the transmission of political discussion? If fruitful
research on social networks is to continue, scholars should try to access
the increasingly available and detailed information gleaned through online
social media sites. Some of this material can be accessed directly (Twitter,
e.g., has established an API that can be scraped with computing software
packages such as Python, but Facebook prohibits this kind of activity). Yet,
these companies may be willing to open their data vaults to partnerships
because while they have data on the accuracy in tie number, strength, and
content of an individual’s social network, they may not know the extent to
which they can gain new insights by looking at the relationship between
observable behavior and political choices. Academics offer much in the way
of theoretical innovation to these partnerships.
EMERGENT THEORETICAL ARGUMENTS
Whether it is online or offline, an individual’s placement within a social network can substantively impact one’s political decision making. One of the
most important developments in social network studies involves the role of
social environments on political decisions: when an individual is confronted
with a set of political choices, her decision is a consequence of her social environment. Within this theoretical framework, social environmental forces can
take on different meanings. One such theory considers a networked individual relying on a social environment for learning about politics through
deliberation and exposure to multiple perspectives, resulting in an individual being more informed to address the current political choice.
Sokhey and McClurg (2012) present a normative investigation of the impact
of an individual’s social network on the “quality of voting decisions.” Quality control for voting decisions is operationalized by using the Lau and
Network Research Experiments
5
Redlawsk (1997) article on “correct voting,” the process of deciding who
to vote for when given all available information on the candidates. In
considering networked individuals making “correct votes,” Sokhey and
McClurg interrogate the work networks accomplish with respect to political
discourse. The social environment can potentially change an individual’s
perspective when interpreting outside political information. Conceivably
networked individuals will be exposed to more information but it is unclear
if that information immobilizes them from the electoral politics or helps
mature initial political preferences (Mutz, 2002). With these competing
questions in mind, Sokhey and McClurg test two different hypotheses: (i)
political disagreement within social networks will lead to more “incorrect”
votes and (ii) increased political expertise in one’s social network will result
in more “correct” votes. Analyzing data from the 1992 Cross-National Study
and the 2000 American National Election Study, they find political disagreement to be an overwhelmingly important predictor of “correct” voting with
more disagreement resulting in a high chance of incorrect voting. More
substantively, the impact of social networks appears to be as a resource
for clear signals about candidates rather than a tool for enhancing political
information. Unlike theories of networks providing information shortcuts
espoused by Downs (1957), “correct” voting is not a consequence for learning more about the political issue; rather, it is receiving and responding to
clear social cues from other members of the network. Ultimately, the social
environment of the individuals studied provided political information cues
that directly influenced varying political choices.
While political information represents one type of social environment
consequence, social pressure emerges as a second theoretical explanation
for a networked individual’s behavior. Social pressure through conformity
functions by establishing a given behavior as necessary for association with
other members of a social network. For inclusion in the group in question, an
individual must share the act or face ostracizing from the perceived audience
which established the behavior as both normal and necessary. As compared
to earlier voter mobilization field experiments where social pressure could
work to shame people into participating, cueing tangible relationships
between networked people can produce the social conformity to maintain
membership within a self-selected community. Sinclair (2012) and Sinclair
et al. (2013) investigate this argument through a quasi-experiment where
subjects are randomly assigned to be contacted by local canvassers (people
from the same neighborhood) versus canvassers from other communities.
It is theorized that turnout represents the social norm that contacted people
will need to conform to for network membership.
In effect, social conformity as a type of social pressure is particularly persuasive among networked individuals and works in a two-tiered process.
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
First, networked individuals need to be engaged by other members in the
network to identify the social norm being applied, with that norm emerging as the new standard for group inclusion. Voter mobilization campaigns
using local volunteers signal the importance of turnout in terms of representation for a specific set of people within a given geographic or social space.
The second tier of the process necessary for social persuasion through conformity involves the credible threat of accountability from that community.
As Sinclair et al. outline, compliance is hypothesized to increase because contacted individuals more than likely will interact with each other in the future.
Local canvassers turned neighbors can potentially raise the issue of turnout
in future interactions creating a need to conform; this conformity pressure is
unfounded if the canvasser is not embedded in one’s network as the individual does not share the same “in-group” norms as he and the canvasser come
from different groups.
Sinclair et al. conducted a field experiment with the Strategic Concepts in
Organizing and Policy Education, a Los Angeles–based grassroots political
group working on turnout and civic engagement in low-income neighborhoods. Local canvassing was defined as a volunteer contacting a resident in
the same zip code as the volunteer. Individuals contacted by volunteers from
a different zip code represented nonlocal canvassers, although both groups
used the same scripts when contacting individuals. Consistent with previous
literature, those contacted by any canvasser reported higher turnout rates
than individuals in the control group who received no contact (Green & Gerber, 2008). Furthermore, results confirm the hypothesis that local canvassers
are more effective in turning out the vote than nonlocal canvassers, suggesting networked persuasion to be a viable campaign strategy and meaningful
implication for social pressure on political decision making. The nature of the
experiment isolated the difference between local and nonlocal canvassers by
utilizing the same messaging cue for social inclusion. This experiment articulates the power of social networks as vehicles to create conformity for maintained membership rather than shaming. The networked individual reacts
by consequence of the network exerting pressure in making a decision to
participate, not mere mobilization as detailed in prior behavior literature.
Sinclair et al.’s emphasis on social environments as influential through
social pressure and McClurg and Sokhey’s highlighting networks as spaces
for political information represent two of the largest theories in networked
research. Both theories place emphasis on the individual connected to the
social network but the range of the theories provides scholars different
areas of research on the strength of either theory on a particular political
decision. Sinclair et al. highlight that people do not necessarily need to be
influenced through their closest social networks as much as strategically
located networks with differentiated communication patterns. McClurg and
Network Research Experiments
7
Sokhey, through using observational data, rely on close social ties to define
information transfers and specific political decisions, whereas Sinclair et al.
concern themselves with the decision to participate at all. Both theories
address different approaches to the impact of social networks but the two
theories converge with respect to individuals making political decisions
influenced by their social environment.
CUTTING-EDGE RESEARCH
Social networking relationships can have substantive impact on political
decision making, but measuring these effects is difficult using observational
data. Too many external confounders can confuse the relationship between
network diffusion and alternative explanations for altered behavior in
response to one’s social environment. Eluding these concerns requires
methodological tools, allowing researchers to stabilize all other explanatory
differences while isolating the impact of networked behavior. Experimental
studies provide the best tool to accomplish this across online and offline
political networks.
One useful experiment attempting to quantifiably measure the impact of
social pressure on political behavior was conducted by Gerber, Green and
Larimer (2008) before the 2006 Michigan primary election. Operationalizing
social pressure involves four different treatments testing different degrees
of social pressure on voter mobilization: appeals to civic duty, appeals that
their voting behavior will be monitored by researchers, appeals featuring
an individual’s voting record will be shared with others in their household,
and appeals that their voting record will be shared with their neighbors.
Gerber et al. propose that social norms become causally influential through
three processes: (i) recognition of the norm, (ii) acceptance of the norm in
question, and (iii) enforcement of norms through fear of exclusion from
the norm-making association. Each treatment group interacts with different
levels of social norm influence with civic duty appeals cueing awareness of a
norm with no accountability while exposure to neighbors cues norms, their
internalization, and accountability. Results show an 8% increase in turnout
among individuals exposed to the neighbors’ treatment; this finding is not
only the highest of any treatment in their experiment, it exceeds exposure to
phone calls and is comparable to in-person mobilization efforts. The innovation in their approach focused on social pressure and message content
rather than simple contact disconnected from an individual’s interpersonal
engagement within a network. This type of large field experiment offers a
departure point for future research on political behavior by incorporating an
individual’s placement within a social sphere governed by specific norms
into a given decision-making calculus.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Incorporating new data resources from social media spaces can also
illuminate the role of an individual’s environment on political behavior.
Bond et al. (2012) present a credible system for identifying and inferring
social ties between individuals through observable online behavior on
Facebook. Available data enabled researchers to confidently predict an
individual’s best friends (strong ties) through compiling public interactions
on Facebook with 82% accuracy. Surprisingly, traditional relationships
between individuals such as shared history and attributes did not strongly
impact the model’s predictive power. This assumes that the social behavior
of networked individuals will be received by others and can potentially alter
future observed behavior.
Bond et al. empirically investigated this by conducting a mobilization
experiment with 61 million Facebook users during the 2012 Congressional
campaign. Users were divided into three groups: a control group exposed to
no message prompts; an information message group exposed to a prompt
to vote, information on polling centers, and a button to click to declare
that individual voted; the final group represented the “social message”
group featuring the same prompts as the information group except the
voting button also displayed six randomly selected Facebook friends who
clicked the button indicating their decision to vote. The design allows for
users to report to friends their behavior interpreted as political expression
and inclusion within a group doing a social norm—voting. Although their
estimated indirect effects are small (approximately 0.01% points), their
results show exposure to social messages positively increased turnout
directly by approximately 60,000 voters and diffused throughout the network for an indirect increase of nearly 280,000 voters (Bond et al., 2012). The
design allowed for self-reporting to be confirmed with publically available
voting records effectively combining observational data with experimental
methods. Where Traud et al. (2011) find no effect of online mobilization, new
data sources allow researchers the ability to go beyond traditional sample
sizes to detect real effects of social-based mobilization. New evidence from
online quasi-experiments provides credible results supporting claims of
online networks tangibly and substantively impacting political behavior.
Observational evidence fails to serve our purposes for understanding
offline network construction and diffusion, too. Empirical claims founded
upon observational data cannot effectively parse out differences between
(i) the universal social environment an individual and her network is
exposed to, (ii) self-selection biases as people with shared preferences tend
to connect with each other, and (iii) unconsidered externalities distorting
the effect of the network on individual behavior. David Nickerson (2008)
compares these problems with prior literature’s insistence to dismiss the
power of “interpersonal influence,” an assumption of an atomistic voter—an
Network Research Experiments
9
individual who makes political decisions without a connection to his
social network. If, however, political behavior can be influenced or caused
by relationships within a social network, research should be focused on
measuring individuals not directly exposed to treatments for behavior (i.e.,
voter mobilization campaigns).
Nickerson proposed a series of field experiments to measure these effects
by considering the contagious nature of get out the vote campaigns among
residents of two person households, focusing on the behavior of the person
who lives with but does not receive the mobilization treatment. Uniquely,
these experiments avoid the worrisome assumptions of an atomistic voter
and external influencing factors by simultaneously conducting a placebo
experiment with some households exposed to a pitch to recycle instead of
a get out the vote message. In addition, experiments were conducted in
two different cities (Denver and Minneapolis) to acquire a high number of
treated and placebo cases. Turnout of people who did not answer the door
in two-person households can then be measured as the research design
accounts for peer effects, similar interests, and self-selection with the placebo
experiment. The placebo protocol (recycling pitch) establishes the necessary baseline for individual and secondary networked behavior; the only
difference between treatment and control is the exposure to a voter mobilization prompt. Subsequent conclusions from the experiment do not require
assumptions on baseline voter rates for each individual in a household; a
simple calculation of differences in average turnout for secondary household
members between voter-mobilized treatments and recycling treatments
reveals the depth of contagious political behavior through social networks.
Nickerson’s results show that the get out the vote campaign produced significantly higher turnout and pooled results from both experiments show a
statistically significant effect on nontreated individual’s voting behavior. The
methodological approach to answering questions of social network behavior
is as important as substantive findings on altered behavior. Use of placebo
experimental controls allows for researchers to confidently isolate the diffusion of political behavior through networks without costly assumptions and
risk of externalities embedded in observational study. Imposing limits on the
size of the network, in this case a two-person network, allows for more manageable analysis without concerns of unaccounted for influences within a
network.
Experiments similar to Nickerson (2008) become problematic when
attempting to generalize findings to different cases. As he points out, there
remain questions on the extent that voting contagion can occur without an
exogenous shock similar to a mobilization campaign. Taking place amid
lower salience elections, there is uncertainty if more salient elections change
the contagious nature of voting. Furthermore, it is unclear whether similar
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
effects of voting behavior can be found in larger social networks with
multiple types of interpersonal connections. As referenced in the theoretical
argument debate earlier, Nickerson acknowledges difficulties in specifying
the process that transfers behavior (either lower costs through greater
information or social pressure by a housemate).
One concern raised by Nickerson requires special attention and involves the
stable unit treatment value assumption (SUTVA), a core tenant stated when
claiming causal relationships between variables in experimental studies. By
focusing on individuals, Nickerson argues previous experiments designed
to measure mobilization (Gerber & Green, 2000a, 2000b; Michelson, 2003;
Nickerson, 2007) did not accurately account for spillover effects of treatments
to unassigned subjects. Given the contagious nature of voting with a single
individual assigned to treatment impacting behavior of nonassigned subjects, experimenters may worry that treatment is “spilling over” to other individuals. Sinclair et al. (2012) investigate the value of multilevel experiment
designs applied to a large-scale mobilization experiment testing the impact
of social pressure on local voters. Multilevel experiments randomly assign
subjects to treatment and control plus randomly assign the treatment to other
subjects within an individual’s social network. In this context, implementing
a multilevel experiment on social pressure expand beyond Nickerson (2008)
by considering the spillover effects in a given household and among an individual’s neighborhood. Despite ignoring the mechanism of indirect influence
on other parts of a social network, Sinclair et al. argue that individuals can
be influenced differently by neighbors than compared with housemates.
Sinclair et al. apply the Gerber, Green, and Larimer’s social pressure mailings to a 2009 Illinois special election and find mixed results of spillover
within populations. Within households, spillover effects are found although
not at the levels reported by Nickerson (2008); differences in turnout levels are attributable to in-person conversations being more likely to prompt
dialog than postcards. Interestingly, no evidence of spillover effects between
individuals within zip-code-defined neighborhoods. Multilevel experimentation allowed for the identification of a null finding of significance given
the difficulty in mapping possible pathways of social influence within complex social networks. Sinclair et al.’s use of precise multilevel experiments
identifies areas of research to be focused on (household transmission of mobilization stimuli) and areas where limited resources do not need to be spent
in bulk (neighborhood diffusion of mobilization treatments).
Regardless of the online or offline nature of social network influence on
political behavior, credible empirical evidence suggests these relationships
do matter for determining how individuals will express themselves politically. Investigations of social network influence using observational data
consistently make questionable assumptions based on correlations between
Network Research Experiments
11
seemingly connected people and their shared behavior. Observational data
falls short, however, in accounting for other factors that could explain
correlated behavior limiting causal claims. Social network research has
increasingly relied on experimental studies to create an environment to
control for unobserved correlates and measure exogenous shock diffusion through social networks. In particular, experimental strategies allow
researchers three different types of manipulations for measurement: (i)
creating an exogenous shock and monitor its diffusion, (ii) alter/control
information flows through the network, or (iii) randomize individual
placements within the network (Nickerson, 2011). Researchers will need
to carefully consider the question they hope to answer when determining
which experimental method to use, but experimental studies provide
the more comprehensive tools available to confidently identify causal
relationships between networked individual’s political behaviors.
FUTURE RESEARCH WITH EXPERIMENTAL PROCESSES
Experimental research is frequently concerned with questions of external
validity, which means there is great potential for new directions for research
on questions previously unstudied using a methodological approach built
for causal claims. One such area involves questions of interest-based voting
as alluded to by Sokhey and McClurg. Most experimental studies focus
on turnout or political communication through social networks, but few
begin to identify the mechanisms surrounding interest building and mobilization in electoral politics. Growths in multilevel experiments now allow
researchers to explore theoretically stable and well-defined social networks
to understand how different types of political information are communicated and when political behavior appeals will be effective. Online data from
social media sites greatly reduce the cost of conducting experiments and
allow researchers new resources to extend beyond geographic boundaries
(zip codes or neighborhood lines) when defining unique networks and their
strength.
Another important area for scholarly growth involves understanding
young people’s involvement and engagement in politics. Literature on
political engagement begins with youth apathy and transitions toward the
importance of socioeconomic status for understanding differences in adult
consumption and participation in electoral contexts. Without traditional
predictors of participation meaningfully available in younger populations,
access to new media can provide insights in how young people form their
political identities. Advanced observational and experimental research on
social networks can reveal the influences of environment on developing
political ideologies, reveal changes in issue salience as individuals mature,
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
and capture the different ways young people communicate about politics.
New data and direct experimentation can enhance our understanding
of how a single individual decides to be political before obtaining more
conventional political resources. There are many questions left to be investigated with respect to social networks, political participation, and how we
are at once influencing others while also being influenced by those near and
far. Access to new data allows researchers to test hypotheses to improve our
understandings of being interconnected, political people.
REFERENCES
Bond, R. M., Fariss, C. J., Jones, J. J., Kramer, A. D. I., Marlow, C., Settle, J. E., & Fowler,
J. H. (2012). A 61-million-person experiment in social influence and political mobilization. Nature, 489, 295–298. doi:10.1038/nature11421
Downs, A. (1957). An economic theory of democracy. New York, NY: Harper.
Gerber, A. S., & Green, D. P. (2000a). The effect of canvassing, telephone calls, and
direct mail on voter turnout: A field experiment. American Political Science Review,
94(3), 653–663.
Gerber, A. S., & Green, D. P. (2000b). The effect of canvassing, telephone calls, and
direct mail on voter turnout: A field experiment. American Political Science Review,
102(1), 33–48.
Gerber, A., Green, D., & Larimer, C. (2008). Social Pressure and Voter Turnout:
Evidence From A Large-Scale Field Experiment. American Political Science Review,
102(1), 33–48.
Green, D. P., & Gerber, A. S. (2008). Get Out The Vote: How to Increase Voter Turnout
(Second ed.). Washington, DC: Brookings Institution Press.
Lau, R., & Redlawsk, D. (1997). Voting correctly. American Political Science Review,
91(3), 585–598.
Michelson, M. (2003). Getting out the Latino vote: How door-to-door canvassing influences voter turnout in rural central California. American Political Science
Review, 25(3), 247–263. doi:10.1023/A:1025167607369
Mutz, D. (2002). The consequences of cross-cutting networks for political participation. American Political Science Review, 46(4), 838–855.
Nickerson, D. W. (2007). Quality is job one: Professional and volunteer voter mobilization calls. American Journal of Political Science, 51(2), 269–282. doi:10.1111/
j.1540-5907.2007.00250.x
Nickerson, D. W. (2008). Is voting contagious? Evidence from two field experiements.
American Political Science Review, 102(1), 49–57. doi:10.1017/S0003055408080039
Nickerson, D. W. (2011). Social networks and political context. In J. Druckman, D.
Green, J. Kuklinski & A. Lupia (Eds.), Cambridge handbook of experimental political
science (pp. 273–286). Cambridge, England: Cambridge University Press.
Sinclair, B. (2012). The social citizen: Peer networks and political behavior. Chicago, IL:
University of Chicago Press.
Network Research Experiments
13
Sinclair, B., McConnell, M., & Michelson, M. R. (2013). Local Canvassing: The
Efficacy of Grassroots Voter Mobilization. Political Communication, 30(1), 42–57.
Sinclair, B., McConnell, M., & Green, D. P. (2012). Detecting spillover effects: Design
and analysis of multi-level experiments. American Journal of Political Science, 56,
1055–1069.
Sokhey, A. E., & McClurg, S. D. (2012). Social networks and correct voting. The Journal
of Politics, 74(3), 751–764. doi:10.1017/S0022381612000461
Traud, A. L., Kelsic, E. D., Mucha, P. J., & Porter, M. A. (2011). Comparing community
structure to characteristics in online collegiate social networks. SIAM Review, 53(3),
526–543. doi:10.1137/080734315
ALLEN L. LINTON II SHORT BIOGRAPHY
Allen L. Linton II is a PhD student in political science at the University of
Chicago. His research interests are youth political participation and socialization, the impact of new/social media on participation and governance, and
local politics.
BETSY SINCLAIR SHORT BIOGRAPHY
Betsy Sinclair is an associate professor of political science at Washington
University in St Louis. Her research focuses on the social foundations of
participatory democracy in American politics, with an emphasis on how an
individual’s social network influences political choices and behaviors.
RELATED ESSAYS
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and
Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
To Flop Is Human: Inventing Better Scientific Approaches to Anticipating
Failure (Methods), Robert Boruch and Alan Ruby
Repeated Cross-Sections in Survey Data (Methods), Henry E. Brady and
Richard Johnston
Ambulatory Assessment: Methods for Studying Everyday Life (Methods),
Tamlin S. Conner and Matthias R. Mehl
Models of Nonlinear Growth (Methods), Patrick Coulombe and James P.
Selig
Quantile Regression Methods (Methods), Bernd Fitzenberger and Ralf
Andreas Wilke
The Evidence-Based Practice Movement (Sociology), Edward W. Gondolf
Meta-Analysis (Methods), Larry V. Hedges and Martyna Citkowicz
The Use of Geophysical Survey in Archaeology (Methods), Timothy J.
Horsley
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Longitudinal Data Analysis (Methods), Todd D. Little et al.
Structural Equation Modeling and Latent Variable Approaches (Methods),
Alex Liu
Data Mining (Methods), Gregg R. Murray and Anthony Scime
Remote Sensing with Satellite Technology (Archaeology), Sarah Parcak
Quasi-Experiments (Methods), Charles S. Reichard
Digital Methods for Web Research (Methods), Richard Rogers
Virtual Worlds as Laboratories (Methods), Travis L. Ross et al.
Modeling Life Course Structure: The Triple Helix (Sociology), Tom Schuller
Content Analysis (Methods), Steven E. Stemler
Person-Centered Analysis (Methods), Alexander von Eye and Wolfgang
Wiedermann
Translational Sociology (Sociology), Elaine Wethington
-
Network Research Experiments
ALLEN L. LINTON II and BETSY SINCLAIR
Abstract
This essay attempts to lay the foundation of modern social networks research with
a jolt toward innovative ways to create data or finding ways to access newly available data to address meaningful political questions. We focus on outlining potential
new resources for data, discuss the emergent theoretical arguments involving political networks, and present some current empirical estimates for the magnitude of
the effects of political networks. With the rise of social media and new technology,
ordinary citizens socialize online with old friends from elementary school, siblings
across the country, and local neighbors. While these relationships have long been
part of the social fabric of ordinary life, the ability to observe these exchanges directly
and on a daily basis is new, for both researchers and citizens. Records of our social
interactions have the potential to transform our academic understanding of the relationship between communication among family, friends, and coworkers and how we
become informed about politics and act politically. Whether the relationship occurs
on or offline, the social element of the relationship can be incredibly vital in understanding the way individuals react and interact with their political environments.
Processing and understanding these interactions, however, can be difficult without
knowing where to look for new information, what patterns to look for, and how to
interpret data in the context of other findings on the effects of social and political
networks. We conclude by considering the new and exciting directions this research
may take in the future.
INTRODUCTION
With the rise of social media and new technology, ordinary citizens socialize
online with old friends from elementary school, siblings across the country,
and local neighbors. While these relationships have long been part of the
social fabric of ordinary life, the ability to observe these exchanges directly
and on a daily basis is new, for both researchers and citizens. Records
of our social interactions have the potential to transform our academic
understanding of the relationship between communication among family,
friends, and coworkers and how we become informed about politics and
act politically. Whether the relationship occurs through an “offline” meeting
in a neighborhood or a Skype conversation across the world, the social
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
element of the relationship can be incredibly vital in understanding the way
individuals react and interact with their political environments. Processing
and understanding these interactions, however, can be difficult without
knowing where to look for new information, what patterns to look for, and
how to interpret data in the context of other findings on the effects of social
and political networks. This brief essay attempts to lay the foundation of
modern social networks research with a jolt toward innovative ways to
create data or finding ways to access newly available data to address meaningful political questions. We focus on outlining potential new resources
for data, discuss the emergent theoretical arguments involving political
networks, and present some current empirical estimates for the magnitude
of the effects of political networks. We conclude by considering the new and
exciting directions this research may take in the future.
FOUNDATIONAL RESEARCH
With the growth and now ubiquitous nature of online social media, the quantity and depth of social ties between individuals is recorded. This data has
enormous potential to yield insights into the relationships between an individual’s social networks and political choices but is as of yet largely inaccessible to the academic community. Here we describe the potential of this
data and some of the ways in which researchers can leverage these resources
to focus on questions involving social networks and political behaviors and
choices.
NEW DATA RESOURCES
Facebook, one of the most popular social networking sites, allows users the
ability to create online profiles about themselves and share these profiles
within a user-controlled network of other people called “friends.” Facebook
user data can provide researchers with a wealth of information germane to
discussions of political networks and political information. First, one can
find information on “friends” to see their relationship with a user in the
online world and in the offline world. Traditionally, the dichotomy between
on and offline social spheres suggested a barrier between the people and
quality of interactions shared in these spaces. Facebook data provides insight
on the type of relationship with the user (school friends, coworkers, family
members, etc.) and its depth with respect to shared communication. It is
possible to reasonably measure how often people interact with others online
by counting the frequency a person shares stories, comments, or “likes”
different types of communication. It is possible to perform a similar measure
for offline interaction using instances where users are tagged in pictures
Network Research Experiments
3
together. Where self-reporting of certain social networking questions left
doubt about the veracity of responses, records of varying types of connections can help researchers map ties to each other and accurately predict the
strength of those ties.
A second and related strength of online social networking sites is the
insight into the type of political communications shared between users.
Tracking Facebook posts (status updates and information shared on a
user’s wall) and responses to posts can reveal the political heterogeneity of
friendship networks and sources of political information that are identified
as credible. We can now figure out not only where people are getting their
political information but what stories appeal to users from these sources
and what subpopulations respond to different types of political information.
Therefore, we can identify a liberal person who maintains a homogeneous
liberal network and quantify differences between that user’s friends who
respond to security versus environmental politics. Beyond the complexity
of Facebook, Twitter has emerged as an avenue that researches can utilize
to track political information. Twitter defines itself as “a real-time information network that connects you to the latest stories, ideas, opinions and
news about what you find interesting.” Researchers can recreate networks
between users to determine the salient political stories and establish their
transmission through a network at low costs. Potentially, sites such as
Twitter can be a preeminent source for disentangling political information
consumption and constructing clustering patterns based on partisanship,
issue salience, and node centrality within a network.
Structured websites such as Facebook and Twitter should not overshadow
the development of more open-source information-based networking sites
such as Reddit. Users collectively vote on discussions and stories, allowing
people to control the popularity and salience of conversation on the site.
Similar to other social networking sites, Reddit is not restricted to politics
but user’s collective political discourse can reveal several important findings
for scholars. First, users on Reddit are not bonded by friendship; therefore,
“popular” political topics emerge without the same filters worrisome in
mediated relationships. An unmediated community is more susceptible to
cross-partisan interaction and emerging topics may truly reflect the general
population as compared to just friendship networks. Second and consistent
with the theme throughout this section, researchers can identify the role
of different types of media sources on public discourse. While users can
generate topics independent of a traditional news source, many users begin
conversations by linking to a traditional news source or popular blog. These
open source sites allow scholars to access data on the types of information
that resonate with broader audiences by utilizing the voting system for
story popularity. Although not as popular as Facebook or Twitter, Reddit is
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
quickly emerging as a destination for sharing and conversing about political
information.
Regardless of the platform, online social networks should be utilized to
further construct and disaggregate the commotion innate in social network
studies. Networking sites such as Facebook, Twitter, Reddit, and YouTube
remove barriers to mapping and researching social networks by allowing
access to quantifiable connections between users. These companies now
possess the data necessary to answer a myriad of political questions.
Establishing partnerships between these companies and academics would
allow us to understand who is central to information diffusion by seeing
how many people interact with a type of communication, for example, and
other types of questions as well. Are there different levels of social pressure
when viewing how users respond to interactions from family (immediate
vs distant) and friends (close vs far)? What is the relationship between
network structure and the transmission of political discussion? If fruitful
research on social networks is to continue, scholars should try to access
the increasingly available and detailed information gleaned through online
social media sites. Some of this material can be accessed directly (Twitter,
e.g., has established an API that can be scraped with computing software
packages such as Python, but Facebook prohibits this kind of activity). Yet,
these companies may be willing to open their data vaults to partnerships
because while they have data on the accuracy in tie number, strength, and
content of an individual’s social network, they may not know the extent to
which they can gain new insights by looking at the relationship between
observable behavior and political choices. Academics offer much in the way
of theoretical innovation to these partnerships.
EMERGENT THEORETICAL ARGUMENTS
Whether it is online or offline, an individual’s placement within a social network can substantively impact one’s political decision making. One of the
most important developments in social network studies involves the role of
social environments on political decisions: when an individual is confronted
with a set of political choices, her decision is a consequence of her social environment. Within this theoretical framework, social environmental forces can
take on different meanings. One such theory considers a networked individual relying on a social environment for learning about politics through
deliberation and exposure to multiple perspectives, resulting in an individual being more informed to address the current political choice.
Sokhey and McClurg (2012) present a normative investigation of the impact
of an individual’s social network on the “quality of voting decisions.” Quality control for voting decisions is operationalized by using the Lau and
Network Research Experiments
5
Redlawsk (1997) article on “correct voting,” the process of deciding who
to vote for when given all available information on the candidates. In
considering networked individuals making “correct votes,” Sokhey and
McClurg interrogate the work networks accomplish with respect to political
discourse. The social environment can potentially change an individual’s
perspective when interpreting outside political information. Conceivably
networked individuals will be exposed to more information but it is unclear
if that information immobilizes them from the electoral politics or helps
mature initial political preferences (Mutz, 2002). With these competing
questions in mind, Sokhey and McClurg test two different hypotheses: (i)
political disagreement within social networks will lead to more “incorrect”
votes and (ii) increased political expertise in one’s social network will result
in more “correct” votes. Analyzing data from the 1992 Cross-National Study
and the 2000 American National Election Study, they find political disagreement to be an overwhelmingly important predictor of “correct” voting with
more disagreement resulting in a high chance of incorrect voting. More
substantively, the impact of social networks appears to be as a resource
for clear signals about candidates rather than a tool for enhancing political
information. Unlike theories of networks providing information shortcuts
espoused by Downs (1957), “correct” voting is not a consequence for learning more about the political issue; rather, it is receiving and responding to
clear social cues from other members of the network. Ultimately, the social
environment of the individuals studied provided political information cues
that directly influenced varying political choices.
While political information represents one type of social environment
consequence, social pressure emerges as a second theoretical explanation
for a networked individual’s behavior. Social pressure through conformity
functions by establishing a given behavior as necessary for association with
other members of a social network. For inclusion in the group in question, an
individual must share the act or face ostracizing from the perceived audience
which established the behavior as both normal and necessary. As compared
to earlier voter mobilization field experiments where social pressure could
work to shame people into participating, cueing tangible relationships
between networked people can produce the social conformity to maintain
membership within a self-selected community. Sinclair (2012) and Sinclair
et al. (2013) investigate this argument through a quasi-experiment where
subjects are randomly assigned to be contacted by local canvassers (people
from the same neighborhood) versus canvassers from other communities.
It is theorized that turnout represents the social norm that contacted people
will need to conform to for network membership.
In effect, social conformity as a type of social pressure is particularly persuasive among networked individuals and works in a two-tiered process.
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
First, networked individuals need to be engaged by other members in the
network to identify the social norm being applied, with that norm emerging as the new standard for group inclusion. Voter mobilization campaigns
using local volunteers signal the importance of turnout in terms of representation for a specific set of people within a given geographic or social space.
The second tier of the process necessary for social persuasion through conformity involves the credible threat of accountability from that community.
As Sinclair et al. outline, compliance is hypothesized to increase because contacted individuals more than likely will interact with each other in the future.
Local canvassers turned neighbors can potentially raise the issue of turnout
in future interactions creating a need to conform; this conformity pressure is
unfounded if the canvasser is not embedded in one’s network as the individual does not share the same “in-group” norms as he and the canvasser come
from different groups.
Sinclair et al. conducted a field experiment with the Strategic Concepts in
Organizing and Policy Education, a Los Angeles–based grassroots political
group working on turnout and civic engagement in low-income neighborhoods. Local canvassing was defined as a volunteer contacting a resident in
the same zip code as the volunteer. Individuals contacted by volunteers from
a different zip code represented nonlocal canvassers, although both groups
used the same scripts when contacting individuals. Consistent with previous
literature, those contacted by any canvasser reported higher turnout rates
than individuals in the control group who received no contact (Green & Gerber, 2008). Furthermore, results confirm the hypothesis that local canvassers
are more effective in turning out the vote than nonlocal canvassers, suggesting networked persuasion to be a viable campaign strategy and meaningful
implication for social pressure on political decision making. The nature of the
experiment isolated the difference between local and nonlocal canvassers by
utilizing the same messaging cue for social inclusion. This experiment articulates the power of social networks as vehicles to create conformity for maintained membership rather than shaming. The networked individual reacts
by consequence of the network exerting pressure in making a decision to
participate, not mere mobilization as detailed in prior behavior literature.
Sinclair et al.’s emphasis on social environments as influential through
social pressure and McClurg and Sokhey’s highlighting networks as spaces
for political information represent two of the largest theories in networked
research. Both theories place emphasis on the individual connected to the
social network but the range of the theories provides scholars different
areas of research on the strength of either theory on a particular political
decision. Sinclair et al. highlight that people do not necessarily need to be
influenced through their closest social networks as much as strategically
located networks with differentiated communication patterns. McClurg and
Network Research Experiments
7
Sokhey, through using observational data, rely on close social ties to define
information transfers and specific political decisions, whereas Sinclair et al.
concern themselves with the decision to participate at all. Both theories
address different approaches to the impact of social networks but the two
theories converge with respect to individuals making political decisions
influenced by their social environment.
CUTTING-EDGE RESEARCH
Social networking relationships can have substantive impact on political
decision making, but measuring these effects is difficult using observational
data. Too many external confounders can confuse the relationship between
network diffusion and alternative explanations for altered behavior in
response to one’s social environment. Eluding these concerns requires
methodological tools, allowing researchers to stabilize all other explanatory
differences while isolating the impact of networked behavior. Experimental
studies provide the best tool to accomplish this across online and offline
political networks.
One useful experiment attempting to quantifiably measure the impact of
social pressure on political behavior was conducted by Gerber, Green and
Larimer (2008) before the 2006 Michigan primary election. Operationalizing
social pressure involves four different treatments testing different degrees
of social pressure on voter mobilization: appeals to civic duty, appeals that
their voting behavior will be monitored by researchers, appeals featuring
an individual’s voting record will be shared with others in their household,
and appeals that their voting record will be shared with their neighbors.
Gerber et al. propose that social norms become causally influential through
three processes: (i) recognition of the norm, (ii) acceptance of the norm in
question, and (iii) enforcement of norms through fear of exclusion from
the norm-making association. Each treatment group interacts with different
levels of social norm influence with civic duty appeals cueing awareness of a
norm with no accountability while exposure to neighbors cues norms, their
internalization, and accountability. Results show an 8% increase in turnout
among individuals exposed to the neighbors’ treatment; this finding is not
only the highest of any treatment in their experiment, it exceeds exposure to
phone calls and is comparable to in-person mobilization efforts. The innovation in their approach focused on social pressure and message content
rather than simple contact disconnected from an individual’s interpersonal
engagement within a network. This type of large field experiment offers a
departure point for future research on political behavior by incorporating an
individual’s placement within a social sphere governed by specific norms
into a given decision-making calculus.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Incorporating new data resources from social media spaces can also
illuminate the role of an individual’s environment on political behavior.
Bond et al. (2012) present a credible system for identifying and inferring
social ties between individuals through observable online behavior on
Facebook. Available data enabled researchers to confidently predict an
individual’s best friends (strong ties) through compiling public interactions
on Facebook with 82% accuracy. Surprisingly, traditional relationships
between individuals such as shared history and attributes did not strongly
impact the model’s predictive power. This assumes that the social behavior
of networked individuals will be received by others and can potentially alter
future observed behavior.
Bond et al. empirically investigated this by conducting a mobilization
experiment with 61 million Facebook users during the 2012 Congressional
campaign. Users were divided into three groups: a control group exposed to
no message prompts; an information message group exposed to a prompt
to vote, information on polling centers, and a button to click to declare
that individual voted; the final group represented the “social message”
group featuring the same prompts as the information group except the
voting button also displayed six randomly selected Facebook friends who
clicked the button indicating their decision to vote. The design allows for
users to report to friends their behavior interpreted as political expression
and inclusion within a group doing a social norm—voting. Although their
estimated indirect effects are small (approximately 0.01% points), their
results show exposure to social messages positively increased turnout
directly by approximately 60,000 voters and diffused throughout the network for an indirect increase of nearly 280,000 voters (Bond et al., 2012). The
design allowed for self-reporting to be confirmed with publically available
voting records effectively combining observational data with experimental
methods. Where Traud et al. (2011) find no effect of online mobilization, new
data sources allow researchers the ability to go beyond traditional sample
sizes to detect real effects of social-based mobilization. New evidence from
online quasi-experiments provides credible results supporting claims of
online networks tangibly and substantively impacting political behavior.
Observational evidence fails to serve our purposes for understanding
offline network construction and diffusion, too. Empirical claims founded
upon observational data cannot effectively parse out differences between
(i) the universal social environment an individual and her network is
exposed to, (ii) self-selection biases as people with shared preferences tend
to connect with each other, and (iii) unconsidered externalities distorting
the effect of the network on individual behavior. David Nickerson (2008)
compares these problems with prior literature’s insistence to dismiss the
power of “interpersonal influence,” an assumption of an atomistic voter—an
Network Research Experiments
9
individual who makes political decisions without a connection to his
social network. If, however, political behavior can be influenced or caused
by relationships within a social network, research should be focused on
measuring individuals not directly exposed to treatments for behavior (i.e.,
voter mobilization campaigns).
Nickerson proposed a series of field experiments to measure these effects
by considering the contagious nature of get out the vote campaigns among
residents of two person households, focusing on the behavior of the person
who lives with but does not receive the mobilization treatment. Uniquely,
these experiments avoid the worrisome assumptions of an atomistic voter
and external influencing factors by simultaneously conducting a placebo
experiment with some households exposed to a pitch to recycle instead of
a get out the vote message. In addition, experiments were conducted in
two different cities (Denver and Minneapolis) to acquire a high number of
treated and placebo cases. Turnout of people who did not answer the door
in two-person households can then be measured as the research design
accounts for peer effects, similar interests, and self-selection with the placebo
experiment. The placebo protocol (recycling pitch) establishes the necessary baseline for individual and secondary networked behavior; the only
difference between treatment and control is the exposure to a voter mobilization prompt. Subsequent conclusions from the experiment do not require
assumptions on baseline voter rates for each individual in a household; a
simple calculation of differences in average turnout for secondary household
members between voter-mobilized treatments and recycling treatments
reveals the depth of contagious political behavior through social networks.
Nickerson’s results show that the get out the vote campaign produced significantly higher turnout and pooled results from both experiments show a
statistically significant effect on nontreated individual’s voting behavior. The
methodological approach to answering questions of social network behavior
is as important as substantive findings on altered behavior. Use of placebo
experimental controls allows for researchers to confidently isolate the diffusion of political behavior through networks without costly assumptions and
risk of externalities embedded in observational study. Imposing limits on the
size of the network, in this case a two-person network, allows for more manageable analysis without concerns of unaccounted for influences within a
network.
Experiments similar to Nickerson (2008) become problematic when
attempting to generalize findings to different cases. As he points out, there
remain questions on the extent that voting contagion can occur without an
exogenous shock similar to a mobilization campaign. Taking place amid
lower salience elections, there is uncertainty if more salient elections change
the contagious nature of voting. Furthermore, it is unclear whether similar
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
effects of voting behavior can be found in larger social networks with
multiple types of interpersonal connections. As referenced in the theoretical
argument debate earlier, Nickerson acknowledges difficulties in specifying
the process that transfers behavior (either lower costs through greater
information or social pressure by a housemate).
One concern raised by Nickerson requires special attention and involves the
stable unit treatment value assumption (SUTVA), a core tenant stated when
claiming causal relationships between variables in experimental studies. By
focusing on individuals, Nickerson argues previous experiments designed
to measure mobilization (Gerber & Green, 2000a, 2000b; Michelson, 2003;
Nickerson, 2007) did not accurately account for spillover effects of treatments
to unassigned subjects. Given the contagious nature of voting with a single
individual assigned to treatment impacting behavior of nonassigned subjects, experimenters may worry that treatment is “spilling over” to other individuals. Sinclair et al. (2012) investigate the value of multilevel experiment
designs applied to a large-scale mobilization experiment testing the impact
of social pressure on local voters. Multilevel experiments randomly assign
subjects to treatment and control plus randomly assign the treatment to other
subjects within an individual’s social network. In this context, implementing
a multilevel experiment on social pressure expand beyond Nickerson (2008)
by considering the spillover effects in a given household and among an individual’s neighborhood. Despite ignoring the mechanism of indirect influence
on other parts of a social network, Sinclair et al. argue that individuals can
be influenced differently by neighbors than compared with housemates.
Sinclair et al. apply the Gerber, Green, and Larimer’s social pressure mailings to a 2009 Illinois special election and find mixed results of spillover
within populations. Within households, spillover effects are found although
not at the levels reported by Nickerson (2008); differences in turnout levels are attributable to in-person conversations being more likely to prompt
dialog than postcards. Interestingly, no evidence of spillover effects between
individuals within zip-code-defined neighborhoods. Multilevel experimentation allowed for the identification of a null finding of significance given
the difficulty in mapping possible pathways of social influence within complex social networks. Sinclair et al.’s use of precise multilevel experiments
identifies areas of research to be focused on (household transmission of mobilization stimuli) and areas where limited resources do not need to be spent
in bulk (neighborhood diffusion of mobilization treatments).
Regardless of the online or offline nature of social network influence on
political behavior, credible empirical evidence suggests these relationships
do matter for determining how individuals will express themselves politically. Investigations of social network influence using observational data
consistently make questionable assumptions based on correlations between
Network Research Experiments
11
seemingly connected people and their shared behavior. Observational data
falls short, however, in accounting for other factors that could explain
correlated behavior limiting causal claims. Social network research has
increasingly relied on experimental studies to create an environment to
control for unobserved correlates and measure exogenous shock diffusion through social networks. In particular, experimental strategies allow
researchers three different types of manipulations for measurement: (i)
creating an exogenous shock and monitor its diffusion, (ii) alter/control
information flows through the network, or (iii) randomize individual
placements within the network (Nickerson, 2011). Researchers will need
to carefully consider the question they hope to answer when determining
which experimental method to use, but experimental studies provide
the more comprehensive tools available to confidently identify causal
relationships between networked individual’s political behaviors.
FUTURE RESEARCH WITH EXPERIMENTAL PROCESSES
Experimental research is frequently concerned with questions of external
validity, which means there is great potential for new directions for research
on questions previously unstudied using a methodological approach built
for causal claims. One such area involves questions of interest-based voting
as alluded to by Sokhey and McClurg. Most experimental studies focus
on turnout or political communication through social networks, but few
begin to identify the mechanisms surrounding interest building and mobilization in electoral politics. Growths in multilevel experiments now allow
researchers to explore theoretically stable and well-defined social networks
to understand how different types of political information are communicated and when political behavior appeals will be effective. Online data from
social media sites greatly reduce the cost of conducting experiments and
allow researchers new resources to extend beyond geographic boundaries
(zip codes or neighborhood lines) when defining unique networks and their
strength.
Another important area for scholarly growth involves understanding
young people’s involvement and engagement in politics. Literature on
political engagement begins with youth apathy and transitions toward the
importance of socioeconomic status for understanding differences in adult
consumption and participation in electoral contexts. Without traditional
predictors of participation meaningfully available in younger populations,
access to new media can provide insights in how young people form their
political identities. Advanced observational and experimental research on
social networks can reveal the influences of environment on developing
political ideologies, reveal changes in issue salience as individuals mature,
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
and capture the different ways young people communicate about politics.
New data and direct experimentation can enhance our understanding
of how a single individual decides to be political before obtaining more
conventional political resources. There are many questions left to be investigated with respect to social networks, political participation, and how we
are at once influencing others while also being influenced by those near and
far. Access to new data allows researchers to test hypotheses to improve our
understandings of being interconnected, political people.
REFERENCES
Bond, R. M., Fariss, C. J., Jones, J. J., Kramer, A. D. I., Marlow, C., Settle, J. E., & Fowler,
J. H. (2012). A 61-million-person experiment in social influence and political mobilization. Nature, 489, 295–298. doi:10.1038/nature11421
Downs, A. (1957). An economic theory of democracy. New York, NY: Harper.
Gerber, A. S., & Green, D. P. (2000a). The effect of canvassing, telephone calls, and
direct mail on voter turnout: A field experiment. American Political Science Review,
94(3), 653–663.
Gerber, A. S., & Green, D. P. (2000b). The effect of canvassing, telephone calls, and
direct mail on voter turnout: A field experiment. American Political Science Review,
102(1), 33–48.
Gerber, A., Green, D., & Larimer, C. (2008). Social Pressure and Voter Turnout:
Evidence From A Large-Scale Field Experiment. American Political Science Review,
102(1), 33–48.
Green, D. P., & Gerber, A. S. (2008). Get Out The Vote: How to Increase Voter Turnout
(Second ed.). Washington, DC: Brookings Institution Press.
Lau, R., & Redlawsk, D. (1997). Voting correctly. American Political Science Review,
91(3), 585–598.
Michelson, M. (2003). Getting out the Latino vote: How door-to-door canvassing influences voter turnout in rural central California. American Political Science
Review, 25(3), 247–263. doi:10.1023/A:1025167607369
Mutz, D. (2002). The consequences of cross-cutting networks for political participation. American Political Science Review, 46(4), 838–855.
Nickerson, D. W. (2007). Quality is job one: Professional and volunteer voter mobilization calls. American Journal of Political Science, 51(2), 269–282. doi:10.1111/
j.1540-5907.2007.00250.x
Nickerson, D. W. (2008). Is voting contagious? Evidence from two field experiements.
American Political Science Review, 102(1), 49–57. doi:10.1017/S0003055408080039
Nickerson, D. W. (2011). Social networks and political context. In J. Druckman, D.
Green, J. Kuklinski & A. Lupia (Eds.), Cambridge handbook of experimental political
science (pp. 273–286). Cambridge, England: Cambridge University Press.
Sinclair, B. (2012). The social citizen: Peer networks and political behavior. Chicago, IL:
University of Chicago Press.
Network Research Experiments
13
Sinclair, B., McConnell, M., & Michelson, M. R. (2013). Local Canvassing: The
Efficacy of Grassroots Voter Mobilization. Political Communication, 30(1), 42–57.
Sinclair, B., McConnell, M., & Green, D. P. (2012). Detecting spillover effects: Design
and analysis of multi-level experiments. American Journal of Political Science, 56,
1055–1069.
Sokhey, A. E., & McClurg, S. D. (2012). Social networks and correct voting. The Journal
of Politics, 74(3), 751–764. doi:10.1017/S0022381612000461
Traud, A. L., Kelsic, E. D., Mucha, P. J., & Porter, M. A. (2011). Comparing community
structure to characteristics in online collegiate social networks. SIAM Review, 53(3),
526–543. doi:10.1137/080734315
ALLEN L. LINTON II SHORT BIOGRAPHY
Allen L. Linton II is a PhD student in political science at the University of
Chicago. His research interests are youth political participation and socialization, the impact of new/social media on participation and governance, and
local politics.
BETSY SINCLAIR SHORT BIOGRAPHY
Betsy Sinclair is an associate professor of political science at Washington
University in St Louis. Her research focuses on the social foundations of
participatory democracy in American politics, with an emphasis on how an
individual’s social network influences political choices and behaviors.
RELATED ESSAYS
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and
Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
To Flop Is Human: Inventing Better Scientific Approaches to Anticipating
Failure (Methods), Robert Boruch and Alan Ruby
Repeated Cross-Sections in Survey Data (Methods), Henry E. Brady and
Richard Johnston
Ambulatory Assessment: Methods for Studying Everyday Life (Methods),
Tamlin S. Conner and Matthias R. Mehl
Models of Nonlinear Growth (Methods), Patrick Coulombe and James P.
Selig
Quantile Regression Methods (Methods), Bernd Fitzenberger and Ralf
Andreas Wilke
The Evidence-Based Practice Movement (Sociology), Edward W. Gondolf
Meta-Analysis (Methods), Larry V. Hedges and Martyna Citkowicz
The Use of Geophysical Survey in Archaeology (Methods), Timothy J.
Horsley
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Longitudinal Data Analysis (Methods), Todd D. Little et al.
Structural Equation Modeling and Latent Variable Approaches (Methods),
Alex Liu
Data Mining (Methods), Gregg R. Murray and Anthony Scime
Remote Sensing with Satellite Technology (Archaeology), Sarah Parcak
Quasi-Experiments (Methods), Charles S. Reichard
Digital Methods for Web Research (Methods), Richard Rogers
Virtual Worlds as Laboratories (Methods), Travis L. Ross et al.
Modeling Life Course Structure: The Triple Helix (Sociology), Tom Schuller
Content Analysis (Methods), Steven E. Stemler
Person-Centered Analysis (Methods), Alexander von Eye and Wolfgang
Wiedermann
Translational Sociology (Sociology), Elaine Wethington