2. were exogenously determined by the researcher. In the
experiment, egos tend to vote for
their party’s candidate, but they defect with greater frequency
when they receive mes-
sages from members of a different party. This willingness to
listen to the other side has
detrimental effects for the quality of their decisions.
Keywords social networks, correct voting, experiments
It may not be polite to discuss politics, but many people
constantly talk about politics at
home, in their businesses, and while they have an evening out.
Many people are talking and
on occasion they are persuading others to choose a different
candidate. It is well established
that discussion can affect how people view candidates and
ultimately how people vote (e.g.,
Berelson, Lazarsfeld, & McPhee, 1954; Huckfeldt & Sprague,
1995). Research also shows
that discussion can lead voters to choose candidates who are not
from their party (Beck,
2002; Ryan, 2010). This is a curious finding given that
individuals should be able to dismiss
political messages from individuals who are not likeminded.
Individuals who have different
political values will view political events differently (Gerber &
Green, 1999). This should
allow someone to discount the messages sent by a source with
divergent preferences, even
if one believes the discussion partner is totally sincere in his or
her messages.
John Barry Ryan is Assistant Professor, Department of Political
Science, Florida State
University.
3. The author wishes to thank Robert Huckfeldt, Matthew Buttice,
Elizabeth Simas, and Benjamin
Highton for their assistance with this project and the Political
Communication reviewers and editor for
their helpful comments. This research was funded by a grant
from the National Science Foundation
(SES-0817082).
Address correspondence to John Barry Ryan, Department of
Political Science, Florida State
University, 558 Bellamy Building, Tallahassee, FL 32306-2230,
USA. E-mail: [email protected]
100
Persuasive Social Communication 101
Experimental studies have shown that individuals are able to
cope with messages sent
by biased sources (e.g., Ahn, Huckfeldt, & Ryan, 2010; Lupia &
McCubbins, 1998). If the
message seems to serve the interests of the source, then subjects
dismiss the information.
This article argues that subjects have a more difficult time
evaluating messages when they
receive messages from multiple sources. Instead of focusing on
the biases of the politi-
cal discussion partners (the alters), voters (the egos) evaluate
the messages the alters send.
When the ego receives messages that all suggest she or he
should prefer a particular candi-
date, then the ego votes for that candidate. This is true even if
all of the messages were sent
by alters with an incentive to mislead the ego. Because of this,
4. social communication can
be harmful to egos when they are located in networks dominated
by out-partisans.
This article presents the results of a group-based experiment in
which subjects are
placed in parties and contexts as they try to choose between two
candidates. Subjects decide
on a candidate using information provided by two sources:
private information and mes-
sages from their fellow subjects. The private information is
accurate on average, but any
single piece of information may be inaccurate. The socially
communicated information
may come from subjects who are uninformed or biased in favor
of one of the candidates.
The experiment’s abstract nature should lead the subjects to
focus on the strategic incen-
tives of their fellow subjects. Yet, subjects are observed voting
for candidates who benefit
their alters and harming themselves.
Expertise, Biases, and Persuasion
Many scholars view interpersonal influence as a political reality
with potentially beneficial
consequences for a public that is largely under-informed about
politics. The key theoret-
ical breakthrough in this line of reasoning is the two-step flow
hypothesis (Katz, 1957;
Lazarsfeld, Berelson, & Gaudet, 1948), which posits that
opinion leaders pay close atten-
tion to the political world and then pass along information to
individuals who pay much less
attention to politics. Hence, individuals with high information
costs may use interpersonal
5. communication to participate effectively in politics while
reducing the price of participa-
tion (Downs, 1957). This division of labor could explain why a
society marked by low
individual levels of information often appears to respond
sensibly to political events in the
aggregate (Erikson, MacKuen, & Stimson, 2002; Page &
Shapiro, 1992).
On the other hand, political discussion does not necessarily
have beneficial effects for
voters. As Rousseau (1762/1994) noted, if individuals discuss
politics, then they may sup-
port the interests of a particularly persuasive opinion leader to
their own detriment. If an
individual already possesses enough information to make a
proper political decision, then
information from friends and family could only harm the
decision-making process (Lupia
& McCubbins, 1998). The uninformed are especially vulnerable
to potentially harmful
influence (Jackman & Sniderman, 2006). When individuals do
not have access to their
own information, then they need to rely on others as a source of
news about candidates
(Mondak, 1995). Informed individuals are less likely to be
persuaded to vote against their
interests because they are more likely to resist messages that do
not conform to their
preconsiderations (Zaller, 1992).
The literature on the efficacy of discussion as an information
shortcut has centered on
two variables: the expertise and partisan biases of the
discussion partners (e.g., Huckfeldt
& Sprague, 1991; Richey, 2008a; Sokhey & McClurg, 2012).
6. Expert information sources
should possess higher quality information and, for this reason,
might be more persuasive
(Richey, 2008b). The biases of the information source determine
both what information the
source possesses and the messages the source will send. The
information that the discussion
102 John Barry Ryan
partner received when reading the newspaper or watching the
news is likely retained in a
way that reflects his or her biases (Bartels, 2002; Zaller, 1992).
When sending messages
about that information, the discussion partner must summarize
the information and will
necessarily omit certain information biasing the message further
(Downs, 1957). Further,
the discussion partner might purposefully bias the information
in an effort to make his
or her side look better than the facts would suggest (Crawford &
Sobel, 1982; Lupia &
McCubbins, 1998).1 Thus, signals from discussion partners with
divergent preferences may
not just be skewed, but intentionally inaccurate.
If individuals are aware of the biases of their political
discussion partners, then they
can account for them. Individuals do not have to accept what
their alters say because their
alters’ different viewpoints may have led to different
conclusions (Ross, Bierbrauer, &
Hoffman, 1976). For example, a Democrat can discount
criticisms of President Obama
7. from a Republican friend because she knows he is a
conservative. She does not have to
think her friend is lying to understand why he views the world
differently. Ahn, Huckfeldt,
and Ryan (2010) show that people are aware of the potential
consequences of speaking
with someone with different preferences. In their experiment,
subjects were very willing
to speak with experts on the other side of a policy scale but
ignored signals that did not
conform to their beliefs.
Voting for the Other Side
Previous experimental studies show that individuals are able to
cope with signals from
discussion partners with biases that are different from their own
(e.g., Ahn, Huckfeldt, &
Ryan, 2010; Lupia & McCubbins, 1998). Why then do we
observe voters ignoring candi-
dates from their own party in favor of candidates supported by
their friends (Beck, 2002;
Ryan, 2010)? What would explain this divergence between these
experimental and obser-
vational results? One possibility is that elements of the designs
in these experiments lead
subjects to evaluate the messages they receive from other
subjects in a particular way.
The basic experimental design in both experiments involves a
subject (an ego) receiving
information from another subject (an alter).2 The ego is given
information about the expertise
and the biases of the alter. The subjects have financial
incentives to hold “correct beliefs”—
about coin flips or computer-generated candidates—and thus are
8. motivated to evaluate
whether to believe the social messages being sent (Petty &
Cacioppo, 1986). That is, they do
not simply follow what a supposedly informed alter tells them,
but they consider the alter’s
biases. If the alter has an incentive to mislead, then they
dismiss the message the alter sends.
In these experiments, subjects are given two pieces of
information along with the mes-
sage the source sends: the source’s biases and the source’s
knowledge level. They can easily
judge the credibility of the message with the information about
the source’s characteristics.
This article employs a different experimental design. In this
experiment, subjects receive
messages from three other subjects instead of just one. Subjects
receive the same informa-
tion for each of the three sources, and thus they receive six
pieces information about the
sources instead of just two.
The experimental subjects still want to hold “correct beliefs,”
but they may not want to
pay too great a cost to determine whether or not their beliefs are
correct.3 Subjects, there-
fore, employ a heuristic to quickly evaluate all of the messages
sent by the alters (Kahneman
& Tversky, 1973; Sniderman, 1993; Tversky & Kahneman,
1973). This article argues that
when subjects are confronted with messages from multiple
sources, they compare the mes-
sages. If the messages are all in accord, the subjects follow the
messages. If the messages
are divergent, the subjects stick with their original beliefs.
9. Persuasive Social Communication 103
This argument is derived from threshold models in sociology
(Granovetter, 1978;
Rolfe, 2005), Festinger’s (1954) theory of social comparison,
and the theory of autoregres-
sive influence (Huckfeldt, Johnson, & Sprague, 2004). These
different theories have subtle
differences, but a common thread is that an individual is more
likely to adopt a belief when
that belief is prevalent in the individual’s social network. Using
this heuristic, individuals
assume that the more people performing a particular behavior,
the greater the probability
that it is the “right” thing to do. This will, on occasion, cause an
individual to vote “incor-
rectly” for, as Cialdini (1984) notes, “[this method] provides a
convenient shortcut . . . but,
at the same time, makes one who uses the shortcut vulnerable to
the attacks of profiteers
who lie in wait along its path” (p. 117).
How does this change the probability that alters with different
preferences will per-
suade an ego? In the dyadic situation, an ego will dismiss an
alter with different preferences
because the ego views her or his messages as biased. When
others send similar messages,
then all of those messages corroborate each other. That is, a
Democrat may not believe a
negative message about President Obama from a Republican. If
three Republicans all send
the same message, however, the Democrat might find that
message more credible and might
10. even be persuaded to defect (i.e., vote for a candidate from
another party).
Experimental Design
Subjects in the experiment participate in groups of nine as they
take part in an election
between two candidates.4 The candidates, Adams and Bates,
represent the A and B parties,
respectively. Three subjects are assigned to the A party. Three
subjects are assigned to the
B party. Three subjects are independents.
Adams and Bates are proposing competing payoffs. Subjects
will receive the global
benefit proposed by the winning candidate plus an additional
party payoff. Subjects receive
a party bonus when the candidate from their party wins the
election. If the candidate from
the other party wins the election, subjects incur a penalty.
Independents receive neither a
bonus nor a penalty regardless of the election outcome.
Subjects know their individual party bonus and penalty, but
they are unaware of the
payoffs proposed by the candidates. To determine the payoff
proposals, subjects receive
randomly drawn private information. Subjects are assigned a
private information level, and
some subjects do not receive any private information at all.
Subjects also receive informa-
tion from three of their fellow subjects. The partisan
preferences and information levels of
these discussion partners vary. Some subjects receive
information from three well-informed
subjects, others from poorly informed subjects. Some subjects
11. primarily receive informa-
tion from members of their own party; others receive
information primarily from members
of the other party. At the end of each experimental period,
subjects use their private and
social information to determine which candidate will provide
them with the higher payoff
and then they vote for that candidate.
Parties and Candidates
Adams and Bates are proposing global benefits for all subjects.
The payoffs are indepen-
dently and randomly drawn from identical, uniform distributions
with a lower bound of
20 experimental currency units (ECUs) and an upper bound of
100 ECUs.5 The payoffs
are drawn from the same distribution and, therefore, the
expected value of Adams’s and
Bates’s proposed payoffs is equal at 60 ECUs.
104 John Barry Ryan
Recall that at the end of each experimental round, subjects are
awarded the payoff
proposed by the winning candidate as well as either a party
bonus or penalty depending
on the outcome of the election. An individual’s partisan payoff
is randomly drawn from a
uniform distribution with a maximum of 20 cents and a
minimum of 10 cents. Independents
receive no partisan payoff regardless of the outcome of the
election.
12. On average, subjects in party A are better off if Adams is
elected and subjects in party
B are better off if Bates is elected because of these partisan
payoffs. Because the expected
value of both candidates’ payoffs is the same, independents
should be indifferent between
the candidates on average.
Timeline of the Experiment
Each subject is aware of the distribution from which each
candidate’s global benefit is
drawn and of her or his individual party payoff. They do not
know any candidate’s global
benefit in a particular election. The subjects must use private
and social information to
determine the candidates’ global benefits. The experiment
proceeds over the following four
stages.
• Stage 1 (Private Information Stage): Subjects receive private
information about the
candidates’ global payoffs. Using this information, they make
initial estimates of the
candidates’ global benefits.
• Stage 2 (First Social Information Stage): Subjects send
messages about the size
of candidates’ global payoffs to one another. They update their
estimates of the
candidates’ global benefits.
• Stage 3 (Second Social Information Stage): Subjects again
send messages about the
size of candidates’ global payoffs to one another. They again
update their estimates
13. of the candidates’ global benefits.
• Stage 4 (Voting and Payoffs): Subjects vote for the candidate
they believe will offer
them the greatest total benefit. Subjects are then paid based on
the benefits offered
by the candidate who received the most votes. The subjects are
not paid based on
whether or not they voted correctly. Voting correctly, however,
would increase the
probability that a subject will be paid more.
Stage 1: Private Information Stage
The first task for subjects is to estimate the global benefits that
the candidates offer based
on unique information given to each subject.
The nine subjects are assigned an information level from 0 to 4
that measures the num-
ber of pieces information about each candidate a subject will
receive. Only one subject
receives the maximum four pieces of information, while there
are two subjects at the other
four information levels. This includes two subjects who do not
receive any private informa-
tion about the candidates. Subjects are assigned to an
information level based on one of the
five different information treatments outlined in Table 1A.
Each piece of information contains signals about both
candidates. For example, a sub-
ject with an information level of one receives one signal about
Adams and one signal about
Bates. These signals are independently and randomly drawn
from a uniform distribution
14. centered on the candidate’s true benefit and extending 25 ECUs
above and below that true
benefit.6 This means that on average the signals accurately
represent the candidate’s true
Persuasive Social Communication 105
Table 1
Subjects’ private information levels and discussion networks
A. Private information levels
Number of pieces of information received by subject
Treatment A1 A2 A3 I4 I5 I6 B7 B8 B9
1 1 0 0 2 1 3 4 3 2
2 4 0 3 2 3 1 1 2 0
3 0 4 1 3 1 2 2 0 3
4 1 3 3 0 4 0 2 1 2
5 3 2 2 4 0 0 1 3 1
B. Network providing social information
Sender
Recipient Alter 1 Alter 2 Alter 3 Network type
A1 A2 I4 B9 Heterogeneous
A2 A1 A3 I5 Homogeneous A (in-group)
A3 I6 B7 B8 Homogeneous B (out-group)
I4 A2 A3 I5 Homogeneous A (out-group)
I5 I6 B7 B9 Homogeneous B (out-group)
I6 A1 I4 B8 Heterogeneous
15. B7 I4 B8 B9 Homogeneous B (in-group)
B8 A1 A3 I5 Homogeneous A (out-group)
B9 A2 I6 B7 Heterogeneous
position, but any particular signal may be as likely to largely
over- or underestimate the
candidate’s true benefit as it is to be on the mark.7
Based on these private signals, subjects are asked to estimate
the candidates’ benefits.
These initial estimates are the subjects’ judgmental priors about
the candidates. They will
be used to determine how the subjects would have voted if they
had been asked to vote
without receiving information from their fellow subjects.
Stage 2: First Social Information Stage
Subjects next share information with three of their fellow
subjects. The subjects send a
signal about their estimate for the global benefits offered by
each candidate. Subjects are
allowed to send any integer value for each candidate from 0 to
100 ECUs.
In this stage, subjects are alters passing along information to
egos. Subjects provide
information to one subject from each party and an independent.
Subjects know the par-
tisanship and the information level of each ego.8 They are told
that they do not have to
provide identical information to each subject, but they are not
encouraged to misrepresent
their beliefs.9 Subjects may send messages strategically,
however, because they know the
information level and partisanship of each ego. Next, subjects
16. become egos receiving the
messages that the alters provided in the previous stage. Subjects
receive information from
106 John Barry Ryan
the network of alters listed in Table 1B. These ego networks are
combined with the infor-
mation levels to place egos into several network “treatments” in
which the partisan makeup
and the information level of the network are manipulated.
Ego networks in this experiment take on one of three types:
heterogeneous, homoge-
neous A, and homogeneous B. In heterogeneous networks, there
is one member of each
party. In homogeneous networks, there are two members of
either party A or party B and
an independent. Partisan subjects, therefore, may receive
messages from a majority of like-
minded subjects, a heterogeneous network, or a network without
any subjects who share
their biases.
The alters providing information may also be well informed (all
alters have three or
four pieces of information), poorly informed (all alters have no
information or one piece
of information), or something in between. When they receive
this social information, sub-
jects are asked to estimate candidate positions again. They are
reminded of their previous
estimate and may update their estimate based on the social
information they have just
17. received.
Stage 3: Second Social Information Stage
Subjects provide social information a second time. The process
is the same as in stage 2.
Subjects provide to and receive information from the same set
of subjects as before. This
second information-sharing period allows subjects to
incorporate the social messages they
previously received into the messages they are sending now.
After receiving the second
round of social information, subjects make a third and final
estimate.
Stage 4: Voting and Payoffs Stage
After making their final estimates, subjects vote for their
favored candidate. The outcome
of the election is revealed to the subjects, as are the true
benefits of the candidates. Payoffs
are awarded to the subjects based on the proposed payoff of the
winning candidate and
the subjects’ partisan payoffs. Subjects are not paid based on
whether or not they voted
correctly, but voting correctly does increase the probability that
they will receive the larger
of the benefits.10 Subjects then participate in a new campaign
with new, randomly drawn
candidate benefits. The subjects’ parties, partisan pay,
information levels, and networks
remain the same.
Subjects participate in as many elections as they can complete
within 1 hour. In the
analyses that follow, I use data from the first seven elections of
18. each experimental ses-
sion. There was a great deal of variation in the number of
elections subjects completed.11
Capping the number of elections at seven allows for balance
across all 15 sessions, but the
substantive conclusions do not change if all elections are
included.
The Design and Strategic Considerations
The experiment described here is more complicated than the
typical economics experi-
ment on social communication (e.g., Lupia & McCubbins,
1998). These complexities were
included to heighten the realism of the experiment. This means,
however, that the exper-
imental setting is too complicated to fully analyze as a
noncooperative game. It is still
important to highlight certain aspects of the strategic setting
facing the subjects. This dis-
cussion assumes all voters are pivotal, which likely means this
overestimates the value of
information.
Persuasive Social Communication 107
The Value of Private Information
Subjects with more information about each candidate should
have more accurate esti-
mates of the candidates’ global payoffs. The gains from having
more information diminish
quickly, however. For example, if the weakest partisans have no
19. information, they should
vote for their own party’s candidate, and they would vote
correctly 72.1% of the time. With
one piece of information about both candidates, the same
subject is able to vote correctly
86.5% of the time. If the subject received the maximum four
pieces of information, the
subject would vote correctly 94% of the time. This 7.5%
improvement with three addi-
tional pieces of information is almost half the improvement
from receiving a single piece
of information.
Private information is less valuable to partisan subjects.
Typically, if partisans simply
vote for the candidate from their party, they will vote for the
candidate who will provide
them with the larger payoff. Information is more valuable to
independents because they are
indifferent between the candidates on average, but the gains
from more information still
diminish quickly. Hence, for all subjects, more information
should lead to better decision
making, but the starkest difference is between the informed and
the uninformed.
The Value of Social Information
Subjects are one of three basic types: an A partisan, a B
partisan, or an independent. If an
alter (the subject sending a social message) is a different type
than the ego (the subject
receiving a social message), the alter has an incentive to send a
biased message. Even
independents have an incentive to send biased messages to
partisans. Partisans want to vote
20. for their party’s candidate the vast majority of the time, while
independents want partisans
to defect half the time. Hence, independents will benefit on
average from sending messages
that undervalue the partisan’s candidate.
It may seem that alters would be compelled to tell the truth
because their messages will
be “verified” when the true global payoffs are revealed at the
end of each period (Lupia &
McCubbins, 1998). The messages are not really verified,
however, because the ego does not
know what information the alter had when the alter sent her or
his message. Because the
private information is noisy, it is possible that an alter will send
a message that truly reflects
her or his beliefs but is inaccurate. It is also possible that the
alter lies based on the infor-
mation she or he has, but her or his message accurately
communicates the candidate’s true
positions. This makes it more difficult for egos to update their
beliefs about the likelihood
that alters are sending truthful messages. This also means that
alters will struggle to build
a reputation as truth tellers in early periods and then send
biased signals in later periods to
take advantage of trusting egos.
Because of all of these complexities, subjects are making
decisions under a situation
of high uncertainty. Given the time and effort it would take to
consider all of the factors,
it is unclear what “rational” behavior would look like (Simon,
1957, 1985). Even though
these elections are less complex than real-world campaigns, the
uncertainty could make it
21. difficult for subjects to realize they should discount information
from biased sources.
Hypotheses
It is important that the reader recall a few elements of the
design to understand the hypothe-
ses. Subjects do not know how much either candidate’s global
benefit is worth. In order to
determine the value of the global benefits, subjects first receive
private information. After
108 John Barry Ryan
receiving this information, subjects twice receive messages
from three of their fellow sub-
jects, and then they vote. The design manipulates the partisan
composition of the network
of three alters sending messages. The first hypothesis is as
follows.
H1: A subject will vote for the candidate from the same party as
the majority of the
members in the subject’s network.
Even if a subject votes for the candidate supported by his or her
social network, this
does not mean that the subject was persuaded by the messages
sent by the network. In order
to truly show persuasion, one must show that the subject was
going to vote for one candidate
and then changed his or her mind following social interaction
(Kenny, 1998).12 I am able
to use the initial estimates of the candidates’ global benefits to
22. determine who the subject
would have voted for if the social communication stages did not
take place. For partisan
subjects, this leads to a particular hypothesis.
H2: Partisan subjects are most likely to change their vote when
their network is dominated
by members of the other party.
It is important to remember that partisan subjects have a
candidate they should favor
in expectation. Following the private information stage, A
partisans should favor Adams
on average, and B partisans should favor Bates on average.
Typically, when a partisan is
persuaded that should mean that the partisan defected, that is,
voted for the other party’s
candidate. Defections should be more frequent when messages
are received from supporters
of the other party.
Networks and Voting
Do the subjects’ votes reflect the partisanship of their networks
as suggested by Hypothesis
1? The results in Table 2A demonstrate that they do, as subjects
are more likely to vote
for Adams as the number of A partisans in a subject’s ego
network increases.13 Subjects
in homogeneous A networks voted for Adams 17 percentage
points more frequently than
subjects in homogeneous B networks.
These results suggest that the network messages are persuading
the subjects. The table
does not show, however, that any subject changed his or her
23. mind following social commu-
nication. Table 2B and Table 2C replicate the analysis in Table
2A but separate the subjects
based on which candidates they initially preferred. I assume that
if subjects had voted after
receiving private information but before receiving any social
information, they would have
chosen the candidate that maximized their payoff based on their
known party pay and their
beliefs about the global benefits. This yields an implied vote in
the absence of social infor-
mation for each subject, which can be compared to the subject’s
actual vote to determine
the effect of social information on her or his decision.
The first key result in Table 2B and Table 2C is that subjects
tended to vote for the
candidate they would have supported based on their initial
estimates. Subjects whose initial
estimates implied a vote for Adams voted for Adams 80% of the
time. Subjects who initially
preferred Bates backed Bates 72% of the time. Overall, subjects
changed their votes less
than a quarter of the time.
When subjects did change their votes, they changed them in the
direction of the party
dominating their social network. Subjects who initially
preferred Adams were 1.8 times
more likely to switch to Bates if they were in a homogeneous B
network than a homo-
geneous A network. Subjects whose initial estimates implied a
vote for Bates were twice
24. Persuasive Social Communication 109
as likely to switch to Adams if they were in a homogeneous A
network compared to a
homogeneous B network.
Table 2 offers support for H1, but does not directly address H2
because it does not show
the frequency with which partisans changed their votes away
from their party’s candidate.
Partisans have a candidate who they are predisposed to favor.
They also have strong reason
to doubt the messages an alter sends trying to convince them to
change their votes.
Table 3 only includes those observations in which partisans’
initial estimates favored
their party’s candidate. As it shows, partisans were more likely
to change away from their
party’s candidate when they received messages from members
of another party. Partisans
changed their vote 17% of the time when located in in-group
ego networks. When located in
out-group networks, they changed their vote away from their
party’s candidate and toward
their network’s candidate 31.8% of the time. It is important to
note that of the 130 occasions
Table 2
Subject votes by network homogeneity
Homogeneous A Heterogeneous Homogeneous B Total
A. All subjectsa
Adams (%) 63.1 58.5 46.1 55.9
25. Bates (%) 36.9 41.5 53.9 44.1
N 306 306 306 918
Subjects 45 45 45 135
B. Subjects who initially preferred Adamsb
Adams (%) 85.3 82.8 74.0 80.8
Bates (%) 14.7 17.2 26.0 19.2
N 163 145 150 458
Subjects 41 36 40 117
C. Subjects who initially preferred Batesc
Adams (%) 35.1 31.4 17.0 27.5
Bates (%) 64.9 68.6 83.0 72.5
N 134 137 147 418
Subjects 34 31 35 100
Note. F tests correct for clustering on subjects and experimental
sessions.
aF = 4.206, p = .016.
bF = 2.378, p = .096.
cF = 3.855, p = .023.
Table 3
Whether partisans who initially preferred their party’s candidate
changed their vote
In-party Heterogeneous Out-party Total
No change (%) 83.0 76.2 68.2 75.9
Change to other party’s candidate (%) 17.0 23.8 31.8 24.1
N 182 181 176 539
Subjects 30 30 30 90
Note. F test corrects for clustering on subjects and experimental
26. sessions. F = 3.855, p = .023.
110 John Barry Ryan
in which an ego moved away from her or his party’s candidate,
that was a correct decision
only 35.4% of the time. Egos receiving messages from alters
with divergent preferences
changed their votes more frequently and usually that was a
mistake.
These results appear to demonstrate that networks have an
effect on how subjects voted.
They suggest that messages sent by biased sources persuaded
subjects unlike subjects in
previous experiments. It is possible, however, that these
differences are the result of actual
differences in payoffs in the different treatments. For example,
note that in Table 2A, there
is only a small difference in the frequency of Adams voting
when comparing homogeneous
A and heterogeneous networks. One possible reason for this is
that Adams actually offered
a better payoff on average even though Adams and Bates had
equal expected benefits.14
The next analyses demonstrate that the findings shown in this
section are the result of the
influence of the network and not simply non-random
assignment.
Individual and Network Determinants of the Vote
The analysis in Table 4 employs mixed-effects logit models in
27. which the dependent vari-
able is coded 1 if the subject voted for Adams and 0 if the
subject voted for Bates.15 The
Table 4
Predicting ego’s vote by network partisanship and messages
controlling for ego’s initial
estimates, partisanship, and information level
Coefficient Z Coefficient Z
Homogeneous A network 1.109 4.89 −0.091 −0.32
Heterogeneous network 0.872 3.92 0.270 1.05
First alter Adams message 0.017 3.58
Second alter Adams message 0.016 3.47
Third alter Adams message 0.011 2.62
First alter Bates message −0.013 −3.30
Second alter Bates message −0.020 −4.39
Third alter Bates message −0.015 −3.58
A party pay 0.066 8.04 0.079 8.43
Initial Adams estimate 0.004 0.70 0.006 0.88
Initial Bates estimate 0.001 0.11 0.003 0.36
Logged subject information 0.316 0.57 0.315 0.50
Adams Estimate ∗ Subject Information 0.036 5.00 0.025 3.15
Bates Estimate ∗ Subject Information −0.040 −5.39 −0.031
−3.76
Constant −0.724 −1.21 −0.116 −0.15
Random effects parameters SD SE SD SE
15 sessions 0.686 0.169 0.507 0.165
135 subjects 0.254 0.29 0.354 0.268
N 918 918
AIC 885.25 791.08
Note. Data are from mixed-effects logit models. The dependent
28. variable was coded 1 if the subject
voted for Adams and 0 if the subject voted for Bates.
Persuasive Social Communication 111
models examine the effect of the ego network controlling for
the subject’s partisanship,
information level, and initial beliefs about the benefits offered
by the two candidates. The
partisan makeup of the ego network is measured with two
dummy variables: homogeneous
A network and heterogeneous network with homogeneous B
networks as the reference
category. In the first model, these are the only network-level
measures included.
The models include several variables measuring the ego’s
characteristics. A party pay
measures the strength of the subject’s attachment to party A. It
ranges from −20 to 20, with
negative values indicating the subject is a member of party B.
Independents are coded at 0,
indicating they are indifferent between the two parties. The
model controls for the subject’s
initial estimates of the candidate’s global payoffs. The inclusion
of the party pay variables
and the initial estimates allow me to take into account how the
subject probably would have
voted if the social communication never took place. The logged
subject information level
is also included in the analysis and interacted with the initial
estimates.16 Subjects with
more information should be more likely to resist persuasion and
vote based on their initial
29. estimates.
The statistically significant interaction effects between the
subject’s information levels
and the initial estimates of the two candidates’ benefits
demonstrates that better informed
subjects place greater weight on their private information when
choosing a candidate. When
the subject has the lowest level of information, the probability
that she or he will vote for
a particular candidate increases by only 9 percentage points
across the full range of pos-
sible estimates.17 At the highest information level, the
difference is 83 percentage points,
indicating that a subject’s vote is almost entirely determined by
the initial estimates of the
candidates’ global benefits.
In the first model, the network heterogeneity measures have
statistically significant
effects indicating support for the first hypothesis. Subjects
within homogeneous A net-
works have a 70% probability of voting for Adams. This
probability drops to 44% if the
subject is in a homogeneous B network. There is a 65%
probability that subjects in hetero-
geneous networks vote Adams. This expectation is statistically
different from homogeneous
B networks, but not homogeneous A networks. Again, this is
likely caused by the fact that
Adams actually did provide better payoffs than Bates on
average.
This first model does not take into account the actual messages
sent by the alters. It is
not clear, therefore, that the networks are influencing the
30. subjects’ vote decisions through
the messages they are sending. For this reason, the second
model in Table 4 includes the
first messages sent by alters.18 Variables such as first alter
Adams message and second alter
Adams message refer to the order in which the messages were
displayed. The partisanship
of the alters sending the messages depends on the partisanship
of the subject’s network.19
The lower AIC value indicates that including the messages that
were sent improves
model fit. All six message variables are statistically significant
in this model, but the net-
work heterogeneity measures are not. This suggests the network
effects from the first table
are explained by the signals sent by the alters. To understand
the size of the effects, Figure 1
displays the expected probabilities of voting for Adams based
on the mean signals sent by
each type of network by the partisanship of the ego. The
messages reflect the biases in the
messages that alters sent. On average, A party alters said that
Adams offered a benefit of
68.9 and Bates offered a benefit of 45.4 ECUs. Independents
sent messages that said that
Adams and Bates offered 57 ECUs. On average, B party alters
said that Adams offered a
benefit of 44 and Bates offered a benefit of 63.2 ECUs.
The probabilities in Figure 1 show that the biased messages that
alters send lead egos
to vote in line with the partisanship of the networks. A partisans
are 14 percentage points
more likely to vote for their candidate when communicating
with fellow partisans than
31. 112 John Barry Ryan
Figure 1. Predicting votes for Adams from mean messages sent
by different network types. Estimates
are from the second model in Table 4. Error bars represent 95%
confidence intervals.
when receiving messages from B partisans. The effect of
network heterogeneity is over
20 percentage points for independents and B partisans.
It is possible that biased messages are less likely to sway
subjects if the subjects are
more confident in their initial estimates—a result suggested by
the greater weight the better
informed place on their initial estimates. Not only will
confidence in the initial estimates
make subjects less likely to rely on social information, but it
will also help them recognize
biased signals. Indeed, subjects were less likely to change their
votes if they received more
informative private signals—that is, signals that significantly
narrowed the range of possible
candidate benefits. A systematic analysis of this question is
beyond the scope of this article,
but it should be considered in future designs.
In sum, while most subjects voted as their private information
suggested, many were
persuaded by alters—even alters with incentives to dissemble.
The clearest evidence that
this had a negative effect on egos comes from examining how
partisans voted. Partisans
32. vote for their party’s candidate 76.5% of the time when the
majority of the ego network
is from the ego’s party. Partisans located in out-party networks
should have voted for their
party’s candidate in about 77% of the observations, but they
actually voted for their party’s
candidate only 62.3% of the time. Based on these results,
subjects are being persuaded
by their networks more often than they should. In fact, results
presented elsewhere demon-
strated that only uninformed independents improved the quality
of their decisions following
social communication (Ryan, 2011). On average, social
information either had no affect or
harmed the decision making of all other types of subjects.
Conclusion
This article presents the results of an experiment designed to
isolate how the biases of alters
in larger networks affect their ability to persuade egos. The
design allows for a comparison
between how subjects voted with how they would have voted if
they had never received
information from alters. The experiment cannot determine the
model that best explains
the social influence process, but it does show that subjects are
persuaded by others, even
those with divergent preferences. Subjects in previous
experiments were better capable
of handling message from biased alters (e.g., Ahn, Huckfeldt, &
Ryan, 2010; Lupia &
McCubbins, 1998). The main difference between those
experiments and this one is that in
this experiment subjects received messages from several alters
instead of only one. It is
33. Persuasive Social Communication 113
possible that the subjects were more accepting of biased alters’
messages when they were
corroborated by another alter—even if that alter had a similar
bias.
To what extent are the results of this experiment externally
valid? There are many
factors besides expertise and partisanship that could make
someone persuasive, from the
intimacy of the relationship (Burt, 1987; Huckfeldt & Sprague,
1991; Kenny, 1998) to the
strength of the argument used (Cobb & Kuklinski, 1997). This
experiment cannot speak
to all of these variables because of its abstract nature. The
experiment strips away all of
the variables that could compel egos to support candidates who
support interests of oth-
ers instead of their own. For example, subjects cannot choose a
candidate whose policies
would hurt them but help their families. Instead, the experiment
focuses subjects on the
strategic incentives of their fellow participants, which should
have caused them to behave
as game-theoretic models predict. Given that, subjects should
have been even less likely to
be persuaded by members of the other party than they are in the
real world, thus providing
a strong argument for the ability of an individual’s network to
shape his or her decision
making.
34. Many studies of political discussion are criticized for being
unable to differentiate
between contextual effects and selection effects (Achen &
Shively, 1995; Johnson, Shively,
& Stein, 2002). If individuals choose discussion partners for
political reasons (MacKuen,
1990), then the networks are endogenously related to opinions,
vote choices, and other
political dependent variables. This article follows other research
that uses experiments to
overcome this issue while providing insights into how
interdependent voters influence one
another (e.g., Ahn, Huckfeldt, Mayer, & Ryan, 2010; Druckman
& Nelson, 2003).
There are elements of the design that do not accurately reflect
real-world behavior.
One of the most important examples is that all subjects in this
experiment provided equal
amounts of social information. In the real world, however,
better informed individuals dis-
cuss politics with much greater frequency. Even with this
deviation from reality, the present
results may be informative. They indicate that expert discussion
partners are influential
because they provide more information, not because they
provide better information. This
article cannot directly address this question, but it does suggest
that future social network
research should attempt to separate out the effects of the
frequency of communication and
expertise of discussion partners.
Notes
1. There are, however, situations in which an alter with a
35. divergent preference may be com-
pelled to provide accurate information to the ego. For example,
as Lupia and McCubbins (1998)
argue, alters will provide accurate information if there is the
possibility that the ego will attempt to
verify the information the alter provides. Further, Calvert
(1985) notes that biased alters who send
signals contrary to their preferences may provide useful
information. For example, if an independent
discusses politics with a Republican who says she is voting for
Barack Obama, then the independent
may infer that Obama has positive characteristics because this
Republican is willing to forgo her
partisan preferences.
2. This is the basic common element in all of the experiments in
Lupia and McCubbins (1998)
and Ahn, Huckfeldt, and Ryan (2010). Lupia and McCubbins
(1998) offer several different treatments
involving manipulation of the alter’s incentives and whether the
alter’s messages will be verified in
some way. In Ahn, Huckfeldt, and Ryan (2010), the egos are
allowed to choose their alter.
3. Often individuals are viewed only as “central” or
“peripheral” processors when attuning
to potentially persuasive information—that is, people either
expend mental effort to evaluate an
attitude object or they do not. This is an oversimplification of
dual process models like the elabo-
ration likelihood model (Petty & Wegener, 1999). First, people
can change from central or peripheral
114 John Barry Ryan
36. processors—and vice versa—depending on the context. Second,
pure central and pure peripheral pro-
cessors are extreme types, and most people in most situations
will lie somewhere on the continuum
between those two end points.
4. The 135 subjects for this experiment were recruited from
undergraduate political science
courses at the University of California, Davis. The subjects
received a $10 show-up fee plus whatever
earnings they accrued during the experiment. The experiment is
programmed using zTree (software
for designing economic experiments) (Fischbacher, 2007).
5. At the end of the experiment, subjects were paid at a rate of 1
ECU equals 1 cent.
6. As a result, the possible signals extend beyond the limits of
possible benefits. For example, if
a candidate offered the lowest possible benefit of 20, then
subjects might receive signals suggesting
that the candidate’s benefit is as low as −5. Subjects are
informed of this possibility.
7. Subjects receiving multiple signals would benefit from
greater variance in the signals. For
example, if a subject received two signals that suggest that
Adams’s benefit was either 45 or 46 ECUs,
then the subject knows that the candidate’s benefit lies
somewhere between 21 and 70 ECUs. All of
the values between that upper and lower bound are equally
likely to produce signals of 45 and 46.
On the other hand, if the signals were 45 and 95, then the
subject would know that the candidate
offers a benefit of 70 ECUs. This is because only 70 ECUs
37. could produce both a signal of 45 and
of 95.
8. Survey evidence suggests that individuals can accurately
identify the preferences of their
political discussion partners (Huckfeldt, Levine, Morgan, &
Sprague, 1998) as well as differentiate
between politically expert and inexpert individuals (Huckfeldt,
2001).
9. The exact instructions to subjects and screen shots of the
experimental treatment are available
at the author’s Web site: http://www.johnbarryryan.com
10. A subject votes correctly if the subject chooses the
candidate with the larger total pay. A sub-
ject is also coded as having voted correctly if the two
candidates offered the exact same total payoff
to the subject. The probabilities of a correct vote for all subject
profiles are available in supplemental
Table SI1 at the author’s Web site.
11. All subjects participate in one practice campaign. In that
practice campaign, all subjects have
a party pay of 15 ECUs and receive two pieces of private
information. The practice campaign is not
used in the analysis. After the practice campaign, one session
participated in five campaigns, one
participated in six, two participated in seven, one participated
in eight, one participated in nine, four
participated in 10, one participated in 11, and four participated
in 12.
12. There is a weakness to this definition of persuasion. It is
difficult to know when an individual
influenced someone who is predisposed to favor the same
38. candidate. For example, imagine a dyad
made up of two Republicans in San Francisco. One may only
continue to vote Republican because
the other Republican counterargues the social messages from
more common Democratic sources, but
this would not count as persuasion using this measure.
13. The F tests in these models are corrected for clustering on
subjects and experimental
sessions.
14. The mean benefit offered by Adams was 60.7 ECUs. The
mean benefit offered by Bates was
57.5 ECUs. This difference is statistically significant at the .1
level. A partisans also received slightly
larger party payoffs—15.5 compared to 14.4 for B partisans.
15. The models include random intercepts to account for the
clustering on subjects nested within
sessions.
16. Some subjects received no information. Because the natural
log of zero is undefined, 1 is
added to the subjects’ information levels and then the log is
taken.
17. When calculating the expected probabilities, all variables
not being discussed are held at
their means.
18. Alters send messages on two occasions. The first and second
messages sent are highly corre-
lated. Hence, both measures are excluded from the model. The
first messages are chosen because they
are sent based solely on the private messages the alters had
previously received. The second messages
39. could include information acquired via social information,
increasing the problem of interdependence
in the observations.
Persuasive Social Communication 115
19. For subjects in homogeneous A networks, the first two
messages are sent by A partisans and
the third message is sent by an independent. For subjects in
heterogeneous networks, the first message
is sent by an A partisan, the second message is sent by an
independent, and the third message is send
by a B partisan. For subjects in homogeneous networks, the first
message is sent by an independent
and the final two messages are sent by B partisans. All
messages are displayed on the screen at the
same time.
References
Achen, C. H., & Shively, W. P. (1995). Cross-level inference.
Chicago, IL: University of Chicago
Press.
Ahn, T., Huckfeldt, R., Mayer, A. K., & Ryan, J. B. (2010).
Politics, expertise, and interdependence
within electorates. In J. E. Leighley (Ed.), The Oxford
handbook of American elections and
political behavior (pp. 278–299). New York, NY: Oxford
University Press.
Ahn, T., Huckfeldt, R., & Ryan, J. B. (2010). Communication,
influence, and informational
asymmetries among voters. Political Psychology, 31, 763–787.
40. Bartels, L. M. (2002). Beyond the running tally: Partisan bias in
political perceptions. Political
Behavior, 24, 117–150.
Beck, P. A. (2002). Encouraging political defection: The role of
personal discussion networks in
partisan desertions to the opposition party and Perot votes in
1992. Political Behavior, 24,
309–337.
Berelson, B. R., Lazarsfeld, P. F., & McPhee, W. N. (1954).
Voting: A study of opinion formation in
a presidential campaign. Chicago, IL: University of Chicago
Press.
Burt, R. S. (1987). Social contagion and innovation: Cohesion
versus structural equivalence.
American Journal of Sociology, 92, 1287–1335.
Calvert, R. L. (1985). The value of biased information: A
rational choice model of political advice.
Journal of Politics, 47, 530–555.
Cialdini, R. B. (1984). Influence: How and why people agree to
things. New York, NY: William
Morrow.
Cobb, M. D., & Kuklinski, J. H. (1997). Changing minds:
Political arguments and political
persuasion. American Journal of Political Science, 41, 88–121.
Crawford, V. P., & Sobel, J. (1982). Strategic information
transmission. Econometrica, 50, 1431–1451.
Downs, A. (1957). An economic theory of democracy. New
York, NY: Harper & Row.
41. Druckman, J. N., & Nelson, K. R. (2003). Framing and
deliberation: How citizens’ conversations
limit elite influence. American Journal of Political Science, 47,
728–744.
Erikson, R. S., MacKuen, M. B., & Stimson, J. A. (2002). The
macro polity. New York, NY:
Cambridge University Press.
Festinger, L. (1954). A theory of social comparison processes.
Human Relations, 7, 117–140.
Fischbacher, U. (2007). z-tree: Zurich toolbox for ready-made
economic experiments. Experimental
Economics, 10, 171–178.
Gerber, A., & Green, D. (1999). Misperceptions about
perceptual bias. Annual Review of Political
Science, 2, 189–210.
Granovetter, M. (1978). Threshold models of collective
behavior. American Journal of Sociology, 83,
1420–1443.
Huckfeldt, R. (2001). The social communication of political
expertise. American Journal of Political
Science, 45, 425–438.
Huckfeldt, R., Johnson, P. E., & Sprague, J. (2004). Political
disagreement: The survival of diverse
opinions within communication networks. New York, NY:
Cambridge University Press.
Huckfeldt, R., Levine, J., Morgan, W., & Sprague, J. (1998).
Election campaigns, social com-
munication, and the accessibility of perceived discussant
42. preference. Political Behavior, 20,
263–294.
Huckfeldt, R., & Sprague, J. (1991). Discussant effects on vote
choice: Intimacy, structure and
interdependence. American Political Science Review, 53, 122–
158.
116 John Barry Ryan
Huckfeldt, R., & Sprague, J. (1995). Citizens, politics, and
social communication: Information and
influence in an election campaign. New York, NY: Cambridge
University Press.
Jackman, S., & Sniderman, P. M. (2006). The limits of
deliberative discussion: A model of everyday
political arguments. Journal of Politics, 68, 272–283.
Johnson, M., Shively, W. P., & Stein, R. M. (2002). Contextual
data and the study of elections and
voting behavior: Connecting individuals to environments.
Electoral Studies, 21, 219–233.
Kahneman, D., & Tversky, A. (1973). On the psychology of
prediction. Psychological Review, 80,
237–251.
Katz, E. (1957). The two step flow of communication: An up-to-
date report on an hypothesis. Public
Opinion Quarterly, 21, 67–81.
Kenny, C. (1998). The behavioral consequences of political
discussion: Another look at discussant
43. effects on vote choice. Journal of Politics, 60, 231–244.
Lazarsfeld, P. F., Berelson, B., & Gaudet, H. (1948). The
people’s choice: How a voter makes up his
mind in a presidential campaign. New York, NY: Columbia
University Press.
Lupia, A., & McCubbins, M. D. (1998). The democratic
dilemma: Can citizens learn what they need
to know? New York, NY: Cambridge University Press.
MacKuen, M. (1990). Speaking of politics: Individual
conversational choice, public opinion and the
prospects for deliberative democracy. In J. A. Ferejohn & J. H.
Kuklinski (Eds.), Information
and democratic processes (pp 59–99). Urbana: University of
Illinois Press.
Mondak, J. J. (1995). Media exposure and political discussion
in U.S. elections. Journal of Politics,
57, 62–85.
Page, B. I., & Shapiro, R. Y. (1992). The rational public: Fifty
years of trends in Americans’ policy
preferences. Chicago, IL: University of Chicago Press.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and
persuasion: Central and peripheral routes
to attitude change. New York, NY: Springer-Verlag.
Petty, R. E., & Wegener, D. T. (1999). The elaboration
likelihood model: Current status and contro-
versies. In S. Chaiken & Y. Trope (Eds.), Dual process theories
in social psychology (pp. 41–72).
New York, NY: Guilford Press.
44. Richey, S. (2008a). The autoregressive influence of social
network political knowledge on voting
behaviour. British Journal of Political Science, 38, 527–542.
Richey, S. (2008b). The social basis of voting correctly.
Political Communication, 25, 366–376.
Rolfe, M. (2005, April). Conditional choice: Social networks
and decision-making. Paper presented
at the annual meeting of the Midwest Political Science
Association, Chicago, IL.
Ross, L., Bierbrauer, G., & Hoffman, S. (1976). The role of
attribution processes in conformity and
dissent: Revisiting the Asch situation. American Psychologist,
31, 148–157.
Rousseau, J.-J. (1762/1994). The social contract or the
principles of political right. New York, NY:
Oxford University Press.
Ryan, J. B. (2010). The effects of network expertise and biases
on vote choice. Political
Communication, 27, 44–58.
Ryan, J. B. (2011). Social networks as a shortcut to correct
voting. American Journal of Political
Science, 55, 752–765.
Simon, H. A. (1957). Administrative behavior (2nd ed.). New
York, NY: Macmillan.
Simon, H. A. (1985). Human nature in politics: The dialogue of
psychology with political science.
American Political Science Review, 79, 293–304.
Sniderman, P. M. (1993). The new look in public opinion
research. In A. Finifter (Ed.), Political sci-
45. ence: The state of the discipline II (pp. 219–245). Washington,
DC: American Political Science
Association.
Sokhey, A. E., & McClurg, S. D. (2012). Social networks and
correct voting. Journal of Politics, 74,
751–764.
Tversky, A., & Kahneman, D. (1973). Availability: A heuristic
for judging frequency and probability.
Cognitive Psychology, 5, 207–232.
Zaller, J. (1992). The nature and origins of mass opinion. New
York, NY: Cambridge University
Press.
Copyright of Political Communication is the property of
Routledge and its content may not be copied or
emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission.
However, users may print, download, or email articles for
individual use.
BUSI 300
Discussion Board Instructions
This document explains how you are to complete your
46. Discussion Board Forum threads and replies. The threads are
due in Modules/Weeks 1, 3, 5, and 7. The replies are due in
Modules/Weeks 2, 4, 6, and 8. The learning theories upon which
this class is based are actualized in the threads. Discussion
Board Forums are designed to assess higher-level critical
thinking skills such as analysis, synthesis, and evaluation.
Threads
In Modules/Weeks 1, 3, 5, and 7 you will select 1 of the 4
questions provided and submit your answer in a substantive*
thread of 250–350 words, citing at least 2 approved sources
using correct and current APA formatting. Approved sources
include the Lehman, C., & DuFrene, D. (2014). BCOM (7th
ed.). Boston, MA: Cengage Learning. (as one of your sources)
and a full-text, peer-reviewed article retrieved from Liberty
University’s online journal database (as your second source).
No general Internet searching is acceptable. Therefore, blogs,
education sites, commercial sites, Wikipedia, About.com,
editorials, abstracts, online books, or book reviews are not
acceptable for this assignment.
*Note: Substantive in this instance means completely answering
the question and providing support with your use of
authoritative sources.
The last portion of your thread must include a complete
reference list properly formatted according to current APA
standards. The reference list does not contribute to your word
count. You are also required to include the persistent link to
this record/permalink within your reference list for any journal
articles used. Within EbscoHost, you will find the “Persistent
link to this Record/Permalink” on the page of the journal
database where you see “Title,” “Authors,” “Source,”
“Document Type,” etc. This permalink typically ends in
“…ehost-live&scope=site.” For assistance with the persistent
link, more instructions are found in Additional Materials in the
Syllabus & Assignment Instructions folder.
Threads are due no later than 11:59 p.m. (ET) on Monday of
47. each assigned module/week. Late submissions will not be
accepted or graded.
Replies
In Modules/Weeks 2, 4, 6, and 8 you will be required to submit
a constructive* reply of 200–300 words to 1 other student’s
thread. You must follow the 2-by-2 format, where you discuss at
least 2 points of agreement and 2 points of disagreement and
why. You must expect to enter into ongoing virtual discussions
with your classmates and be prepared to answer questions about
your own topics put forth during the discussion. Student
interaction in the Discussion Board Forum is key to success in
the course.
Note: *Constructive in this instance means promoting further
development of the thread. The goal is to facilitate the learning
process and add suggestions for improvement that benefit the
student to whom you are responding.
Replies 1, 2, and 3 are due no later than 11:59 p.m. (ET) on
Monday of Modules/Weeks 2, 4, and 6. Reply 4 is due no later
than 11:59 p.m. (ET) on Friday of Module/Week 8, the last day
of the course.
Late postings will not be accepted or graded.
To gain the most from your Discussion Board Forums:
· Read both the threads of your classmates and the articles used.
· Draw your classmates into discussion of topics by asking
questions.
· Expect to spend some time each day reviewing all threads and
replies, even those with which you are not involved.
· The research and writing done for each forum is to prepare
you to research and write upper-level research papers in current
APA format. Each submission, whether thread or replies, must
be written without grammar or spelling errors. In addition, all
threads must contain in-text citations and at least 2 approved
48. sources to substantiate and add credibility to your claims. These
must be properly formatted according to current APA
guidelines.
· Utilize the 3–4 paragraph format (1 introductory paragraph, 1–
2 body paragraphs, and 1 conclusion paragraph), as this works
well for this assignment. By employing the use of this format,
readers are able to quickly discern the direction and
organization of the thread. Written communication is often a
challenge for college students due to the perceived complexity
necessary for demonstration of writing skills. However, simpler
is usually better. A logical flow of material is vital for effective
conveyance of a certain message.
Consult the Additional Information folder to find out how to
access the Liberty University Online Library. Also, consult the
same folder to see examples of a properly formatted in-text
citation and a source within a reference list.
Page 2 of 2