Online persuasion: an examination of gender differences in computer mediates interpersonal influence

Online Persuasion: An Examination of Gender Differences inComputer-Mediated Interpersonal Influence

Rosanna E. Guadagno and Robert B. CialdiniArizona State University

The purpose of this research was to investigate how computer-mediated communica-tion affects persuasion in dyadic interactions. Two studies compared participants’attitudes after hearing a series of arguments from a same-gender communicator viaeither e-mail or face-to-face interaction. In Study 1, women showed less messageagreement in response to e-mail versus face-to-face messages, whereas men showed nodifference between communication modes. Study 2 replicated this finding and exam-ined the impact of prior interaction with the communicator. For women, the conditionthat provided the least social interaction led to the least message agreement. For men,the condition that provided the most social challenge led to the least message agree-ment. Results are interpreted in terms of gender differences in interaction style.

A mounting body of evidence indicates thatcommunication modality influences the charac-ter and effectiveness of the communication pro-cess (see Chaiken & Eagly, 1983, for a review).Although there are many ways in which com-munication modes differ, one dimension may beparticularly relevant to current thinking aboutinterpersonal processes: the extent to which themedium makes prominent (or merely available)various personal and social factors not related tothe message itself (Chaiken & Trope, 1999).Communication modalities that restrict the avail-ability of factors such as those mentioned abovemay be referred to as socially constrained,whereas those that provide access to such cuesmay be termed socially unconstrained.

We argue that this dimension extends fromwritten, entirely text-based modes (e.g., essays,e-mail, newspaper articles) on the socially con-strained side, to voice-based modes (e.g., radioor intercom transmissions), through visuallybased modes (e.g., televised or videotaped pre-sentations), and finally to face-to-face interac-tions (e.g., workplace meetings, corridor con-versations) on the socially unconstrained pole.

It is noteworthy that social psychological re-search on persuasion has rarely examined eithercomputer-mediated or face-to-face interaction,preferring to use other modalities for reasons ofmethodological control and ease. One goal ofthe present research was to redress this disparityby assessing the impact of persuasive appealsdelivered in the ecologically frequent butgrossly understudied contexts of face-to-faceand computer-mediated exchanges.

According to dual-processing models of per-suasion (Chaiken, 1980; Chaiken & Trope,1999; Petty & Cacioppo, 1984), circumstancesthat direct the attention of communication re-cipients toward or away from features of themessage can have decidedly different persua-sive consequences. Thus, socially constrainedand unconstrained communication modes mayproduce different persuasion patterns among in-dividuals focused differentially on message-re-lated or interpersonal aspects of the communi-cation setting.

For example, Chaiken and Eagly (1976) ex-amined how mode of communication affectedmessage processing as well as subsequent atti-tudes. In their study, participants received eithera difficult or an easy to comprehend persuasivemessage through one of three communicationmodalities: videotape, audiotape, or written.The easy message was more effective in thevideotape condition, the socially unconstrainedcommunication modality in which the speaker’scues were most salient. Conversely, the difficult

Rosanna E. Guadagno and Robert B. Cialdini, Depart-ment of Psychology, Arizona State University.

This research was supported by a National Science Foun-dation Graduate Fellowship.

Correspondence concerning this article should be ad-dressed to Rosanna E. Guadagno, Department of Psychol-ogy, Arizona State University, Box 1104, Tempe, Arizona85257-1104. E-mail: [email protected]

Group Dynamics: Theory, Research, and Practice Copyright 2002 by the Educational Publishing Foundation2002, Vol. 6, No. 1, 38–51 1089-2699/02/$5.00 DOI: 10.1037//1089-2699.6.1.38

38

message was more effective in the written com-munication condition, the socially constrainedmode in which source cues were minimized.This study provided clear evidence that differ-ent types of persuasive messages produce dif-ferential degrees of attitude change as a functionof communication medium.

Chaiken and Eagly (1983) conducted a fol-low-up study in which they manipulated likabil-ity of the communicator. As in the initial study,participants received a persuasive messagethrough one of three communication modalities.However, in this case, participants also read apersonal statement from the speaker that madehim sound likable or unlikable. When thespeaker was likable, participants in both video-tape and audiotape conditions evidenced greaterattitude change than participants in the writtencommunication condition. When the speakerwas not likable, attitude change was greatest forparticipants who received the written commu-nication. These results suggest that in the vid-eotape and audiotape conditions (the less so-cially constrained conditions), the personal cuesof the communicator were salient and partici-pants engaged in heuristic processing of thepersuasive message. Conversely, in the writtencommunication condition, in which source cueswere less salient, participants processed themessage systematically.

Similarly, Morley and Stephenson (1977)conducted a series of studies that investigatedthe influence of formality of communicationsystem on negotiation. These studies primarilyinvestigated the persuasive factors involved in atwo-person negotiation that took place eitherover the phone or face-to-face. According to ourterminology, because nonverbal feedback (e.g.,eye contact, body language, facial expression)was not available to participants in the phonecondition, the phone condition was more so-cially constrained than the face-to-face condi-tion. In each negotiation, one participant wasgiven a strong case (i.e., a large number ofhigh-quality arguments) whereas the other wasgiven a weak case to argue. The overall resultsof this series of studies indicted that, as pre-dicted, the strong case was more successful inthe (more socially constrained) phone conditionthan in the (less socially constrained) face-to-face condition. Conversely, the weak case argu-ment was more successful in the face-to-facecondition than in the phone condition. A clear

implication of these results is that social con-straint of the communication modality has animpact on the persuasive factors at work in anegotiation process (Morley & Stephenson,1977). With more social cues available, theresearch participants were less swayed on thequality of their opponent’s position.

In sum, the results of the previously reviewedstudies (Chaiken & Eagly, 1976, 1983; Morley& Stephenson, 1977) suggest that the persua-sive impact of different types of messages ismoderated by the extent to which the commu-nication modality makes salient message-rele-vant versus non-message-relevant (e.g., social)cues.

Computer-Mediated Communication:A New Communication Mode

More recently, a newer communication mo-dality has emerged—computer-mediated com-munication (CMC)—which stands to becomeincreasingly important as a medium for com-munication. A recent survey reported that 71million people in the United States have accessto the Internet (Iconocast, 1999), and the num-bers are increasing. People use the Internet tosend e-mail, participate in real-time interactivegroup discussions, download software, partici-pate in noninteractive discussion (e.g., Usenet),use a remote computer, conduct business trans-actions, and engage in real-time audio or videoconversations (Jones, 1995).

To date, CMC has been highly socially con-strained, restricted for the most part to text-based, impersonal forms. Therefore, we wouldexpect that persuasive messages delivered inthis fashion would produce response patternssimilar to those of other socially constrainedcommunication modalities.

And in fact, such patterns were found inthe research of Kiesler, her colleagues, andothers in studies of group decision making(Dubrovsky, Kiesler, & Sethna, 1991; Hiltz,Johnson, & Turoff, 1986; Kiesler, Siegel, &McGuire, 1984; Siegel, Dubrovsky, Kiesler, &McGuire, 1986). These investigators found thatcompared with face-to-face participants indecision groups, individuals communicatingthrough a computer were more likely to violatesocial norms of politeness and to be focusedmore uniformly on the task. Similarly, a studyby Matheson and Zanna (1989) on face-to-face

39SPECIAL ISSUE: ONLINE PERSUASION

versus CMC persuasion revealed a direct linkbetween social cues and attitude change only inthe face-to-face condition. Several studies haveadditionally reported that participants interact-ing via face-to-face like their discussion part-ners more than those interacting via CMC(Kiesler, Zubrow, Moses, & Geller, 1985;Weisband & Atwater, 1999). Additionally,these measures of partner liking were predictiveof decisions in face-to-face interaction but notin CMC. As with the previously reviewed stud-ies on communication modality, it appears thatindividuals who communicate through this so-cially constrained mode are less focused ontheir partners and more focused on the assignedtask.

Gender Differences in Persuasion

Social roles, especially gender-based roles,are another factor that can influence responsesto persuasion attempts. In a meta-analysis of148 studies, Eagly and Carli (1981) found atendency for women to be more persuadablethan men, but this effect was moderated by theextent to which social factors, such as grouppressures and publicness of response, werepresent. Eagly (1987) and others (Tannen,1990; Wood & Stagner, 1994) attribute thesefindings to different social role expectations formen and women. Men are said to be orientedtoward agency, which often manifests in at-tempts to demonstrate one’s independence fromothers in successful performances. Women, onthe other hand, are said to be more communallyoriented, which often manifests in activities de-signed to foster interpersonal cooperation andrelationship formation and maintenance.

According to linguist Deborah Tannen (1990),men’s communication style in interpersonal in-teraction is based on the perception that ininteractions, a man must strive to achieve in-dependence and avoid failure. Women, on theother hand, have a slightly different percep-tion of their world. This perspective is one inwhich conversations are meant to achieve close-ness and consensus. Thus, according to Tan-nen’s view, when interacting with others, menare interested in establishing independencethrough assertiveness or mastery of their envi-ronment, whereas women are interested in mak-ing connections with other individuals throughcooperation.

Overview of the Present Experiments

As more and more people gain access to theInternet, a greater amount of interpersonal com-munication is taking place through this medium.Moreover, an increasing number of businessdecisions are being made primarily throughtext-based messaging, such as e-mail. The pur-pose of these experiments was to examine (a)the ways in which this relatively new and verysocially constrained communication modalityinfluences the process of a persuasive appealand (b) the impact that this modality has onmale versus female message recipients, whotend to focus differentially on the interpersonalaspects of an exchange.

Study 1

The purpose of our first study was to examinethe way a communicator could persuade a dis-cussion partner to favorably evaluate a counter-attitudinal message, depending on strength ofargument, communication modality, and genderof the target of persuasion. In our design, aconfederate attempted to induce agreement in asame-gender research participant on the meritsof instituting comprehensive exams as a newgraduation requirement. The confederate wasinterviewed by each research participant usingone of two communication modality conditions:e-mail or face-to-face. During the course of theinterview, the confederate used either a strongor a weak set of arguments in favor of theproposal in an attempt to persuade the researchparticipant.

A novel aspect of this paradigm was the useof a confederate to administer the persuasivearguments in an interactive exchange ratherthan having research participants read a writtenmessage on a computer screen or watch a vid-eotaped message. This allowed us to investigatethe impact of persuasive messages in a bilateralexchange context rather than in a unilateralpersuasion agent-to-persuasion target context,which is typical of persuasion research butmight not be representative of the way persua-sion is accomplished in many nonexperimentalsettings.

Given that men and women differ in theirmotivational goals, we expected that women inthe face-to-face condition would express moreagreement with the confederate than would

40 GUADAGNO AND CIALDINI

women in the e-mail condition because the so-cial constraint of e-mail does not easily allowfor the establishment of a connection or bond.For men, however, we expected communicationmode not to make as much of a difference inevaluation of the persuasive message becausemen are more likely to enter interactions with adesire for independence rather than cooperationor bonding. Because communicator characteris-tics matter more for women and are more salientin face-to-face interactions, we also predictedthat personality trait ratings of the confederatewould be related to attitude toward the compre-hensive exams only among women in the face-to-face condition.

We also predicted that across communicationmode and gender, strong arguments would elicitmore message agreement than weak arguments.Finally, because e-mail is a highly socially con-strained communication mode, we expected thatparticipants in this condition would generatemore message-oriented cognitive responses tothe persuasive interaction. Conversely, becauseface-to-face interaction is a socially uncon-strained communication mode, we expectedparticipants to be more focused on the commu-nicator and therefore to record more communi-cator-focused thoughts than participants in theemail conditions.

Method

Participants

Research participants were 159 (80 fe-male, 79 male) undergraduate psychology stu-dents.1 Only those with computer experiencewere eligible for this study.

Design

The experimental design was a 2 (communi-cation mode: face-to-face vs. e-mail) � 2 (ar-gument strength: strong vs. weak) � 2 (genderof dyad: male vs. female) factorial.2 From evi-dence that most undergraduates would not sup-port the institution of senior comprehensive ex-ams,3 the confederate was instructed to presenteither a set of strong or weak arguments (adaptedfrom Petty, Harkins, & Williams, 1980)4 in anattempt to change the participant’s attitude.

Procedure

On arrival, research participants were in-formed that they were to take part in a two-person interview–discussion and opinion sur-vey on behalf of the university administration.They were told that they would be paired with apartner (the confederate) with whom they woulddiscuss the merits of instituting senior compre-hensive exams as an additional graduation re-quirement and that this discussion would bestructured like an interview. One partner wasassigned the task of “interviewing” the otherpartner (always the confederate). At this point,participants were given an informational para-graph describing the comprehensive exam pro-posal and the goal of the interview discussion.

Next, participants were given a personalstatement handwritten by the confederate. Allparticipants read the same statement from theconfederate, which included information suchas his or her year in school, favorite food, andhobbies and an indistinct description of his orher personality. This information was providedto create a uniform initial impression of theconfederate.

Prior to engaging in the interview–discus-sion, participants in the e-mail condition re-ceived training on how to use the computer

1 An additional set of participants were run but excludedfrom the analyses: 12 expressed suspicion, 5 did not under-stand the task, and 4 failed the relevance manipulationcheck. Analyses of the primary dependent measures withthese participants included in the data set did not yielddifferent results than reported.

2 Personal relevance (high vs. low) was also manipulated,but it did not yield any significant effects on the primarydependent measures. Thus, this variable is not discussedfurther. However, relevance was still included in all theanalyses reported here for appropriate partitioning of thevariance.

3 Pretesting of the introductory psychology subject poolindicated the mean favorability rating of this issue was 3.14on a scale ranging from 1 (extremely unfavor) to 9 (ex-tremely favor).

4 The arguments used in this study were adapted fromPetty, Harkins, and Williams (1980). Examples of thestrong arguments emitted by the confederate are as follows:“The quality of teaching is better at schools with exams”and “Average starting salaries are higher for graduates forschools with exams.” Examples of the weak arguments areas follows: “Companies that develop these exams wouldn’tmarket to schools unless they worked” and “Graduate stu-dents have to take comprehensive exams and if undergradsdon’t have to, that’s discrimination.”


program.5 All participants were given a list ofquestions to ask their discussion partner. Onceparticipants were prepared, they were intro-duced to the confederate and the interview–discussion began.

For each question, the confederate was pre-pared to emit a scripted response that containedan assigned (strong or weak) argument. Theconfederate was instructed that during the inter-view–discussion he or she should discuss onlythe comprehensive exam proposal and provideonly statements based on the assigned set ofarguments. However, the confederate was toldthat if the participant went off-task, he or sheshould try to reveal only information consistentwith the information on the personal statementor to state additional opinions on the compre-hensive exams that were consistent with thearguments and overall cover story. At the end ofthe interview–discussion, each participant com-pleted the dependent measures. Next, a suspi-cion check was conducted, and the participantwas debriefed and excused.

Dependent Variables

The main measure was participants’ attitudetoward the comprehensive exam proposal, mea-sured using a scale ranging from 1 (extremelyun ) to 9 (extremely ) on the followingdimensions: workable, valuable, needed, andfavorable.

Next, participants’ cognitive responses to theinterview–discussion were measured using athought-listing exercise. Finally, a series of ad-ditional measures assessed participants’ impres-sion of the confederate on a scale rangingfrom 1 (not at all ) to 9 (very ) onthe 13 different dimensions.

Results

Attitude Measure

Participants’ ratings of the comprehensiveexam proposal on the attitude measures wereaveraged to form one composite of overall atti-tude toward the exam proposal (� � .92). A 2(face-to-face vs. e-mail) � 2 (strong vs. weakargument) � 2 (male vs. female dyad) analysisof variance (ANOVA) revealed two significanteffects. The first was a main effect for argumentstrength, F(1, 143) � 6.64, p � .01, �2 � .044.

Participants in the strong argument conditionrated the comprehensive exam more favorably(M � 5.56) than participants in the weak argu-ment condition (M � 4.76).

As depicted in Table 1, there was also asignificant communication mode by gender in-teraction, F(1, 143) � 6.58, p � .01, �2 � .044.A test of simple effects indicated that this in-teraction was due to the fact that women in theface-to-face condition (M � 5.54) were morefavorable toward comprehensive exams thanwomen in the e-mail condition (M � 4.68),F(1, 156) � 3.87, p � .05, whereas menshowed a nonsignificant trend, F(1, 156) �2.47, p � .12, in the opposite direction.

Partner Ratings

A principal-axis factor analysis with obliquerotation was conducted on the partner trait rat-ings. This analysis indicated that the 13 traits onwhich the confederate was rated could be re-duced to form three distinct factors. Factor 1,labeled “Congenial,” accounted for 49% of thevariance and contained the following traits:approachable, confident, likable, interesting,friendly, sincere, and warm. Factor 2, labeled“Knowledgeable,” accounted for 7% of thevariance and contained the following traits:competent, informed, and credible. Finally,Factor 3, labeled “Sincerity,” accounted for 6%of the variance and contained the followingtraits: modest, honest, and trustworthy. The fac-tor loadings for Congenial ranged from .573 to.879; for Knowledgeable, from .737 to .804; andfor Sincerity, from .323 to .851.

A 2 (face-to-face vs. e-mail) � 2 (strong vs.weak argument) � 2 (male vs. female dyad)ANOVA on the Knowledgeable composite re-vealed a significant main effect for argumentstrength, F(1, 143) � 20.82, p � .01, �2 �.127, indicating that participants perceived theconfederates emitting the strong arguments asmore knowledgeable (M � 7.35) than confed-erates emitting the weak arguments (M � 6.34).Similar ANOVAs on the Congenial and Sincer-ity composites revealed no significant differ-ences by condition.

5 Both the participant’s and the confederate’s names ap-peared on the screen, creating a nonanonymous CMCenvironment.


Correlation coefficients between the partnerevaluation composites and the attitude measureindicated that partner trait ratings were signifi-cantly related to attitude in only one condition.For women in the face-to-face condition, atti-tude was significantly positively correlated withall three factors (for Congenial, r � .406, p �.01; for Knowledgeable, r � .448, p � .01; andfor Sincerity, r � .317, p � .05). For women inthe e-mail condition and for men in both com-munication modalities, there were no significantcorrelations between partner trait ratings andattitude.

Cognitive Responses

Two independent judges who were blind toexperimental condition rated participants’ cog-nitive responses as recorded on the thought list-ing measure for valence (positive, negative, orneutral) as well as focus (communicator, mes-sage, or irrelevant). The judges’ ratings wereaveraged to form a more reliable measure.6

Communicator thoughts. Communicatorthoughts refer to any comments about the con-federate that were recorded by participants onthe thought-listing measure. A 2 (face-to-facevs. e-mail) � 2 (strong vs. weak argument) � 2(male vs. female dyad) ANOVA on the totalnumber of communicator thoughts recorded byparticipants indicated that participants in theface-to-face condition recorded more communi-

cator thoughts (M � 2.02) than participantsin the e-mail condition (M � 1.44),F(1, 143) � 4.83, p � .03, �2 � .033. Thisanalysis revealed an additional significant maineffect for argument strength, indicating that par-ticipants in the weak argument condition(M � 2.01) recorded more communicatorthoughts than participants in the strong argu-ment condition (M � 1.45), F(1, 143) � 4.38,p � .04, �2 � .03.

Message thoughts. Message thoughts referto any comments about the comprehensive ex-ams and the arguments emitted by the confed-erate. A 2 (face-to-face vs. e-mail) � 2 (strongvs. weak argument) � 2 (male vs. female dyad)ANOVA on the total number of messagethoughts recorded by participants did not revealany significant effects. However, an examina-tion of the negative message thoughts revealeda significant main effect for communicationmode. Participants in the e-mail condition re-corded relatively more negative messagethoughts (M � 1.12) than participants inthe face-to-face condition (M � 0.83),F(1, 143) � 3.98, p � .05, �2 � .027.

Unscripted Comments

The interview–discussion transcripts werecontent coded.7 The total number of unscriptedcomments emitted by participants in each ses-sion was counted. A 2 (face-to-face vs. e-mail)� 2 (strong vs. weak argument) � 2 (male vs.female dyad) ANOVA on the total number ofunscripted comments revealed a significantmain effect for communication mode, indicat-ing that participants in the face-to-face condi-tion were more likely to deviate from thescripted questions (M � 6.91) than were partic-ipants in the e-mail condition (M � 1.38),F(1, 132) � 35.74, p � .01, �2 � .213. Thisanalysis also revealed a main effect for gender,indicating that men were more likely to deviatefrom the scripted questions (M � 5.49) than

6 The interrater reliabilities for each type of cognitiveresponse were as follows: for total, r � .88; communicator,r � .99; positive communicator, r � .87; negative commu-nicator, r � .48; message, r � .73; positive message, r �.78; negative message, r � .73; irrelevant, r � .67. Note thatthe lower correlation coefficients occurred in cells in whichthere was a restricted range of responses.

7 Eleven transcripts (10 face-to-face, 1 e-mail) were lostthrough a recording error.

Table 1Mean Attitude Toward the Comprehensive Exam(and Standard Deviations)

Dyad andattitude

Communication mode

Face-to-face E-mail

Study 1

Female 5.54 (1.85) 4.68 (2.10)Male 4.86 (1.99) 4.68 (1.99)

Study 2

FemaleCompetitive 5.33 (1.90) 5.27 (2.10)Independent 5.52 (2.06) 4.25 (2.08)Cooperative 5.21 (0.99) 5.09 (1.64)

MaleCompetitive 4.08 (1.95) 5.57 (1.85)Independent 6.12 (1.47) 5.89 (1.78)Cooperative 5.57 (1.42) 5.50 (1.26)


women (M � 2.55), F(1, 132) � 10.38, p � .01,�2 � .073. These two main effects were quali-fied by a significant two-way communicationmode by gender interaction, F(1, 132) � 3.92,p � .05, �2 � .029. A test of the simple effectsindicated that for participants in the face-to-facecondition, men were more likely to make un-scripted comments (M � 9.29) than werewomen (M � 4.54), F(1, 145) � 9.29, p � .01.However, in the e-mail condition, there was nosignificant gender difference in number of un-scripted comments.

Discussion

This study provides new information on theimpact of strength of argument, participant gen-der, and communication mode on interpersonalinfluence. Our results suggest that the most ef-fective way to persuade an individual differsaccording to the strength of the argument aswell as his or her gender and the mode ofcommunication.

A number of conclusions can be drawn. First,in this study, as in previous research (see Petty& Cacioppo, 1986), the strength of the argu-ment had a significant impact on agreementwith the message, indicating that strong argu-mentation is more persuasive than weak argu-mentation. Additionally, participants were morefocused on the communicator in the weak argu-ment condition as compared with the strongargument condition, as indicated by the findingthat participants in the weak argument conditionrecorded more communicator thoughts than didparticipants in the strong argument condition.

As predicted, participants in the face-to-face condition recorded more communicatorthoughts than did participants in the e-mail con-dition, suggesting that source cues were moresalient in the face-to-face condition than in thee-mail condition. In addition, more negativemessage thoughts were generated in the e-mailcondition as compared with the face-to-facecondition, suggesting a greater message focus inthe e-mail condition as compared with the face-to-face condition. The fact that the communica-tion mode difference occurred only for negativemessage thoughts suggests that e-mail partici-pants may have responded to the message withcounter argumentation that was suppressed forthose in face-to-face interactions.

Perhaps the most intriguing finding in thisstudy is that men and women differed in mes-sage-favorable attitudes depending on whichcommunication mode they used to interactwith the same-gender confederate. Specifically,women in the face-to-face condition reportedmore agreement than did women in the e-mailcondition. However, there was no significantdifference between men in the e-mail and menin the face-to-face condition. These findings donot appear to be due to a tendency for women tobe more persuadable than men, as the womennever exhibited significantly higher levels ofagreement than men in comparable conditions.

We believe that participants responded in amanner consistent with gender-stereotypical ex-pectations. For men, there was no difference inagreement with the message between e-mail andface-to-face perhaps because the differences insocial constraint between the two conditionswere relatively unimportant to the men, whosesocial roles focus more on independence andagency than on relationships. We believe thatwomen, on the other hand, whose roles focusedthem more on relationship formation and coop-eration, aligned their attitudes more with thepersuader’s position in the face-to-face condi-tion because it was in that condition that theirrelationship goals were salient and attainable.This interpretation receives support from thestrong positive correlations between attitude to-ward the exams and persuader personality traitratings that occurred only for women in theface-to-face condition.8

An alternative explanation of these results isthat male and female confederates differed intheir persuasiveness and that these differencesled participants to evaluate the arguments dif-ferently. To explore this possibility, a pair ofraters coded the e-mail transcripts for persua-siveness. All references to participant genderwere removed. An analysis of these data re-vealed no significant gender differences. How-ever, a main effect for argument strength wasrevealed, as in the attitude measure, F(1, 74) �153.60, p � .01, �2 � 1.0.

8 The previous research that reported greater liking inface-to-face interactions as compared with e-mail did notinclude a detailed breakdown of the means by gender com-position of the dyad, so it is difficult to compare theseresults with the previous findings.


Finally, although the interview–discussiontranscripts were coded only for the quantity ofunscripted comments, in informal observationsduring this coding we noted that when the menin face-to-face conditions went off script, theyseemed to be attempting to establish domi-nance, whereas women who went off scripttended to do more bonding. This observation issimilar to an analysis of gender differences inthe content of Internet newsgroup postings con-ducted by Herring (1993) and greatly influencedthe design of Study 2.

Study 2

We conducted a second study to (a) replicatethe gender by communication mode interactionof Study 1 (so as to increase confidence in itsreliability) and (b) modify the pattern throughadditional manipulations designed to shed lighton the conceptual mediation on this basic effect.That is, if it is the case that men are more likelyto see their interactions with others in term ofcompetition whereas women are more likely tosee such interactions in terms of cooperation(Eagly, 1987; Tannen, 1990), we wonderedwhether it would be possible to influence maleand female responses to a communicator byvarying the nature of their prior (competitive orcooperative) interaction. In the socially uncon-strained environment of face-to-face communi-cation, men who have had a prior competitiveinteraction with the communicator should re-spond competitively by rejecting the communi-cator’s argument. This should not be the casefor women, however, for whom prior interac-tion may serve to establish a relationship inwhich cooperation and harmony are sought.Thus, for women, it might be the case thatvarious forms of prior interaction would set thestage for future agreement with the otherby way of relationship-building attempts. Forwomen, then, it would not be a competitiveprior interaction that would lead to rejection ofa communicator’s argument but rather a lack ofmeaningful prior interaction.

To examine these possibilities, we used theface-to-face and e-mail procedures of Study 1 toreplicate the basic finding of that study (thatwomen showed less message agreement in e-mail versus face-to-face modes, whereas men’slevels of agreement did not differ) when weprovided participants with no meaningful prior

interaction. In addition, we gave some partici-pants a competitive experience with the com-municator before the persuasion attempt, andwe gave other participants a cooperative expe-rience with the communicator before the per-suasion attempt.

According to a social roles perspective, onewould expect a prior competitive interaction tohave the most negative effect on the men andthat this would be the case primarily in theface-to-face communication mode, where socialand personal cues are prominent. By this sameaccount, however, one would expect the mostnegative effect on the women’s levels of mes-sage agreement to occur when there had beenthe least amount of prior interaction and thatthis would be the case primarily in the e-mailcondition, where social and personal cues aremost prominent.

Method

Participants

Research participants were 237 (139 fe-male, 98 male) undergraduate psychology stu-dents selected in the same way as in Study 1.9

Design

The experimental design was a 2 (communi-cation mode: face-to-face vs. e-mail) � 2 (gen-der: male vs. female dyad) � 3 (prior interac-tion: competitive vs. cooperative vs. indepen-dent) factorial. The confederate was instructedto present the set of strong arguments used inStudy 1.

Procedure

Participants were told that they would partic-ipate in two separate studies, one examining theway individuals put together numbers and oneproviding feedback on proposed changes to ac-ademic policy.

As in the previous study, the participant andthe confederate arrived at the same place andtime. This time, however, the experimenter

9 An additional 20 participants were excluded from thedata analyses because they expressed suspicion. Analyses ofthe primary dependent measures with these participantsincluded did not yield different results.


brought both participant and confederate to thesame lab room at the same time.

The experimenter then proceeded to presentinstructions for the first study, a number gamethat was designed to manipulate prior interac-tion. There were three versions of the game: onedesigned to induce competition, one designedto induce cooperation, and one that providedequivalent interpersonal exposure without anyexplicit cooperative or competitive interaction.

All participants were presented with the sameintroductory statement and were told that theywould take turns building their own puzzle us-ing three-sided dominos. The instructions thendeviated depending on condition.

Participants in the competitive prior interac-tion condition were told that the purpose of thisstudy was to compare the performance of intro-ductory psychology students with that of edu-cation students (such as the confederate).10

They were each instructed to take turns playinga piece, then take a piece from the other’s pileof pieces. Confederates were instructed to try totake pieces that the participants were likely touse. Finally, participants in this condition weretold that the person who performed the bestwould receive a $25 prize. The prize was of-fered to strengthen the competitive environmentand to motivate participants to attend to thetask.

Participants in the cooperative prior interac-tion condition were told that the purpose of thisstudy was to examine the performance of intro-ductory psychology students partnered with ed-ucation students. They were each instructed totake turns playing a piece and then offer a pieceto their partner. Confederates were instructed totry to offer pieces to the participants that theycould use. Finally, participants in this conditionwere told that the pair who performed the bestwould receive a joint $25 prize. In this case, theprize was offered to strengthen the cooperativeenvironment and to motivate participants to at-tend to the task.

Participants in the independent prior interac-tion condition received only the introductorystatement and were then instructed to take turnsplaying without exchanging any pieces. Theywere also told that the best performing intro-ductory psychology participant would win a$25 prize, as would the best performing educa-tion student. The prize was offered to motivate

participants to attend to the task at the samelevel as participants in the other conditions.

All participants were given a sheet to tracktheir performance. Participants were told thatthis tracking sheet would be used to computetheir performance scores. Finally, the experi-menter set a timer for 5 min and left the con-federate and participant alone in the room toplay the game.

Once the timer rang, the experimenter reen-tered the room and transitioned to the discussionof the comprehensive exam. The experimenterexplained that the participant and confederatewould discuss a potential change to academicpolicy. The participant was asked to pick out ofan envelope one of four possible topics. Theconfederate was asked to pick out of an enve-lope one of two possible roles: the interviewer(the one who asks the questions) or the respon-dent (the one who answers the questions). Inreality, each envelope contained multiple copiesof the same choice: comprehensive exam for thetopic of discussion and respondent as the rolefor the confederate. After these assignmentswere made, the experimenter moved the partic-ipant to a room with a computer and the rest ofthe experiment replicated the procedure fromStudy 1.

Dependent Variables

Attitude toward the comprehensive exam,cognitive responses, and partner trait ratingswere assessed using the same measures as inStudy 1.

As manipulation checks for the nature of theprior interaction (game check), participantswere asked three questions. First, they wereasked to rate the nature of the number-matchinggame on a Likert-type scale ranging from 1(cooperative) to 9 (competitive) with 5 (neithercompetitive or cooperative) as the scale mid-point. Next, participants assessed their partner’scompetitiveness and cooperativeness using aLikert-type scale ranging from 1 � not atall to 9 � very .

To assess computer experience, we askedparticipants to indicate the number of hours

10 The confederate was also introduced as an educationstudent in Study 1. This was done to reduce suspicion dueto the fact that two participants arrived for an experiment forwhich only one was scheduled.


spent using a computer and the Internet formultiple purposes.

Predictions

We predicted that the communication modeby participant gender interaction reported inStudy 1 would be replicated in the independentprior interaction conditions in this study. Spe-cifically, we expected the women in the inde-pendent prior interaction, e-mail condition toshow less message agreement than women inthe face-to-face conditions. In addition, we ex-pected to find no communication mode differ-ence between men across conditions.

The general social role prediction for menwas that there would be no communicationmode or prior interaction difference in agree-ment toward the message except in the face-to-face, competitive prior interaction condition. Inthis condition, we expected less message agree-ment than in the other male conditions, resultingin a 1 versus 5 pattern of results.

For women, the general social role predictionwas for a different 1 versus 5 pattern of results.Because motivation for cooperation and bond-ing should override the competitive nature ofthe prior interaction, we expected women toreport more message agreement in all condi-tions in which some type of interaction oc-curred. Thus, we predicted a 1 versus 5 patternof results, with women in the e-mail, indepen-dent prior interaction condition showing lessmessage agreement as compared with all otherconditions.

Finally, we did not expect to find any genderdifferences in computer experience among ourparticipants.

Results

Manipulation Checks

Game check. A 2 (face-to-face vs. e-mail)� 2 (male vs. female dyad) � 3 (competitivevs. cooperative vs. independent prior interac-tion) ANOVA revealed a significant main effectfor prior interaction condition, indicating thatparticipants perceived the competitive gamecondition (M � 6.16) as more competitive thanthe independent (M � 5.52) and the cooperative(M � 3.00) conditions and that the independentcondition was seen as more competitive than the

cooperative condition, F(1, 225) � 52.01, p �.01, �2 � .316. This analysis also indicated thatmen viewed the game as more competitive thandid women (M � 5.32 vs. M � 4.93),F(1, 225) � 10.84, p � .01, �2 � .046.

Competitive trait rating. A 2 (face-to-facevs. e-mail) � 2 (male vs. female dyad) � 3(competitive vs. cooperative vs. independentprior interaction) ANOVA revealed a signifi-cant main effect for prior interaction condition,indicating that participants perceived their part-ners in the competitive (M � 5.47) and inde-pendent (M � 5.52) prior interaction conditionsas more competitive than those in the coopera-tive prior interaction condition (M � 4.78),F(1, 225) � 4.80, p � .01, �2 � .041. In addi-tion, this analysis revealed a significant maineffect for participant gender, indicating thatmen perceived their partners as more competi-tive than did women (M � 5.32 vs. M � 4.93),F(1, 225) � 11.55, p � .01, �2 � .049.

Cooperative trait rating. A 2 (face-to-facevs. e-mail) � 2 (male vs. female dyad) � 3(competitive vs. cooperative vs. independentprior interaction) ANOVA revealed a signifi-cant main effect for participant gender, indicat-ing that across all conditions, female confeder-ates were rated as more cooperative than maleconfederates (M � 8.33 vs. M � 7.82),F(1, 225) � 9.41, p � .01, �2 � .04. Thus,analyses of these three manipulation checks in-dicated that the situation and the confederatewere perceived accurately in each condition. Inaddition, men perceived the confederate andsituation as more competitive than did women.

Attitude Measure

The reliability of the attitude composite was� � .91.11 The 2 (face-to-face vs. e-mail) � 2(male vs. female dyad) � 3 (competitive vs.cooperative vs. independent prior interaction)ANOVA yielded a significant two-way interac-tion between communication mode and priorinteraction, F(1, 225) � 3.54, p � .03, �2 �

11 Prior attitudes toward the comprehensive exam pro-posal were available for 181 participants (76% of the sam-ple). An analysis of covariance on the attitude measureusing the pretest attitude as the covariate revealed the samepattern of results as without the covariate. Consequently,data on the full sample without the covariance analysis arereported hereafter.


.031; a significant two-way interaction betweenprior interaction and gender, F(1, 225) � 4.14,p � .02, �2 � .036; and a marginal two-waycommunication mode by gender interaction,F(1, 225) � 3.05, p � .08, �2 � .013. See Table 1for a breakdown of means by condition.

In addition, we conducted a more focused setof analyses relative to our specific predictions.Our first major prediction was that the commu-nication mode by gender interaction of Study 1would be replicated in the independent interac-tion condition of Study 2. An analysis of theindependent prior interaction cells revealed anearly significant interaction of communicationmode and participant gender, F(1, 235) � 3.57,p � .06. Simple effects tests indicated that, as inStudy 1, women in the e-mail condition re-ported less agreement than women in the face-to-face condition (M � 4.25 vs. M � 5.52),F(1, 234) � 6.49, p � .01. For the men in theindependent condition, the difference betweene-mail and face-to-face was not significant,F(1, 234) � 0.01, ns. Thus, the results ofStudy 1 were replicated.

The second major prediction was that mes-sage agreement would not differ among menexcept for those in the competitive prior inter-action, face-to-face condition, which shouldshow the least agreement. A 1 versus 5 contrasttesting this hypothesis proved significant,F(1, 225) � 11.65, p � .01. An additional ex-amination of the attitude measure for menacross condition revealed that in addition to theabove results, there was no communicationmode difference between men within the coop-erative condition, F(1, 225) � 0.008, ns, or inthe independent condition, F(1, 225) � 0.17,ns. For men in the competitive condition, how-ever, there was less agreement in the face-to-face condition as compared with men in thee-mail condition (M � 4.08 vs. M � 5.57),F(1, 225) � 5.91, p � .02. Thus, for men, thetype of interaction did not have an impact onagreement unless they were initially forced tocompete, then later placed in a face-to-face in-teraction where their prior competitor espousedhis views.

For women, a different picture emerged inkeeping with the third major prediction. Theleast agreement occurred among those in theindependent prior interaction condition. Thedifference in attitude toward the exams betweene-mail and face-to-face was not significant if the

prior interaction was cooperative or competi-tive, F(1, 225) � 0.032 and F(1, 225) � 0.019,respectively. The hypothesized 1 versus 5 con-trast between the independent, face-to-face con-dition and all other female conditions was sig-nificant, F(1, 225) � 6.94, p � .01. Thus, forwomen, the absence of any type of prior inter-action minimized their willingness to agree withtheir discussion partner. See Table 1 for a pre-sentation of the means for this measure.

Partner Ratings

The partner trait ratings were analyzed usinga confirmatory factor analysis to test whetherthe factors established by the exploratory factoranalysis in Study 1 generalized to this sample.The model fit reasonably well, CFI � .928,�2(1, N � 62) � 161.34, p � .01.

Congenial. A 2 (face-to-face vs. e-mail)� 2 (male vs. female dyad) � 3 (competitivevs. cooperative vs. independent prior interac-tion) ANOVA revealed a significant main effectfor gender, indicating that women rated the con-federate as more congenial than did men (M� 7.03 vs. M � 6.30), F(1, 225) � 21.57, p �.01, �2 � .087.

In addition, women in the independent, e-mail condition rated their discussion partner asless likable than did women in any other con-dition, F(1, 225) � 10.53, p � .01. Thus, thesame 1 versus 5 pattern that appeared in theattitude measure also appeared in ratings ofpartner congeniality. For men, there were nodifferences in partner ratings by condition.

Knowledgeable. A 2 (face-to-face vs. e-mail)� 2 (male vs. female dyad) � 3 (competitivevs. cooperative vs. independent prior interac-tion) ANOVA revealed a significant main effectfor gender, indicating that women rated the con-federate as more knowledgeable than did men(M � 7.50 vs. M � 6.95), F(1, 225) � 6.27,p � .01, �2� .027.

Sincerity. A 2 (face-to-face vs. e-mail) � 2(male vs. female dyad) � 3 (competitive vs.cooperative vs. independent prior interaction)ANOVA revealed a significant main effect forgender, indicating that women rated the confed-erate as more sincere than did men (M � 7.50vs. M � 6.95), F(1, 225) � 16.21, p � .01,�2 � .067.

In addition, women in the independent, e-mail condition rated their discussion partner as


less sincere than did women in any other con-dition, F(1, 225) � 7.03, p � .01. Thus, thesame 1 versus 5 pattern that appeared in theattitude measure also appeared in ratings ofpartner sincerity. For men, there were no differ-ences in partner ratings by condition.

Finally, an examination of the correlationsbetween attitude toward the comprehensive ex-ams and communicator trait ratings did not re-veal any significant differences by condition,contrary to the results reported in Study 1.

Cognitive Responses

An analysis of the message and communica-tor thoughts did not reveal the communicationmode differences found in Study 1—that face-to-face interaction produced more communica-tor thoughts than did CMC interaction, F(1,225) � 0.14, ns. This suggests that the priorinteraction eliminated the differences in mes-sage processing typically found in the persua-sion literature.

An examination on the positive communica-tor thoughts showed that women in the inde-pendent, e-mail condition recorded fewer posi-tive thoughts about their discussion partnerthan did women in any other condition,F(1, 225) � 6.94, p � .01. Thus, the same 1versus 5 pattern that appeared in the attitudemeasure also appeared in ratings of positivethoughts about their partner. For men, therewere no differences in positive communicatorthoughts by condition.

Computer Experience

The items on the computer experience mea-sure were summed to form one composite (� �.68). An ANOVA on this measure revealed nosignificant effects for gender, communicationmode, or prior interaction. Thus, it appears thatthe gender differences in persuasion cannot beexplained by gender differences in computer use.

Discussion

The results of this study replicated the findingin Study 1 that without a prior meaningful in-teraction, women taking part in a persuasiveexchange via e-mail agreed with a communica-tor less than women taking part in the sameexchange in a face-to-face setting. In addition,we demonstrated that this result was not due togender differences in computer experience. In-

stead this difference looks to be associated withgender differences in interaction style: womenare motivated to form bonds, whereas men aremotivated to compete if necessary to achieveindependence. The finding that women reportedthe same level of message agreement in allconditions involving prior interaction with theconfederate, even when this prior interactionwas competitive in nature, bolsters this interpre-tation. This interpretation is additionally sup-ported by the predicted 1 versus 5 patternwherein women in the e-mail, independent con-dition reported lower ratings on attitude towardthe comprehensive exam, partner congeniality,partner sincerity, and positive thoughts aboutthe partner as compared with women in theother conditions. It is not surprising that womenchose to bond rather than compete, in thatwomen feel more comfortable cooperating, evenin a competitive environment (Anderson &Morrow, 1995). Additionally, women will chooseto bond with other women, especially in timesof stress (Taylor et al., 2000). Finally, previousresearch on gender-stereotypical behavior indi-cates that women will reject imposed roles ifthey do not agree with them (Cialdini, Wosin-ska, Dabul, Whetstone-Dion, & Heszen, 1998).

Men, on the other hand, did not appear to befocused on establishing a cooperative bond withthe confederate. Instead, they evaluated the ar-guments for what they were and showed nodifferences in attitude toward the exams unlessthey had competed previously, and then tookpart in the face-to-face discussion with the priorrival. Although men showed no universal ten-dency for competition, it appears that they canbe pushed to compete and that the competitive,face-to-face condition spurred them to do so,decreasing their willingness to align their atti-tudes with their competitor.

General Discussion

Taken together, these two studies shed lighton the impact of interactive CMC on interper-sonal influence. For women, having any priorinteraction with a communicator enhances thelevel of agreement relative to that occurring inimpersonal e-mail interactions. For men, onlyan intensely competitive environment led to lessagreement.

However, certain unanswered questions re-main and deserve further investigation. First, it


is possible that the communication mode differ-ence in message processing typically found inthe persuasive communication literature (thatface-to-face interaction produces more commu-nicator-relevant thoughts than CMC interac-tion) may extend only to short-term interactionswith strangers. The fact that there was no dif-ference in cognitive responses in Study 2 sug-gests that a prior relationship with the commu-nicator superseded the communication mode. Inaddition, Walther and Burgoon (1992) foundthat modality effects in impression formationwere eliminated after a group interacted viaCMC over an extended period of time. Thus, insituations where an individual attempts to per-suade a person he or she knows, there may beno difference in the amount of systematic orheuristic processing. Future research should ex-amine this phenomenon in real-world settings.12

It is additionally possible that women in the face-to-face conditions agreed more with the persua-sive message because it facilitated bonding and acomfortable interaction environment. It would beinteresting to test the duration of their attitudetoward the exams. If their reported opinions werejust a function of public conformity, then we wouldexpect that their agreement with the messagewould fade faster over time than it would for men.

Finally, it is possible that our results mayhold true only for same-gender pairings. Same-gender pairs were used in the present research toreduce additional error variance that may haveoccurred as the result of mixed-gender pairings.Our results might not replicate as strongly inother contexts such as a mixed-gender situation,because mixed-gender pairs display less gen-der-stereotypical behavior than do same-genderpairs (Carli, 1989; Deaux & Major, 1987) andevaluate each other differently (see Kiesler etal., 1985). Future research on this phenomenonshould be conducted on mixed-gender dyads.

12 In addition, these results may not generalize to con-texts in which the CMC is completely anonymous. Researchindicates that in-group identity becomes more salient whenthe CMC is anonymous (see Postmes, Spears, & Lea, 1998,for a review). Social identities such as gender become moresalient and may serve as a heuristic cue and may lead tomore agreement with in-group members and less agreementwith out-group members. However, because participants inthis study were not anonymous, an increased salience ofsocial categories was not an important feature of the CMCenvironment we created.

References

Anderson, C. A., & Morrow, M. (1995). Competitiveaggression without interaction: Effects of compet-itive versus cooperative instructions on aggressivebehavior in video games. Personality and SocialPsychology Bulletin, 21, 1020–1030.

Carli, L. L. (1989). Gender differences in interactionstyle and influence. Journal of Personality andSocial Psychology, 56, 565–576.

Chaiken, S. (1980). Heuristic versus systematic in-formation processing and the use of source versusmessage cues in persuasion. Journal of Personalityand Social Psychology, 39, 752–766.

Chaiken, S., & Eagly, A. H. (1976). Communicationmodality as a determinant of message persuasive-ness and message comprehensibility. Journal ofPersonality and Social Psychology, 34, 606–614.

Chaiken, S., & Eagly, A. H. (1983). Communicationmodality as a determinant of persuasion: The roleof communicator salience. Journal of Personalityand Social Psychology, 45, 241–265.

Chaiken, S., & Trope, Y. (Eds.). (1999). Dual-pro-cess theories in social psychology. New York:Guilford Press.

Cialdini, R. B., Wosinska, W., Dabul, A. J., Whet-stone-Dion, R., & Heszen, I. (1998). When rolesalience leads to role rejection: Modest self-pre-sentation among women and men in two cultures.Personality and Social Psychology Bulletin, 24,473–481.

Deaux, K., & Major, B. (1987). Putting gender intocontext: An interactive model of gender-relatedbehavior. Psychological Review, 94, 369–389.

Dubrovsky, V. J., Kiesler, S., & Sethna, B. N.(1991). The equalization phenomenon: Effects incomputer-mediated and face-to-face decision-making groups. Human-Computer Interaction, 6,119–146.

Eagly, A. H. (1987). Sex differences in social behav-ior: A social-role interpretation. Hillsdale, NJ:Erlbaum.

Eagly, A. H., & Carli, L. L. (1981). Sex of research-ers and sex-typed communications as determinantsof sex differences in influenceability: A meta-anal-ysis of social influence studies. Psychological Bul-letin, 90, 1–20.

Herring, S. C. (1993). Gender and democracy incomputer-mediated communication. ElectronicJournal of Communication, 3(2). Retrieved 2000from http://www.cios.org/www/ejcmain.htm

Hiltz, S. R., Johnson, K., & Turoff, M. (1986). Ex-periments in group decision making: Communica-tion process and outcome in face-to-face versuscomputerized conferences. Human Communica-tion Research, 13, 225–252.

Iconocast. (1999). Internet users at a glance. Re-trieved from http://www.iconocast.com


Jones, S. G. (Ed.). (1995). Cybersociety: Computer-mediated communication and community. Thou-sand Oaks, CA: Sage.

Kiesler, S., Siegel, J., & McGuire, T. W. (1984).Social psychological aspects of computer-medi-ated communication. American Psychologist, 39,1123–1134.

Kiesler, S., Zubrow, D., Moses, A. M., & Geller, V.(1985). Affect in computer-mediated communica-tion: An experiment in synchronous terminal-to-terminal discussion. Human-Computer Interac-tion, 1, 77–107.

Matheson, K., & Zanna, M. P. (1989). Persuasion asa function of self-awareness in computer-mediatedcommunication. Social Behaviour, 4, 99–111.

Morley, I. E., & Stephenson, G. M. (1977). Thesocial psychology of bargaining. London: Allen &Unwin.

Petty, R. E., & Cacioppo, J. T. (1984). The effects ofinvolvement on responses to argument quantityand quality: Central and peripheral approaches topersuasion. Journal of Personality and Social Psy-chology, 46, 69–81.

Petty, R. E., & Cacioppo, J. T. (1986). Communica-tion and persuasion: Central and peripheralroutes to attitude change. New York: Springer-Verlag.

Petty, R. E., Harkins, S. G., & Williams, K. D.(1980). The effects of group diffusion of cognitiveeffort on attitudes: An information-processing

view. Journal of Personality and Social Psychol-ogy, 38, 81–92.

Postmes, T., Spears, R., & Lea, M. (1998). Breachingor building social boundaries? SIDE effects ofcomputer-mediated communication. Communica-tion Research, 25, 689–715.

Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire,T. W. (1986). Group processes in computer-medi-ated communication. Organizational Behavior andHuman Decision Processes, 37, 157–187.

Tannen, D. (1990). You just don’t understand: Menand women in conversation. New York: BallantineBooks.

Taylor, S. E., Klein, L. C., Lewis, B. P., Gruenewald,T. L., Gurung, R. A. R., & Updegraff, J. A. (2000).Biobehavioral responses to stress in females:Tend-and-befriend, not fight-or-flight. Psycholog-ical Review, 107, 411–429.

Walther, J. B., & Burgoon, J. K. (1992). Relationalcommunication in computer-mediated communi-cation. Human Communication Research, 19, 50–88.

Weisband, S., & Atwater, L. (1999). Evaluating selfand others in electronic and face-to-face groups.Journal of Applied Psychology, 84, 632–639.

Wood, W., & Stagner, B. (1994). Why are somepeople easier to influence than others? In S. Shavitt& T. C. Brock (Eds.), Persuasion (pp. 149–174).Boston: Allyn & Bacon.


Business

Online persuasion: an examination of gender differences in computer mediates interpersonal influence