Situational influences on the use of communication technologies; a meta-analysis and...

As communication technologies such as e-mail have become indispensable for business communication, the study of the use and effects of such technologies is increasingly relevant. In this article, the authors present a model of the different factors influencing e-mail use in organizations. Building

on three theoretical appreaches concerning the adoption and use of communication technologies, they present a meta-analysis of recent studies about e-mail use in organizations. The results of the analysis form the basis for a model explaining e-mail use. A first validation of the model using survey methodology is presented.

Keywords: conmmunication technologies: media choice theory: e-mail; meta-analysis; organizational communication

**********

In today's organizations, computer technologies play an increasingly important part as media for business communication. At present, technologies such as e-mail are no longer considered as advanced innovations in the way we communicate--they have become relatively common, in the sense that their use in organizational communication is no longer a question. Such technologies are being used, on a large and still increasing scale and, as such, have reached a stage of maturity almost comparable to that of the telephone. E-mail has now become "ubiquitous in contemporary organizations" (Minsky & Marin, 1999, p. 194) and it has become virtually impossible to consider organizational communication without attention to computer-based communication technologies (Rice & Gattiker, 2001). One can argue that such maturity makes the study of the use of such technologies obsolete-it's there, it's being used, period. On the other hand, it can be argued that the fact that these technologies are now so pervasive makes such studies more relevant. Given that communication technologies are an integral part of our organizational communication landscape, explaining the extent to which they are used and with what effect is important.

Since the late 1970s, a considerable research tradition concerning the adoption and use of such technologies has emerged in which emphases and conceptualizations have shifted over time. In this article, we present a comparison of different theoretical approaches that have emerged as part of this research tradition. Our aim is to construct a model that is based on these approaches and integrates useful insights from each approach to determine its relative value in explaining today's use of communication technologies. In surveying these theoretical approaches, we focus on the use of e-mail because it has become the most widely used communication technology over the past decade (Katz & Rice, 2002; Minsky & Marin, 1999). The model we present is based on a meta-analysis of recent empirical studies concerning e-mail use. We also discuss the results of an exploratory survey study in which the results of this meta-analysis are partially tested. The central research questions addressed in this article are (a) Which theoretical approaches about the use of communication technologies in organizations can be identified? (b) How can these approaches be integrated in a model explaining the use of e-mail in organizations? and (c) What can be concluded concerning the relative value of each of these approaches?

To answer the first research question, we identify three theoretical approaches that can be distinguished in the study of organizational communication technologies. The first of these includes "contingency theories" in which the fit between medium and task is considered the primary explanation for people's media choices. The second approach embodies "subjectivistic" theories that stress the importance of the social context for media choice processes. Finally, the third approach is "situational" theories that emphasize the importance of a number of specific characteristics of communication technology and of users' experience and expertise in perceiving these characteristics. Building on these three theoretical approaches, we present a meta-analysis of recent empirical studies concerning e-mail use in organizations, leading to an integrated model explaining e-mail use and an answer to the second research question. An initial test of the results of the meta-analysis was conducted in an exploratory study in the Dutch branch of an international consultancy firm. Based on both meta-analysis and the initial test, we then reflect on the relative significance of each of the theoretical approaches and thus answer the third research question.

ORGANIZATIONAL E-MAIL USE IN THEORY

Over the past 25 years, much theorizing and research has taken place concerning the use of communication technologies in organizations. In the discussion that follows, we identify three relevant theories: contingency, subjectivistic, and situational.

Contingency Theories

Contingency theories have as their central premise that users aim to achieve an optimal fit between the characteristics of the communication task they perform, on one hand, and the characteristics of the media they apply to perform this task on the other. Social presence theory (Short, Williams, & Christie, 1976) states that communication media differ in the degree to which they can communicate (or simulate) the social presence of the communication partners through the use of social cues (both verbal and nonverbal cues). If a medium can communicate only a few social cues, the theory states, communication partners will not experience each other's social presence, and they will pay less attention to each other in the interaction.

Building on these insights, media richness theory (Daft & Lengel, 1984, 1986) proposes that not all communication media are equally suited for the information requirements that various tasks create. Daft and Lengel (1984) distinguish two different information requirements: uncertainty, the lack of information, which creates the need for more information, and equivocality, the absence of clear definitions of situations. Equivocality does not require more information; instead, richer information is necessary. According to Daft and Lengel (1986), richer information has a higher capacity "to change understanding within a time interval" (p. 560). Daft and Lengel go on to suggest that "communication transactions that can overcome different frames of reference or clarify ambiguous issues to change understanding in a timely interval are considered rich" (p. 560). Media are more suitable for equivocal information tasks, or have a higher degree of media richness, if they score higher on the following four criteria: (a) the possibility of instant feedback, (b) the medium's ability to convey multiple cues, such as body language, facial expressions, tone of voice, and so on, (c) the use of natural language to convey subtleties and nuances, and (d) the personal focus of the medium.

Daft and Lengel's central argument is that an optimal fit between task and medium can be achieved and that individuals whose media choice corresponds with this optimum perform better. Media high in richness are better suited for tasks that have a high degree of equivocality. If there is a low degree of equivocality, however, a lean medium is predicted to be more effective. On the basis of the criteria previously discussed, communication technologies such as e-mail are generally considered to be relatively "lean" media. Thus, based on contingency theories, such media are expected to be primarily used for noncomplex, information lean tasks.

Media richness theory has spawned numerous studies investigating its central premises. For instance, McGrath and Hollingshead (1993) proposed explicit taskmedia fit hypotheses and found support for the importance of the fit between task and medium for communication performance. Rice, Hughes, and Love (1987) reported similar results. Likewise, Rice, Grant, Schmitz, and Torobin (1990) found that the degree to which a person perceives a medium to be appropriate to his or her task influences his or her evaluation of that medium and, thus, his or her adoption and use of it. Kraut, Rice, Cool, and Fish (1998) discovered that, consistent with media richness theory, people whose work lacked routine used video telephony more than those with more routine tasks. However, inconsistent with media richness theory, these researchers also found that managers who had "people management" jobs did not use rich media more than lean media. Examination of the extant research shows, at best, only partial support for the central premises of media richness theory. Hollingshead, McGrath, and O'Connor (1993), for instance, did find that face-to-face groups outperformed computer-mediated groups for negotiation and intellective tasks but also discovered no significant differences on generative and decision-making tasks. Laboratory experiments conducted by Kinney and Watson (1992), Kinney and Dennis (1994), and Valacich, Mennecke, Wachter, and Wheeler (1994) each failed to support assumptions concerning the influence of task equivocality and media richness on the completion of various communication and decision-making tasks. Finally, a similar experiment by Suh (1999) did not find any task-medium interaction effects on decision time or decision quality.

Trevino, Daft, and Lengel (1990) present an addition to media richness theory in which they acknowledge that media characteristics are not entirely objective: Media also have a symbolic value, and this symbolic value can lead to media choices that are not optimal in terms of the fit between task and medium. For instance, face-to-face communication symbolizes commitment and personal interest. The symbolic value is believed to be an important reason for face-to-face interaction in circumstances where another medium would have been a more optimal fit. Nevertheless, the central premise of the theory is that there is an optimal possible fit between task and medium and that users aim to achieve this fit.

Subjectivistic Theories

The central premises of the contingency theories have received considerable criticism. The focus of the criticism has been brought by Fulk, Steinfield, Schmitz, and Power (1987); also see Fulk, Schmitz, and Steinfield (1990) in their social influence theory. Their critique focuses on three key assumptions about the medium and the user. The first assumption is about objectivity. In media richness theory, communication media are assumed to have certain inherent characteristics that can be described objectively. Task perceptions are also presumed to be largely based on objective features. Fulk et al. (1987) disagree, stating that perceptions of features of both media characteristics and task features are largely subjective and determined by the social context of the user. The second assumption is saliency. Media richness theory assumes that all media characteristics and task features, as well as their differences, are known to users. Fulk et al. state that a user's attentiveness to specific work and media characteristics is affected by social influence from his or her working environment--not all of these features are equally salient to individuals. The third assumption is the choice-making processes. Choice is assumed to be a rational, cognitive process of evaluating all these salient, inherent features of communication media and consequently choosing the optimal alternative. Fulk et al.'s view is that choices are made within a historical and social context and only subjectively and retrospectively rational: Choices are rationalized after they have been made.

Research has also identified two other factors that affect the way a user's social environment exerts influence on media choice. First, overt statements about characteristics of media or tasks from coworkers, supervisors, or other members of social networks of which the individual is a part can influence choice. Second, vicarious learning from observing the experiences of others and modeling behavior according to the consequences of certain choices by members of a user's environment affect media choice.

A second perspective on the importance of a user's social environment for his or her media choice concerns the role of that environment in enhancing the utility of a medium. As Kraut et al. (1998) argue, social environment plays two roles. First, it makes up the normative environment (in terms of coworkers' attitude toward and use of a new medium--in line with social influence theory). Second, social environment also strongly determines the medium's utility (its added value in terms of communication tasks). Utility is, of course, strongly dependent on the number of useful contacts who use a certain medium. This is related to the concept of critical mass (Markus, 1987, 1990). For any collective action to become valuable to the members of a social system, a sufficient amount of these members need to take part in this action (Oliver, Marwell, & Teixeira, 1985). Thus, for a new technology to become truly useful, it is necessary that enough members of a social system accept and use it to make the technology worthwhile. The fact that we are dealing with, in this case, an interactive medium increases the importance of a critical mass. Once enough members begin using a particular medium, that medium becomes sufficiently widespread for the majority of members to adopt. This is the point where the rate of adoption rises sharply and where the critical mass occurs. Thus, the number of coworkers using e-mail is not only affected by variables associated with social influence, media use is also affected by the perceived enhanced utility of the medium; that is, the higher the relative number of coworkers using e-mail, the more useful the medium becomes to users.

The utility factor strongly resembles what Davis (1989) calls the "perceived usefulness" of an information and communication technology. Davis notes that perceived usefulness is "the degree to which a person believes that using a particular system would enhance his or her job performance" (p. 320). Perceived usefulness is a key variable in Davis's Technology Acceptance Model, which seeks to explain the acceptance and use of communication technologies in organizations. The other key variable is perceived ease of use, or "the degree to which a person believes that using a particular system would be free of effort" (p. 320). According to the model, both perceived usefulness and perceived ease of use are positively related to a person's attitude toward a system, to his or her intention to use this system, and ultimately, to his or her actual use of it. The model belongs to the subjectivistic theoretical approach because it stresses the importance of perceptions, attitudes, and intentions (all of which are subject to social influence) over any objective match between system and task.

Situational Theories

Although social influence theory, critical mass theory, and the technology acceptance model have proven to be valuable in explaining media use not in line with contingency theory, additional study has prompted further theorizing on the factors that influence the use of communication technologies in organizations. Many different theories have emerged that have several key notions in common: They stress the role of communication technologies' capacity to overcome certain situational constraints. In addition, they acknowledge that experience with a medium strongly influences the way characteristics of that medium are defined. Users, taking into account their individual communication styles and requirements as well as the situational demands of their tasks and organizational environments, use communication technologies, such as e-mail, in ways that best fit these styles, requirements, and demands. The opportunities offered by the technology, on the other hand, depend strongly on the user's experience and expertise with it and the organizational and social context. As users' expertise in using e-mail grows, perceptions of its relative richness change.

The importance of user experience in determining a medium's perceived richness is supported by Carlson and Zmud (1999). Their channel expansion theory identifies certain experiences as important in shaping how an individual perceives a medium's richness. These experiences include experience with the channel, experience with the messaging topic, experience with the organizational context, and experience with other communication participants. Through these experiences, communication participants develop associated knowledge bases that may be used to more effectively encode and decode rich messages on a channel. In this way, an objectively lean channel can become increasingly appropriate and usable for rich communication. In a similar vein, Walther (1996) argues that computer-mediated communication (CMC) media possess certain characteristics that make them very well suited for interpersonal communication (traditionally the domain of rich media). Walther's argument is that "cues filtered out" approaches such as media richness theory have always had a blind spot for these characteristics and also that they pay insufficient attention to the influence of time in the process. Walther concludes, on the basis of a considerable number of studies, that time is an important factor in determining the degree to which CMC media are used for interpersonal (i.e., rich) communication. The perception of a medium's richness, in other words, changes with time and with the experience and expertise that users have with using that medium.

Markus (1994b) offers further evidence for Walther's (1996) conclusion and also stresses the importance of the context in which communication technologies are used. Although a technology does possess certain inherent characteristics, use and effects of such technologies are strongly determined by how and why it is used in practice. Markus (1994a) applies a social definition perspective, which argues that a medium's appropriateness for a given task evolves from the experience the organization has with that medium, and such social definitions of appropriateness may not conform to objective interpretations. Fulk (1993) provides more insight into the social definition process, as she shows how, in addition to experience, shared meanings and behavioral patterns in work groups are more important predictors of the use of communication technology than are media richness-related variables such as task routineness. D'Ambra, Rice, and O'Connor (1998) also question the objective character of media richness, as they show that task equivocality and media richness are not unidimensional; instead, they interact with other variables that may affect media choice.

A number of studies have further investigated other influences on media choice, specifically in terms of the value of communication technologies in overcoming certain communication barriers. Dennis, Valacich, Speier, and Morris (1998) stress the importance of a medium's synchronicity (i.e., the number of simultaneous conversations supported by the medium, the degree to which the message is available for reexamination, and the ability to edit or polish a message before transmission) over its objective richness. As discussed before, Kraut et al. (1998) demonstrated how a medium's utility (i.e., its added value in terms of communication tasks) is a strong influence on media choice. With regard to utility, the fact that e-mail facilitates communication independent of time and place may, in many situations, be a more important characteristic than its richness. For instance, Dimmick, Kline, and Stafford (2000) state that "e-mail is used to obviate the barriers posed by time zones and work schedules even though it may not be the optimal medium for conveying or expressing feelings" (p. 241).

Clearly, theory and empirical research concerning the use of communication technologies in organizations leads to a broad range of influences about the use of e-mail. These influences range from task characteristics to user characteristics and from medium characteristics to the social environment. In the following section, we will present a meta-analysis that integrates these different variables (and different theoretical approaches). We then present a model that is proposed to explain e-mail use.

META-ANALYSIS

The theoretical discussion in the previous section answers the first of our research questions. The second question concerns the possibility of integrating the different perspectives into a model explaining the use of e-mail. To answer this question, we present a model for explaining e-mail use based on a meta-analysis of previous research concerning e-mail use in organizations.

Meta-analysis is a method of literature research that is based on an exact, accurately stated question and pays sufficient attention to the way the data in original studies were gathered, interpreted, and analyzed (Bouwman & Neijens, 1991). The aim of a meta-analysis is to conduct the literature research in an objective and scientifically accepted way, by extracting the data (variables and relationships) from the studies and analyzing these data in a way that is imitable and replicable. A meta-analysis of previous research is a compilative analysis of relationships between variables that were found in earlier studies with regard to a certain subject. The analysis aims to give as complete a picture as possible of such variables and relationships produced by research in a particular field. Attention is paid both to relationships that have been studied often but sometimes with contradictory results and to those variables that have not received much attention. A meta-analysis can be a quantitative analysis, compiling correlations found between the variables in question in an effort to create an aggregate measure of correlation, or the procedure can be more exploratory in an effort to establish the relevant variables and relationships in a given research area. The result of the latter approach is often the presentation of hypotheses offered for further examination.

The meta-analysis reported here does focus on quantitatively tested relationships and on the strength of the relationships. Yet because the measures used to test these relationships are quite diverse, we chose not to compile them. Hence, our meta-analysis focuses on identifying the relevant variables and relationships in recent research on e-mail use in organizations. Our goal is to build a model concerning these vital links. We selected studies published in peer-reviewed journals in the areas of communication science, organization science, and management. We specifically focused on studies concerning variables influencing the use of e-mail and on studies in which relationships were explicitly defined and empirically tested. We analyzed 17 studies from one published dissertation and the following journals: Organization Science, Communication Research, MIS Quarterly, Journal of Business Communication, Information & Management, Social Networks, Human Communication Research, Academy of Management Journal, and Journal of Strategic Information Systems. In total, 73 relationships were defined. The results of this meta-analysis are reported in Table 1. For every bivariate relationship found, this table reports the following:

* The variables found (Variable 1 & Variable 2). The names of these variables have been slightly adapted in some cases to reduce the diversity to an acceptable level. For example, both task complexity and required social presence have been renamed as task equivocality.

* The category to which the variables belong (Carl And Cat2). These categories will be discussed in more detail below.

* The direction of the relationship, either positive or negative.

* The measure used to test the strength of the relationship, including Pearson correlation, beta, F-value, t-value, or different. Although the meta-analysis focused on studies in which such quantitative testing of relationships was reported, a small number of relationships has been included that were not tested quantitatively. For these, "none" is reported in this column.

* The strength of the relationship, that is, the value of this measure. For the few relationships that were not tested quantitatively, no value is reported.

* The source from which the relationship was derived: that is, the study in which it was found.

The results shown in Table 1 are the basis for our model for the explanation of email use. The model is presented in Figure 1. The meta-analysis produced a large number of variables. To reduce the complexity of the results, a limited number of factors were identified that contain these variables.

First, the meta-analysis indicates that two variables play a mediating role in explaining e-mail use. These variables appear related to e-mail use and a number of the other variables. These variables can be described as two dimensions of the perceived applicability of e-mail. The first is the range of tasks for which e-mail is perceived to be appropriate (based on contingency theories), ranging from lean to rich communication tasks. The second is the perceived usefulness of e-mail (the subjectivistic and situational theories, especially the technology acceptance model), determined by the perceived value of e-mail for an individual's work. Together with these variables, a number of factors influencing e-mail use were identified on the basis of this meta-analysis. Characteristics of the tasks people perform are hypothesized to be of influence--based both on contingency and situational theories. In line with contingency theories, a task's equivocality is expected to be relevant. Theories concerning situational constraints lead to the expectation that perceived time pressure and geographical dispersedness of the task will also play an important role.

A second set of factors that were identified are based on subjectivistic theories. We hypothesize that a user's social environment is relevant, as evidenced by coworkers' statements and behavior and critical mass mechanisms. A number of characteristics of the medium itself can also be expected to influence the perceived applicability and use of e-mail. The extent to which e-mail is perceived to help obviate barriers of time and space is relevant in terms of situational theories, as well as its richness (in line with contingency theories) and perceived ease of use and compatibility (in line with subjectivistic theories). Finally, characteristics of the users that the meta-analysis indicates as important influences are their innovativeness, their experience with e-mail, their expertise in using the medium (in line with situational theories), their age, and their organizational position (the latter in line with contingency theories, assuming managers have more complex tasks and thus will use e-mail less).

Figure 1 shows that the meta-analysis did indeed produce a mix of the theoretical approaches previously identified. Variables from contingency, subjectivistic, and situational theories are integrated in the model. An important finding from our meta-analysis is that the perceived applicability of e-mail (in terms of the range of tasks for which it is perceived to be appropriate and its perceived usefulness for a user's tasks) is a central variable in explaining e-mail use. To validate the model, and to further determine the relative value of each of the theoretical approaches, an exploratory study was conducted in which the model was tested by means of a survey in an organization. In the following sections, we discuss the setting and the results of the study. The study analysis is followed by a discussion of the model.

[FIGURE 1 OMITTED]

STUDY SITE AND METHOD

In the Dutch branch of an international consultancy firm, a questionnaire was distributed in which variables from our model were measured. The organization is an interesting site for such a study because it is an information-intensive organization with much variance in the kinds of tasks performed--from temporary staffing to strategic consultancy and from technological development to sales. Much time-critical communication takes place, both within and outside the organization. In addition, the organization has considerable experience with e-mail. In fact, e-mail is considered to be a primary means of communication in the organization.

First, the organization's management was contacted to get their explicit support for the study. Second, two contact persons were approached, and these contacts distributed the survey within their local branches and account teams. One of these local branches had a strong external focus, whereas the other one had a more internal orientation. The externally focused branch was primarily engaged in temporary staffing, whereas the other branch was made up primarily of engineers engaged in multimedia and Internet development. The survey was distributed to 162 persons via e-mail. Ninety-two respondents (57% response rate) returned completed questionnaires. Of the respondents, 83% (n = 76) were male, 17% (n = 16) female. The average age of respondents was 34.5 years, with 50% of the respondents 32 or younger and only 20% 40 or older. The average duration of employment with the organization was 4.5 years, with 50% having been with the organization for 3 years or less, and only 10% having been with the organization for 10 years or more. Most respondents (65%) described their function as consultant, 11.0% as manager, 6.5% as account director, and 5.5% as system developer. Finally, 20% described their job level as management.

Respondents were informed that the aim of the study was to better understand email use in the organization and the value of e-mail to the organization. The fact that the organization's top management endorsed the study was emphasized. In the questionnaire, different items were used to measure each of the variables in the model. The items used to measure the range of tasks for which e-mail is perceived to be appropriate are based on Rice and Case's (1983) list of communication tasks, which in turn is based on Short et al.'s (1976) measures of social presence. These items are listed in Table 2. Respondents were asked to indicate (on a 5-point scale) to what degree they used e-mail for these tasks.

Although the factor loadings indicate that the scale is made up of two dimensions, the scale is used as having but one scale in these analyses. These dimensions are lean (Factor 1) and rich (Factor 2) communication tasks, respectively. Because this variable is meant to encompass exactly those two dimensions, the scale is an appropriate measurement of the variable. Table 3 shows that the other scales used in the analyses are all unidimensional. These scales are made up of statements. Respondents were asked to indicate to what degree they agreed or disagreed with the statements on a 5-point scale. For each scale, Table 3 shows the statements used, the mean score and standard deviation for the scale (on a 1-5 range) and the reliability score for the scale.

For the variables "coworker evaluations" and "coworker use," no reliable scales could be constructed. Thus, single items were used to measure these variables. These items are also presented in Table 3. For e-mail's richness, we asked respondents to indicate how high they thought e-mail scored (on a 1-10 scale) on each of the four dimensions of media richness, as presented in Table 4. Table 4 lists the items used, the mean score and standard deviation (on a 1-10 range), and the reliability score (Cronbach's alpha) for the scale.

A user's age was measured by asking respondents to indicate their age in years. Experience with e-mail was measured by asking respondents to indicate the number of years they had used e-mail. Organizational position was measured by asking users their formal function. This was subsequently reduced to a dichotomous variable on which managers were coded as 2, and nonmanagers coded as 1. Use of email was measured by asking respondents to indicate the numbers of messages they sent and received, respectively, each day.

Unfortunately, problems with the operationalization of variables caused three variables not to be measured: a user's innovativeness, the medium's perceived ease of use, and the medium's compatibility. Therefore, the results of this empirical study should be interpreted with caution. They provide only a partial test of the mode.

RESULTS

To test the model derived from the meta-analysis, a structural equation model was constructed using AMOS (Arbuckle & Wothke, 1999), a software package that supports such data analysis techniques. Structural equation modeling enables the testing of a set of regression equations simultaneously, providing both parameter statistics for each equation and indices that indicate the fit of the model to the original data. Based on the data previously discussed, the structural equation model that optimally fits these data and has the strongest explanatory power is presented in Figure 2.

[FIGURE 2 OMITTED]

In the model in Figure 2, three endogenous (or downstream) variables are distinguished: perceived range of tasks, perceived usefulness, and e-mail use. In addition, nine exogenous (or upstream) variables are defined. The standardized regression coefficient for each relationship is indicated by the number near the arrow symbolizing the relationship. For each endogenous variable, the proportion of variance explained in the model is indicated above the variable: [R.sup.2] = .31 for range of tasks, .48 for perceived usefulness, and .29 for e-mail use. For the model, three statistics are found to be relevant (Arbuckle & Wothke, 1999; University of Texas, 2002). The first relevant statistic is the chi-square value. This value indicates the absolute fit of the model to the data. For the model in Figure 2, the chi-square test of overall model fit is 15.8 with 16 degrees of freedom, returning a probability value of .469 that a chi-square value this large would be obtained by chance if the null hypothesis that the model fits the data is true. So, the null hypothesis is supported and we can conclude that the model fits the data. The second statistic is the TuckerLewis Index (TLI). The TLI is an example of a relative fit statistic, which is less sensitive to sample size and nonnormality in variable distribution. TLI values close to 1 indicate a very good fit (Arbuckle & Wothke, 1999, p. 409), and for the model in Figure 3, this value is 1.000. The third relevant statistic is the Root Mean Square Error of Approximation (RMSEA). In general, the closer the RMSEA value is to 0, the better the fit. Because the model in Figure 4 has an RMSEA of .000, this statistic indicates a very good fit of the model.

[FIGURES 3-4 OMITTED]

It should be kept in mind that the empirical validation of the model is based on a single exploratory study. In addition, not all of the model variables were studied. Despite these limitations, a number of preliminary conclusions can be drawn on the basis of both regression coefficients and fit indices. First, the empirical model is very similar to the model that resulted from the meta-analysis. Preliminary conclusions drawn from the analysis are that perceived applicability of e-mail is an important influence on e-mail use and is in turn influenced by different variables--from task characteristics to medium characteristics and from social environment to user characteristics.

The first dimension of the perceived applicability of e-mail, the range of tasks for which it is perceived to be appropriate, is explained by expertise and skill with the medium, by the extent to which coworkers use the medium, and by two characteristics of the medium; that is, its ability to overcome barriers of geography and its perceived richness. Users who have more expertise in using e-mail, whose coworkers use e-mail to a larger extent, and who perceive e-mail as a relatively rich medium that helps them overcome barriers of geography tend to use e-mail for a wider range of tasks--both lean and rich communication tasks. The second dimension of e-mail's perceived applicability, its perceived usefulness, is explained by the geographical dispersedness of a user's task, the extent to which a critical mass of users exists in his or her environment, e-mail's perceived ability to overcome limitations of time, and again, its perceived richness. These variables are positively related to this dimension of e-mail's applicability.

Finally, e-mail use is explained by its perceived applicability. The wider the range of tasks for which it is perceived to be appropriate, and the more useful it is perceived to be for the execution of a user's tasks, the more it is used. E-mail use is also explained by critical mass (if such a critical mass of users exists, it is used more), by the geographical dispersedness of a user's task, and by age--older persons use e-mail less than younger persons. These results confirm the expectations from the meta-analysis, but a final relationship in the empirical model directly contradicts these expectations. Experience with e-mail (the number of years a user has been using e-mail) is negatively related to e-mail use. Contrary to situational theories that predict that experience is positively related to the use of communication technologies, it is interesting that our data show that the longer a person uses e-mail, the less the person tends to use it.

DISCUSSION

The three research questions presented in the introduction were (a) Which theoretical approaches toward the use of communication technologies in organizations can be identified? (b) How can these be integrated in a model explaining the use of e-mail in organizations? and (c) What can be concluded concerning the relative value of each of these approaches?

The first question was answered in the theoretical section, where three approaches were identified: contingency theories, subjectivistic theories, and situational theories. Concerning the second question, the results of the meta-analysis and the empirical validation of these results in an exploratory study indicate that the use of electronic mail is determined by a combination of individual user characteristics, medium characteristics, users' social environment, and task characteristics, integrating perspectives from different approaches. Finally, concerning the relative value of the three theoretical approaches, we found that the contingency theories offer little explanation of e-mail use, whereas subjectivistic and situational theories tend to be more prominent in the model.

With regard to the contingency theories, although the meta-analysis indicated that task equivocality influences e-mail use, the results from the research are mixed. Two studies found a positive influence of this variable on e-mail use, but three other studies found a negative influence. Our exploratory study found no influence. In addition, both meta-analysis and our survey study found that perceived richness of the medium was positively related to its applicability. This finding seemingly contradicts what is expected using media richness theory. Indeed, our analyses found little support for the media richness approach.

In this study, social environment was found to be of influence. In the meta-analysis, various studies were discussed in which coworker evaluations and coworker use positively influenced perceived applicability and actual use of e-mail. Critical mass was also shown to influence e-mail use. In our survey study, environment was found to be a significant predictor of e-mail applicability and use--both in terms of vicarious learning and critical mass. In line with Kraut et al.'s (1998) findings, we discovered that the social environment acts both as a normative context (through vicarious learning) and as a utility-enhancing context (through critical mass) for media choice decision making. This finding also supports Davis's (1989) emphasis on the importance of perceived usefulness in explaining system use. The meta-analysis further pointed toward the importance of compatibility. Rogers (1983) defines compatibility as "the degree to which an innovation is perceived as being consistent with the existing values, past experiences, and needs of potential adopters" (p. 15). Although this factor was not tested as part of the survey study, we expect that it is related to the social environment or, as Kraut et al. (1998) note, as a normative context.

Situational factors were found to be very prominent influences on e-mail use. The ability of e-mail to overcome barriers of time and space clearly influences its perceived applicability, as well as a user's expertise with the medium. This applicability is, by nature, more dependent on situational constraints on communication tasks than on any objective fit between task and medium. The meta-analysis indicated that both time pressure and geographical dispersedness associated with a user's task influenced e-mail applicability and use and that the perceived ability of e-mail to help overcome such constraints (speed and geographical reach) was an important reason to use e-mail. Finally, although it was not tested in the survey study, e-mail's perceived ease of use can also be related to these situational theories: As the technological accessibility of a medium is perceived to be higher, its contribution in overcoming such constraints of time and place will be realized more easily.

To more rigorously determine the model's value, a much more elaborate empirical test is needed to further explore the relative value of each of the theoretical approaches identified in this article. A multiple case study (comparing different organizations) would he a fruitful avenue of research. Using different organizations (e.g., differing in size, market, primary process, and technological advancement) to validate the model would enable us to draw more precise conclusions on its external validity and the model's explanatory power in different organizational contexts. Moreover, such studies should engage in a more elaborate research design that might include interviews, discourse analysis, and observations. Another way to more elaborately test the model would be a longitudinal case study. Examining the importance of time with regard to experience, expertise, and the changing perceptions of applicability of e-mail would be an interesting way to further explore these issues. Studying e-mail use within one organization, but over a longer period of time, might show a shift in how important different factors are; on the basis of our discussion, it is to be expected that variables related to the matching of task and medium (i.e., contingency factors) are important determinants of e-mail use in the early stages. Yet as users learn to apply the medium to overcome constraints, regardless of task equivocality and medium richness, situational factors become the more important determinants. Following the development of e-mail use over a longer period of time would allow researchers to test these assumptions. In addition, further research should include communication technologies other than email. Such research might point toward other factors influencing applicability and use.

LIMITATIONS

Although our model is primarily based on a literature review and meta-analysis, the survey testing of the model presented here was exploratory and thus limited. The survey measure was made up of a single exposure, in which the model was not completely tested--three variables that emerged from the meta-analysis were not measured: a user's innovativeness and the medium's perceived ease of use and compatibility. Another limitation of the survey study is the possible self-selection of respondents: distributing a survey via e-mail may mean that people who use email less did not participate. However, this limitation seems less problematic because the organization's management maintained that everybody used (had to use) e-mail. Nevertheless, the results from the survey study should be interpreted with caution. The meta-analysis also has some limitations. We had a very narrow focus in our analysis, where a broader focus might further enrich the model.

Despite the inherent limitations, the results of our meta analysis and survey study warrant further exploration. Contrary to what the meta-analysis indicated, our respondents' experience was negatively related to e-mail use. More experienced users sent fewer messages than less experienced users. One explanation could be that more experienced users use e-mail more efficiently, are more focused than less experienced users, and thus need to send fewer messages--but that does not seem to be a very plausible explanation. A more plausible explanation is that more experienced users have discovered that e-mail is not as useful as they initially thought and thus see less use for it. Yet this explanation may be contradicted by the fact that perceived usefulness is not related to experience. The exact nature of this relationship remains puzzling, but it is important to realize that actual expertise (i.e., the ability to communicate efficiently and effectively via e-mail) is a more important determinant of its perceived applicability and use than the mere number of years one has been using it. Exactly how experience influences e-mail use is a subject that could be studied in greater detail.

Table 1. Results of Meta-Analysis

Cat 1            Variable 1       Cat 2            Variable 2

User             age              range: complex   use for
                                                     socioemotional
                                                     tasks
User             age              range: routine   use for routine
                                                     tasks
Medium           compatibility    use              e-mail adoption
Medium           compatibility    use              e-mail use
Social           coworker         perceived        e-mail attitude
  environment      evaluations      usefulness
Social           coworker         use              e-mail use
  environment      evaluations
Social           coworker         use              quantity of
  environment      evaluations                       communication
Social           coworker         use              e-mail use
  environment      evaluations
Social           coworker         use              e-mail use
  environment      evaluations
Social           coworker use     use              e-mail use
  environment
Social           coworker use     use              e-mail use
  environment
Social           coworker use     use              quantity of
  environment                                        communication
Social           coworker use     use              e-mail use
  environment
Critical mass    critical mass    use              e-mail adoption
User             experience       range: routine   use for routine
                                                     tasks
User             experience       use              total messages
                                                     processed
User             experience       use              total messages
                                                     processed
User             experience       use              # of messages sent
User             experience       use              e-mail adoption
User             experience       use              e-mail use
User             expertise and    perceived        e-mail attitude
                   skill            usefulness
User             expertise and    use              use of coil Clear
                   skill                             media
User             expertise and    use              e-mail use
                   skill
User             expertise and    use              e-mail use
                   skill
Medium           geography        use              e-mail adoption
Medium           geography        use              e-mail use
Task             geography        use              e-mail adoption
Task             geography        use              e-mail adoption
Task             geography        use              e-mail use
User             innovativeness   use              e-mail use
User             innovativeness   use              quantity of
                                                     communication
User             innovativeness   use              e-mail use
Medium           perceived ease   perceived        e-mail attitude
                   of use           usefulness
Medium           perceived ease   perceived        e-mail attitude
                   of use           usefulness
Medium           perceived ease   use              e-mail use
                   of use
Medium           perceived ease   use              e-mail use
                   of use
Medium           perceived ease   use              e-mail adoption
                   of use
Medium           perceived ease   use              quantity of
                   of use                            communication
Medium           perceived ease   use              quantity of
                   of use                            communication
Medium           perceived ease   use              e-mail use
                   of use
Medium           perceived ease   use              e-mail use
                   of use
Medium           perceived ease   use              e-mail use
                   of use
Medium           perceived ease   use              use of coll electr
                   of use                            media
Medium           perceived ease   use              e-mail use
                   of use
Medium           perceived ease   use              e-mail adoption
                   of use
User             perceived ease   use              e-mail use
                   of use
User             position         range: complex   use for
                                                     socioemotional
                                                     tasks
User             position         range: routine   use for routine
                                                     tasks
User             position         range: routine   use for routine
                                                     tasks
User             position         use              # of messages sent
User             position         use              # of messages sent
User             position         use              e-mail use
User             position         use              e-mail use
Range: all       range: all       use              e-mail use
range: complex   range: complex   use              e-mail use
range: complex   range: complex   use              e-mail use
range: routine   range: routine   use              e-mail use
range: routine   range: routine   use              e-mail use
range: routine   range: routine   use              e-mail use
Medium           richness         perceived        e-mail attitude
                                    usefulness
Medium           speed            use              e-mail use
Task             task             use              use of coll electr
                   equivocality                      media
Task             task             use              e-mail adoption
                   equivocality
Task             task             use              e-mail adoption
                   equivocality
Task             task             use              e-mail use
                   equivocality
Task             task             use              e-mail use
                   equivocality
Task             time             range: complex   use for complex
                                                     tasks
Task             time             use              # of messages sent
Task             time             use              e-mail adoption
Task             time             use              e-mail use

Cat 1            Direction   Measure   Strength   Source

User             negative    beta          .25    Ku, 1996
User             negative    beta          .18    Ku, 1996
Medium           positive    corr          .37    van den Hooff, 1997
Medium           positive    corr          .18    van den Hooff, 1997
Social           positive    corr          .35    Trevino, Webster, &
  environment                                       Stein, 2000
Social           positive    none           -     Markus, 1994a, 1994b
  environment
Social           positive    corr          .19    Trevino & Webster,
  environment                                       1992
Social           positive    corr          .31    Trevino, Webster, &
  environment                                       Stein, 2000
Social           positive    t-value      3.39    Wijayanayake & Higa,
  environment                                       1999
Social           positive    none           -     Markus, 1994a, 1994b
  environment
Social           positive    corr          .23    Minsky & Marin, 1999
  environment
Social           positive    corr          .23    Trevino & Webster,
  environment                                       1992
Social           positive    f-value    257.0     Wijayanayake & Higa,
  environment                                       1999
Critical mass    positive    beta          .70    Rice, Grant, Schmitz,
                                                    & Torobin, 1990
User             positive    beta          .30    Ku, 1996
User             positive    corr          .18    Carlson & Zmud, 1999
User             positive    corr          .18    Carlson & Zmud, 1999
User             positive    beta          .22    Ku, 1996
User             positive    beta          .39    Rice, Grant, Schmitz,
                                                    & Torobin, 1990
User             positive    f-value     16.6     Wijayanayake & Higa,
                                                    1999
User             positive    corr          .52    Trevino, Webster, &
                                                    Stein, 2000
User             positive    gamma         .221   Jarvenpaa & Staples,
                                                    2000
User             positive    corr          .35    Minsky & Marin, 1999
User             positive    corr          .26    Trevino, Webster, &
                                                    Stein, 2000
Medium           positive    gamma         .064   Straub & Karahanna,
                                                    1998
Medium           positive    none           -    Dimmick, Kline, &
                                                    Stafford, 2000
Task             negative    gamma         .119   Straub & Karahanna,
                                                    1998
Task             negative    gamma         .240   Straub & Karahanna,
                                                    1998
Task             positive    t-value      2.24    Wijayanayake & Higa,
                                                    1999
User             positive    corr          .38    Minsky & Marin, 1999
User             positive    corr          .25    Trevino & Webster,
                                                    1992
User             positive    corr          .34    van den Hooff, 1997
Medium           positive    corr          .74    Trevino, Webster, &
                                                    Stein, 2000
Medium           positive    corr          .75    Trevino, Webster, &
                                                    Stein, 2000
Medium           positive    corr          .295   Adams, Nelson, &
                                                    Todd, 1992
Medium           positive    corr          .37    Minsky & Marin, 1999
Medium           positive    gamma         .054   Straub & Karahanna,
                                                    1998
Medium           positive    corr          .37    Trevino & Webster,
                                                    1992
Medium           positive    corr          .27    Trevino & Webster,
                                                    1992
Medium           positive    corr          .31    van den Hooff, 1997
Medium           positive    con           .347   Adams, Nelson, &
                                                    Todd, 1992
Medium           positive    gamma         .42    Gefen & Straub, 1997
Medium           positive    gamma         .305   Jarvenpaa & Staples,
                                                    2000
Medium           positive    corr          .46    Minsky & Marin, 1999
Medium           positive    corr          .48    van den Hooff, 1997
User             positive    corr          .32    van den Hooff, 1997
User             negative    beta          .25    Ku, 1996
User             positive    beta          .29    Ku, 1996
User             negative    beta          .19    Ku, 1996
User             positive    beta          .18    Ku, 1996
User             negative    beta          .19    Ku, 1996
User             positive    K-S Z        1.85    Markus, 1994a, 1994b
User             positive    none           -     Rice & Shook, 1990
Range: all       positive    beta          .55    Rice, 1993
range: complex   positive    none           -     Markus, 1994a, 1994b
range: complex   positive    beta          .59    Rice, 1993
range: routine   positive    none           -     Markus, 1994a, 1994b
range: routine   positive    beta          .46    Rice, 1993
range: routine   positive    corr          .34    van den Hooff, 1997
Medium           positive    corr          .60    Trevino, Webster, &
                                                    Stein, 2000
Medium           positive    none           -     Dimmick, Kline, &
                                                    Stafford, 2000
Task             positive    gamma         .255   Jarvenpaa & Staples,
                                                    2000
Task             negative    beta          .32    Rice, Grant, Schmitz,
                                                    & Torobin, 1990
Task             negative    gamma         .476   Straub & Karahanna,
                                                    1998
Task             positive    corn          .35    van den Hooff, 1997
Task             negative    t-value      2.36    Wijayanayake & Higa,
                                                    1999
Task             positive    beta          .23    Ku, 1996
Task             positive    beta          .18    Ku, 1996
Task             negative    gamma         .089   Straub & Karahanna,
                                                    1998
Task             negative    t-value      5.45    Wijayanayake & Higa,
                                                    1999

Table 2. Range of Tasks: Items, Factor Loadings, and Scale Statistics

                                                           Cronbach's
Item                  Factor 1   Factor 2     M      SD      Alpha

Range of tasks                              30.81   6.37      .82
  (scale)
Exchanging              .682
  information
Time-sensitive          .416
  information
Asking questions        .674
Exchanging opinions     .720
Staying in touch        .668
Decision making                    .505
Confidential                       .601
  information
Resolving conflict                 .797
Getting to know                    .716
  someone
Negotiations                       .869

REFERENCES

Adams, D. A., Nelson, R. R., & Todd, P. A. (1992). Perceived usefulness, ease of use and usage of information technology: A replication. MIS Quarterly, 16(2), 227-247.

Arbuckle, J. L., & Wothke, W. (1999). Amos 4.0 user's guide. Chicago: Smallwaters Corporation.

Bouwman, H., & Neijens, P. (1991). Een meta-analyse van videotex-literatuur: Een aanzet tot een acceptatiemodel voor de consumentenmarkt [A meta-analysis of videotex literature: A model for consumer market acceptance]. Massacommunicatie, 19, 134-148.

Carlson, J. R., & Zmud, R. W. (1999). Channel expansion theory and the experiential nature of media richness perceptions. Academy of Management Journal, 42(2), 153-170.

Daft, R. L., & Lengel, R. H. (1984). Information richness: A new approach to managerial behavior and organizational design. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior 6 (pp. 191-233). Homewood, IL: JAI.

Daft, R. L., & Lengel, R. H. (1986). Organizational information requirements, media richness and structural design. Management Science, 32, 554-569.

D'Ambra, J., Rice, R. E., & O'Connor, M. (1998). Computer mediated communication and media preference. An investigation of the dimensionality of perceived task equivocality and media richness. Behaviour & Information Technology, 17(3), 164-174.

Davis, E R. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319-340.

Dennis, A. R., Valacich, J. S., Speier, C., & Morris, M. G. (1998). Beyond media richness: An empirical test of media synchronicity theory. In Proceedings of the 31st Hawaii International Conference on System Sciences (HICSS'98) (pp. 48-57). Los Alamitos, CA: IEEE Computer Society.

Dimmick, J., Kline, S., & Stafford, L. (2000). The gratification niches of personal e-mail and the telephone. Communication Research, 27(2), 227-248.

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

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

Fulk, J., Steinfield, C. W., Schmitz, J., & Power, J. G. (1987). A social information processing model of media use in organizations. Communication Research, 14, 529-552.

Gefen, D., & Straub, D. W. (1997). Gender differences in the perception and use of e-mail: An extension to the technology acceptance model. MIS Quarterly, 21(4), 389-400.

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

Jarvenpaa, S. L., & Staples, D. S. (2000). The use of collaborative electronic media for information sharing: An exploratory study of determinants. Journal of Strategic Information Systems, 9, 129-154.

Katz, J. E., & Rice, R. E. (2002). Social consequences of internet use: Access, involvement and interaction. Cambridge, MA: MIT Press.

Kinney, S. T., & Dennis, A. (1994). Reevaluating media richness: Cues, feedback and task. In Proceedings of the Hawaii International Conference on System Sciences (pp. 21-30). Los Alamitos, CA: IEEE Computer Society.

Kinney, S. T., & Watson, R. T. (1992). The effect of medium and task on dyadic communication. Proceedings of the International Conference on Information Systems, pp. 107-117.

Kraut, R. E., Rice, R. E., Cool, C., & Fish, R. S. (1998). Varieties of social influence: The role of utility and norms in the success of a new communication medium. Organization Science, 9(4), 437-453.

Ku, L. (1996). Social and nonsocial uses of electronic messaging systems in organizations. Journal of Business Communication, 33(3), 291-325.

Markus, M. L. (1987). Toward a "critical mass" theory of interactive media: Universal access, interdependence and diffusion. Communication Research, 14(5), 491-511.

Markus, M. L. (1990). Toward a "'critical mass" theory of interactive media. In J. Fulk & C. W. Steinfield (Eds.), Organizations and communication technology (pp. 194-218). Newbury Park, CA: Sage.

Markus, M. L. (1994a). Electronic mail as the medium of managerial choice. Organization Science, 5(4), 502-527.

Markus, M. L. (1994b). Finding a happy medium: Explaining the negative effects of electronic communication on social life at work. ACM Transactions on Information Systems, 12(2), 119-149.

McGrath, J. E., & Hollingshead, A. B. (1993). Putting the group back in group support systems: Some theoretical issues about dynamic processes in groups with technological enhancements. In L. M.

Jessup & J. S. Valacich (Eds.), Group support systems: New perspectives (pp. 78- 96). New York: Macmillan.

Minsky, B. D., & Marin, D. B. (1999). Why faculty members use e-mail: The role of individual differences in channel choice. Journal of Business Communication, 36(2), 194-217.

Oliver, E, Marwell, G., & Teixeira, R. (1985). A theory of critical mass I. Interdependence, group heterogeneity and the production of collective action. American Journal of Sociology, 91, 522-556.

Rice, R. E. (1993). Media appropriateness: Using social presence theory to compare traditional and new organizational media. Human Communication Research, 19(4), 451-484.

Rice, R. E., & Case, D. (1983). Electronic message systems in the university: A description of use and utility. Journal of Communication, 33, 121-152.

Rice, R. E., & Gattiker, U. (2001). New media and organizational structuring. In F M. Jablin & L. L. Putnam (Eds.), The new handbook of organizational communication (pp. 554-581). Thousand Oaks, CA: Sage.

Rice, R. E., Grant, A. E., Schmitz, J., & Torobin, J. (1990). Individual and network influences on the adoption and perceived outcomes of electronic messaging. Social Networks, 12, 27-55.

Rice, R. E., Hughes, D., & Love, G. (1987). Usage and outcomes of electronic messaging at an R&D organization: Situational constraints, job level and media awareness. Office: Technology and People, 5(2), 141-161.

Rice, R. E., & Shook, D. (1990). Relationships of job categories and organizational levels to use of communication channels, including electronic mail: A meta-analysis and extension. Journal of Management Studies, 27(2), 195-229.

Rogers, E. M. (1983). Diffusion of innovations. New York: Free Press.

Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecommunications. London: John Wiley.

Straub, D. W., & Karahanna, E. (1998). Knowledge worker communications and recipient availability: Toward a task closure explanation of media choice. Organization Science, 9(2), 160-175.

Suh, K. S. (1999). Impact of communication medium on task performance and satisfaction: An examination of media-richness theory, Information & Management, 35, 295-312.

Trevino, L. K., Daft, R. L., & Lengel, R. H. (1990). Understanding manager's media choices: A symbolic interactionist perspective. In J. Fulk & C. W. Steinfield (Eds.), Organizations and communication technology (pp. 71-94). Newbury Park, CA: Sage.

Trevino, L. K., & Webster, J. (1992). Flow in computer-mediated communication; Electronic mail and voice mail evaluation and impacts. Communication Research, 19(5), 539-573.

Trevino, L. K., Webster, J., & Stein, E. W. (2000). Making connections: Complementary influences on communication media choices, attitudes and use. Organization Science, 11(2), 163-182. University of Texas. (2002). Structural equation modeling using AMOS: An introduction. Retrieved January 2002 from http://www.utexas.edu/cc/stat/tutorials/amos/index.html

Valacich, J. S., Mennecke, B. E., Wachter, R. M., & Wheeler, B. C. (1994). Extensions to media richness theory: A test of the task-media fit hypothesis. In Proceedings of the Hawaii International Conference on System Sciences (pp. 11-20). Los Alamitos, CA: IEEE Computer Society.

van den Hooff, B. J. (1997). Incorporating electronic mail: Adoption, use and effects of electronic mail in organizations. Amsterdam: Otto Cramwinckel.

Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal and hyper-personal interaction. Communication Research, 23(1), 3-43.

Wijayanayake, J., & Higa, K. (1999). Communication media choice by workers in distributed environments. Information & Management, 36, 329-338.

Bart van den Hooff (Ph.D., University of Amsterdam, 1997) is an assistant professor in the Amsterdam School of Communications Research ASCoR at the University of Amsterdam. Jasper Groot (MA, University of Amsterdam, 2003) is an internal communication coordinator at Demon Internet, Amsterdam. Sander de Jonge (MA, University of Amsterdam, 2002) is a consultant at LogicaCMG, Amstelveen. Correspondence concerning this article should be addressed to Dr. Bart van den Hooff, Assistant Professor, University of Amsterdam, Amsterdam School of Communications Research ASCoR, Kloveniersburgwal 48, 1012 CX Amsterdam, the Netherlands: e-mail: b.j.vandenhooff@uva.nl.