Recent research suggests that rational choice and

18
194 Why Faculty Members Use E-Mail: The Role of Individual Differences in Channel Choice Barbara D. Minsky Daniel B. Marin Louisiana State University, Baton Rouge Recent research suggests that rational choice and social influence theories provide complementary explanations for individual selection and use of communication media in organizations. Focusing specifically on e-mail selection and use, our study builds on this research by investigating the determinative role of individual differences. We find that individual differences influence e-mail use directly, as well as influence the relation between other predictor variables and e-mail use. Specifically, favorable attitudes toward innovation and change, computer self-effi- cacy, and computer experience directly and positively influence e-mail use, and in addition, attitudes toward innovation and change influence (moderate) the rela- tionship between social context and e-mail selection and use. These findings point to the need for a more comprehensive and complex model of the process determin- ing the selection and use of e-mail. lectronic mail is ubiquitous in contemporary organizations. E-mail lib- erates the communicator from the time and space constraints of other media, allowing communication between two or more people widely separated geographically, each of whose messages may be received and responded to when convenient. The breadth of current research on electronic communication argues the embeddedness of electronic communication, particularly e-mail, in the continuing evolution of organizational life. For example, researchers have explored attitudinal responses to voice mail at Syncrude Canada Ltd. (Beswick & Reinsch, 1987); challenges posed for e-mail research by ethi- cal and intellectual property issues (Howard, 1993); the organization-shap- ing force of the genre repertoire developed by a community of computer language designers in their e-mail communications (Orlikowski & Yates, 1994); the efficacy of technology-use mediation in helping to adapt new communication technology to its organizational context (Orlikowski, Yates, Okamura, & Fujimoto, 1995); the privacy, accuracy, and intellectual property issues raised by technological advances in business communica- tion (Herschel & Andrews, 1997); and the virtues of computer-mediated versus face-to-face communication (Bordia, 1997). In 1995, Organization Science devoted an entire issue to electronic communication. Researchers have documented the significant organizational opportu- nities e-mail presents. Rapid transmission of large files increases com- munication velocity, supports collaborative work, and sustains both strong and weak ties among communicators (Wellman, Salaff, Dimitrova, Garton, Gulia, & Haythornthwaite, 1996). E-mail makes it easier for organizations to access and process information. E-mail encourages distribution. © 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized by Juan Pardo on November 14, 2007 http://job.sagepub.com Downloaded from

description

lectronic mail is ubiquitous in contemporary organizations. E-mail lib- other media, allowing communication between two or more people widely separated geographically, each of whose messages may be received and responded to when convenient. Two major explanatory theories seem applicable: rational choice theory and social influence theory. Rational choice theory posits that individuals choose communication media by matching the medium’s inherent objec-

Transcript of Recent research suggests that rational choice and

Page 1: Recent research suggests that rational choice and

194

Why Faculty Members Use E-Mail:The Role of Individual Differences inChannel Choice

Barbara D. MinskyDaniel B. MarinLouisiana State University, Baton Rouge

Recent research suggests that rational choice and social influence theories providecomplementary explanations for individual selection and use of communicationmedia in organizations. Focusing specifically on e-mail selection and use, ourstudy builds on this research by investigating the determinative role of individualdifferences. We find that individual differences influence e-mail use directly, aswell as influence the relation between other predictor variables and e-mail use.Specifically, favorable attitudes toward innovation and change, computer self-effi-cacy, and computer experience directly and positively influence e-mail use, and inaddition, attitudes toward innovation and change influence (moderate) the rela-tionship between social context and e-mail selection and use. These findings pointto the need for a more comprehensive and complex model of the process determin-ing the selection and use of e-mail.

lectronic mail is ubiquitous in contemporary organizations. E-mail lib-erates the communicator from the time and space constraints ofother media, allowing communication between two or more people widelyseparated geographically, each of whose messages may be received andresponded to when convenient.

The breadth of current research on electronic communication arguesthe embeddedness of electronic communication, particularly e-mail, in thecontinuing evolution of organizational life. For example, researchers haveexplored attitudinal responses to voice mail at Syncrude Canada Ltd.(Beswick & Reinsch, 1987); challenges posed for e-mail research by ethi-cal and intellectual property issues (Howard, 1993); the organization-shap-ing force of the genre repertoire developed by a community of computerlanguage designers in their e-mail communications (Orlikowski & Yates,1994); the efficacy of technology-use mediation in helping to adapt newcommunication technology to its organizational context (Orlikowski,Yates, Okamura, & Fujimoto, 1995); the privacy, accuracy, and intellectualproperty issues raised by technological advances in business communica-tion (Herschel & Andrews, 1997); and the virtues of computer-mediatedversus face-to-face communication (Bordia, 1997). In 1995, OrganizationScience devoted an entire issue to electronic communication.

Researchers have documented the significant organizational opportu-nities e-mail presents. Rapid transmission of large files increases com-munication velocity, supports collaborative work, and sustains both strongand weak ties among communicators (Wellman, Salaff, Dimitrova,Garton, Gulia, & Haythornthwaite, 1996). E-mail makes it easier for

organizations to access and process information. E-mail encourages

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 2: Recent research suggests that rational choice and

195

increased participation, more egalitarian participation, and less central-ized leadership (Hollingshead, McGrath, & O’Connor, 1993). More equalparticipation improves the quality of ideas (Finholt, Sproull, & Kiesler,1990; Wellman et al., 1996). E-mail facilitates cross-organizational com-munication and improves customer service. In an era of down-sizing, e-mail may help managers handle broader spans of control. Although e-maildoes not have the same effect in every organization, researchers agreethat e-mail is significantly changing life in organizations.

Much research has focused on the determinants of communicationmedia use in organizations. What factors influence a person’s choice touse and how widely to use particular media? Why does one person chooseto use e-mail at least once a week, while another uses it seldom, if ever?Two major explanatory theories seem applicable: rational choice theoryand social influence theory. Rational choice theory posits that individualschoose communication media by matching the medium’s inherent objec-tive characteristics and the objective requirements of the communicationtask (Fulk, Schmitz, & Steinfield, 1990). Social influence theory, in con-trast, argues that channel choice is a function not only of objective char-acteristics of the medium or the task, but also, and perhaps even moreso, of individual perceptions conditioned by the social context of mediaand task (Webster & Trevino, 1995). Both theories have contributedgreatly to our understanding of media selection and use in general andof e-mail selection and use in particular. However, both theories focus onfactors external to the individual communicator: rational choice on thecharacteristics of the communication task and the communication

medium, social influence on the social context in which the communica-tion occurs. Neither theory purports to explain why in a particular set ofcircumstances one person may use a particular communication mediummore frequently than another person uses it. This question is our researchquestion: Why in the same circumstances might one person use e-mail fre-quently while another person uses it infrequently, or not at all?

Reviewing theoretical developments in organizational communication,Fulk and Boyd (1991) suggested, &dquo;Beyond rational and social influence fac-tors, other important forces operate at both the individual and organiza-tional level&dquo; (p. 413). Their suggestion is consistent with literature focus-ing on technological innovation and its diffusion (Nelson & White, 1990;Rogers, 1983), which claims that the way people perceive and use newtechnologies may be explained, at least in part, by individual differences:the configurations of traits, including personality traits and inclinationsand demographic factors, particular to the individual. Accordingly, we pro-pose adding individual differences to those factors specified by rationalchoice and social influence theories to more fully explain the decision touse e-mail.We have attempted to build upon extant rational choice and social

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 3: Recent research suggests that rational choice and

196

influence research by investigating the role of individual differences in thechoice and use of e-mail. Our findings indicate that individual differencesdo directly influence choice and use of e-mail and, in addition, may mod-erate the influences of both rational choice and social context. In this

paper we will (a) briefly review the business communication literature one-mail and some prominent rational choice and social influence theories,(b) draw upon the innovation literature to propose and discuss individualdifference factors, (c) present our hypotheses, (d) present and discuss theresults of our investigation, (e) indicate limitations of the study, and (f)assess the contribution of our results to an understanding of the deter-minants of e-mail use in organizations and indicate some implications forfuture research. What emerges, as one considers in succession rational

choice, social influence, and individual difference, is a progressively morecomprehensive and more complex view of the process of selecting andusing communication media in organizations.

Literature Review

Rational choice and social influence theories have been used byresearchers to explain channel choice. Researchers have suggested a rolefor individual differences in the dissemination and adoption of innovation.

Rational Choice Theories

Rational choice theory posits that individuals choose communicationmedia by matching the medium’s inherent objective characteristics andthe objective requirements of the communication task (Fulk, Schmitz, &

Steinfield, 1990). One of the earliest models posited that choice dependedon the social presence of the medium. Short, Williams, and Christie(1976) proposed a continuum based upon the degree of social presenceeach medium provides; they suggested that the medium providing thedegree of social presence required by the particular communication taskwould then be chosen. Trevino, Lengel, and Daft (1987) ranged mediaalong a continuum according to information richness, a measure of theopportunity the medium affords communicators to interpret equivocalmessages. Face-to-face contact is the richest medium because face, voice,and body cues allow interpretation of equivocal message content. Much asdoes social presence, information richness declines along a continuumfrom face-to-face through telephone, e-mail, and various other kinds ofwritten documents. By ranging the various communication media along asingle continuum, both the social presence model and the informationrichness model propose unidimensional, single-criterion, concepts ofhuman rationality in the choice and use of communication media.

Fulk, Schmitz, and Steinfield (1990) noted that although the socialpresence model received &dquo;moderate support&dquo; in laboratory tests, it

explained &dquo;only a small proportion of the variance in media-related behav-

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 4: Recent research suggests that rational choice and

197

ior&dquo; and was not readily generalizable (p. 118); they noted the informa-tion richness model has received only weak empirical support.Researchers’ inability to find strong empirical support for these modelssuggests that their unidimensionalality and objectivist perspective lead toan insufficiently complex model.

Social Influence Theories

Social influence theories set aside the assumptions of objectivity andfull cognition and look to the social context to identify factors that mayaffect a person’s perception, choice, and use of communication media(Webster & Trevino, 1995).

Social influence theories do not deny that choosing whether and howoften to use a medium may be rational and reasonable given communi-cators’ perceptions, whether objectively correct or not, of the characteris-tics of communication media and tasks. Gutek (1990) found that workgroups requiring communication within the work group or between thework group and other work groups might collaborate more effectively ifthey used a computer system that fitted their task structure. This findingopens the prospect of a rationality, but not necessarily the narrowly con-ceived objectivity of rational choice theories, in the fitting of technologyto task in the determination of communication technology. Further,Hunter and Allen (1992) found that people were more likely to continueto use e-mail when the system was easy to use and when people perceivedbenefits of using it. Insofar as the perceived benefits of using e-mail maybe social or political, this finding also points to social influence and amore broadly conceived rationality. Sproull and Kiesler (1986) found thatgeographic, organizational, and situational variables-features of the com-munication situation, including the relationships among senders andreceivers, the topic of the communication, the social norms-influencedperceived social context. Perceived social context, in turn, might influencethe choice of communication medium through cognitive interpretationsand behavior (Fulk, 1993). Thus, situational variables such as a supervi-sor’s use of e-mail, a peer group’s use, or a group norm of using e-mail,which might be functions of geographic location, could influence an

employee to use e-mail.Under social influence theories, &dquo;media perceptions are, in part, sub-

jective and socially constructed&dquo; (Fulk, Schmitz, & Steinfield, 1990, p.

121). Social influence theories thus present us with a more complex modelof channel choice. The communicator is not merely matching communi-cation task and media, but is influenced, consciously or unconsciously, bysocial relationships, organizational structures, and local norms in theselection and use of communication media.

The social influence theory of technology use (Fulk, 1993; Webster &

Trevino, 1995) draws on social learning theory and social information pro-

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 5: Recent research suggests that rational choice and

198

cessing theory. Fulk (1993) pointed out that technology-related behaviorsand attitudes, including media choice, may result from processes of sociallearning in the work place. Bandura (1977) suggested that one learns anduses behaviors based upon what one sees modeled within social groupings.Observed behaviors of others influence the observer to emulate those

behaviors. Rogers (1983) found that similarities between the potential userand the observed person as well as the potential user’s view of this personaffect the potential user’s willingness to adopt the new behaviors. In thecase of e-mail, this might mean that behaviors of co-workers and superiorsaffect perception and hence use of e-mail through social modeling.

Social information processing theory (Salancik & Pfeffer, 1978) positsmechanisms by which peers influence behavior and attitudes. Social infor-mation consists of comments and observations made by people whoseviews an individual considers relevant. Thus, co-workers’ overt statementsmay influence media choice. Social information may bear on media choice

and use as aspects of the environment become more or less salient or asenvironmental clues are interpreted by co-workers. Social information

helps a person identify what other people in organizations consider impor-tant. Fulk (1993) suggested that social information influences both atti-tudes towards and use of communication media.

Webster and Trevino (1995) argued that rational choice and socialinfluence theories &dquo;are complementary, not competing&dquo; (p. 1564). Socialinfluence theories, we propose, do not deny rationality but provide a mul-tidimensional and therefore more complete, and perhaps more realistic,view of the processes of choosing and using communication media by indi-vidual members of organizations. However, some researchers have indi-cated the need to move &dquo;beyond rational and social influence factors [to]other important forces&dquo; to account for the observation attributed to Riceand Case (1983) that &dquo;some media are favored regardless of circum-stances&dquo; (Fulk & Boyd, 1991, p. 413).

The Role of Individual Differences in Adopting InnovationSocial influence theories regard perception as a social construct. How-

ever, a person’s perception of a particular communication medium may bea function not only of social context and rationality but of the combinedinfluences of these and of traits intrinsic to the person, such as person-ality traits, inclinations, and demographic factors. The behavior flowingfrom this confluence, moreover, may be inconsistent with the behaviorpredicted by rational choice and social influence theories. A person char-acterized by a strong disinclination to change may refuse to adopt e-maildespite demonstrable benefit and social pressure, preferring instead totalk face-to-face even about trivial matters amenable to e-mail. Some such

possibility, at least, is urged by Staw and Ross’ (1985) investigation of a&dquo;dispositional approach to job attitudes&dquo; (p. 469). They observed not only

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 6: Recent research suggests that rational choice and

199

&dquo;a much stronger case for dispositional effects than has been presentedfor many years in organizational psychology&dquo; but also &dquo;not ... nearly asstrong situational effects as one would predict from the social informationprocessing or job design perspectives&dquo; (p. 477).

The innovation literature supports the theory that individual differ-ences matter. Rogers (1983) found favorable attitudes toward change pos-itively associated with favorable attitudes toward innovation. He sug-

gested that early adopters of innovation tend to be people who are morefavorably inclined towards change and science, who possess greater abil-ity to deal with uncertainty and risk, and who are less fatalistic.

Nelson and White (1990) found that people’s own attitudes and values,in addition to those of their peers, leaders, and supervisors, influencetheir adoption of innovation. The diversity of factors they identified

argues the complexity of adopting an innovation. Attitudes toward changein general and toward working with computers in particular were eitherpositively or negatively related to a variety of factors such as faith in man-agement, time pressure, role conflict, responsibility for people, careerprogress, job overload, job scope, stress, group openness, group morale,group pressure, group goal clarity, and group cohesion. In additionNelson and White observed that computer anxiety tends to diminish andmore positive attitudes toward computers develop as people gain experi-ence with computers.

Research HypothesesJust as social influence theory, with its broader conception of ration-

ality, may explain the weaknesses in empirical support for rational choicetheories, so individual differences such as an inherent disinclination tochange might account for defects in the explanatory force of social influ-ence. Such predispositions may turn out to be relatively stable (Staw &

Ross, 1985). If so, we would expect individual differences to interact withsome of the expected effects of both rational choice and social context. Aperson’s general disinclination to change would work against using a newtechnology, the dictates of rational choice and social influence notwith-standing.

Although individual differences have, as we suggest above, received theattention of researchers in the fields of job attitudes and dissemination ofinnovation, their impact on e-mail use has received only limited researchattention (Fulk, Schmitz, & Steinfield, 1990). We propose that insofar asperceptions, bearing on the choice to use e-mail, may be a function of indi-vidual differences as well as of external social influences, individual dif-ferences will play a role in the determination of e-mail use.We believe that e-mail presents an especially likely case of the impor-

tance of individual differences to adopting innovations. Arguably, techno-logical innovation, especially computer-based technology, purports to help

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 7: Recent research suggests that rational choice and

200

people be more effective and efficient in a variety of tasks. E-mail is morethan an innovative technology; it affects communication mode or style,which becomes a part of the person, an aspect of personal image. Thus,we can reasonably expect not only rational choice and social influence butalso intrinsic individual differences to play roles in determining whether,to what extent, and in what circumstances a person will use e-mail. Sucha model may explain behavior that would be regarded as anomalous byrational choice and social influence theories.

Our first two hypotheses recognize the role of a broadly conceivedrationality, not necessarily grounded in the assumption of an objectivereality, as a determinant of e-mail use by people in organizations. If useof a system seems to be beneficial, it is rational to use it. If the systemis easy to use, it is rational to use it. These hypotheses flow from the worknoted above of Gutek (1990) and of Hunter and Allen (1992).

H 1: Perception of the benefits of using e-mail is positively relatedto the use of e-mail.

H2: The perception that it is easy to use an e-mail system is posi-tively related to the use of that system.

Our third hypothesis examines the influence of social context on e-mailuse, based on the findings by Sproull and Kiesler (1986) and Fulk (1993).A group’s leader’s use of e-mail provides a social context which encour-ages the use of e-mail by group members.

H3: The use of e-mail by people within a group is positively relatedto the use of e-mail by the group leader.

The rest of our hypotheses concern the effects of individual differencesboth alone and interacting with the effects of rational choice and socialinfluence. The next two hypotheses, prompted by the innovation literatureof Rogers (1983) and Nelson and White (1990), concern the impact of atti-tudes toward science and change on e-mail use. People with an intrinsicinclination for science, we suspect, will be more at ease with the computertechnology associated with e-mail. The same, we suspect, will be true forpeople with a favorable attitude toward change and innovation, since, rel-atively speaking, e-mail does represent change and innovation.

H4: Working in the sciences is positively related to the use of e-mail.

H5: A favorable attitude toward change and innovation is posi-tively related to e-mail use.

In our sixth hypothesis, we propose that attitudes toward change actnot only directly but also indirectly, moderating the effects of rationalchoice and social influence. This is an important distinction, we think,because it opens the possibility of explaining why in some cases a personmay use e-mail infrequently even though it is an appropriate medium for

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 8: Recent research suggests that rational choice and

201

the communication task and is socially validated. The moderating effectsof individual difference might help to explain such an anomaly.

H6: A favorable attitude toward change and innovation is posi-tively associated with intensification of the positive effects ofrational choice and social influence on the use of e-mail.

Two additional hypotheses are prompted by Rogers’ (1983) work on thediffusion of innovation, specifically his observation, noted above, con-

cerning the inclination of less fatalistic people toward change and inno-vation. Here, as our discussion above proposes, we translate Rogers’ ideainto the concept of locus of control and apply it directly (H7) and indi-rectly (H8) to the issue of e-mail use.

H7: An internal locus of control is positively related to e-mail use.H8: Locus of control positively or negatively moderates the effects

of social influence on e-mail use. That is, for example, weexpect an internal locus of control to intensify positive socialinfluence effects and diminish negative social influence effectson e-mail use.

Related to these hypotheses concerning locus of control is the followinghypothesis, grounded in the premise that people characterized by anexternal locus of control and hence a strong desire to fit in, as indicatedby the social desirability construct, will be more affected by social context.

H9: A high score on a social desirability scale is positively associ-ated with e-mail use in social contexts which favor its use.

Our last two hypotheses concern another individual difference investi-gated by Nelson and White (1990): the relation between experience withcomputers and attitudes toward computers. Our focus is the attitude,specifically the degree of confidence vis-a-vis computers, as a direct (H10)and indirect (H 11) influence on e-mail use. We suppose that people char-acterized by high computer self-efficacy will be more likely to use e-mailand that, for example, a person with high computer self-efficacy will bemore likely to use e-mail even in a social context inimical to its use.

H 10: High computer self-efficacy is positively associated with e-mailuse.

H 11: High computer self-efficacy positively or negatively moderatesthe effects of social influence and rational choice on e-mail use.

Method

Our sample was 163 faculty members in two colleges in a large stateuniversity: 60% from the College of Arts and Sciences and 40% from theCollege of Basic Science. Of these, 73% were male, 27% female; 57% werebetween the ages of 30 and 50, 41% were over 50, and only 2% were under

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 9: Recent research suggests that rational choice and

202

30. Almost all, 96%, had desktop computers in their offices, and over 86%had computers at home. Length of time that e-mail was available in thecollege varied by department from approximately 6 months to over a year.

We chose these two groups assuming faculty in the College of BasicScience (which at this university includes the hard sciences, such as chem-istry, physics, and biology) were more &dquo;science-oriented&dquo; than faculty inthe College of Arts and Sciences, which includes humanities such as his-tory, philosophy, foreign language, and speech and social sciences, suchas anthropology, psychology, and sociology. Also we hoped that these twogroups might provide a particularly wide range of individual inclinationsand hence of e-mail selection and use outcomes. (Of course, we acknowl-

edge the fact that research methodology in the social sciences and evenin humanities such as history is or is becoming more quantitative andmore like the methodology in the natural sciences.)We delivered surveys to the associate dean of each of the Colleges.

Each dean wrote a cover letter to department chairs and mailed the sur-veys to the chairs for distribution to faculty members. The surveys werereturned directly to the researcher to lessen the possibility that individu-als would feel peer and/or supervisor pressures in responding to thesurvey questions. A total of 188 surveys were mailed to College of BasicScience faculty; 51 were returned. A second request resulted in the returnof 14 more, for a response of 34% from Basic Science. A total of 292 sur-

veys were mailed to the College of Arts and Science; 98 were returned fora 34% response from Arts and Science. Nonrespondents can be comparedto late respondents (Armstrong & Overton, 1977). In our survey the

second request respondents were demographically similar to the otherrespondents, as well as representative with respect to age and gender ofthe entire faculty. A second request was not sent to the College of Artsand Sciences.

Measures

To ascertain an individual’s level of e-mail use, we adapted Ku’s (1996)scale of social and nonsocial uses of electronic messaging systems inorganizations (a = .8871). Responses on a 5-point Likert scale (1 = Never,2 = Seldom, 3 = Sometimes, 4 = Often, 5 = Always) tapped frequency ofe-mail use for nine activities, such as exchanging routine information,scheduling meetings, coordinating project activities, and negotiating. (SeeAppendix A, section II, for the full list.)

To measure general e-mail use, we adapted Ku’s (1996) measure of elec-tronic messaging systems. Respondents answered &dquo;Yes&dquo; or &dquo;No&dquo; to 15

questions (e.g., &dquo;Do you use e-mail at least once a day?&dquo; &dquo;Do you like usinge-mail?&dquo;). We used a combination of items as surrogate measures for

social context and rational choice. Social context was measured with the

following items: &dquo;Do you use e-mail to keep in touch with others?&dquo; &dquo;Does

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 10: Recent research suggests that rational choice and

203

your chair and/or dean use e-mail on a regular basis?&dquo; &dquo;Does e-mail facil-

itate your committee work?&dquo; Rational choice was measured with these

items: &dquo;Is e-mail an efficient method of communication?&dquo; &dquo;Is e-mail a con-

venient method of communication?&dquo; &dquo;Do you have access to more infor-

mation with e-mail?&dquo; (See Appendix A, Section III for the full prompt.)We used Campeau and Higgins’ (1995) 10-item scale to assess com-

puter confidence. The measure tapped magnitude and strength by askingindividuals to first respond &dquo;Yes&dquo; or &dquo;No&dquo; and then to indicate their level

of confidence in their ability on a 10-point Likert scale (1 = Not at all con-fident to 10 = Totally confident). The survey stated, &dquo;I could use the soft-

ware package in my teaching/administrative duties ...&dquo; and then posed10 scenarios such as &dquo;... if there were no one around to tell me what to

do as I go&dquo; or &dquo;if someone had helped me get started.&dquo; (See Section IV,Appendix A for the full prompt.)

Ettlie and O’Keefe (1982) determined that the best way to predictinnovative behavior is to assess attitudes toward change in general as wellas those toward a particular innovation. We adapted their scale (a = .865).A 5-point Likert scale (1 = Strongly agree to 5 = Strongly disagree) with15 items measured attitudes towards change and innovation. Sampleitems included &dquo;I try new ideas and new approaches to problems.&dquo; &dquo;Peoplewho depart from the accepted university routine should not be punished.&dquo;(See Appendix A, Section V for the full list.)We adapted Rotter’s (1966) scale to measure locus of control (a = .70).

People with an external locus of control might be more easily influencedby the social context. Respondents indicated which of each of seven pairedstatements was truer. (See Appendix A, Section VII.) For example,respondents were asked which of this pair was truer: &dquo;Without the rightbreaks, one cannot be an effective leader&dquo; or &dquo;Capable people who fail tobecome leaders have not taken advantage of their opportunities.&dquo;We used the seven items suggested by Fischer and Fick (1993; a =

.792) from the Marlowe-Crowne Social Desirability Scale to assess indi-vidual differences. This dual-use scale measures need for approval and sit-uational demands. Sample items in this True/False scale include &dquo;I liketo gossip at times&dquo; and &dquo;I’m always willing to admit when I make a mis-take.&dquo; (See Appendix A, Section VIII.) A high score indicates high needfor approval, which would increase the influence of social context. Forexample, if one’s peers and leader used e-mail, an individual scoring highin social desirability would be even more likely to use e-mail.

Data AnalysisWe created correlation matrices to determine if any of the variables of

interest were significantly correlated with one another. Hypotheses 1-3

and 5-11 were tested by analyzing the correlation matrix: Is there a rela-tionship between e-mail use on the one hand and attitudes toward change

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 11: Recent research suggests that rational choice and

204

and innovation, locus of control, social desirability, computer self-efficacy,leader e-mail use, ease of learning, or recognition of benefits on the other?Hypotheses 6 and 11 were also tested with multiple regression. We usedmoderated regression to test the moderator hypotheses: Will a favorableattitude toward change, or locus of control, or computer self-efficacy mod-erate the effects of rational choice and social influence? As suggested byBaron and Kenny (1986), we first entered the main effects and thenentered the interaction term. The two models were then analyzed to deter-mine whether the model with the interaction term explained significantlymore variance. We performed one-way ANOVA to test hypothesis #4: Doesworking in the sciences presage e-mail use?

Results and Discussion

Hypotheses 1 and 2 were supported: Analysis of the correlation matrixshows that people who perceive e-mail as beneficial and easy to learn touse are more likely to use e-mail (r = .46 and r = .37 respectively. SeeTable 1 for the means, standard deviations, and correlations of the vari-ables. Table 2 indicates which hypotheses were supported.) Although oursurvey data do not allow us say whether these perceptions are objectivelycorrect, our results do suggest that rationality-broadly conceived-plays apositive role in the decision to use e-mail.

Hypothesis 3 was supported: Group members are more likely to use e-mail when the group’s leaders also use it (r = .23). Thus, social context,as well as rationality, appears to play a role in determining e-mail use.Arguably, an effect here assigned to social context may be attributable torationality broadly conceived. If the group’s leader is an e-mail user, itmakes sense (is rational) for an individual member to use it. Followingthe leader, as it were, is rational. Social influence and rationality con-verge. These results support Webster and Trevino’s (1995) argument thatrational choice and social influence theories provide complementaryexplanations for media choice. Rational choices and social influences aresometimes mutually reinforcing. On the other hand, were a person to per-ceive e-mail as easy to use, but still not use it, the explanation might liein social influences militating against its use.

Hypothesis 4 was supported: ANOVA indicated that college was signif-icant for both computer self-efficacy and e-mail usage. Insofar as predilec-tion for science is signaled by college, individuals with a predilection forscience were more likely to use e-mail.

Hypotheses 5 and 10 were supported: A favorable attitude toward

change and innovation (r = .38) and high computer self-efficacy (r = .35)were positively associated with e-mail use, as revealed by analysis of thecorrelation matrix. This result is intuitively appealing: such individual dif-ferences may underlie perceptual differences which lead to variant inter-actions with new technologies; people favorably disposed toward innova-

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 12: Recent research suggests that rational choice and

205

M

111.

=xa

(Aa0Icd

td

,-1&dquo;0

o 3~ ’&dquo;9 a

.2*4°§AIpa

irA

9;

Ixons o..... 1.00 00 0(2) w-5 -5

..........ce..........r . r.ce ceQ U

= 5*a arJ’.J rJ’.J~_ ~_e eo 0

~ -t~’S ’?Q)Q)S S11o 0

r *

v

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 13: Recent research suggests that rational choice and

206

Table 2

Hypotheses Which Were Supported

tion and change and confident as well as experienced in their engage-ments with computers are likely to be more disposed to use e-mail. Hereindividual difference factors suggested by innovation literature were asso-ciated with e-mail use, opening the prospect of a more comprehensive andcomplex model of the determination process that may explain individualbehavior that seems anomalous under the theories of rational choice and

social influence.

Hypothesis 6 was partially supported: The regression model includingattitude toward change and innovation, the social context, and rationalchoice explained a significant amount of the variance (adjusted R2 = .40;F = 29.823; p = .000) in e-mail use. A favorable attitude toward changeand innovation moderated the influence of social context but did not mod-

erate that of rational choice on e-mail use. This finding, interestingly, sug-gests that a person’s favorable regard of change and innovation maystrengthen (or in some cases override) the effects of social context,although it does not alter more rationally determined behavior. Perhapsanother way of putting this is to say that, in the presence of favorableregard of change and innovation, rationality is more compelling thansocial context in the determinative process.

However, Hypotheses 7, 8, 9, and 11 were not supported. Neither inter-nal locus of control nor high social desirability scores seemed to be posi-tively related to e-mail use and neither internal locus of control nor highcomputer self-efficacy seemed to moderate the effects of rational choiceor social influence respectively on e-mail use. The regression modelincluding locus of control, social context, and rational choice explained asignificant amount of variance (RZ = .37, F = 20.291, p = .000) in e-mail

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 14: Recent research suggests that rational choice and

207

use, but locus of control did not play a significant role in the equation.Locus of control did not moderate the relationship between e-mail use andsocial context. The regression model including computer self-efficacy,social context, and rational choice did explain a significant amount of thevariance in e-mail use, but computer self-efficacy did not moderate therelationship between e-mail use and rational choice or that between e-mailuse and social influence.An interesting case here is that of locus of control, which, prompted

by Rogers (1983), we expected to directly, as well as indirectly, affect e-mail use. Our negative results may signal a distinction to be drawn, as wesuggest above, between e-mail use as a personal communication mediumand other technological innovations. Conceivably an internal locus of con-trol in conjunction with a disinclination toward change and innovationwould work against e-mail use, because mode of communication consti-tutes part of a person’s self-image, which one’s internal locus of controldictates is to be preserved against external social forces. Such an expla-nation would be consistent with Staw and Ross’ (1985) observation thatdisposition (i.e., prior attitude) was a stronger predictor than someaspects of social context. However, our present data do not allow us toinvestigate this interesting possibility.

While neither locus of control nor social desirability were directlyrelated to e-mail use, we found, though we had not predicted it, that locusof control moderated the influence of rational choice. Perhaps people withan external locus of control are more alert to their social environments,which may include favorable perceptions of e-mail, and consequently aremore likely to base a rational decision to use e-mail on those perceptionsand people with an internal locus of control, being less alert to their

social environment, are less likely find rational grounds for e-mail use inthat environment.

However, though not predicted, locus of control did moderate the influ-ence of rational choice on e-mail use (the change in F is significant, p =

.015). The regression model including computer self-efficacy, the socialcontext, and rational choice did explain a significant amount of the vari-ance (adjusted R2 = .37, F = 22.689, p = .000) in e-mail use. But computerself-efficacy moderated neither the relationship between e-mail use andrational choice nor that between e-mail use and social influence.

Limitations

First, the nature of our sample means that we cannot generalize to cor-porate populations. University faculty are bright and are often earlyadopters of computer technology (Rice & Williams, 1984). Almost all par-ticipants had computers both at their offices and at their homes. Thesample turned out to be more homogenous than we desired. However, anexamination of the scores on the various scales revealed they fell within

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 15: Recent research suggests that rational choice and

208

. normal curves, thus suggesting that range restriction was not a problem.In addition, although some of the moderating effects hypothesized werenot revealed, we think it is possible that because our sample was a highlyeducated group of university professors, the two personality scales mayhave proved to be too transparent and may have distorted the result.Second, all of the variables investigated were assessed using single-sourceself-reports. We attempted to ameliorate some of these effects by usingvaried rating scales, etc. A factor analysis suggested that common methodvariance was not an issue.

Implications for Future Research and ConclusionsOur findings suggest that the explanations provided by the theories of

rational choice and social influence can be enhanced through considera-tion of the effects of individual differences. Our results support therational choice model in the sense that individuals seem to make rationalchoices to use e-mail based upon their perceptions of e-mail and of thebenefits it offers. However, our results neither justify nor discourageacceptance of the objectivist perspective associated with the rationalchoice model.

Nelson and White (1990) found individuals with an internal locus ofcontrol had more positive attitudes towards computers than those with anexternal locus of control. We have suggested that attitudes towards e-mailmay not be quite the same as attitudes towards computer technology ingeneral. Computers can be used to assist people in being more effectiveand efficient, while e-mail, a means of communication, capable of bearingon interpersonal relationships, may thereby engage individual personalityattributes. Our investigation has just begun to tap the importance of indi-vidual differences in this particular context.

Kersten and Phillips (1992) suggested numerous behaviors demonstratedby e-mail users were also impression management behaviors: ingratiation,self-promotion, intimidation, exemplification, and supplication. Hence theuse of e-mail may play a role in the development, maintenance, and distri-bution of power in an organization. We suggest that as computer technologybecomes more common, research should investigate the ways people in

organizations may reinvent technologies in accordance with their uniqueindividual personalities to reach general social goals. Our somewhat per-plexing results regarding the personality attributes of locus of control andsocial desirability reinforce the need for future research along these lines.Other individual difference factors need to be identified and investigated todetermine their effect on e-mail use. It is possible-we think likely-that someof these might have both a main effect on e-mail use and a moderator effecton the relationships between e-mail and social context and rational choice.

Webster and Trevino (1995) recommended that &dquo;system designers mayneed to look beyond general job level to specific work people do and the

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 16: Recent research suggests that rational choice and

209

type of messages they need to send&dquo; (p. 1565). Our investigation promptsus to recommend that system designers may also need to consider indi-vidual characteristics of the users as well as those of the media, commu-nication task, and social context.

Despite what appear to us as limitations in our research, as indicatedabove, we propose that our results do point to the interaction of individ-ual differences with rational choice and social influence factors and henceto a new and rewarding direction for future research. Perhaps even moresignificant revelations will flow from the interactions between innovationresearch and the rational choice and social influence models. Electroniccommunication media have historically been associated with the evolutionof the modern corporation. Given its accelerating rate of evolution, elec-tronic communication promises to be perpetually innovative. This promptsthe further observation that the contributions of innovation research canbe expected to be particularly important to the advancement of our knowl-edge in this area.

In short, our investigation suggests the desirability of a more compre-hensive and complex model of the determining process of e-mail selectionand use by individuals in organizations, of advance from the simple, uni-dimensional (social presence, information richness) models proposed byrational choice theories and the more broadly conceived multidimensional,but still rationalistic, models proposed by social influence theories tomodels comprehensive and complex enough to accommodate a range ofapparently anomalous human behaviors, without denying force to ration-ality and social influence in the determination of human behavior.

NOTES

Barbara D. Minsky is a doctoral candidate in Business Administration major-ing in Organizational Behavior/Human Resource Management in the William W.& Catherine M. Rucks Department of Management, E.J. Ourso College of Busi-ness Administration at Louisiana State University. Her research interests are

related to the area of perceptual differences and include leadership/LMX, careerissues, and values.

Daniel B. Marin is Assistant Professor in the William W. & Catherine M. Rucks

Department of Management, E.J Ourso College of Business Administration atLouisiana State University. He has published fiction, nonfiction, poetic translation,literary criticism, and articles on organizational evolution and business communica-tion. His current research interests are business ethics and business communication.

An earlier version of this paper was presented at the Southern ManagementAssociation Meeting in November, 1998, in New Orleans, LA.

Send correspondence to Daniel B. Marin, Louisiana State University, WilliamW. & Catherine M. Rucks Department of Management, E.J. Ourso College of Busi-ness Administration, Baton Rouge, LA 70803-6312 <[email protected]>.

REFERENCES

Armstrong, J. S., & Overton, T.S. (1977, August). Estimating nonresponse bias inmail surveys. Journal of Marketing Research, 14(3), 396-402.

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 17: Recent research suggests that rational choice and

210

Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall.Baron, R. M., & Kenny, D. A. (1986, December). The moderator-mediator variable

distinction in social psychological research: Conceptual, strategic, and statisti-cal considerations. Journal of Personality and Social Psychology, 51(6), 1173-1182.

Beswick, R. W., & Reinsch, N. L., Jr. (1987, Summer). Attitudinal responses tovoice mail. The Journal of Business Communication, 24(3), 23-35.

Bordia, P. (1997, January). Face-to-face versus computer-mediated communication:A synthesis of the experimental literature. The Journal of Business Commu-nication, 34(1), 99-120.

Campeau, O. R., & Higgins, C. A. (1995, June). Computer self-efficacy: Develop-ment of a measure and initial test. MIS Quarterly, 19(2), 189-211.

Ettlie, J. E., & O’Keefe, R. D. (1982, April). Innovative attitudes, values, and inten-tions in organizations. Journal of Management Studies, 19(2), 163-182.

Finholt, T., Sproull, L., & Kiesler, S. (1990). Communication and performance inad hoc task groups. In J. Galegher, R. E. Kraut, & C. Egido (Eds.), Intellectualteamwork: Social and technological foundations of cooperative work (pp. 291-325). Hillsdale, NJ: Lawrence Erlbaum.

Fischer, O. G., & Fick, C. (1993, Summer). Measuring social desirability: Shortforms of the Marlowe-Crowne social desirability scale. Educational and Psy-chological Measurement, 53(2): 417-424.

Fulk, J. (1993, October). Social construction of communication technology. Acad-emy of Management Journal, 36(5), 921-950.

Fulk, J., & Boyd, B. (1991, June). Emerging theories of communication in organ-izations. Journal of Management, 17(2), 407-446.

Fulk, J., Schmitz, J. A., & Steinfield, C. W. (1990). A social influence model oftechnology use. In J. Fulk & C. Steinfield (Eds.), Organizations and communi-cation technology (pp. 117-142). Newbury Park, CA: Sage.

Gutek, B. A. (1990). Work group structure and information technology: A struc-tural contingency approach. In J. Galegher, R. E. Kraut, & C. Egido (Eds.),Intellectual teamwork: Social and technological foundations of cooperativework (pp. 63-78). Hillsdale, NJ: Lawrence Erlbaum.

Herschel, R. T. & Andrews, P. H. (1997, April). Ethical implications of technological advances on business communication. The Journal of Business Communi-cation, 34(2), 160-170.

Hollingshead, A. B., McGrath, J. E., & O’Connor, K. M. (1993, August). Grouptask performance and communication technology: A longitudinal study of com-puter-mediated versus face-to-face work groups. Small Group Research, 24(3),307-333.

Howard, T. (1993, June). The property issue in e-mail research. The Bulletin,56(2), 40-41.

Hunter, J., & Allen, M. (1992, August). Adaptation to electronic mail. Journal ofApplied Communication Research, 2(3), 254-274.

Kersten, L., & Phillips, S. R. (1992). Electronic identities: The strategic use of e-mail for impression management. Paper presented at the Annual Meeting ofthe Western Speech Communication Association.

Ku, L. (1996, July). Social and nonsocial uses of electronic messaging systems inorganizations. The Journal of Business Communication, 33(3), 297-325.

Nelson, D. L., & White, M. A. (1990, Fall). Management of technological innova-

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from

Page 18: Recent research suggests that rational choice and

211

tion: Individual attitudes, stress, and work group attributes. The Journal ofHigh Technology Management Research, 1(2), 137-148.

Orlikowski, W., & Yates, J. (1994). Genre repertoire: The structuring of commu-nicative practices in organizations. Administrative Science Quarterly, 39(4),541-574.

Orlikowski, W., Yates, J., Okamura, K. , & Fujimoto, M. (1995). Shaping electroniccommunication: The metastructuring of technology in the context of use. Orga-nization Science, 6(4), 423-444.

Rice, R. E., & Case, D. (1983, Winter). Electronic message systems in the uni-versity : A description of use and utility. Journal of Communication, 33(1), 153-161.

Rice, R. E., & Williams, F. (1984). Theories old and new: The study of new media.In Rice and Associates (Eds.), The new media (pp. 55-80). Beverly Hills, CA:Sage.

Rogers, E. M. (1983). Diffusion of innovations. New York: The Free Press.Rotter, J. B. (1966, January). Generalized expectancies for internal versus exter-

nal control of reinforcement. Psychological Monographs, 80(1), 1-28.

Salancik, G. R., & Pfeffer, J. (1978, June). A social information processingapproach to job attitudes and task design. Administrative Science Quarterly,23(2), 224-253.

Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecom-munications. London: Wiley.

Sproull, L., & Kiesler, S. (1986, November). Reducing social context cues: Elec-tronic mail in organizational communication. Management Science, 32(11),1492-1512.

Staw, B. M., & Ross, J. (1985, August). Stability in the midst of change: A dispo-sitional approach to job attitudes. Journal of Applied Psychology, 70(3), 469-480.

Trevino, L. K., Lengel, R. H., & Daft, R. L. (1987, October). Media symbolism,media richness, and media choice in organizations. Communication Research,14(5), 553-574.

Webster, J., & Trevino L. K. (1995, December). Rational and social theories ascomplementary explanations of communication media choices: Two policy cap-turing studies. Academy of Management Journal, 38(6), 1544-1572.

Wellman, B., Salaff, J., Dimitrova, D., Garton, L., Gulia, M., & Haythornthwaite,C. (1996). Computer networks as social networks: Collaborative work, telework,and virtual community. In J. Hagan and K. S. Cook, (Eds.), Annual Review ofSociology, 22, 213-239.

distribution.© 1999 Association for Business Communication. All rights reserved. Not for commercial use or unauthorized

by Juan Pardo on November 14, 2007 http://job.sagepub.comDownloaded from