To Switch or Not To Switch Understanding Social Influence i.docx

To Switch or Not To Switch: Understanding Social
Influence in Online Choices
Haiyi Zhu*, Bernardo A. Huberman, Yarun Luon
Social Computing Lab
Hewlett Packard Labs
Palo Alto, California, USA
[email protected]; {bernardo.huberman, yarun.luon}@hp.com
ABSTRACT
We designed and ran an experiment to measure social
influence in online recommender systems, specifically how
often people’s choices are changed by others’
recommendations when facing different levels of
confirmation and conformity pressures. In our experiment
participants were first asked to provide their preferences
between pairs of items. They were then asked to make
second choices about the same pairs with knowledge of

others’ preferences. Our results show that others people’s
opinions significantly sway people’s own choices. The
influence is stronger when people are required to make their
second decision sometime later (22.4%) than immediately
(14.1%). Moreover, people seem to be most likely to
reverse their choices when facing a moderate, as opposed to
large, number of opposing opinions. Finally, the time
people spend making the first decision significantly predicts
whether they will reverse their decisions later on, while
demographics such as age and gender do not. These results
have implications for consumer behavior research as well as
online marketing strategies.
Author Keywords
Social influence, online choices, recommender systems.
ACM Classification Keywords
H.5.3 [Information Interfaces and Presentation]: Group and
Organization Interfaces – Collaborative computing, Web-
based interaction; K.4.4 [Computers and Society]:

Electronic Commerce – Distributed commercial
transactions.
General Terms
Experimentation.
INTRODUCTION
Picture yourself shopping online. You already have an idea
about what product you are looking for. After navigating
through the website you find that particular item, as well as
several similar items, and other people’s opinions and
preferences about them provided by the recommendation
system. Will other people’ preferences reverse your own?
Notice that in this scenario there are two contradictory
psychological processes at play. On one hand, when
learning of other’s opinions people tend to select those
aspects that confirm their own existing ones. A large
literature suggests that once one has taken a position on an
issue, one’s primary purpose becomes defending or
justifying that position [21]. From this point of view, if

others’ recommendations contradict their own opinion,
people will not take this information into account and stick
to their own choices. On the other hand, social influence
and conformity theory [8] suggest that even when not
directly, personally, or publicly chosen as the target of
others’ disapproval, individuals may choose to conform to
others and reverse their own opinion in order to restore their
sense of belonging and self-esteem.
To investigate whether online recommendations can sway
peoples’ own opinions, we designed an online experiment
to test how often people’s choices are reversed by others’
preferences when facing different levels of confirmation
and conformity pressures. We used Rankr [19] as the study
platform, which provides a lightweight and efficient way to
crowdsource the relative ranking of ideas, photos, or
priorities through a series of pairwise comparisons. In our
experiment participants were first asked to provide their
preferences between pairs of items. Then they were asked

to make second choices about the same pairs with the
knowledge of others’ preferences. To measure the pressure
to confirm people’s own opinions, we manipulated the time
before the participants were asked to make their second
decisions. And in order to determine the effects of social
pressure, we manipulated the number of opposing opinions
that the participants saw when making the second decision.
Finally, we tested whether other factors (i.e. age, gender
and decision time) affect the tendency to revert.
Our results show that other people’s opinions significantly
sway people’s own choices. The influence is stronger when
people are required to make their second decision later
* Haiyi Zhu is currently a PhD candidate in Human
Computer Interaction Institute at Carnegie Mellon
University. This work was performed while she was a
research intern at HP labs.
Permission to make digital or hard copies of all or part of this
work for

personal or classroom use is granted without fee provided that
copies are
not made or distributed for profit or commercial advantage and
that copies
bear this notice and the full citation on the first page. To copy
otherwise,
or republish, to post on servers or to redistribute to lists,
requires prior
specific permission and/or a fee.
CHI 2012, May 5-10, 2012, Austin, TX, USA.
Copyright 2012 ACM xxx-x-xxxx-xxxx-x/xx/xx...$10.00.
(22.4%) rather than immediately (14.1%) after their first
decision. Furthermore, people are most likely to reverse
their choices when facing a moderate number of opposing
opinions. Last but not least, the time people spend making
the first decision significantly predicts whether they will
reverse their decisions later on, while demographics such as
age and gender do not.
The main contribution of the paper is that we designed and

ran an experiment to understand the mechanisms of social
influence in online recommender systems. Specifically, we
measured the impact of others’ preferences on people’s own
choices under different conditions. The results have
implications for consumer behavior research and online
marketing strategies.
RELATED WORK
Confirming Existing Opinions
Confirmation of existing opinions is a long-recognized
phenomenon [21]. As Francis Bacon stated several
centuries ago [2]:
“The human understanding when it has once
adopted an opinion (either as being received
opinion or as being agreeable to itself) draws all
things else to support and agree with it. Although
there be a greater number and weight of instances
to be found on the other side, yet these it either
neglects and despises, or else by some distinction

sets aside and rejects”
This phenomenon can be explained by Festinger’s
dissonance theory: as soon as individuals adopt a position,
they favor consistent over inconsistent information in order
to avoid dissonance [11].
A great deal of empirical studies supports this idea (see [21]
for a review). Many of these studies use a task invented by
Wason [30], in which people are asked to find the rule that
was used to generate specified triplets of numbers. The
experimenter presents a triplet, and the participant
hypothesizes the rule that produced it. The participants then
test the hypothesis by suggesting additional triplets and
being told whether it is consistent with the rule to be
discovered. Results show that people typically test
hypothesized rules by producing only triplets that are
consistent with them, indicating hypothesis-determined
information seeking and interpretation. Confirmation of
existing opinions also contributes to the phenomenon of

belief persistence. Ross and his colleagues showed that
once a belief or opinion has been formed, it can be very
resistant to change, even after learning that the data on
which the beliefs were originally based were fictitious [25].
Social conformity
In contrast to confirmation theories, social influence
experiments have shown that often people change their own
opinion to match others’ responses. The most famous
experiment is Asch’s [1] line-judgment conformity
experiments. In the series of studies, participants were
asked to choose which of a set of three disparate lines
matched a standard, either alone or after 1 to 16
confederates had first given a unanimous incorrect answer.
Meta-analysis showed that on average 25% of the
participants conformed to the incorrect consensus [4].
Moreover, the conformity rate increases with the number of
unanimous majority. More recently, Cosley and his
colleagues [10] conducted a field experiment on a movie

rating site. They found that by showing manipulated
predictions, users tended to rate movies toward the shown
prediction. Researchers have also found that social
conformity leads to multiple macro-level phenomenons,
such as group consensus [1], inequality and unpredictability
in markets [26], unpredicted diffusion of soft technologies
[3] and undermined group wisdom [18].
Latané proposed a theory [16] to quantitatively predict how
the impact of the social influence will increase as a function
of the size of the influencing social source. The theory
states that the relationship between the impact of the social
influence (I) and the size of the influencing social source (N)
follows a negative accelerating power function
[16]. The theory has been empirically
supported by a meta-analysis of conformity experiments
using Asch’s line-judgment task [4].
There are informational and normative motivations
underlying social conformity, the former based on the
desire to form an accurate interpretation of reality and

behave correctly, and the latter based on the goal of
obtaining social approval from others [8]. However, the two
are interrelated and often difficult to disentangle
theoretically as well as empirically. Additionally, both
goals act in service of a third underlying motive to maintain
one’s positive self-concept [8].
Both self-confirmation and social conformity are extensive
and strong and they appear in many guises. In what follows
we consider both processes in order to understand users’
reaction to online recommender systems.
Online recommender systems
Compared to traditional sources of recommendations -
peers such as friends and coworkers, experts such as movie
critics, and industrial media such as Consumer Reports,
online recommender systems combined personalized
recommendations sensitive to people’s interests and
independently reporting other peoples’ opinions and
reviews. One popular example of a successful online

recommender system is the Amazon product recommender
system [17].
Users’ reaction to recommender system
In computer science and the HCI community, most research
in recommender systems has focused on creating accurate
and effective algorithms for a long time (e.g. [5]). Recently,
researchers have realized that recommendations generated
by standard accuracy metrics, while generally useful, are
not always the most useful to users.[20] People started
building new user-centric evaluation metrics [24,32]. Still,
there are few empirical studies investigating the basic
psychological processes underlying the interaction of users
with recommendations; and none of them addresses both
self-confirmation and social conformity. As mentioned
above, Cosley and his colleagues [10] studied conformity in
movie rating sites and showed that people’s rating are
significantly influenced by other users’ ratings. But they did

not consider the effects of self-confirmation or the effects
of different levels of social conformity pressures. Schwind
et al studied how to overcome users’ confirmation bias by
providing preference-inconsistent recommendations [28].
However, they represented recommendations as search
results rather than recommendations from humans, and thus
did not investigate the effects of social conformity.
Furthermore, their task was more related to logical
inference rather than purchase decision making.
In the area of marketing and customer research, studies
about the influence of recommendations are typically
subsumed under personal influence and word-of-mouth
research [27]. Past research has shown that word-of-mouth
plays an important role in consumer buying decisions, and
the use of internet brought new threats and opportunities for
marketing [27,14,29]. There were several studies
specifically investigating social conformity in product
evaluations [7,9,23]. Although they found substantial

effects of others’ evaluations on people’s own judgments,
the effects were not always significantly stronger when the
social conformity pressures are stronger
1
. In Burnkrant and
Cousineau’s [7] and Cohen and Golden’s [9] experiments,
subjects were exposed to evaluations of coffee with high
uniformity or low uniformity. Both results showed that
participants did not exhibit significantly increased
adherence to others’ evaluation in the high uniformity
condition (although in Burnkrant’s experiments, the
participants significantly recognized the difference between
high and low uniformity). On the other hand, in Pincus and
Waters’s experiments (college students rated the quality of
one paper plate while exposed to simulated quality
evaluations of other raters), it was found that conformity
1
Unlike other conformity experiments such as line-
judgment where the pressure of social conformity is

manipulated by increasing the number of “unanimous”
majority, experiments about social influence in product
evaluation [7, 9, 23] usually manipulate the pressure of
social influence by changing the degree of uniformity of
opinions. As discussed in [9], since it is seldom that no
variation exists in the advice or opinions in reality, the latter
method is more likely to stimulate participants’ real
reactions. We also use the latter method in our experiment
by manipulating the ratio of opposing opinions versus
supporting opinions.
effects are stronger when the evaluations are more
uniform[20].
In summary, while previous research showed that others’
opinions can influence people’s own decisions, none of that
research addresses both the self-confirmation and social
conformity mechanism that underlie choice among several
recommendations. Additionally, regarding to the effects of
increasing social conformity pressures, experiments using

Asch’s line-judgment tasks supported that people are more
likely to be influenced when facing stronger social
pressures, while the findings of studies using product
evaluation tasks were mixed.
Our experiments address how often people reverse their
own opinions when confronted with others people
preferences, especially when facing different levels of
confirmation and conformity pressures. The hypothesis is
that people are more likely to reverse their minds when the
reversion causes less self-inconsistency (the confirmation
pressure is weaker) or the opposing social opinions are
stronger (the conformity pressure is stronger).
a. Comparing baby pictures, not showing others’
preferences
b. Comparing loveseats, showing others’ preferences
Figure 1. Example pairwise comparisons in Rankr

EXPERIMENTAL DESIGN
We conducted a series of online experiments. All
participants were asked to go to the website of Rankr [19]
to make a series of pairwise comparisons with or without
knowing others people’s preferences (Figure 1). The
pictures were collected from Google Images. We wanted to
determine whether people reverse their choices by seeing
others’ preferences.
Basic idea of the experiment
Participants were asked to provide their preferences
between the same pair of items twice. The first time the
participant encountered the pair, they made a choice
without the knowledge of others’ preferences. The second
time the participant encountered the same pair, they made a
choice with the knowledge of others’ preferences. Social
influence is measured as whether people switched their
choices between the first time and second time they

encountered the same pair.
To manipulate the pressure to confirm people’s own
opinions, we changed the time between the two decisions.
In the short interval condition, people first compared two
pictures on their own and they were then immediately asked
to make another choice with available information about
others’ preferences. When the memories were fresh,
reversion leads to strong inconsistency and dissonance of
other people’s choices with their own previous ones (strong
confirmation pressure). However, in the long interval
condition, participants compared pairs of items in the
beginning of the test followed by several distractor pairs
and then followed again by the same pairs the participant
had previously compared but with augmented information
of others’ preferences. In this case, participants’ memories
of their previous choices decay, so the pressure to confirm
their own opinions is less explicit.
To manipulate the social pressure, we changed the number

of opposing opinions that the participants saw when making
the second decision. We selected four levels: the opposing
opinions were twice, five times, ten times, or twenty times
as many as the number of people who supported their
opinions.
In the following section, the details of experimental
conditions are discussed.
Conditions
The experimental design was 2 (baby pictures and loveseat
pictures) X 3 (short interval, long interval and control) X 4
(ratio of opposing opinions versus supporting opinions: 2:1,
5:1, 10:1 and 20:1). Participants were recruited from
Amazon’s Mechanical Turk and were randomly assigned
into one of six conditions (baby–short, baby-long, baby–
control, loveseat-short, loveseat-long and loveseat-control)
and made four choices with different levels of conformity
pressure.
In the baby condition, people were asked to compare

twenty-three or twenty-four pairs of baby pictures by
answering the question “which baby looks cuter on a baby
product label”. Note that the Caucasian baby pictures in
Figure 1 are examples. We displayed baby pictures from
different races in the experiment. In the loveseat condition,
the question was “your close friend wants your opinion on a
loveseat for their living room, which one do you suggest”;
and people also needed to make twenty-three or twenty-four
choices.
In the short interval condition, people first compared two
pictures on their own and they were then immediately asked
to make another choice with available information about
others’ preferences. Furthermore, we tested whether people
would reverse their first choice under four levels of social
pressure: when the number of opposing opinions was twice,
five times, ten times, and twenty times as many as the
number of people who supported their opinions. The
numbers were randomly generated

2
. Except for theses eight
experimental pairs, we also added fourteen noise pairs and
an honesty test composed of two pairs (twenty-four pairs in
total, see Figure 2 for an example). In this condition, noise
pairs also consisted of consecutive pairs (a pair with social
information immediately after the pair without social
information). However, others’ opinions were either
indifferent or in favor of the participants’ choices. We
created an honesty test to identify participants who cheated
the system and quickly clicked on the same answers. The
test consisted of two consecutive pairs with the same items
but with the positions of the items exchanged. Participants
needed to make the same choices among these consecutive
two pairs in order to pass the honesty test. The relative
orders of experimental pairs, noise pairs, and honesty test in
the sequence and the items in each pair were randomly
assigned to each participant.
In contrast with the short interval condition where people

were aware that they reversed their choices, in the long
interval condition we manipulated the order of display and
the item positions so that the reversion was less explicit.
People first compared pairs of the items without knowing
others’ preferences, and then on average after 11.5 pairs
later we showed the participants the same pair (with the
positions of items in the pair exchanged) and others’
opinions. Similarly, with the short interval condition we
showed eight experimental pairs to determine whether
people reversed their previous choices with increasing
pressures of social influence. Additionally, we showed
thirteen noise pairs (nine without others’ preferences and
four with others’ preferences) and performed an honesty
test (see Figure 2 for an example).
2
We first generated a random integer from 150 to 200 as
total participants. Then we generate the number of people
holding different opinions according to the ratio. Here are a

few examples: 51 vs 103 (2X), 31 vs156 (5X), 16 vs 161
(10X) and 9 vs 181(20X).
By increasing the time between two choices, we blurred
people’s memories of their choices in order to exert a subtle
confirmation pressure. However, as people proceeded with
the experiment they were presented with new information
to process. This new information may lead them to think in
a different direction and change their own opinions
regardless of social influence. In order to control for this
confounding factor, we added a long interval control
condition, where the order of the pairs were the same as
with the long interval condition but without showing the
influence of others.
Procedures
We conducted our experiment on Amazon Mechanical Turk
(mTurk) [15]. The recruiting messages stated that the
objective of the survey was to do a survey to collect

people’s opinions. Once mTurk users accepted the task they
were asked to click the link to Rankr, which randomly
directed them to one of the six conditions. This process was
invisible to them.
First, the participants were asked to provide their
preferences about twenty-three or twenty-four pairs of
babies or loveseats. They were then directed to a simple
survey. They were asked to report their age, gender and
answer two 5-Likert scale questions. The questions were as
follows. “Is showing others' preferences useful to you?”
“How much does showing others' preferences influence
your response?” After filling out the survey, a unique
confirmation code was generated and displayed on the
webpage. Participants needed to paste the code back to the
mTurk task. With the confirmation code in hand we
matched mTurk users with the participants of our
experiments, allowing us to pay mTurk users according to
their behaviors. We paid $0.35 for each valid response.

Participants
We collected 600 responses. Of this number, we omitted 37
responses from 12 people who completed the experiment
multiple times; 22 incomplete responses; 1 response which
did not conform to the participation requirements (i.e. being
at least 18 years old); and 107 responses who did not pass
the honesty test. These procedures left 433 valid
participants in the sample, about 72% of the original
number. According to participant self-reporting, 40% were
females; age ranged between18 to 82 with a median age of
27 years. Geocoding
3
the ip addresses of the participants
revealed 57% were from India, 25% from USA, with the
remaining 18% of participants coming from over 34
different countries.
The numbers of participants in each condition were as
follows. Baby-short: 72; baby-long: 91; baby-control: 49;
loveseat-short:75; loveseat-long:99; loveseat-control:47.

4
People spent a reasonable amount of time on each decision
(average 6.6 seconds; median 4.25 seconds).
3
MaxMind GeoLite was used to geocode the ip addresses
which self-reports a 99.5% accuracy rate.
4
Among the 600 responses, originally 20% were assigned
for baby-strong; 20% for baby-weak; 10% for baby-control;
20% for loveseat-strong; 20% for loveseat-weak and10%
for loveseat-control. The valid responses in short interval
conditions were fewer than the ones in long interval
conditions because the short interval condition had a higher
failure rate in the honesty test. The reason might be that
short interval condition had more repetitive pairs, fewer
new items and more straightforward patterns, leading to
boredom and casual decisions, which in turn caused failure
in the honesty tests.

Experimental
pair
Experimental pair displaying
others preferences which are
against people’s previous
choice
Pair
Short interval
Long interval
Long interval control
Pair i, j Pair i, j Pair i, j
Figure 2. Example displaying orders in each condition.

Pair i, j
Pair displaying
others’ preferences
1, 2 2, 1 3, 4 3, 4 5, 6 5, 6 7, 8 7, 8 9,10 9,10 11,12
1
11,12
1
13,14
1
13,14
1
15,16
1
15,16
1
17,18
1

17,18
1
19,20
1 1
19,20
1
21,22
1
21,22
1
23,24
1
23,24
1
1, 2 3, 4 5, 6 6, 5 7, 8 9,10 11,12 13,14 15,16
1
17,18
1

19,20
1
21,22
1
23,24
1
25,26
1 1
27,28
1
8, 7 14,13
1
29,30
1
31,32
1
33,34
1

2, 1 35,36
1
16,15
1
1, 2 3, 4 5, 6 6, 5 7, 8 9,10 11,12 13,14 15,16
1
17,18
1
19,20
1
21,22
1
23,24
1
25,26
1 1
27,28
1

8, 7 14,13
1
29,30
1
31,32
1
33,34
1
2, 1 35,36
1
16,15
1
Honesty test
Honesty test
Honesty test

Among the 433 responses, 243 left comments in the open-
ended comments section at the end of the experiments.
Most of them said that they had a good experience when
participating in the survey. (They were typically not aware
that they were in an experiment).
Measures
knowing others’ opinion.
opinions to supporting opinions.
ple spent in
making each decision.
-reported usefulness of others’ opinions.
-reported level of being influenced.
RESULTS
1. Did people reverse their opinions by others’
preferences when facing different confirmation
pressures?

Figure 3. Reversion rate by conditions.
Figure 3 shows the reversion rate as a function of the
conditions we manipulated in our experiment. First, we
found out that content does not matter, i.e., although baby
pictures are more emotionally engaging than loveseat
pictures, the patterns are the same. The statistics test also
shows that there is no significant difference between the
baby and the loveseat results (t(431)=1.35, p=0.18).
Second, in the short interval condition, the reversion rate
was 14.1%, which is higher than zero (the results of the t-
test is t(146)=6.7, p<0.001).
Third, the percentage of people that reversed their opinion
was as high as 32.5% in the long interval condition,
significantly higher than the long interval control condition
(10.1%) which measures the effects of other factors leading
to reversion regardless of the social influence during the
long interval. T-test shows that this difference is significant:

t(284)=6.5, p<0.001. We can therefore conclude that social
influence contributes approximately to 22.4% of the
reversion of opinions observed.
To summarize the results, in both the long and the short
interval conditions, others’ opinions significantly swayed
people’s own choices (22.4% and 14.1%
5
). The effect size
of social influence was larger when the self-confirmation
pressure was weaker (i.e., the time between the two choices
is larger).
2. Were people more likely to reverse their own
preferences when more people are against them?
Figure 4. Reversion rate by the ratio of opposing opinions.
Table 1. Linear regression predicting the reversion percent.
Note that the squared ratio of opposing opinions has a
significant negative value (-0.142, p=0.045).
Interestingly, we saw an increasing and then decreasing
trend when the opposing opinions became exponentially

stronger (from 2X, 5X, 10X to 20X). The condition with
the most uniform opposing opinions (20X) was not more
effective in reversing people’s own opinions than the
moderate opposing opinions (5X and 10X). The statistical
5
In order to calibrate the magnitude of our results, we point
out that our results are of the same magnitude as the classic
line-judgment experiments. According to a 1996 meta-
analysis of line-judgment experiment consisting of 133
separate experiments and 4,627 participants, the average
conformity rate is 25% [4].
0.00%
5.00%
10.00%
15.00%
20.00%
25.00%
30.00%

35.00%
40.00%
Baby Loveseat
Short Interval
Long Interval
Long Interval
Control
0.00%
5.00%
10.00%
15.00%
20.00%
25.00%
30.00%
35.00%
40.00%
2X 5X 10X 20X
Long
Interval

Short
Interval
Predictors Coef. Std.
Err.
P
value
Condition
(1-long interval;
0-short interval)
.839 .059 <.001
Ratio of opposing opinions .563 .184 .038
Square ratio of opposing
opinions
-.142 .049 .045
Intercept -2.42 .151 <.001
Adjusted R square 0.96
R
e
v
e

rs
io
n
R
e
v
e
rs
io
n
The ratio of opposing opinions to supporting opinions
test is shown in Table 1. Note that the squared ratio of
opposing opinions has a significant negative value (Coef. =
-0.142, p<0.05), suggesting that the returning effect is
statistically significant.
These results might be explained by Brehn’s finding of
psychological reactance [6]. According to Brehn, if an
individual’s freedom is perceived as being reduced or

threatened with reduction, he will become aroused to
maintain or enhance his freedom. The motivational state of
arousal to reestablish or enhance his freedom is termed
psychological reactance. Therefore if the participants
perceived the uniform opposing opinions as a threat to their
freedom to express their own opinions, their psychological
reactance might be aroused to defend and confirm their own
opinions.
These results can also be explained in terms of Wu and
Huberman’s findings about online opinion formation [31].
In their work they used the idea of maximizing the impact
that individuals have on the average rating of items to
explain the phenomenon that later reviews tend to show a
big difference with earlier reviews in Amazon.com and
IMDB.
We can use the same idea to explain our results. Social
influence in product recommendations is not just a one-way
process. People are not just passively influenced by others’

opinions but also want to maximize their impact on other
people’s future decision making (e.g., in our experiments,
according to our recruiting messages, participants would
assume that their choices would be recorded in the database
and shown to others; in real life, people like to influence
their friends and family). We assume that the influence of
an individual on others can be measured by how much his
or her expression will change the average opinion. Suppose
there are supporting opinions and opposing opinions,
and that . A person’s choice c (0 indicates
confirming his or her own choice, 1 indicates conforming to
others) can move the average percentage of opposing
opinions from to .
So the influence on the average opinion is
. A simple derivation shows that to maximize the
influence on average opinion, people need to stick to their
own choices and vote for the minority. Then their influence

gain will be stronger when the difference between existing
majority opinions and minority ones is larger. Therefore,
the motivation to exert influence on other people can play a
role in resisting the social conformity pressure and lead
people to confirm their own decisions especially when
facing uniform opposing opinions.
3. What else predicts the reversion?
We used a logistic regression model to predict the decision-
level reversion with the participants’ age, gender, self-
reported usefulness of the recommendation system, self-
reported level of being influenced by the recommendation
system and standardized first decision time (as shown in
Table 2). Note that standardized first decision time = (time
in this decision – this person’s average decision time) / this
person’s standard deviation. So “first decision time” is an
intrapersonal variable.
The results showed that age and gender do not significantly
predict reversion (p=0.407, p=0.642). Self-reported

influence level has a strong prediction power (Coef. = 0.334,
p <0.001), which is reasonable. The interesting fact is that
decision time, a simple behavioral measure, also predicts
reversion very well (Coef. = 0.323, p<0.001). The longer
people spent on the decisions, the more equivalent the two
choices are for them. According to Festinger’s theory [12]:
the more equivocal the evidence, the more people rely on
social cues. Therefore, the more time people spend on a
choice, the more likely they are to reverse this choice and
conform to others later on.
DISCUSSION
On one hand, the phenomena we found (i.e., the returning
effects of strong influence pressure) are quite different from
the classic line-judgment conformity studies [1,16]. Notice
that these experiments used questions with only one single
correct answer [1]. In contrast, in our experiment we
examined the social influence on people’s subjective

preferences among similar items, which might result in
such different phenomena. On the other hand, our findings
reconcile the mixed results of studies investigating social
conformity in product evaluation tasks (which are often
subjective tasks) [7,9,23]. Due to the returning effects,
strong conformity pressure is not always more effective in
influencing people’s evaluations compared to weak
conformity pressure.
Additionally, in our experiment we did not look at social
influence in scenarios where the uncertainty level is quite
high. For example, in settings such as searching
recommendations for restaurants or hotels where people
have never been to, the level of uncertainty is high and
people need to rely on other cues. However, in our
experimental setting people can confidently make their
Predictors Coef. Std.
Err.
P>|z|

Condition
(1-long interval;
0-short interval)
1.26 .152 <.001
Age -.006 6.89e-3 .407
Gender .067 .143 .642
Self-reported usefulness .164 .070 .020
Self-reported influence level .334 .072 <.001
Std. first decision time .323 .065 <.001
Log likelihood -657.83
Table 2. Logistic regression predicting the reversion.
choice by comparing the pictures of the settings. Therefore,
some caution is needed when trying to generalize our
results to other settings with high uncertainty.
Regarding the tasks used in the experiment, the choice of
“which baby looks cuter on a product label” involves

emotions and subjective feelings. Alternatively similar
choices include the preferences of iTune music or
YouTube's commercial videos. In contrast, the other type of
question, “which loveseat is better”, is less emotionally
engaging. In this case people are more likely to consider the
usability factors such as color and perceived
comfortableness when making these types of choices.
Results showed that the impact of social influence on these
two different types of choices are similar and consistent,
which suggests the general applicability of the results.
There might be concern about cultural differences between
the global nature of the respondents and the Western-style
nature of the tasks. We point out that this intercultural
preference difference inherent in the participant is assumed
consistent throughout the test and therefore should not
affect a particular person’s preference changes and will not
confound the results.
During our research, we invented an experimental paradigm

to easily measure the effects of social influence on people’s
decision making (i.e., the reversion) and manipulate
conditions under which people make choices. This
paradigm can be extended to scenarios beyond those of
binary choices, to the effect of recommendations from
friends as opposed to strangers and whether social influence
varies with different visualizations for the recommendations.
LIMITATIONS & FUTURE WORK
In our experiments, we examined whether people reverse
their choices when facing different ratios of opposing
opinions versus supporting opinions (2X, 5X, 10X and
20X). In order to further investigate the relationship
between the ratio of the opposing opinions and the tendency
to revert, it would be better to include more fine-grained
conditions in the ratio of opposing opinions. The ideal
situation would be a graph with the continuous opposing
versus supporting ratio as the x-axis and the reversion rate
as the y-axis.

Also, additional manipulation checks or modification of the
design of the experiment would be needed to establish
whether processes such as psychological reactance or the
intent to influence others have been operating. For example,
the degree of perceived freedom in the task could be
measured. And it would be revealing to manipulate whether
or not people’s choices would be visible to other
participants to see whether the intention of influencing
others takes effect.
Regarding the usage of Mechanical Turk as a new source of
experimental data, we agree with Paolacci and his
colleagues that, “workers in Mechanical Turk exhibit the
classic heuristics and biases and pay attention to directions
at least as much as subjects from traditional sources” [22].
Particularly, compared to recruiting people from a college
campus, we believe the use of Mechanical Turk has a lower
risk of introducing the Hawthorne Effect (i.e., people alter
their behaviors due to the awareness that they are being

observed in experiments), which is itself a form of social
influence and might contaminate our results.
In our experiment, we used several methods such as an
honest test and IP address checking to further ensure that
we collected earnest responses from Mechanical Turk
workers. The average time they spent on the task, the
statistically significant results of the experiment and the
comments participants left all indicate that our results are
believable. However, there is still a limitation of our
honesty test. On one hand, the honesty test (the consecutive
two pairs with positions of items switched) was unable to
identify all the users who tried to cheat the system by
randomly clicking on the results, which added noise in our
data. On the other hand, the honesty test might also exclude
some earnest responses. It is possible that, immediately
after people made a choice, they regret it.
CONCLUSION & IMPLICATION
In this paper, we present results of a series of online

experiments designed to investigate whether online
recommendations can sway peoples’ own opinions. These
experiments exposed participants making choices to
different levels of confirmation and conformity pressures.
Our results show that people’s own choices are significantly
swayed by the perceived opinions of others. The influence
is weaker when people have just made their own choices.
Additionally, we showed that people are most likely to
reverse their choices when facing a moderate, as opposed to
large, number of opposing opinions. And last but not least,
the time people spend making the first decision
significantly predicts whether they will reverse their own
later on.
Our results have three implications for consumer behavior
research as well as online marketing strategies. 1) The
temporal presentation of the recommendation is important;
it will be more effective if the recommendation is provided
not immediately after the consumer has made a similar

decision. 2) The fact that people can reverse their choices
when presented with a moderate countervailing opinion
suggests that rather than overwhelming consumers with
strident messages about an alternative product or service, a
more gentle reporting of a few people having chosen that
product or service can be more persuasive than stating that
thousands have chosen it. 3) Equally important is the fact
that a simple monitoring of the time spent on a choice is a
good indicator of whether or not that choice can be reversed
through social influence. There is enough information in
most websites to capture these decision times and act
accordingly.
ACKNOWLEDGMENTS
We thank Anupriva Ankolekar, Christina Aperjis, Sitaram
Asur, Dave Levin, Thomas Sanholm, Louis Yu, Mao Ye at
HP labs and the members of the Social Computing Group at
Carnegie Mellon University for helpful feedback.

REFERENCES
1. Asch, S.E. (1956) Studies of independence and
conformity: I. A minority of one against a unanimous
majority. Psychological Monographs, Vol 70(9), 1956,
70.
2. Bacon, F., (1939), Novum organum. In Burtt, E.A. (Ed),
The English philosophers from Bacon to Mill (pp.24-
123). New York: Random House. (Original work
published in 1620).
3. Bendor. J., Huberman, B.A., Wu,F., (2009)
Management fads, pedagogies, and other soft
technologies, Journal of Economic Behavior &
Organization, Volume 72, Issue 1, October 2009, Pages
290-304
4. Bond, R, Smith, P. B., Culture and conformity: A meta-
analysis of studies using Asch's (1952b, 1956) line
judgment task. Psychological Bulletin, Vol 119(1), Jan
1996, 111-137.
5. Breese JS, Heckerman D, Kadie C. (1998) Empirical
analysis of predictive algorithms for collaborative

filtering. Proceedings of the 14th Conference on
Uncertainty in Artificial Intelligence. Morgan
Kaufmann: San Francisco, CA.
6. Brehm, J.W. (1966) A Theory of Psychological
Reactance. New York: Academic Press.
7. Burnkrant, R.E., and Cousineau, A., Informational and
Normative Social Influence in Buyer Behavior, Journal
of Consumer Research, Vol. 2, No. 3 (Dec., 1975), pp.
206-215.
8. Cialdini, R.B., and Goldstein,N.J., (2004) Social
Influence: Compliance and Conformity. Annual Review
of Psychology. Vol. 55: 591-621.
9. Cohen, Joel B.; Golden, Ellen (1972) Informational
social influence and product evaluation. Journal of
Applied Psychology, Vol 56(1), Feb 1972, 54-59.
10. Cosley. D., Lam, S.K, Albert., I, Konstan., J.A., and
Riedl. J. (2003). Is seeing believing?: how recommender
system interfaces affect users' opinions. In Proceedings
of the SIGCHI conference on Human factors in
computing systems (CHI '03).
11. Festinger, L. (1957). A theory of cognitive dissonance.

Stanford University Press.
12. Festinger, L., (1954). A theory of social comparison
processes. Human Relations 7, 117–140.
13. Gerard, H.B., Wilhelmy, R.A., & Conolley, E.S.(1968)
Conformity and group size. Journal of Personality and
Social Psychology, 1968, 8, 79-82.
14. Hennig-Thurau, T., Gwinner, K. P., Walsh, G., Gremler,
D. D., (2004) Electronic word-of-mouth via consumer-
opinion platforms: What motivates consumers to
articulate themselves on the Internet? Journal of
Interactive Marketing.
15. Kittur, A., Chi, E.H., and Suh, B. (2008) Crowdsourcing
user studies with Mechanical Turk. In Proceeding of the
twenty-sixth annual SIGCHI conference on Human
factors in computing systems (CHI '08). ACM, New
York, NY, USA, 453-456.
16. Latané, B.,(1981) The psychology of social impact.
American Psychologist, Vol 36(4), Apr 1981, 343-356.
17. Linden, G.; Smith, B.; York, J.; , Amazon.com
recommendations: item-to-item collaborative filtering,
Internet Computing, IEEE , vol.7, no.1, pp. 76- 80,

Jan/Feb 2003
18. Lorenz. J., Rauhut, H., Schweitzer, F., and Helbing, D.
How social influence can undermine the wisdom of
crowd effect. PNAS 2011 108 (22) 9020-9025.
19. Luon, Y., Aperjis, C., and Huberman, B.A., Rankr: A
Mobile System for Crowdsourcing Opinions.
http://www.hpl.hp.com/research/scl/papers/rankr/rankr.p
df
20. McNee, S.M., Riedl, J., and Konstan, J.A. (2006). Being
accurate is not enough: How accuracy metrics have hurt
recommender systems. CHI Extended Abstracts, 1097-
1101.
21. Nickerson, R. S.(1998) Confirmation bias: A ubiquitous
phenomenon in many guises. Review of General
Psychology, Vol 2(2), Jun 1998, 175-220.
22. Paolacci, G, Chandler, J, and Ipeirotis, P.G. (2010)
"Running experiments on Amazon Mechanical Turk."
Judgment and Decision Making, vol 5, no 5.
23. Pincus, S.; Waters, L. (1977) K. Informational social
influence and product quality judgments. Journal of
Applied Psychology, Vol 62(5), Oct 1977, 615-619.

24. Pu, P., Chen, L., and Hu, R. (2011). A user-centric
evaluation framework for recommender systems. In
Proceedings of the fifth ACM conference on
Recommender systems (RecSys '11). ACM, New York,
NY, USA, 157-164.
25. Ross, L., Lepper, M. R., & Hubbard, M. (1975).
Perserverance in self perception and social perception:
Biased attributional processes in the debriefing
paradigm. Journal of Personality and Social
Psychology, 32, 880-892.
26. Salganik, M.J., Dodds, P.S., and Watts. D.J.,
Experimental Study of Inequality and Unpredictability
in an Artificial Cultural Market. Science 10 February
2006: Vol. 311 no. 5762 pp. 854-856.
http://www.hpl.hp.com/research/scl/papers/rankr/rankr.pdf
http://www.hpl.hp.com/research/scl/papers/rankr/rankr.pdf
27. Senecal, S., Nantel, J., The influence of online product
recommendations on consumers' online choices, Journal
of Retailing, Volume 80, Issue 2, 2004, Pages 159-169.
28. Schwind, C., Buder, J., and Hesse, F.W. 2011. I will do
it, but i don't like it: user reactions to preference-

inconsistent recommendations. In Proceedings of the
2011 annual conference on Human factors in computing
systems (CHI '11).
29. Stauss, B. (1997). Global Word of Mouth. Service
Bashing on the Internet is a Thorny Issue. Marketing
Management, 6(3), 28 –30.
30. Wason, P. C. (1960) On the failure to eliminate
hypotheses in a conceptual task.The Quarterly Journal
of Experimental Psychology, Vol 12, 1960, 129-140.
31. Wu, F., and Huberman, B.A., (2010) Opinion formation
under costly expression. ACM Trans. Intell. Syst.
Technol. 1, 1, Article 5 (October 2010), 13 pages.
32. Xiao, B. and Benbasat, I. (2007). Ecommerce Product
Recommendation Agents: Use, Characteristics, and
Impact. MIS Quarterly 31(1), 137-209.
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journal
Code=psif20

Social Influence
ISSN: 1553-4510 (Print) 1553-4529 (Online) Journal homepage:
https://www.tandfonline.com/loi/psif20
Moral conformity in online interactions: rational
justifications increase influence of peer opinions
on moral judgments
Meagan Kelly, Lawrence Ngo, Vladimir Chituc, Scott Huettel &
Walter
Sinnott-Armstrong
To cite this article: Meagan Kelly, Lawrence Ngo, Vladimir
Chituc, Scott Huettel & Walter
Sinnott-Armstrong (2017) Moral conformity in online
interactions: rational justifications
increase influence of peer opinions on moral judgments, Social
Influence, 12:2-3, 57-68, DOI:
10.1080/15534510.2017.1323007
To link to this article:
https://doi.org/10.1080/15534510.2017.1323007
Published online: 28 Apr 2017.
Submit your article to this journal
Article views: 776
View related articles
View Crossmark data
Citing articles: 1 View citing articles

https://www.tandfonline.com/action/journalInformation?journal
Code=psif20
https://www.tandfonline.com/loi/psif20
https://www.tandfonline.com/action/showCitFormats?doi=10.10
80/15534510.2017.1323007
https://doi.org/10.1080/15534510.2017.1323007
https://www.tandfonline.com/action/authorSubmission?journalC
ode=psif20&show=instructions
https://www.tandfonline.com/action/authorSubmission?journalC
ode=psif20&show=instructions
https://www.tandfonline.com/doi/mlt/10.1080/15534510.2017.1
323007
https://www.tandfonline.com/doi/mlt/10.1080/15534510.2017.1
323007
http://crossmark.crossref.org/dialog/?doi=10.1080/15534510.20
17.1323007&domain=pdf&date_stamp=2017-04-28
17.1323007&domain=pdf&date_stamp=2017-04-28
https://www.tandfonline.com/doi/citedby/10.1080/15534510.201
7.1323007#tabModule
https://www.tandfonline.com/doi/citedby/10.1080/15534510.201
7.1323007#tabModule
Social influence, 2017
Vol. 12, noS. 2–3, 57–68
https://doi.org/10.1080/15534510.2017.1323007
Moral conformity in online interactions: rational justifications
increase influence of peer opinions on moral judgments
Meagan Kellya,b†, Lawrence Ngoa,c,d,g†, Vladimir Chituch,
Scott Huettele,f,g and
Walter Sinnott-Armstronga,b,e
aKenan institute for ethics, Duke university, Durham, nc, uSa;

bDepartment of Philosophy, Duke university,
Durham, nc, uSa; cMedical Scientist Training Program, Duke
university School of Medicine, Durham, nc, uSa;
dDepartment of neurobiology, Duke university School of
Medicine, Durham, nc, uSa; ecenter for cognitive
neurosciences, Duke university, Durham, nc, uSa; fDepartment
of Psychology and neuroscience, Duke
university, Durham, nc, uSa; gBrain imaging analysis center,
Duke university, Durham, nc, uSa; hSocial
Science Research institute, Duke university, Durham, nc, uSa
ABSTRACT
Over the last decade, social media has increasingly been used as
a
platform for political and moral discourse. We investigate
whether
conformity, specifically concerning moral attitudes, occurs in
these
virtual environments apart from face-to-face interactions.
Participants
took an online survey and saw either statistical information
about
the frequency of certain responses, as one might see on social
media
(Study 1), or arguments that defend the responses in either a
rational
or emotional way (Study 2). Our results show that social
information
shaped moral judgments, even in an impersonal digital setting.
Furthermore, rational arguments were more effective at eliciting
conformity than emotional arguments. We discuss the
implications of
these results for theories of moral judgment that prioritize
emotional
responses.

Introduction
People conform to a blatantly erroneous majority opinion, even
on a simple perceptual
task (Asch, 1956). Although a large body of research in social
psychology has elucidated
some of the varying conditions under which conforming
behavior occurs – such as social
setting, type of judgment, number and group membership of the
confederates – contention
remains about exactly what the conditions are (Bond & Smith,
1996).
Changes in how people interact socially – from synchronous in-
person conversations
to asynchronous and abstract digital communication – present
new environments for con-
formity. Research predating the development of anonymous
online settings suggests that,
without direct, face-to-face contact, there won’t be the same
level of pressure to conform
(e.g., Allen, 1966; Deutsch & Gerard, 1955; Levy, 1960).
Furthermore, early research dur-
ing the development of online spaces suggest that, without
nonverbal cues such as body
KEYWORDS
conformity; morality;
reasoning; emotion; social
media
ARTICLE HISTORY
Received 26 august 2016
accepted 18 april 2017
© 2017 informa uK limited, trading as Taylor & francis Group

CONTACT Walter Sinnott-armstrong [email protected],
[email protected]
†equal contribution.
mailto: [email protected]
mailto: [email protected]
http://www.tandfonline.com
17.1323007&domain=pdf
58 M. KELLY ET AL.
language or prosody, digital communication will alter the ways
in which we exchange infor-
mation, communicate norms, and exert persuasive influence
(Bargh & McKenna, 2004).
Nonetheless, in certain online contexts, other studies have
shown that laws of social influ-
ence, such as the foot-in-the-door technique, still hold in purely
virtual settings (Eastwick
& Gardner, 2009), and merely providing participants with
numerical consensus information
can change prejudicial beliefs about various racial groups
(Stangor, Sechrist, & Jost, 2001)
and the obese (Puhl, Schwartz, & Brownell, 2005). This
suggests that, while there may have
been initial doubts about the extent of conformity in anonymous
online contexts, these new
virtual spaces remain susceptible to social influence.
Prior research has also raised questions about whether
conformity operates differently
within certain domains, such as moral or evaluative judgments.
Traditional philosophical

views (e.g., Aristotle, 1941; Kant, 1996) emphasize that moral
judgments should ideally be
free from social influences, depending only one’s own
judgment. In line with this ideal, more
recent psychological experimentation suggests that people at
least sometimes are less likely
to conform when they have a strong moral basis for an attitude
(Hornsey, Majkut, Terry, &
McKimmie, 2003). In contrast, however, other studies have
shown that at least some moral
opinions can be influenced by social pressure in small group
discussions (Aramovich, Lytle,
& Skitka, 2012; Kundu & Cummins, 2012; Lisciandra, Postma-
Nilsenová, & Colombo,
2013), and information about the distribution of responses
elicits conformity in deonto-
logical, but not consequentialist, responses to the Trolley
problem (Bostyn & Roets, 2016).
Taking these ideas together, we were interested in whether the
mere knowledge of others’
opinions online would produce conformity regarding moral
issues, particularly in online
contexts.
Study 1: impersonal statistics influence moral judgments
In Study 1, we examined participants’ sensitivity to anonymous
moral judgments regard-
ing ethical dilemmas. We presented participants with two
stories, along with statistical
information about how other participants had responded. Unlike
other research providing
distributions of responses (e.g., Bostyn & Roets, 2016) this
information is similar to what
users might see on a social media website like Twitter,
Facebook, or Reddit, where users

can see numerical information about how other users reacted to
some opinion (e.g., ‘15
users liked this post’ or ‘35 users favorited this tweet’). While
we provide no information
about what proportion of participants responded this way to
each scenario, this mirrors
the experience of being in an online context where we are
unaware of how many users have
seen a post without reacting.
Method
Participants
Participants were recruited through the online labor market
Amazon Mechanical Turk
(MTurk) and redirected to Qualtrics to complete an online
survey. All participants provided
written informed consent as part of an exemption approved by
the Institutional Review
Board of Duke University. Each participant rated one of two
scenarios; 302 participants
rated Scenario A, while 290 participants rated Scenario B.
Participants were restricted to
those located in the US with a task approval rating of at least
80%. Although no demographic
SOCIAL INFLUENCE 59
information was collected on our participants specifically, a
typical sample of MTurk users
is considerably more demographically diverse than an average
American college sample
(36% non-White, 55% female; mean age = 32.8 years,
SD = 11.5; Buhrmester, Kwang, &

Gosling, 2011). Numerous replication studies have also
demonstrated that data collected
on MTurk is reliable and consistent with other methods (Rand,
2012). Participants were
compensated $.10 for their involvement.
Materials
Participants were randomly assigned to one of two scenarios.
Scenario A, one of Haidt’s
classic moral scenarios, describes a family that eats their dead
pet dog (Haidt, Koller, & Dias,
1993). Scenario B involves the passengers of a sinking lifeboat
that sacrifice an overweight,
injured passenger. (See Table 1 for full text of scenarios.) These
scenarios were chosen partly
because they fall under different moral foundations (Haidt &
Graham, 2007). Because the
foundations have been shown to exhibit dissimilar properties in
other studies (e.g., Young
& Saxe, 2011), we were interested in how the degree of
conformity might vary in a scenario
involving harm violations versus purity violations.
Procedure
Participants read an ethical dilemma and were asked how
morally condemnable the agent’s
actions were. Ratings were made on an 11-point Likert scale
from 0 (completely morally
acceptable) to 10 (completely morally condemnable).
Participants were randomly assigned to
one of three conditions in this survey. Two of the conditions
contained a prime to induce
conformity by providing an established opinion about the
scenario. The form of that prime
mirrored that seen on many social media websites (e.g.,
Facebook): it described the number

of people who provided a given rating when viewing a similar
scenario. For Scenario A,
participants read the following: ‘58 people who previously took
this survey rated it as morally
condemnable [acceptable]’. Participants read an identical
statement for Scenario B, except
they were told that 65 people previously took the survey. To
ensure that no deception was
used, these numbers of people had indeed rated these scenarios
that way in a previous
experiment.
The final condition served as a baseline and contained no prime;
participants merely read
and rated the moral dilemma. This design was repeated in
separate samples for scenarios
A and B. While the core of the paradigm remained constant
throughout our experiments,
the survey from Study 1 Scenario B also contained a follow-up
question measuring level of
confidence and a catch question about details from the scenario.
Table 1. Scenarios detailing moral violations in the purity
(Scenario a) and harm (Scenario B) domains.
Scenario
a a family’s dog was killed by a car in front of their house.
They had heard that dog meat was delicious, so
they cut up the dog’s body and cooked it and ate it for dinner
B a cruise boat sank. a group of survivors are now overcrowding
a lifeboat, and a storm is coming. The
lifeboat will sink, and all of its passengers will drown unless
some weight is removed from the boat.
nobody volunteers. Ten passengers are so small that two of

them would have to be thrown overboard
to save the rest. However, one passenger is very large and
seriously injured. if the ten small passengers
throw the very large passenger overboard, then he will drown
but the others will survive. They throw
the large passenger overboard
60 M. KELLY ET AL.
Results
We performed a one-way ANOVA on moral ratings by condition
for each scenario. In
Scenario A, moral ratings differed significantly across three
conditions, [F(2, 299) = 3.78,
p = .024, �2
p
= .025]. Post-hoc Tukey tests of the three conditions indicated
that the con-
demnable group (M = 7.09, SD = 2.98) gave significantly higher
ratings (more condemnable)
than the acceptable group (M = 5.80, SD = 3.67), p = .019,
d = .39 (Figure 1). Comparisons
between the baseline group (M = 6.26, SD = 3.47) and the other
two groups were not sig-
nificant. The same results were obtained for Scenario B: moral
ratings differed significantly
across three conditions, [F(2,287) = 4.28, p = .015, �2
p
= .029]. Post-hoc Tukey tests of the

three conditions indicated that the condemnable group
(M = 6.08, SD = 2.90) gave signifi-
cantly higher ratings (more condemnable) than the acceptable
group (M = 4.82, SD = 3.08),
p = .010, d = .42 (Figure 1). Comparisons between baseline
group (M = 5.43, SD = 2.97)
and the other two groups were not significant. For illustrative
purposes all figures show the
average difference from baseline for each condition.
Discussion
We found manipulations containing sparse statistical data about
other participants’ attitudes
were effective in inducing conformity in moral judgments.
Though early research in con-
formity suggested that face-to-face interactions were critical,
and both philosophical and
psychological writing on moral judgments suggest it should be
free from social influence,
these results show that all that is required to induce conformity
in moral judgments is to
provide statistical information about how others responded.
Even subtle social information
in anonymous contexts seems to affect moral judgments.
Figure 1. Statistical information about other participants’ moral
judgments significantly influences
individual responses.
note: error bars represent standard errors. *p < .05.
SOCIAL INFLUENCE 61
Having observed conformity to manipulations containing only

statistical information,
we were next interested in how different kinds of arguments,
specifically emotional and
rational arguments, might be more or less effective at
influencing moral judgments.
Study 2: rational arguments elicit more conformity than
emotional
arguments
Having observed conformity to primes using mere statistical
information, we were inter-
ested in whether the effect could be strengthened by the
addition of different types of argu-
ments: those containing emotionally charged language to appeal
to participants’ feelings or
arguments using reasoning referring to consequences or moral
principles. The distinction
between emotional and rational arguments reflects some of the
core predictions put forth
by prominent psychological models of moral judgment. In the
Social Intuitionist Model
(SIM), for example, ‘moral intuitions (including moral
emotions) come first and directly
cause moral judgments’ (Haidt, 2001, p. 814), while reasoning
is purely a post hoc defense
of those emotional intuitions. The SIM predicts that moral
conformity would only manifest
by altering others’ emotional intuitions, thus in order to change
what people think about a
moral issue, they must first change how they feel.
This prediction is supported by a host of studies that measure
changes in moral opinions
after manipulating emotions and reasoning (for a review, see
Avramova & Inbar, 2013).

For example, inducing positive emotions through funny videos
(Valdesolo & DeSteno,
2006), encouraging emotion regulation (Feinberg, Willer,
Antonenko, & John, 2012), and
prompting longer reflection (Paxton & Greene, 2010) all
generated less harsh moral judg-
ments. Furthermore, moral outrage from one scenario may spill
over into harsher judgments
of subsequent scenarios (Goldberg, Lerner, & Tetlock, 1999),
and emotion drives higher
ascription of intentionality in cases involving negative
consequences (Ngo et al., 2015).
Recent work utilizing virtual reality also demonstrates a
discrepancy between hypothetical
moral judgments and moral decisions taken in virtual
environments, and this discrepancy
seems modulated by emotional responses (Francis et al., 2016;
Patil, Cogoni, Zangrando,
Chittaro, & Silani, 2014). Other work, for example, suggests
that emotions are instrumental
for driving moral behavior (for a review, see Teper, Zhong, &
Inzlicht, 2015). Therefore, this
literature suggests that emotional manipulations would be
particularly effective in swaying
moral attitudes.
In accordance with these findings, we hypothesized that
arguments appealing to partic-
ipants’ emotions would affect their judgments more than
arguments citing abstract princi-
ples, rights, or reasons. To test this hypothesis, we gave
participants emotional or rational
justifications for why the dilemma was either morally
acceptable or morally condemnable
according to previous participants.

Method
Participants
Again, participants were recruited online from Amazon
Mechanical Turk and redirected
to a survey on Qualtrics. Scenario A was rated by 506
participants, and 496 participants
rated Scenario B. All participant restrictions and compensation
rates were identical to
Study 1. To ensure that participants interpreted the stimuli as
intended, we recruited 160
62 M. KELLY ET AL.
additional subjects via Amazon Mechanical Turk, two of which
were dropped for failing
an attention check.
Procedure
Once more, participants were presented with a vignette
describing a moral violation and
asked how morally wrong they believed the agent’s actions were
on a scale from 0 (com-
pletely morally acceptable) to 10 (completely morally
condemnable). However, in this experi-
ment, participants were randomly assigned either to a baseline
or one of four experimental
conditions. The four experimental conditions arose from a 2 × 2
between-subjects factorial
design with statistical norm (condemnable vs. acceptable) as
one IV, similar to Study 1,
and argument type (emotional vs. rational) as the other. The
condemnable emotional
argument in Scenario B, for instance, stated: ‘75 people who

previously took this survey
rated it as morally condemnable and said something similar to
“Those barbaric passen-
gers committed a horrible murder!”’ Analogously, the
condemnable rational argument
in Scenario B was:
75 people who previously took this survey rated it as morally
condemnable and said some-
thing similar to ‘The passengers do not have the right to judge
who gets thrown off. Whether
someone is large or small, injured or uninjured, it is never okay
to take a life.’ (See Table 2 for
full text of Study 2 manipulations.)
The baseline condition contained no manipulations. Again, this
paradigm was repeated for
scenarios A and B. The content used for the arguments
represents a combination of indi-
vidual replies to a previous survey’s free response question
prompting participants to either
explain the rationale behind their rating or describe their
emotional response to the scenario.
To ensure that these naturalistic responses were interpreted as
either rational or emotional
by our subjects, we presented participants in our post hoc test
with one random argument
Table 2. Study 2 manipulations representing actual participant
responses from a prior study.
Condition Scenario A Scenario B
acceptable rational fifty-eight people who previously took this
survey rated it as morally acceptable,
and said something similar to ‘The family

did not cause the dog any harm; it was
already dead. Many cultures eat dogs, and
they should not let food go to waste.’
Seventy-five people who previously took
this survey rated it as morally acceptable,
and said something similar to ‘The pas-
sengers did what they had to do to save
the most human lives. The injured man
may not have survived anyways.’
acceptable emotional fifty-eight people who previously took
this
survey rated it as morally acceptable, and
said something similar to ‘i feel bad for
the poor family! They must have been
starving to have to make this decision!’
this survey rated it as morally acceptable,
and said something similar to ‘i feel bad
for the passengers because they had to
make an extremely stressful choice!’
condemnable emotional fifty-eight people who previously took
this
survey rated it as morally condemnable,
and said something similar to ‘i feel
completely disgusted that this sick family
would eat a beloved pet!’
this survey rated it as morally condemna-
ble, and said something similar to ‘Those
barbaric passengers committed a horrible
murder. i am sickened by what they did!’

condemnable rational fifty-eight people who previously took
this
survey rated it as morally condemnable,
and said something similar to ‘You are
supposed to respectfully mourn and
honor a dead pet’s body with a proper
burial, not abuse it.’
this survey rated it as morally condemn-
able, and said something similar to ‘The
passengers do not have the right to judge
who gets thrown off. Whether someone
is large or small, injured or uninjured, it is
never okay to take a life.’
SOCIAL INFLUENCE 63
from Scenario A and another from Scenario B in a within-
subjects design. Participants
rated these arguments on a scale from 1 (‘Not at all rational
[emotional]’) to 7 (‘Extremely
rational [emotional]’).
In order to compare the magnitude of conformity based on
whether participants were
conforming to condemnable information or acceptable
information, we converted the raw
moral ratings into a conformity index to account for the fact
that the acceptable and con-
demnable conditions moved participant’s responses in opposite
directions. This allows us
to compare the magnitude of conformity based on whether

participants were conforming
to condemnable information or acceptable information.
To construct the conformity index, we calculated the difference
in moral ratings from
the baseline and sign-normalized for condition. Thus, positive
scores represented agree-
ment with the provided statistical norm, or conformity, while
negative scores represented
disagreement with the statistical norm, or non-conformity/anti-
conformity. First, we sub-
tracted the average of the baseline condition from each moral
rating and took the absolute
value of that number (see Figure 2 for the raw differences from
the baseline). Next, based
on condition, we assessed whether the difference from the
baseline represented conform-
ity or non-conformity. On the moral rating scale, higher
numbers corresponded to more
condemnable ratings. Therefore, if a rating in the condemnable
condition was greater than
the baseline, it remained positive to represent conformity. If a
rating in the condemnable
condition was less than the baseline, it was made negative to
represent non-conformity.
There were no ratings in either scenario or for any condition
that was exactly at the baseline.
The opposite was done for the acceptable condition, where
ratings below the baseline repre-
sented conformity (and thus stayed positive), while ratings
above the baseline represented
non-conformity (and thus made negative).
Figure 2. Rational arguments have a stronger effect on
participants’ moral judgments than emotional
arguments.

note: error bars represent standard errors.
64 M. KELLY ET AL.
Results
Our post hoc test of argument type revealed that, on the whole,
participants rated the rational
arguments as more rational (M = 4.71, SD = 1.86) than
emotional (M = 4.20, SD = 2.07,
t(314) = 2.28, p = .02, d = .26) on a 7-point scale. Similarly,
participants rated emotional
arguments as more emotional (M = 5.67, SD = 1.32) than
rational (M = 4.13, SD = 1.88,
t(314) = 8.37, p < .0001, d = .94) on a 7-point scale. This
suggests that the participants in
our main experiment interpreted our stimuli as intended.
To test the role of argument type and statistical norm, we
conducted a 2 (argument
type: emotional vs. rational) × 2 (statistical norm: condemnable
vs. acceptable) between
subjects ANOVA. Starting with the raw scores of Scenario A
(see Figure 2), we found a
main effect of statistical norm [F(1, 401) = 15.89, p < .001, �2
p
= .038], replicating the results
of Experiment 1. There was no main effect, however, of type of
argument [F(1, 401) = 1.18,
p = .28, �2
p

= .003], though the interaction between argument type and
norm was significant
[F(1, 401) = 5.94, p = .02, �2
p
= .015].
To explore directly the extent to which each condition elicited
conformity, we conducted
a 2 × 2 ANOVA using the conformity index. In Scenario A,
there was a main effect of
argument type [F(1, 401) = 5.94, p = .015, �2
p
= .015], such that the conformity index was
significantly greater for rational arguments (M = 1.09,
SD = 3.42) than for emotional argu-
ments (M = .27, SD = 3.49). There was also a significant main
effect of statistical norm [F(1,
401) = 5.48, p = .02, �2
p
= .013], such that acceptable judgments elicited more
conformity
(M = 1.07, SD = 3.46) than condemnable judgments (M = .28,
SD = 3.45) . There was no
significant interaction, however, between statistical norm and
argument type [F(1, 401) =
1.18, p = .28, �2
p
= .003] for the conformity index.
A similar pattern of results obtained for Scenario B. Starting

with the raw scores, we found
a main effect of statistical norm [F(1, 394) = 10.53, p = .001,
�2
p
= .026], again replicating
the results of Experiment 1. There was no main effect, however,
of type of argument [F(1,
401) = .92, p = .337, �2
p
= .002], though the interaction between argument type and
norm
was significant [F(1, 401) = 7.18, p = .008, �2
p
= .018].
To explore directly the extent to which each condition elicited
conformity, we conducted
a 2 × 2 ANOVA using the conformity index. There was a main
effect of argument type [F(1,
394) = 7.18, p = .008, �2
p
= .018], such that the conformity index was significantly
greater for
rational arguments (M = .86, SD = 2.97) than for emotional
arguments (M = .08, SD = 2.81).
Here, acceptable judgments (M = .65, SD = 2.92) were no more
prone to conformity than
condemnable ones (M = .29, SD = 2.91) [F(1, 394) = 1.60,
p = .21, �2

p
= .004]. Again, there
was no significant interaction between statistical norm and
argument type [F(1, 394) = .92,
p = .34, �2
p
= .002].
Discussion
When presented with either rational or emotional justifications
for moral judgments, par-
ticipants conformed more to the rational justifications. These
results are inconsistent with
our second hypothesis and with predictions made more broadly
by the SIM (Haidt, 2001),
because our participants responded more to appeals citing
reasons than to appeals citing
emotions. This is unexpected given the body of literature
demonstrating that manipulations
of emotion are powerful tools in shaping judgment (Valdesolo
& DeSteno, 2006; Feinberg
SOCIAL INFLUENCE 65
et al., 2012; Paxton & Greene, 2010). Furthermore, the SIM
suggests that moral judgments
can only be affected by changing moral intuitions, though the
model may be consistent
with these findings, since post hoc reasoning of one person, via
the ‘reasoned persuasion’
link in the model, may still impact the judgments of others. The

reasoned persuasion link,
however, remains largely unspecified, and it makes no
predictions or claims about how
that persuasion works, nor what kinds of persuasion should be
most effective. We discuss
potential explanations for our findings in the following section.
General discussion
In this paper we have shown that participants readily conformed
to subtle statistical manip-
ulations of their moral judgments. Furthermore, we have
provided some evidence that
arguments appealing directly to participants’ emotions did not
induce conformity as strongly
as rational appeals.
In the literature on conformity, some studies have drawn a
distinction between nor-
mative social motivations to conform, which are characterized
by a desire to avoid social
isolation, and informational motivations, which are based on a
need to be correct (Deutsch
& Gerard, 1955). Several features of our experiments suggest
that the nature of conformity
in this context may be due to informational rather than social
factors. First, the context of
our experiments is much less personal than in other studies,
which include face-to-face
social interaction. Given the lack of social interaction and the
lack of possibility for social
feedback, the likelihood that participants are responding to
direct social pressure seems low.
Further, a previous study has shown that participants rely more
heavily epistemologically
on their peers when the answer to a question is more ambiguous

and open to interpreta-
tion (Stangor et al., 2001). The nature of moral judgment can be
quite ambiguous, and the
stimuli in this experiment were designed to evoke competing
intuitions. Therefore, our
participants seem to be interpreting the number of supporters as
evidence for the correct
judgment about a very difficult moral question.
Additionally, contrary to the SIM and other literature on
emotional manipulation, our
emotional primes were not as successful in inducing conformity
as their rational counter-
parts. These results do accord well, however, with recent critics
of the SIM, such as those
questioning the link between disgust and moral judgment (e.g.,
Landy & Goodwin, 2015;
Johnson et al., 2016). Our results also fit into a burgeoning
literature exploring the role of
reasoning in moral judgment. Moral reasoning, this research
suggests, can set the bound-
aries of what we consider moral (Royzman, Landy, & Goodwin,
2014), aid in discounting
intuitions with no justifications, and correct for bias (see Paxton
& Greene, 2010, for a
review). Furthermore, controlling for demographic factors, the
willingness to engage in
rational thinking predicts wrongness judgments of purity
violations like Scenario A of our
study (Pennycook, Cheyne, Barr, Koehler, & Fugelsang, 2014).
Supporters of SIM may argue that perhaps these primes failed to
make participants feel
any emotions, or perhaps participants counterreacted to what
they saw as excessive expres-
sions of emotion. Even if that were the case, the arguments used

were real responses given
by participants and represent ecologically valid instances of
emotional persuasion in many
online settings, where the expression of emotion is done
through written words rather than
the ‘emotional’ stimuli explored in other studies (e.g.,
Valdesolo & DeSteno, 2006). Given
the limitations of the expression of emotion through online
media, our data suggest that
66 M. KELLY ET AL.
the more effective tactic for persuasion regarding moral
judgments, whether on the smaller
scale between individuals or the larger scale of public opinion,
may be rational appeals to
abstract principles rather than expressions of emotions. It is
worth noting, however, that
our stimuli hardly capture the full breadth of emotional and
rational arguments available.
Future work might explore whether this pattern holds more
broadly, or only for the stimuli
in the present study.
Today, in contrast with Asch’s time, more of our social
interactions and, consequently,
discussions on matters of morality and politics are conducted
across digital screens rather
than face-to-face. Though it is reasonable to predict that the
influence we have on each oth-
er’s opinions would be greatly diminished in this detached
world, it appears that the power
of social influence is retained. The exact consequences of an
increasingly interconnected

virtual web of people, ideas, and opinions remain to be seen.
Future research may eluci-
date whether the robustness of conformity online will lead to
good or bad consequences,
whether it be through the facilitation of advances in knowledge
as with ‘The Wisdom of
Crowds’ effect (Golub & Jackson, 2010) or an amplification of
erroneous noise through a
‘Groupthink’ phenomenon (Esser, 1998).
Acknowledgments
We thank Phil Costanzo for his helpful feedback.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Allen, V. L. (1966). Situational factors in conformity. Advances
in Experimental Social Psychology,
2, 133–175.
Aramovich, N. P., Lytle, B. L., & Skitka, L. J. (2012). Opposing
torture: Moral conviction and resistance
to majority influence. Social Influence, 7, 21–34.
Aristotle. (1941). Ethica Nicomachea. In R. McKeon (Ed.), The
basic works of Aristotle (pp. 935–1126).
New York: Random House.
Asch, S. E. (1956). Studies of independence and conformity: A
minority of one against a unanimous
majority. Psychological Monographs: General and Applied,
70(9), 1–70.

Avramova, Y. R., & Inbar, Y. (2013). Emotion and moral
judgment. Wiley Interdisciplinary Reviews:
Cognitive Science, 4, 169–178. doi:10.1002/wcs.1216
Bargh, J. A., & McKenna, K. Y. (2004). The internet and social
life. Annual Review of Psychology, 55,
573–590.
Bond, R., & Smith, P. B. (1996). Culture and conformity: A
meta-analysis of studies using Asch’s
(1952b, 1956) line judgment task. Psychological Bulletin, 119,
111.
Bostyn, D. H., & Roets, A. (2016). An asymmetric moral
conformity effect subjects conform to
deontological but not consequentialist majorities. Social
Psychological and Personality Science.
doi:10.1177/1948550616671999
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s
mechanical turk a new source of
inexpensive, yet high-quality, data? Perspectives on
Psychological Science, 6, 3–5.
Deutsch, M., & Gerard, H. (1955). A study of normative and
informational influences upon individual
judgment. The Journal of Abnormal and Social Psychology, 51,
629–636.
Eastwick, P. W., & Gardner, W. L. (2009). Is it a game?
Evidence for social influence in the virtual
world. Social Influence, 4, 18–32.
https://doi.org/10.1002/wcs.1216
https://doi.org/10.1177/1948550616671999

SOCIAL INFLUENCE 67
Esser, J. K. (1998). Alive and well after 25 years: A review of
groupthink research. Organizational
Behavior and Human Decision Processes, 73, 116–141.
Feinberg, M., Willer, R., Antonenko, O., & John, O. P. (2012).
Liberating reason from the passions:
Overriding intuitionist moral judgments through emotion
reappraisal. Psychological Science, 23,
788–795.
Francis, K. B., Howard, C., Howard, I. S., Gummerum, M.,
Ganis, G., Anderson, G., & Terbeck, S.
(2016). Virtual morality: Transitioning from moral judgment to
moral action? PLoS One, 11,
e0164374. doi:10.1371/journal.pone.0164374
Goldberg, J. H., Lerner, J. S., & Tetlock, P. E. (1999). Rage and
reason: The psychology of the intuitive
prosecutor. European Journal of Social Psychology, 29, 781–
795.
Golub, B., & Jackson, M. O. (2010). Naive learning in social
networks and the wisdom of crowds.
American Economic Journal: Microeconomics, 2, 112–149.
Haidt, J. (2001). The emotional dog and its rational tail: A
social intuitionist approach to moral
judgment. Psychological Review, 108, 814.
Haidt, J., & Graham, J. (2007). When morality opposes justice:
Conservatives have moral intuitions
that liberals may not recognize. Social Justice Research, 20, 98–

116.
Haidt, J., Koller, S. H., & Dias, M. G. (1993). Affect, culture,
and morality, or is it wrong to eat your
dog? Journal of Personality and Social Psychology, 65, 613.
Hornsey, M. K., Majkut, L., Terry, D. J., & McKimmie, B. M.
(2003). On being loud and proud:
Non-conformity and counter-conformity to group norms. British
Journal of Social Psychology,
42, 319–335.
Johnson, D. J., Wortman, J., Cheung, F., Hein, M., Lucas, R. E.,
Donnellan, M. B., … Narr, R. K.
(2016). The effects of disgust on moral judgments testing
moderators. Social Psychological and
Personality Science, 7, 640–647.
Kant, I. (1996). Kant: The metaphysics of morals. (M. J. Gregor
Ed.). Cambridge: Cambridge University
Press.
Kundu, P., & Cummins, D. D. (2012). Morality and conformity:
The Asch paradigm applied to moral
decisions. Social Influence, 8, 268–279.
doi:10.1080/15534510.2012.727767
Landy, J. F., & Goodwin, G. P. (2015). Does incidental disgust
amplify moral judgment? A meta-
analytic review of experimental evidence. Perspectives on
Levy, L. (1960). Studies in conformity. The Journal of
Psychology, 50, 39–41.
Lisciandra, C., Postma-Nilsenová, M., & Colombo, M. (2013).
Conformorality. A study on group

conditioning of normative judgment. Review of Philosophy and
Psychology, 4, 751–764.
Ngo, L., Kelly, M., Coutlee, C. G., Carter, R. M., Sinnott-
Armstrong, W., & Huettel, S. A. (2015). Two
distinct moral mechanisms for ascribing and denying
intentionality. Scientific Reports, 5.
Patil, I., Cogoni, C., Zangrando, N., Chittaro, L., & Silani, G.
(2014). Affective basis of judgment-
behavior discrepancy in virtual experiences of moral dilemmas.
Social Neuroscience, 9,
94–107.
Paxton, J. M., & Greene, J. D. (2010). Moral reasoning: Hints
and allegations. Topics in Cognitive
Science, 2, 511–527.
Pennycook, G., Cheyne, J. A., Barr, N., Koehler, D. J., &
Fugelsang, J. A. (2014). The role of analytic
thinking in moral judgements and values. Thinking &
Reasoning, 20, 188–214.
Puhl, R. M., Schwartz, M. B., & Brownell, K. D. (2005). Impact
of perceived consensus on stereotypes
about obese people: A new approach for reducing bias. Health
Psychology, 24, 517.
Rand, D. G. (2012). The promise of mechanical turk: How
online labor markets can help theorists
run behavioral experiments. Journal of Theoretical Biology,
299, 172–179.
Royzman, E. B., Landy, J. F., & Goodwin, G. P. (2014). Are
good reasoners more incest-friendly? Trait

cognitive reflection predicts selective moralization in a sample
of American adults. Judgment and
Decision Making, 9, 175.
Stangor, C., Sechrist, G. B., & Jost, J. T. (2001). Changing
racial beliefs by providing consensus
information. Personality and Social Psychology Bulletin, 27,
486–496.
Teper, R., Zhong, C., & Inzlicht, M. (2015). How emotions
shape moral behavior: Some answers
(and questions) for the field of moral psychology. Social and
Personality Psychology Compass, 9(1),
1–14. doi:10.1111/spc3.12154
https://doi.org/10.1371/journal.pone.0164374
https://doi.org/10.1080/15534510.2012.727767
https://doi.org/10.1111/spc3.12154
68 M. KELLY ET AL.
Valdesolo, P., & DeSteno, D. (2006). Manipulations of
emotional context shape moral judgment.
Young, L., & Saxe, R. (2011). When ignorance is no excuse:
Different roles for intent across moral
domains. Cognition, 120, 202–214.
AbstractIntroductionStudy 1: impersonal statistics influence
moral
judgmentsMethodParticipantsMaterialsProcedureResultsDiscuss
ionStudy 2: rational arguments elicit more conformity than
emotional
argumentsMethodParticipantsProcedureResultsDiscussionGener
al discussionAcknowledgmentsDisclosure statementReferences

Contents lists available at ScienceDirect
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
Full length article
Deindividuation effects on normative and informational social
influence
within computer-mediated-communication
Serena Coppolino Perfumia,b,∗ , Franco Bagnolic, Corrado
Caudekd, Andrea Guazzinie
a Department of Sociology, Stockholm University, S-106 91,
Stockholm, Sweden
b Department of Educational Sciences and Psychology,
University of Florence, 50135, Florence, Italy
c Department of Physics and Astronomy and Center for the
Study of Complex Dynamics (CSDC), University of Florence,
50019 Sesto Fiorentino, also INFN sec, Florence,
Italy
d Department of Neuroscience, Psychology, Drug Research and
Children's Health (NEUROFARBA) – sect. Psychology,
University of Florence, 50135, Florence, Italy
e Department of Educational Sciences and Psychology and
Center for the Study of Complex Dynamics (CSDC), University
of Florence, 50135, Florence, Italy
A R T I C L E I N F O

Keywords:
Social influence
Conformity
Computer-mediated-communication
Anonymity
Deindividuation
A B S T R A C T
Research on social influence shows that different patterns take
place when this phenomenon happens within
computer-mediated-communication (CMC), if compared to face-
to-face interaction. Informational social influ-
ence can still easily take place also by means of CMC, however
normative influence seems to be more affected by
the environmental characteristics. Different authors have
theorized that deindividuation nullifies the effects of
normative influence, but the Social Identity Model of
Deindividuation Effects theorizes that users will conform
even when deindividuated, but only if social identity is made
salient.
The two typologies of social influence have never been studied
in comparison, therefore in our work, we
decided to create an online experiment to observe how the same
variables affect them, and in particular how
deindividuation works in both cases. The 181 experimental
subjects that took part, performed 3 tasks: one
aiming to elicit normative influence, and two semantic tasks
created to test informational influence. Entropy has
been used as a mathematical assessment of information
availability.
Our results show that normative influence becomes almost
ineffective within CMC (1.4% of conformity) when
subjects are deindividuated.

Informational influence is generally more effective than
normative influence within CMC (15–29% of con-
formity), but similarly to normative influence, it is inhibited by
deindividuation.
1. Introduction
With the diffusion of social networking platforms, the social
and
information seeking-related human behaviors have been affected
by the
“new” environment. Information seeking increasingly takes
place on
social media platforms, relying on what a users' contacts and
followed
pages share (Zubiaga, Liakata, Procter, Hoi, & Tolmie, 2016).
Because of this filtering and selection, the users' knowledge-
building
process could be severely biased and polarized.
For example, a study shows that 72% of participants (college
stu-
dents) trusted links sent by friends, even if they contained
phishing
attempts (Jagatic, Johnson, Jakobsson, & Menczer, 2007).
The recent debate on fake news, highlighted the potential link
be-
tween the increase in their spread, and the structure of social
networks
as well as their embedded algorithms, which turned these
environments
into “echo chambers”, in which users are selectively exposed to

information, and tend to filter the information in order to
reinforce
their positions (confirmation bias), rather than to find
alternatives (Del
Vicario et al., 2016).
These factors highlight the importance of studying the effects of
social influence within computer-mediated-communication, in
order to
understand which environmental factors can enhance its effects.
Social norms exist also in online environments, but the users'
per-
ception of them can be different according to the platform, to
anon-
ymity and the social ties among contacts. Therefore, compliance
to
social norms can emerge in different ways, than those
observable in
face-to-face interaction.
Also, information-seeking behavior can be affected by online
en-
vironments: on one side we observe its interrelation with social
norms,
especially when it takes place on social media platforms, and
users
gather information on the basis of what they read on their
personal
newsfeed. However, we also observe how users can rely on
opinions
https://doi.org/10.1016/j.chb.2018.11.017
Received 29 March 2018; Received in revised form 9 October
2018; Accepted 7 November 2018

∗ Corresponding author. Department of Sociology, Stockholm
University, S-106 91, Stockholm, Sweden.
E-mail address: [email protected] (S. Coppolino Perfumi).
Computers in Human Behavior 92 (2019) 230–237
Available online 13 November 2018
0747-5632/ © 2018 Elsevier Ltd. All rights reserved.
T
http://www.sciencedirect.com/science/journal/07475632
https://www.elsevier.com/locate/comphumbeh
mailto:[email protected]
http://crossmark.crossref.org/dialog/?doi=10.1016/j.chb.2018.11
.017&domain=pdf
expressed by unknown actors, as it happens on platforms like
TripAdvisor.
The present study, using online experiments, aims to separate
norms-oriented social influence from information-oriented
social in-
fluence, in order to observe which elements and environmental
factors
have an effect on both typologies and which are peculiar for
each.
1.1. Theoretical framework
A major understanding on the functioning of social influence
came

about thanks to the pioneering works of Sherif (1937) and then
Asch
(1951, 1955, 1956). The authors studied how the physical
presence of
other people can lead experimental subjects to conform their
judgment
to the one of the others. They used two different types of tasks:
while in
Asch conformity experiments, guessing the correct answer could
be
straightforward (Asch, 1955, 1956; Asch & Guetzkow, 1951),
Sherif
used the autokinetic effect, so a more ambiguous task, to test
the effects
of social influence (Sherif, 1937). From these experiments, two
typol-
ogies of social influence have been identified, called
“normative” when
people conform in order to satisfy a need to belong and comply
to social
norms, as observed in Asch's experiments, and “informational”
when
the subjects lack on information in order to perform a task, as
observed
in the autokinetic experiment (Deutsch & Gerard, 1955).
According to
this theorization proposed by Deutsch and Gerard (1955), we
can say
that we are able to observe normative social influence in Asch's
con-
formity experiments, because the task is relatively easy and the
sub-
jects, when interviewed after taking part to the experiment
stated that
they were able to spot the correct answer, but conform in order
not to

break the social norms and be group outsiders. Instead, given
that the
task presented in the autokinetic experiment is more ambiguous,
as it is
based on a visual illusion, in this case we can say that subjects
conform
because they are unsure on how to proceed.
While, as observed in these classical studies, to elicit
conformity in
face-to-face situations, the physical presence of other people
and being
exposed to their judgment can be enough, things go differently
when
people interact online, especially for normative social influence.
Indeed, it is still unclear which elements can have the power to
lead
people to conform during computer-mediated-communication.
Deindividuation, namely the diminished perception of one's per-
sonal traits (Zimbardo, 1969), has been identified as a potential
key
element in the discourse on normative influence.
The original deindividuation model was proposed by Zimbardo
in
1969, and the author identified a series of variables that
according to
him can lead to a deindividuation state. The variables
considered by
Zimbardo are for example anonymity, arousal, sensory overload,
novel
or unstructured situations, involvement in the act, and the use of
al-
tering substances (Zimbardo, 1969). Several other authors

suggest that
if people interact while being in a deindividuation state,
normative
social influence can disappear (Deutsch & Gerard, 1955;
Latané, 1981;
Lott & Lott, 1965; Short, Williams, & Christie, 1976). This
happens
because there is not the possibility to identify the interlocutors,
due to a
lack of actual or perceived proximity, and consequently,
deindividua-
tion should lighten the pressure to act according to social norms
(Latané, 1981).
Furthermore, a study which tested antinormative behavior by
counterposing deindividuation to the presence of an explicit
aggressive
social norm, showed that subjects were actually more aggressive
when
deindividuated, rather than when exposed to the explicit norm,
so in
this case, deindividuation resulted to be more powerful in
leading to
antinormative behavior (Mann, Newton, & Innes, 1982).
A significant advancement in explaining the functioning of
norma-
tive social influence in online environments is represented by
the
contribution provided by the Social Identity Model of
Deindividuation
Effects (SIDE Model), that takes the concept of deindividuation
and
expands it, explaining its link and implications on social
influence in
online environments (Spears, Postmes, Lea, & Wolbert, 2002).

The authors theorize that deindividuation is indeed likely to
occur
in online environments, but it can become a powerful tool to
trigger
conformity: given that while deindividuated, subjects have a
dimin-
ished perception of their personal traits, if the group the
subjects are
interacting with is made salient, then the subjects will be more
likely to
conform (Spears, Postmes, & Lea, 2018).
This happens because combining a lack of relevance of one's
per-
sonality with an enhancement of the importance of the
interlocutors,
will lead the subjects to identify at the group level, and
consequently to
comply to the social norms. The experimental results seem to
confirm
the predictions presented by the SIDE Model (Lee, 2004;
Postmes,
Spears, Sakhel, & De Groot, 2001), but it is not clear what
happens
when users are deindividuated but the group saliency is not
enhanced.
On the matter of informational influence during computer-medi-
ated-communication instead, studies have focused on different
aspects.
As aforementioned, a visible example of informational influence
in
online environments is represented by users making choices on
the

basis of reviews or ratings provided by other unknown users
while
using platforms such as Tripadvisor, Uber or Airbnb (Liu &
Zhang,
2010), but other examples show that it can take place easily also
in
other ways.
A study conducted by Rosander and Eriksson (2012), shows that
users facing a general knowledge quiz in which they were
exposed to
histograms showing the distribution of the answers provided by
other
unknown users, conformed in high percentages (52%).
While many studies on online consumers behavior focused on
fac-
tors such as the perceived importance of feedback (Liu &
Zhang, 2010)
on informational influence, or on the conjunct effect of
informational
and normative influence on behavior when subjects interact
without
personal contact (LaTour & Manrai, 1989), no study tried to
isolate it,
and point out the environmental factors that could be able to
enhance
or diminish the compliance of users in this case. Furthermore,
no study
tested the effects of deindividuation on informational influence.
In order to test and fulfill the predictions developed based on
the
literature, we developed an experimental framework aiming to
study
separately the two typologies of social influence during

computer-
mediated-communication.
On one side, we reduced group saliency to test how
deindividuation
works on both typologies of social influence and controlled the
possible
interactions between some psychological dimensions and the
operative
variables.
On the other side, we calculated the items entropy to test if task
ambiguity increases informational-based compliance. The
environ-
mental factors that we decided to manipulate and study in
relation to
both typologies of social influence are anonymity and physical
isola-
tion, as their combination can trigger deindividuation.
1.2. Overview and predictions
To test online normative influence, we replicated Asch's
conformity
experiment (Asch, 1955, 1956; Asch & Guetzkow, 1951) on a
web-
based platform, while to test online informational influence we
created
two linguistic tasks of increasing ambiguity, designed adopting
the
same structure of the “classical” Asch's items. Task ambiguity
was
measured by calculating the items' entropy, and in this way, we
were
able to assess the subjects' lack of information. The diversity of
the

tasks, allowed us to measure the interaction between anonymity,
phy-
sical isolation, and degree of ambiguity, in relation to the
behavior of
the experimental subjects. Considering the literature, we could
for-
mulate the following predictions:
• H1) Diminished effectiveness of normative influence due to
the
combination of a deindividuation state given by anonymity and
physical isolation, and minimum levels of group saliency, as
theo-
rized by several authors (Deutsch & Gerard, 1955; Latané,
1981;
Lott & Lott, 1965; Short et al., 1976) and hypothesized by the
SIDE
S. Coppolino Perfumi et al. Computers in Human Behavior 92
(2019) 230–237
231
Model (Postmes et al., 2001).
• H2) There is no specific evidence to build on, on the potential
re-
lationship between deindividuation and informational influence
(if
separated by normative influence), but we expect it to have the
same inhibitory effect it has on normative influence (Lee,
2007). The
effect of the anonymity and physical isolation variables alone
will

To Switch or Not To Switch Understanding Social Influence i.docx

To Switch or Not To Switch Understanding Social Influence i.docx

Recommended

Recommended

More Related Content

Similar to To Switch or Not To Switch Understanding Social Influence i.docx

Similar to To Switch or Not To Switch Understanding Social Influence i.docx (20)

More from juliennehar

More from juliennehar (20)

Recently uploaded

Recently uploaded (20)

To Switch or Not To Switch Understanding Social Influence i.docx