Psychology of Popular Media Culture © 2013 American Psychological Association

2013, Vol. 2, No. 1, 51– 60 2160-4134/13/$12.00 DOI: 10.1037/a0030592

Social Influence Online: The Impact of Social Validation and

Likability on Compliance

Rosanna E. Guadagno, Nicole L. Muscanell, Lindsay M. Rice, and Nicole Roberts

University of Alabama

Text-based communication via the Internet has provided new opportunities to study
social influence and persuasion. Specifically, Guadagno and Cialdini (2005) con-
tend that the effectiveness of social influence attempts have yet to be thoroughly
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

investigated online. To test Guadagno and Cialdini’s contention, the present study
examined whether the social influence principles of likability and social validation
This document is copyrighted by the American Psychological Association or one of its allied publishers.

impacted individuals’ willingness to comply with a request when the setting is

online. Results revealed that social validation affected compliance, but communi-
cator likability did not. Thus, our results indicate that contrary to previous work in
offline contexts, not all social influence principles are effective online. Explana-
tions for these differences are discussed.

Keywords: online persuasion, social influence, social validation, likability, Internet

Interpersonal interactions in contemporary so- are effective when using the aforementioned
ciety reflect the increasing use of technology, par- types of CMC, whereas some (e.g., authority)
ticularly text-based, stand alone, computer- are not. Furthermore, they found that others
mediated communication (CMC) such as e-mail, (e.g., liking) are effective in some online con-
texting, blogging, and instant messaging. texts but not others. The present investigation
Among these interactions, the aforementioned builds on previous literature to examine the
types of CMC have become a modality for effectiveness of two different social influence
social influence attempts. Thus, the effective- principles: likability and social validation on
ness of social influence—the scientific study of compliance— behavior change without pres-
attitude and behavior change due to real or sure—and whether these compliance tactics ex-
imagined pressure (Cialdini, 2009)—in comput- tend to an online context such as the types of
er-mediated environments has been of interest CMC mentioned earlier in the text.
to research and influence practitioners alike.
One such area of interest pertains to the effec- Online Social Interaction1
tiveness of Cialdini’s six principles of influ-
ence—authority, reciprocity, scarcity, social McKenna and Bargh (2000) have long con-
validation, likability, and commitment and con- tended that there are four aspects of online
sistency— online, particularly in text-based in- social interaction that make it unique: time and
teractions in which the communicator is distant place, anonymity, physical appearance, and
from the target of influence. Guadagno and physical distance. For instance, people are able
Cialdini (2005) conducted a literature review to control when, where, how, and with whom
and concluded that, thus far, some influence they interact with online. As a result, individu-
principles (e.g., commitment and consistency) als exert greater control over their interactions.
Furthermore, unlike interactions through other
media, instantaneous responses are not norma-
tive in CMC. With CMC, individuals may for-
Rosanna E. Guadagno, Nicole L. Muscanell, Lindsay M. mulate and edit responses over any given
Rice, and Nicole Roberts, Department of Psychology, Uni-
versity of Alabama.
Correspondence concerning this article should be ad-
1
dressed to Rosanna E. Guadagno, Ph.D., National Science Because work on Facebook and other social networking
Foundation, 4201 Wilson Boulevard, Suite 995N, Arling- sites is still emerging, we limit our literature review to CMC
ton, VA 22230. E-mail: [email protected] in the presocial networking era.

51
 52 GUADAGNO, MUSCANELL, RICE, AND ROBERTS

amount of time so that the target individual world without having to leave the comfort of
receives an articulate communication. With re- their home.
gard to anonymity, communicating with some- Given that more and more people are com-
one via CMC may create a buffer against actual municating through CMC, it is also likely that
knowledge of a person’s identity. As a result, an the normative expectations for online interac-
individual can choose to present him-/herself tions are unique, and therefore warrant further
genuinely or alter and/or conceal certain aspects investigation. To some extent, these norms are
of their identity. For instance, CMC allows for still being established (Kiesler, Siegel, &
the creation of screen names and avatars that McGuire, 1984; Bagozzi, Dholakia, & Pearo,
can provide a mask for the user because iden- 2007). Furthermore, online communication may
tifiers such as gender, age, and identity can be lack nonverbal feedback, and this deficiency
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

hidden or changed. Because of this greater abil- could make it more difficult to interpret mes-
ity to control visual anonymity online, it has sages communicated via CMC (Derks, Bos, &
This document is copyrighted by the American Psychological Association or one of its allied publishers.

been proposed as one of several factors that can von Grumbkow, 2008). Also, social influence
lead to engagement in both more normative and may be less affected by social hierarchy or
antinormative behaviors (McKenna & Bargh, attractiveness because such information can be
2000; Postmes, Spears, Sakhel, & De Groot, hidden or is not as apparent online (Edinger &
2001; Sassenberg & Postmes, 2002). Specifi- Patterson, 1983). Combined, these factors may
cally, anonymity can lead to a deindividuated or make certain types of CMC less personal than
disinhibited state (state of lowered self- other forms of communication because individ-
awareness), leading to neglect of one’s usual uals are directed toward the message content
personal standards for behavior, and ultimately rather than the speaker. Compared with more
increased antinormative behavior (Mendels, traditional forms of communication, social
1999). In contrast, anonymity can also lead to norms and social standards may be less salient
more focus on group identity and relevant group among individuals communicating online
norms, increasing group conformity (Postmes et (Chaiken & Eagly, 1983; Kiesler, Siegal, &
al., 2001). For instance, it has been found that McGuire, 1984; Guadagno & Cialdini, 2002,
individuals who are identifiable may conform 2005, 2007). Taken together, these factors sug-
more to group norms compared with anony- gest that individuals process information
mous individuals via CMC if there are clear uniquely when it is presented via CMC.
sanctions for breaking the norm (Sassenberg &
Postmes, 2002). Social Influence Online
Another unique aspect of online communica-
tion is the minimized importance of physical Given the evidence presented earlier in the
appearance, something that is traditionally a text pertaining to the unique characteristics of
powerful determinant of initial attraction online communication, it seems likely that in-
(Bargh & McKenna, 2004; Belmore, 1987). As fluence attempts may or may not be effective in
a result, individuals may interact with others online settings. However, there is a dearth of
online while being unconcerned about their ap- research examining this question. Work by
pearance. Additionally, status cues are often Cialdini (2009) indicates that there are six uni-
missing online, and therefore, there is no visible versal principles of social influence: authority,
social hierarchy (McKenna & Bargh, 2000). reciprocity, scarcity, social validation (also
This may produce more equal participation called social proof), likability, and commitment
among members of an online group, regardless and consistency. These principles serve as heu-
of their status. Thus, individuals are more likely ristic cues for decision making. When process-
to form impressions of people by the words ing heuristically, individuals can use certain
written and not what they can see about a per- cues, rules of thumb, or surface features to
son. Finally, the fourth aspect of online com- determine whether to comply with a request.
munication, physical distance, illustrates that For instance, the personal characteristics of a
the geographical distance between individuals communicator (e.g., attractiveness, expertise,
has less influence on their choice of communi- likability) are factors that influence the extent to
cation partners or groups. Individuals can easily which individuals targeted for an influence at-
meet and interact with people anywhere in the tempt are swayed by the individual attempting
 SOCIAL INFLUENCE ONLINE 53

to influence them (i.e., the influence agent). It is when liking was manipulated, women, but not
noteworthy that these are some of the same men, were less swayed by a confederate through
features that are less salient in text-based CMC. an e-mail interaction. Furthermore, this effect
Guadagno and Cialdini (2005) reviewed the was stronger when the confederate was also
literature on social influence online, and they unlikable. Because the latter study examined
focused on two domains of social influence: persuasion rather than compliance as the out-
compliance (i.e., behavior change without pres- come measure, it may be that liking in online
sure)2 and persuasion (i.e., attitude change with- contexts is affected by the type of social influ-
out pressure). The purpose of this literature re- ence under examination. However, given the
view was to investigate the extent to which limited number of studies available, the effect
compliance and persuasion had been examined of liking on social influence online is still an
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

in online environments. The review of the on- open question.

line compliance literature was framed by Cial- Overall, considering that only three of the six
This document is copyrighted by the American Psychological Association or one of its allied publishers.

dini’s (2009) six principles of influence. Guad- principles have been examined online, and of
agno and Cialdini found that the effectiveness those three, one produces contradictory evi-
of some, but not all, of these six principles of dence, more research should be conducted to
influence had been examined in an online con- further examine how the principles of influence
text. The researchers hypothesized that compli- affect compliance in an online context. The
ance tactics may be more or less effective online present study focuses on two of the Cialdini’s
because these social influence attempts create a (2009) six principles: likability and social vali-
context in which the influence agent is more dation, and examines how these principles in-
distant than he or she would be in attempts fluence compliance rates when the request oc-
delivered in other contexts (Dubrovsky, Kiesler, curs online.
& Sethna, 1991; Guéguen & Jacob, 2002).
After reviewing the literature, Guadagno and Liking
Cialdini (2005) concluded that the effectiveness
of three of the six principles of influence— Research reviewed by Cialdini (2009) indi-
authority, commitment and consistency, and cates that likable people are more influential.
liking— on compliance had been examined Factors that produce liking include physical at-
solely in an online context (i.e., in the absence tractiveness and/or similarity. Both of these fac-
of a face-to-face condition). Specifically, re- tors are powerful determinants of initial liking
search indicates that authority cues are largely but may not be as important with online inter-
ignored in online interactions, whereas commit- actions owing to the lack of visual and social
ment and consistency tactics such as the foot- cues. Compliments are also effective at increas-
in-the-door technique are effective online ing liking because flattery typically increases
(Dubrovsky, Kiesler, & Sentha, 1991; Guéguen liking toward the flatterer. Other factors that can
& Jacob, 2001; Markey, Wells, & Markey, increase liking are familiarity and association.
2001; Guégen, 2002; Petrova, Cialdini, & Sills, As indicated earlier in the text, Guadagno and
2007). Cialdini (2002) showed that liking was an im-
Research on the effect of liking on compli- portant factor in the persuasiveness of an argu-
ance in online contexts is less prevalent. Fur- ment for women but not for men. The research-
thermore, the results of the limited existing lit- ers explained this finding in terms of social role
erature are conflicting. For instance, Guéguen, theory (Eagly, 1987; Eagly, Wood, & Diekman,
Jacob, and Morineau (2010) demonstrated that 2000) and argued that text-based CMC interfere
if an e-mail solicitation is sent by someone with with the ability of women to form a bond with
the same name as the recipient, the recipient is the influence agent. Guadagno and Cialdini
more likely to comply with the request. Thus, (2007) replicated this finding and also demon-
this study demonstrated that liking, operation-
alized as similarity, increases compliance to a 2
request. Contrary to these results, Guadagno and Please note that compliance is conceptually different
from conformity because the former occurs willingly and
Cialdini (2002, 2007) demonstrated that liking an the latter occurs under conditions of social pressure. Please
influence agent was not always an effective means see Cialdini and Goldstein (2004) for a review that thor-
of influencing a target via e-mail. Specifically, oughly differentiates the two.
 54 GUADAGNO, MUSCANELL, RICE, AND ROBERTS

strated that cooperation or similarity eliminated our study that individuals will comply more
this effect for women. Furthermore, their work when social validation suggests that many oth-
demonstrated that text-based CMC actually fa- ers are compliant.
cilitated influence between men who had re-
cently competed with each other or were told
they were highly dissimilar. The Present Study

The purpose of the current study was two-

Social Validation
fold. First, we wanted to examine whether two
Cialdini’s (2009) principle of social valida- specific influence principles, likability and so-
tion concerns social norms and the idea that cial validation, would be effective in an online
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

people often look to the behavior of others to setting. Second, we sought to determine
decide how to behave across situations, espe- whether there would be evidence of an additive
This document is copyrighted by the American Psychological Association or one of its allied publishers.

cially ambiguous contexts. Specifically, this effect of both likability and social validation on
principle is based on social evidence (descrip- compliance in a computer-mediated context.
tive norms) indicating that individuals consider McKenna and Bargh’s (2000) framework of
an action more appropriate when they see others online behavior and the work of Guadagno and
reacting similarly to the situation. For example, colleagues (Guadagno & Cialdini, 2002, 2005,
Asch’s (1951) conformity studies demonstrate 2007; Okdie & Guadagno, 2008) suggest that
that social validation of how others are behav- the influence of a likable communicator may not
ing (many others publicly pick an incorrect an- be strong online (especially among women) ow-
swer) can then lead to others to conform to the ing to the decreased salience of the influence
same behavior (an individual picks the incorrect agent. This work and the work by Postmes and
answer, even knowing it is incorrect). Postmes, colleagues (e.g., Postmes et al., 1998) suggest
Spears, and Lea (1998) studied the effect of that the influence of social norms will likely be
anonymity on adherence to social norms and effective online if participants feel anonymous.
named the resulting processes the Social Iden- Furthermore, research indicates that when com-
tity Model of Deinviduation Effects (Lea & paring audible or visually salient communica-
Spears, 1991). One of the key tenets of the tions with a written communication, the audio
Social Identity Model of Deinviduation Effects or visual communication was more influenced
is that the depersonalization that is seen with by the characteristics of the communicator
CMC can cause users to be more sensitive to (Chaiken & Eagly, 1983). Thus, likability may
norms conveyed by salient groups. Under con- not be a particularly powerful compliance tactic
ditions of anonymity, individuals look more to- online because the communicator is less salient,
ward a group for normative direction rather than and other message features may become more
following their internal standards for behavior. important.
Postmes et al. (2001) demonstrated the afore- Owing to fewer visual and aural cues, social
mentioned phenomenon pertaining to the con- validation may be particularly influential online.
ditions under which individuals adhere to group With increased anonymity online, deindividua-
norms in an online environment. Specifically, tion may occur. Deindividuated participants are
they reported that when anonymous, group not typically swayed by their internal standards
members displayed behavior consistent with the (Matheson & Zanna, 1989). As discussed ear-
salient norm (efficiency or prosocial behavior). lier, Postmes et al. (1998) found that when
Group members interacting nonanonymously deindividuation occurs online, individuals tend
did not adhere to group norms. Their second to rely less on their individual standards of
study replicated and expanded this finding by behavior and instead identify more with group
illustrating that, within the group, the norms norms. This suggests that social validation will
were also socially transmitted. This suggests be effective online because by identifying more
that when individuals are interacting anony- with standards of their group, individuals are
mously, as they often do online, they will be looking toward others for validation of their
influenced by the group norms that are most own actions. Thus, our first research question
salient. Because there will be no clearly speci- was: Does social validation information in-
fied sanctions for noncompliance, we predict in crease compliance with a request?
 SOCIAL INFLUENCE ONLINE 55

Based on the literature reviewed earlier in the a majority opinion or group. Thus, there were
text, we sought to examine the influence of three levels of communicator likability (likable,
social validation and likability on compliance unlikable, and no likability manipulation con-
when requests to volunteer were presented trol). Social validation was manipulated by the
through online communication. Owing to the presence of comments from other fictitious stu-
rationale described earlier in the text, we ex- dents in response to the blog author’s request.
pected to find that social validation would be There were either several students offering or
influential in an online context. Conversely, ow- those students refusing (based on social valida-
ing to the mixed results and McKenna and tion condition) to volunteer to help in their
Bargh’s (2000) work demonstrating that when comments. In the control condition, comments
interacting online, an individuals’ personal from other students were absent. Thus, there
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

characteristics are less salient, we predicted that were three levels of social validation (willing to
likability may not be as influential. Specifically, help, unwilling to help, and no social validation
This document is copyrighted by the American Psychological Association or one of its allied publishers.

the work reviewed earlier in the text suggests information control). The primary dependent
that, owing to decreased communicator salience variable was a measure of willingness to volun-
inherent in text-based CMC, communicator lik- teer time for a campus fundraiser.
ability cues may either be reduced or be less
effective. However, given that there is conflict- Participants
ing evidence concerning the effectiveness of
communicator likability in computer-mediated Participants were 249 (64 men, 185 women)
contexts, we sought to examine the effect of introductory psychology students from a large
communicator likability on compliance. Thus, public university. It is typical for most psychol-
our second research question was: Does com- ogy experiments, including social psychology,
municator likability increase compliance with a to include a convenience sample containing
request? mostly undergraduate students (Peterson,
2001). Students received partial course credit
for participation in the study.
Method
Stimuli
Design
Blogs. Participants read the blog online.
The experimental design was a 3 (Likability Each participant was randomly assigned to one
of communicator: Likable vs. Unlikable vs. No of the nine conditions. As stated earlier in the
Likability Control) ⫻ 3 (Social Validation: text, likability was manipulated by the fictitious
Willing to Help vs. Others Refusing to Help vs. blog author’s comments pertaining to support of
No Response Control) between-subjects facto- the University football team. In the likable con-
rial. Participants were randomly assigned to dition, the request was for help with campus
read one of nine blog entries in which a ficti- clothing drive for the needy and included a
tious student (the blog author) asked for volun- statement “ROLL TIDE,” a commonly used
teers to help with a university-related fund- statement of support for the University football
raiser. Because data were collected at a large and other athletic teams. An image of the Uni-
Football University, communicator likability versity’s sports logo was also included in the
was manipulated by the presence of pro- or layout of the blog. In the unlikable condition,
anti-football sentiments made by the blog au- the sports logo and statement were not included,
thor or the absence of such comments. Previous and the same request for help was included, but
experiments on social influence and persuasion the fictitious blogger indicated that his blog was
have used similar methods of manipulating a place to share information about campus
communicator likability (Chaiken, 1980; Cial- events unrelated to football. In the neutral con-
dini, 2009; Sinclair, Moore, Mark, Soldat, & dition, no such mention of football was made.
Lavis, 2010). That is, likability can be success- As described earlier in the text, social vali-
fully manipulated by portraying the communi- dation was manipulated by the presence or ab-
cator as one who is similar and/or who supports sence of comments from other fictitious stu-
a majority opinion or group versus someone dents. When social validation was present, six
who is not similar and/or who does not support comments from fictitious students either agreed
 56 GUADAGNO, MUSCANELL, RICE, AND ROBERTS

to volunteer time (e.g., “I’d like to help out! conditions, read the corresponding blog, com-
When will the clothing drive be?” and “I can pleted the dependent measures, and were then
volunteer and also bring clothes.”). When the debriefed.
social validation condition was low, the ficti-
tious student comments indicated, for example, Dependent Measures
“Sorry wish I could help but I’m already in- Likability manipulation check. The par-
volved in so much other stuff” and “I would if ticipants rated the likability of the blog owner
I had the time!” In the no social validation using a Likert scale (1 ⫽ not at all likable to
condition, there were no comments in response 9 ⫽ very likable).
to the request for volunteers. Social validation manipulation check. To
The user names for the fictitious students examine perceived social validation, partici-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

appeared as respondent’s first initial and last pants rated the extent to which they thought
name. We selected this style to appear neutral
This document is copyrighted by the American Psychological Association or one of its allied publishers.

other students were willing to volunteer. The

so as not to be confounded with the likability participants were asked to rate other students on
condition (see Figure 1 for a picture of an ex- a Likert scale (1 ⫽ extremely unwilling to 9 ⫽
ample blog). extremely willing).
Procedure
Results
All participants completed this study by ac-
cessing the Internet from a location of their A series of univariate analyses of variance
choice. Participants were told they would be were conducted to examine the manipulation
participating in a study requiring them to eval- checks for blog owner likability and social val-
uate a blog. Once participants gave consent, idation, and to determine whether the indepen-
they were randomly assigned to one of the nine dent variables had an influence on individuals’

Figure 1. Example blog: neutral likability and high social validation.

 SOCIAL INFLUENCE ONLINE 57

willingness to volunteer. Participant gender was Table 2

also examined, but no significant differences Social Validation Manipulation Check: Ratings of
were found on any of our measures. how Much Participants Were Willing to Volunteer
Experimental condition M SD
Manipulation Checks
Control (n ⫽ 88) 5.47a 1.38
There was a significant main effect for com- Low social validation (n ⫽ 82) 3.40b 1.85
municator likability on likability ratings of the High social validation (n ⫽ 79) 6.99c 2.13
blog owner (F(2, 246) ⫽ 3.75, p ⫽ .025, ␩2p ⫽ abc
Different superscripts denote significant differences.
.03). Post hoc tests using Fischer’s Least Sig-
nificant Difference (LSD) indicated that indi-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

viduals in the likable condition rated the blog General Discussion

owner significantly higher in likability than
This document is copyrighted by the American Psychological Association or one of its allied publishers.

those in the unlikable condition (see Table 1). In the current study, we examined the influ-
There was also a significant main effect of ence of social validation and likability on com-
the social validation condition on how much pliance with a request presented online, focus-
the individuals perceived that others were ing on text-based CMC in the form of a blog.
willing to volunteer (F(2, 246) ⫽ 106.27, p ⬍ We found that, similar to other communication
.001, ␩2p ⫽ .464). Post hoc tests using Fisher’s contexts, social validation is also influential in
LSD indicated that individuals in the high online text-based contexts. Furthermore, partic-
social validation condition perceived that oth- ipants in the study were influenced by the di-
ers were more willing to help compared with rection of the comments of others indicating a
those in the low social validation condition willingness or unwillingness to comply. Specif-
(see Table 2). ically, participants in the high and low social
validation conditions were more likely to follow
Main Analyses the comments of fictitious individuals and vol-
The impact of communicator likability and unteer more or less hours respectively. Thus,
social validation on willingness to volunteer participants who read a blog with comments
revealed a main effect of social validation from fictitious others agreeing to help, also
(F(2, 240) ⫽ 3.31, p ⬍ .05, ␩2p ⫽ .027). Post agreed to volunteer more hours themselves
hoc tests using Fischer’s LSD indicated that compared with those who read the blog with
participants in the high social validation con- comments refusing to help.
dition were willing to volunteer for more The results of this study also indicate that
hours compared with those in the low social although likability of the fictitious blog owner
validation condition. Both the main effect of was perceived, it did not impact compliance in
likability (F(2, 240) ⫽ .62, p ⫽ .54, ␩2p ⫽ willingness to volunteer. Participants reported
.005) and the interaction between social val- the communicator to be likable or unlikable in
idation and likability (F(2, 240) ⫽ 1.54, p ⫽ accordance with their assigned experimental
.19, ␩2p ⫽ .025) were not significant in terms conditions, but communicator likability was not
of their effect on participants’ willingness to as influential. Thus, perceived likability did not
volunteer (see Table 3). affect participant compliance. This is consistent
with the literature indicating that, online, a com

Table 1
Likability Manipulation Check: Ratings of the Table 3
Likability of the Blog Owner Mean Number of Hours Participants Were Willing
to Volunteer by Social Validation Condition
Experimental
condition M SD Experimental condition M SD
Control (n ⫽ 81) 6.99ab 1.37 Control (n ⫽ 88) 2.39ab 2.06
Unlikable (n ⫽ 81) 6.60a 1.56 Low social validation (n ⫽ 82) 2.06a 2.36
Likable (n ⫽ 87) 7.20b 1.30 High social validation (n ⫽ 79) 2.92b 2.38
ab ab
Different superscripts denote significant differences. Different superscripts denote significant differences.
 58 GUADAGNO, MUSCANELL, RICE, AND ROBERTS

municator’s perceived likability does not have nally, there is a paucity of research on social
the same effect as it would in interactions in influence online with the current generation of
which he/she is more salient (Guadagno & Cial- young adults. This makes the generalizability of
dini, 2002, 2005, 2007). Therefore, one reason much of the previous literature cited earlier in
communicator likability may not have been in- the text questionable. Future research should
fluential is that in text-based interactions, the also examine whether this generation of Face-
communicator is less salient. Future research book-, Pintrest-, and Twitter-using young adults
should further examine this issue. The effective- respond to the Internet in the way their prede-
ness of the social validation manipulation sug- cessors did. Indeed, there is ample evidence that
gests that in online interactions, the influence of they do not (see Guadagno, Muscanell, & Pol-
others affects our decisions. This supports ideas lio, in press).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

presented by Guadagno and Cialdini (2005) in The results of this study also have limitations
their review of online persuasion and compli- that need to be addressed. First, it is unknown
This document is copyrighted by the American Psychological Association or one of its allied publishers.

ance and also with Postmes et al. (2001). whether these results generalize to other forms
of social influence such as persuasion and con-
Limitations and Future Directions formity. Second, given that we used mostly
female college students as our sample, it is
With the increasing popularity of various unknown whether our results will generalize to
forms of CMC accessible via the Internet, it is the global adult population. Certainly, the work
important for researchers to understand how of Larry Rosen would suggest our results might
factors unique to online contexts affect individ- not replicate in different generations (Rosen,
ual’s social interactions. The present investiga- 2011). However, the work of Guadagno et al.
tion focused on furthering our understanding of (in press) suggest the results will generalize
online social influence processes, and suggests within generations. These issues should be ad-
that some social influence principles may ex- dressed in future research.
tend to some online contexts and produce There are also implications for clinical work
changes in compliance rates, whereas others over the Internet. For instance, Barak, Hen,
may not. It is important to note that it is an open Boniel-Nissim, and Shapira (2008) conducted a
empirical question as to whether these results meta-analysis on therapeutic counseling online
would generalize to other more interactive on- and found that the effects were quite similar to
line technologies such as Facebook (see Mus- those found in more traditional face-to-face set-
canell & Guadagno, 2012, for a description of tings. Furthermore, Finn and Barak (2010) con-
how this generation uses social networking ducted a qualitative study of e-therapy and
sites). Future research should continue to con- found it to be effective yet lacking in ethnical
sider how delivering influence attempts via considerations. Applied to the present research,
other online contexts, such as social networking there remains an open question as to the role of
sites, impact compliance with an influence at- counselor likability on therapy outcomes. Given
tempt. This is a particularly important issue that communicator salience is low online (Gua-
because Facebook advertisements often feature dagno & Cialdini, 2002), our results suggest
social validation information such as the name that liking in a one-on-one therapeutic interac-
and/or number of friends who have pressed tion is not a mechanism for therapeutic success.
the “like” button. For instance, how many Social validation in the sense of group counsel-
“likes” will it take before a person decides to ing, however, may be quite successful online.
visit a restaurant? Or on bookseller sites such as Future research should further examine the is-
Amazon.com, how many positive reviewers are sue of social validation and communicator lik-
necessary to influence someone to buy a book or ability in a clinical context.
other product? Does the use of the “like” button Finally, the present investigation has impli-
provide social validation information? Has the cations for organizational behavior as well as
“like” button changed the cultural meaning of the characteristics of the communicator. Our
the concept of liking something when online results suggest that when needing to influence in
and offline? Moreover, does it create a virtual the real world, use of online communication
space where social democracy rules? Future will be successful if it is social validation based
research should examine these questions. Fi- but will not if the influence attempt is liking
 SOCIAL INFLUENCE ONLINE 59

based. Thus, if a communicator has strong or- Cialdini, R. B., & Goldstein, N. J. (2004). Social
ganizational support for his or her agenda, com- influence: Compliance and conformity. Annual Re-
municating it via asynchronous text-based view of Psychology, 55, 591– 621. doi:10.1146/
CMC will likely be successful. Conversely, if annurev.psych.55.090902.142015
the communicator is unlikable, although their Derks, D., Bos, A. E. R., & von Grumbkow, J.
(2008). Emoticons and online message interpreta-
attempts at influence may not be greatly im-
tion. Social Science Computer Review, 26, 379 –
pacted, they should probably use other means of 388. doi:10.1177/0894439307311611
social influence, as likability is still an impor- Dubrovsky, V. J., Kiesler, S., & Sethna, B. N. (1991).
tant factor for online impression management The equalization phenomenon: Status effects in
and formation (Boyd & Ellison, 2007; Okdie, computer-mediated and face-to-face decision-
Guadagno, Bernieri, Geers, & Mclarney- making groups. Human-Computer Interaction, 6,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Vesotski, 2011). Thus, these results have impli- 119 –146. doi:10.1207/s15327051hci0602_2
cations for social influence in the workplace and Eagly, A. H. (1987). Sex differences in social behav-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

when and how technological influence attempts ior: A social-role interpretation. Hillsdale, NJ:
may trump interpersonal influence attempts. Fu- Erlbaum.
ture research should further explore these Eagly, A. H., Wood, W., & Diekman, A. B. (2000).
issues. Social role theory of sex differences and similari-
ties: A current appraisal. In T. Eckes & H. M.
Trautner (Eds.), The developmental social psychol-
References ogy of gender (pp. 123–174). Mahwah, NJ:
Erlbaum.
Asch, S. E. (1951). Effects of group pressure upon Edinger, J. A., & Patterson, M. L. (1983). Nonverbal
the modification and distortion of judgment. In H. involvement and social control. Psychological
Guetzkow (Ed.), Groups, leadership and men. Bulletin, 93, 30 –56. doi:10.1037/0033-2909.93
Pittsburgh, PA: Carnegie Press. .1.30
Bagozzi, R. P., Dholakia, U. M., & Pearo, L. R. K. Finn, J., & Barack, A. (2010). A descriptive study of
(2007). Antecedent and consequences of online e-counsellor attitudes, ethics, and practice. Coun-
social interactions. Media Psychology, 9, 77–114. seling and Psychotherapy Research, 10, 268 –277.
Barak, A., Hen, L., Boniel-Nissim, M., & Shapira, N. Guadagno, R. E., & Cialdini, R. B. (2002). Online
(2008). A comprehensive review and a meta- persuasion: An examination of gender differences
analysis of the effectiveness of internet-based psy- in computer-mediated interpersonal influence.
chotherapeutic interventions. Journal of Technol-
Group Dynamics: Theory Research and Practice
ogy in Human Services, 26, 109 –160.
Special Issue on Internet Research, 6, 38 –51. doi:
Bargh, J. A., & McKenna, K. Y. A. (2004). The
10.1037/1089-2699.6.1.38
Internet and social life. Annual Review of Psychol-
Guadagno, R. E., & Cialdini, R. B. (2005). Online
ogy, 55, 573–590. doi:10.1146/annurev.psych.55
persuasion and compliance: Social influence on the
.090902.141922
Belmore, S. M. (1987). Determinants of attention Internet and beyond. In Y. Amichai-Hamburger
during impression formation. Journal of Experi- (Ed.), The Social Net: The social psychology of the
mental Psychology: Learning, Memory, and Cog- Internet (pp. 91–113). New York: Oxford Univer-
nition, 13, 480 – 489. doi:10.1037/0278-7393.13.3 sity Press.
.480 Guadagno, R. E., & Cialdini, R. B. (2007). Persuade
Boyd, D. M., & Ellison, N. B. (2007). “Social net- him by email, but see her in person: Online per-
work sites: Definition, history, and scholarship.” suasion revisited. Computers in Human Behavior,
Journal of Computer-Mediated Communication, 23, 999 –1015. doi:10.1016/j.chb.2005.08.006
13, article 11. Guadagno, R. E., Muscanell, N. L., & Pollio, D. E.
Chaiken, S. (1980). Heuristic versus systematic in- (in press). The homeless use Facebook?! Similar-
formation processing and the use of source versus ities of social network use between college stu-
message cues in persuasion. Journal of Personality dents and homeless young adults. Computers in
and Social Psychology, 39, 752–766. Human Behavior.
Chaiken, S., & Eagly, A. H. (1983). Communication Guégen, N. (2002). Foot-in-the-door technique and
modality as a determinant of persuasion: The role computer-mediated communication. Computers in
of communicator salience. Journal of Personality Human Behavior, 18, 11–15.
and Social Psychology, 45, 241–256. doi:10.1037/ Guéguen, N., & Jacob, C. (2001). Fund-raising on the
0022-3514.45.2.241 web: The effect of electronic foot-in-the-door on
Cialdini, R. B. (2009). Influence: Science and prac- donation. CyberPsychology and Behavior, 4, 705–
tice. New York: William Morrow. 709. doi:10.1089/109493101753376650
 60 GUADAGNO, MUSCANELL, RICE, AND ROBERTS

Guéguen, N., & Jacob, C. (2002). Solicitation by e-mail In K. St. Amant & S. Kelsey (Eds.), Handbook of
and solicitor’s status: A field study of social influence on research on computer mediated communication
the web. CyberPsychology and Behavior, 5, 377– (pp. 477– 491). Hershey, PA: IGI Global.
383. doi:10.1089/109493102760275626 Okdie, B. M., Guadagno, R. E., Bernieri, F. J., Geers,
Guéguen, N., Jacob, C., & Morineau, T. (2010). A. J., & Mclarney-Vesotski, A. R. (2011). Getting
What is in a name? An effect of similarity in to know you: Face-to-face vs. online interactions.
computer-mediated communication. Electronic Computers in Human Behavior, 27, 153–159.
Journal of Applied Psychology, 6, 1– 4. Peterson, R. A. (2001). On the use of college students
Kiesler, S., Siegel, J., & McGuire, T. W. (1984). in social science research: Insights from a second
Social psychological aspects of computer-medi- order meta-analysis. Journal of Consumer Re-
ated communication. American Psychologist, 39, search, 28, 450 – 461.
Petrova, P. K., Cialdini, R. B., & Sills, S. J. (2007).
1123–1134. doi:10.1037/0003-066X.39.10.1123
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Consistency-based compliance across cultures.

Lea, M., & Spears, R. (1991). Computer-mediated
Journal of Experimental Social Psychology, 43,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

communication, de-individuation and group deci- 104 –111. doi:10.1016/j.jesp.2005.04.002

sion-making. International Journal of Man Ma- Postmes, T., Spears, R., & Lea, M. (1998). Breaching
chine Studies, 34, 283–301. doi:10.1016/0020- or building social boundaries? SIDE effects of
7373(91)90045-9 computer-mediated communication. Communica-
Markey, P. M., Wells, S. M., & Markey, C. N. tion Research, 25, 689 –715. doi:10.1177/
(2001). Personality and social psychology in the 009365098025006006
culture of cyberspace. In S. P. Shohov (Ed.), Ad- Postmes, T., Spears, R., Sakhel, K., & De Groot, D.
vances in Psychology Research (Vol. 9, pp. 103– (2001). Social influence in computer-mediated
124). Huntington, NY: Nova Science Publishers. communication: The effects of anonymity on
Matheson, K., & Zanna, M. P. (1989). Persuasion as group behavior. Personality and Social Psychol-
a function of self-awareness in computer-mediated ogy Bulletin, 27, 1242–1254. doi:10.1177/
communication. Social Behaviour, 4, 99 –111. 01461672012710001
McKenna, K. Y. A., & Bargh, J. A. (2000). Plan 9 Rosen, L. (2011). Teaching the igeneration. Educa-
from cyberspace: The implications of the Internet tional Leadership, 68, 10 –15.
for personality and social psychology. Personality Sassenberg, K., & Postmes, T. (2002). Cognitive and
and Social Psychology Review, 4, 57–75. doi: strategic processes in small groups: Effects of an-
10.1207/S15327957PSPR0401_6 onymity of the self and anonymity of the group on
Mendels, P. (1999, July 21). The two faces of on-line social influence. British Journal of Social Psychol-
anonymity. New York Times. Retrieved from: ogy, 41, 463– 480.
http://www.nytimes.com/library/tech/99/07/cyber/ Sinclair, R. C., Moore, S. E., Mark, M. M., Soldat,
A. S., & Lavis., C. A. (2010). Incidental moods,
articles/21anonymity.html
source likeability, and persuasion: Liking moti-
Muscanell, N. L., & Guadagno, R. E. (2012). Make
vates message elaboration in happy people. Cog-
new friends or keep the old: Gender and person-
nition and Emotion, 24, 940 –961.
ality differences in social networking use. Comput-
ers in Human Behavior, 28, 107–112. Received September 5, 2011
Okdie, B. M., & Guadagno, R. E. (2008). Social Revision received July 31, 2012
influence and computer mediated communication. Accepted August 1, 2012 䡲