Attraction to Computer-Mediated Social Support (Walther & Boyd)

A revised version of this work appears in C. A. Lin & D. Atkin (Eds.), 2002. Communication technology and society: Audience adoption and uses (pp. 153-188). Cresskill, NJ: Hampton Press.

Although social support is traditionally conceived as taking place in close personal relationships via face-to-face interaction, social support is also exchanged in numerous virtual communities through computer-mediated communication across the Internet and other electronic networks. These venues substantially alter the communication process and the sources of support upon which participants may draw. This research reviews social support and computer-mediated communication literature to identify potential reasons that make electronic support attractive, reflecting disadvantages of traditional support, advantages of electronic support, and their complementarity. Based on this review, an empirical study among Usenet news participants confirmed several dimensions of electronic support attractiveness: Social Distance, Anonymity, Interaction Management, and Access.

One of the most interesting and controversial phenomena in the recent explosion in computer-mediated communication (CMC) and Internet use involves the development of "virtual communities" in which people meet, share interests, and exchange social support via text-based messages on computer networks. As Rheingold (1993) describes them, "virtual communities are social aggregations that emerge from the Net when enough people carry on those public discussions long enough, with sufficient human feeling, to form webs of personal relationships... " (p. 5). While the "net" has its share of idle and banal chat, it is also a viable and vital conduit for valuable information and interpersonal affect. Although critics have raised questions about the "communal" nature of the Internet and its sub-networks, and the validity or utility of these personal relationships (e.g. Critical Art Ensemble, 1996; Fernback & Thompson, 1995; Kraut et al., 1998), there is a great deal of social support activity taking place on-line, appealing to its users.

Popular descriptions of Internet capabilities, and limited previous research about CMC support, provide numerous possible benefits that CMC may render in the support process. Up to now we do not know whether any or all of these may be at the heart of users' attraction to the medium, and whether it is a single benefit or the array of benefits that make the medium useful. In contrast to a view of unlimited bounty and a bonanza of benefits from on-line support, research on traditional social support offers "two well-established empirical generalizations," according to Ford, Babrow, and Stohl (1996, p. 189): "First, not all ostensibly supportive social interactions are experienced as supportive...Second, the supportee's perception of the quality or substance of social support is a better predictor of successful coping than the sheer number or quantity of support at one's disposal." For these reasons, a systematic exploration into the perceived attractions to online social support is especially warranted.

We will examine questions previously raised asking why face-to-face (FtF) personal relationships may offer disadvantages for social support in some contexts. In addition, we will revisit arguments about the advantages of weak-tie networks, as happen to occur over computer networks, in their affordances for support seekers and givers. We also consider the special capabilities of computer-mediated communication, and how these features also facilitate social support processes. This review led to an empirical investigation among Usenet support users, attempting to assess the social and technological dimensions of online support that make it attractive to users, the results of which conclude this study.

Social support communication is traditionally considered to be the exchange of verbal and nonverbal messages conveying emotion, information, or referral, to help reduce someone’s uncertainty or stress, and "whether directly or indirectly, communicate to an individual that she or he is valued and cared for by others" (Barnes & Duck, 1994, p. 176). Conventional social support is thought of as taking place most often within established, multi-dimensional, and primarily dyadic close personal relationships (Cutrona & Suhr, 1992; Leatham & Duck, 1990), although it may also take place in therapeutic relationships (e.g. Robinson, 1988) as well as through informal social networks (Albrecht & Adelman, 1987).

In growing numbers, social support is being exchanged via computer-mediated communication, in relatively large networks among people who do not know each other and do not communicate FtF. The communication process is thus changed in some dramatic respects. While traditionally one may turn to a target person in seeking support, in electronic interaction support seekers go to virtual "spaces" where it is unlikely that one knows anyone or has heard about any person in particular, but where it is probable that messages left by other people with certain expertise and common experience are found. Unlike FtF support relationships, most CMC support exchange begins by discussing the topic of concern, immediately and often in very personal terms, rather than leading up to these concerns after establishing relationships based on other commonalities: Electronic social support develops among strangers whose only connection is their common affliction or concern over a source of personal discomfort, in a "uniplex," rather than "multiplex," relationship.

Electronic Venues. There are now many electronic venues in which people exchange social support solicitations and responses. These include various "spaces" available through the world-wide Internet, and through proprietary commercial on-line systems. In these cases, the communication networks were established for other reasons, and support spaces evolved. Additionally, some private electronic support systems have developed with the specific and exclusive purpose of providing peer-to-peer and/or expert advice on personal, medically-related, or other issues (e.g. Scheerhorn, Warisse, & McNeilis, 1995). Steve Harris (1996) keeps an up-to-date list on the World Wide Web of hundreds of on-line sources for "emotional support" on his World Wide Web site, Emotional Support on the Internet. These sources include Usenet newsgroups, electronic mailing lists, discussion groups available through commercial CMC providers such as Prodigy and America OnLine, as well as real-time chat sessions through Internet Relay Chat and several "MUDs" (multi-user discussions). Harris lists these services as venues for emotional support as well as information support, devoted to discussion of a wide range of illnesses, social problems, psychological issues, and other topics around which users may seek company. They are places in which to share feelings and experiences.

Most dedicated electronic support spaces work via asynchronous communication. These venues primarily work on a "bulletin board"- type arrangement of messages (see Rafaeli, 1986; Rapaport, 1991). That is, users post messages that are stored on-line for some period of time for others to read, with readers logging in and responding at different times. Users need not be joined at the same time in such systems. Alternatively, mailing lists ("distribution lists") form another method to communicate asynchronously with many others. List systems involve multiple-addressing remailers, so when someone sends a message to the one official electronic mail address for the group, a host computer system re-sends the mail message out to everyone who has subscribed to the list. Whereas a "bulletin board" keeps messages in virtual space until perused, list-based systems replicate the messages to each subscriber’s electronic mailbox.

In addition to such asynchronous systems, fewer, synchronous "chat" spaces, allow participants to type messages to one another simultaneously and read them in "real time." Hybrid systems exist as well; one such system is the SANCTUARY multi-user domain—a chat space for survivors of physical abuse--that also contains an asychronous bulletin board system. The chat capacity provides a valuable environmental ambiance: according to Moursun (1997, p. 61), "in spite of having only words and one's own imagination to create an environment, the sense of being in a familiar place, interacting with people you know and care about, is astonishingly vivid in SANCTUARY." The bulletin board facilitates other social support exchanges: in order of frequency; expressions of companionship (talking about problems, sharing experiences), information, positive feedback, motivational support, and belongingness (Moursun, 1997).

Perhaps the largest and most accessible asynchronous venues are the support spaces transmitted via Usenet. While Usenet was not originally part of the Internet, it is now commonly accessible through the Internet in such a seamless way that they are almost indistinguishable to users. Usenet feeds are sorted into topical "newsgroups," with topics now reaching the ten thousands, ranging from scientific discussions; fan clubs for sports teams, music groups, and soap operas; technology user groups; and sound files and picture files, from television themes to erotica. Within each newsgroup, people post written messages, and others read and sometimes reply. Chains of replies (that is, messages with the same subject line) are known as "threads," and many readers’ computer systems sort these messages by thread and then by date and time of posting. Thus, following a thread, a continuous discussion takes shape.

Like the Internet proper, participation in these discussion groups may be international in scope. Among these are several newsgroups particularly focused on support, often indicated by name (e.g. alt.support.depression). Some of these support groups are estimated to be very heavily trafficked, with thousands of readers, if not writers, participating. Table 1 lists the ten most popular Usenet support-related groups and their readership estimates as of March, 1995 (after which time the data were no longer updated for public dissemination; see also Galagher, Sproull, & Kiesler, 1998).

One should not infer that social support takes place only within electronic groups with names making support an obvious theme. Indeed, conversations about any kind of problem may raise issues of insecurity and need for information and comfort. For instance, in the alt.sex related Usenet groups, so maligned as prurient by the press and politicians, it has been found that

This is also not to say that the entire Internet is nothing but a therapy couch, but rather that various types of support arise in sometimes unexpected places.

Functions. It is simple to find examples of social support in many forms on public networks. While several typological approaches have been made to defining the scope of support activities (see for review Barbee & Cunningham, 1995), we will describe social support by means of the list Cutrona and Suhr (1992) offer. These activities include informational support, emotional support, esteem support, tangible aid, and social network support. This typology is not only sound, but it also provided a useful classification scheme for Braithwaite, Waldron, and Finn (1999) to content analyze the types of support messages in an analysis of an electronic support group for the disabled, findings of which are also reflected below. Readers are encouraged to verify for themselves the extent of each activity taking place in on-line venues which they may observe.

Informational support comes in the form of advice, factual input, and feedback. Such messages may discuss symptomatology, medications, legal issues, etc. Information support helps people make decisions and attributions, and judge actions. According to Braithwaite et al.'s (1999) examination, this type of support message was one of the two most frequently exchanged on-line. The other was emotional support, which is given through expressions of caring, concern, empathy, and sympathy, according to Cutrona and Suhr (1992). These messages are less about decisional issues and more about psychological well being. Examples of emotional support would include statements of affection, emotional understanding, and statements geared towards relieving pain and stress.

Esteem support comes through expressing admiration and/or understanding another's worth. This can come through complimentary statements about another's skills and abilities, or through expressions of genuine admiration and telling someone that s/he is a good or normal person. This type of support appears highly personal and is psychologically oriented. Tangible aid is support in the form of actual physical assistance. It provides needed goods and services. An example of tangible aid would be buying groceries for a friend who was sick. We should expect that this type of support is very frequent among computer-mediated support groups, whose members tend not to share the requisite geographic proximity needed to exchange goods or physical services, other than typing to one another. Indeed, Braithwaite et al’s (1999) analysis found this type of support among the least frequently exchanged. At the same time, it is not impossible. Rheingold (1993) describes how one user community activated enough expertise and finances to mount a medical airlift for an injured colleague, and how even the beginnings of "the Well"--network home to his "virtual community"--relied on financial contributions and credit from early users to upgrade and maintain the system. While such anecdotes may stand out because of their very rarity, they point out that tangible aid may indeed be mobilized, even if it is not common, via electronic networks.

The final type of support, social network support, involves directing or referring someone to a another person or group of people who share a common set of experiences or expertise. This type of support allows an individual to feel tied in to a larger community.

Electronic social support entails some changes in the delivery of these functions. For instance, the tangible aid function is obviously mitigated when there is no shared physical space in which to exchange tangible artifacts and services. Other functions may emerge which are facilitated by network dynamics, as will be shown. While the basic functions of social support change somewhat in CMC, the means by which people can access these functions have radically expanded.

Group Support. In addition to personal, dyadic relationships, social support is exchanged FtF in support groups. Since electronic social support is exchanged in network "spaces" involving numerous participants, it is therefore instructive to consider how this milieu parallels a support group or therapy group. According to Braithwaite et al. (1999, p. 126), "the majority of face-to-face self help groups have also been duplicated online." Off-line groups of this nature are formed explicitly to help their members cope with and/or solve their common concern. Robinson (1988) suggests several characteristics of these groups, many of which are evidently occurring on-line, and as we have seen informally in the comments of our respondents, including common experience of the members: all members of self-help groups face the same problem issue; mutual help and support: the reassurance that groups meet regularly and provide mutual aid; the helper principle: those providing support to others may be the ones who, in doing so, actually derive the greatest benefit; differential association: emphasizes a healthy self-concept, and hastens members’ separation from previous deviant self-concepts; collective belief: members may draw validation from one another; and the importance of information: like information support, above, promoting "factual understanding of the problem as opposed to intrapsychic understanding" (p. 119).

In several ways, support groups or self-help groups seem most comparable to that which occurs in electronic social support, yet differences, again, exist. Primarily, the aspect of regular FtF meetings in these support groups entails a drastic difference. To have regular meetings implies spatio-temporal and sociological arrangements that are not relevant in electronic space. Face-to-face support groups assume that participants can attend a common location, both because they are geographically close, and within a geographic area have sufficient community to achieve critical mass; that they are ambulatory and mobile; and that they can schedule such meetings into their lives. These preconditions are not insignificant. And they are negated by electronic social support. According to Galagher et al. (1998), online support groups function much the same way as FtF support groups, but with the additional benefits of greater access (to be discussed below), and greater confidentiality:

Aside from formal groups, however, day-to-day FtF social support takes place within personal relationships. In the following, we will examine some disadvantages posited to accompany support exchange in close personal relationships, identify how CMC support brings support partners in contact with people outside such relationships, and how these different types of connections affect support exchange.

While the feeling of social support may be a cornerstone to psychic health and functioning, some commentators have pointed out that the exchange of social support ironically may bring with it serious harmful effects to autonomy and relational well-being. In particular, according to La Gaipa, support exchange encapsulates a struggle between autonomy and dependence, both for providers as well as receivers of social support. When the exchange of social support leads to dependence, unhealthy outcomes result. It is our contention that support exchange in electronic venues, and the types of social systems they encompass, attenuate some of the negative effects liable to appear in FtF personal relationships. In the following, we review several aspects of the "dark side of support" (Albrecht, Burleson, & Goldsmith, 1994), and contrast them to electronic support, leading to hypothetical attractions to on-line support.

Mismatching. Difficulties can arise when efforts to help one in need are not well matched to the recipient’s needs, and these difficulties can affect relationships adversely (see for review Goldsmith, 1992). According to La Gaipa (p. 124); "when, for example, members of one’s social network have no personal experience with a particular crisis, such as cancer, their efforts to be helpful are likely to be strained and clumsy." Different types of support may be most helpful at different stages in a problem, yet inexpert sources may be more or less directive or blindly reassuring than needed at various points in dealing with a problem. Should they persist, their efforts at assistance may be ineffective, annoying, or counterproductive.

Close relationship partners may be likely to offer unqualified acceptance of a member with a problem. If that partner is seeking accurate feedback for social comparison, close relations may be less forthcoming than less intimate associates. The obligation to spare feelings of a close partner may be mitigated among less closely-tied participants in electronic support. In many cases, friends may be of no help in relation to a particular problem one may experience. As friendship relations tend to feature normatively bounded levels of intimacy, time, and commitment, friends may be reluctant or inappropriate sources to go into much depth in the case of a personal issue. When they do, they may be ineffectual (La Gaipa, 1990). As on-line social support is exchanged among those with similar experiences or concerns, without the multiple concerns and filters that close relationships impose, these issues may be ameliorated.

Gender role, also, seems to affect what kinds of support messages are given and what will be perceived as effective. Research by Mickelson, Helgeson, and Weiner (1995) found that women provide more emotional support messages than do men, and moreso in mixed-sex dyads than in same-sexed interactions, in FtF conversations. And whether preferred or not, more negatively-oriented and trivializing messages are directed toward males who discuss their problems FtF. In the on-line social support group examined by Egdorf and Rahoi (1994), however, very little communication was coded as negative. It is unknown whether the medium alters the kinds of messages that are produced by men and women, or whether they adhere to the same patterns as they do FtF, and gender demographics account for the overall tone of support. Our sample of respondents was disproportionately female, yet we cannot infer with certainty whether that ratio describes the population of social support participants on-line or across groups. While Mickelson (1997) also found a higher proportion of females (52.3%) than males (47.7%) online in a study of electronic social support for parents of developmentally disabled children, a comparison group of non-electronic support group members was 98% female. Gender effects may pervade social support.

Costs of giving. To provide social support to others may risk "high costs associated with caring," according to La Gaipa (1990, p. 126), or a "drainage of resources, particularly when the need for support is great or extends of a long period of time," according to Albrecht et al. (1994, p. 433). These commentators reviewed several studies showing that those with numerous network links and who are more supportive may experience great stress and emotional strain. If the provision of support becomes extended and emotionally demanding, "caregiver burden" may result, leading to numerous mental-health problems such as depression, sleeplessness, anxiety, and frustration. Other symptoms may include emotional exhaustion, impatience, and irritation. These outcomes would appear to be counterfunctional to the support relationship in the long run. Costs of receiving. One of the great costs in soliciting and accepting social support in FtF personal relationships has to do with the presentation of self and the self-evaluations that follow the admission of a problem or uncertainty. According to Albrecht et al. (1994, p. 433), asking for help may make people "appear weak or less competent"; undesirable information may be disclosed in seeking support; and "the problem for which help is needed results in stigmatization." (We will address the special affordances of CMC for stigma management below.)

Another cost of disclosing has to do with the judgments providers may make about other social actors related to the problem. La Gaipa gives the example of a mother’s hostility toward her daughter’s husband should the daughter seek counsel about a marital issue. This is not the case of a support receiver wishing to avoid negative social judgments, but wishing to forestall negative judgments of other close network members. Such paradoxes can inhibit potential recipients from seeking support in the first place.

A further cost of receiving social support is the potential to be called on to reciprocate. There is a norm of reciprocity in close relationships. To deny reciprocation is to violate relational equity; yet to be "on call" to provide support when one is ill-disposed, unable, or inexpert, may be equally disturbing (see for review Albrecht et al., 1994). The potential reciprocity debt, like the obligation to provide support in close relationships mentioned above, may be an unattractive aspect of FtF support exchange in close personal relationships.

However, more costly may be the relational dependency that may accompany a support seeker's admission of concern, fear, or incompetence, and the acceptance of advice from others (La Gaipa, 1990, p. 136). So resentful of this reduced autonomy accompanying support needs, persons may reject otherwise useful advice in order to restore a balance of relational power.

In support among strangers, aspects of CMC social support may ameliorate a number of the disadvantages associated with support in close relationships. For one thing, one need not build a relationship to a point of a problem disclosure as one may in FtF relationships. In CMC support venues, it is expected that one is seeking support. Relational dependency and obligation are less likely to be a factor in this environment. It is quite easy to extricate oneself from the support group--one simply quits writing, and leaves the business to the many others whose messages fill the void.

One way to view the way that CMC reduces the costs of support through personal relationships is to assume that not only the medium, but the sources of support have different characteristics. One contention of the present work is that the exchange of social support in virtual support "spaces" on computer networks constitute weak tie, rather than strong tie, relationships (as has also been discussed by Egdorf & Rahoi, 1994; Wellman & Gulia, 1999). These are also relationships within which interactions typically revolve around more than one kind of topic, that is, they are multiplex. Strong ties are typified by friends, loved ones, work associates with whom one also socializes, confidantes, family, etc. It is these kinds of relationships in which social support is usually exchanged, according to most sources (see for review Albrecht & Adelman, 1987; Leatham & Duck, 1990).

Weak tie relationships differ from strong ties in several structural and social respects. Structurally, in off-line realms, weak ties are not originally known to a person directly, but may be known through secondary associations (e.g., a friend of a friend) in a social network, or through some social agency. Weakly tied persons are themselves embedded in other social networks. In the case of electronic support networks, we would argue that the association to other people is offered by the agency of the electronic space rather than a secondary social link; people come to the space for support, rather than coming to meet a specific person. In social terms, the overlap of weak ties onto one’s primary network of close ties is relatively small. That is, while our closest associates or strong ties are often likely also to know one another, weak ties are less likely to know and interact with each other, or with our primary network members. As a result, when information rests with a weak tie, it is less likely to be commonly known among strong ties.

Acknowledging that contact with weak ties is often accomplished via mediated communication channels such as telephones or electronic bulletin boards, Adelman et al. attribute special facilitation to restricted channels. They assert that "Restricted communication channels...provide individuals with a comparatively anonymous way to disclose highly personal or potentially embarrassing information" (p. 133).

Expertise. It is the dynamic of the separation of close from weak networks, and the greater number of weak links to strong ones, which undergirds the "strength of weak ties" hypothesis (Granovetter, 1973). This contention in its simplest form is that there is often more information to be found in the extended network than in the primary one. This is due to several factors. There are more numerous individuals available through weakly-linked associations (secondary links, tertiary links, etc.) than there are within one’s primary network of close ties. Moreover, there tends to be greater diversity among weak ties than close ones; close ties tend to be homogenous in many respects, while different people know more different things. Thus locating and using expertise is highly facilitated by the technology as well as the loose confederation of users in a single support space, where, just by numbers alone, there may be more expertise than in one’s primary network. Thus, the following research question:

Stigma management. By stigma management we refer to the discussion of incidents or topics with others in such a way that people whom one does not want cognizant of one’s concern with that issue, do not become aware of it (Thompson & Siebold, 1978). For instance, it may be an important and legitimate question to ask, "does anyone else experience sexual dysfunction as a side-effect of certain antidepressants?", yet it is not a question likely to be asked in the office meeting, nor desired to be gossiped about there either. It is a common question in depression-related electronic support groups, however. Information which is imparted to weak links is less likely to feed back to one’s close network as extensively as information shared within a close, more densely-connected link. As Adelman et al. (1987, p. 131) surmised, "This network distance enhances perceived anonymity and allows people to seek information and support without having to deal with the uncertainty of how those in primary relationships might respond." In this way, they argue, weak links facilitate "low-risk discussions about high-risk topics" (p. 133). Thus the following research question is posed:

Validation and intimacy. Another benefit of weak tie interaction is its capability to foster community, according to Adelman et al. Two processes relate to this dynamic. First, they allow participants to conduct "reality checks" by comparing their own reactions and feelings to others who are different but who have had similar experiences. To detect similar reactions among others would help attribute normality to one’s reactions. Indeed, in on-line support, one may observe numerous disclosures about critical events and participants’ psychological and physical reactions to them. Many messages begin or end with "me, too" (see also Galagher et al. 1998). As Adelman et al. project, "Self-evaluation is facilitated by comparison to weak ties because they provide a greater variety of information and thus a better ability to judge how typical or normal our own behavior is" (p. 135).

A secondary effect seems to accompany this level of self-disclosure, as a great degree of intimacy seems to develop through electronic social support. Here, however, it seems to be the result of a different dynamic: To discuss these conventionally personal issues requires some serious self-disclosure, and to observe self-disclosure may be to infer significant intimacy. Validation by others often follows the disclosure of personal issues, manifesting highly rewarding outcomes from what Millar and Rogers (1976) call a "vulnerability pattern," on-line. Among weak ties, "discussion about intimate topics occurs...within highly limited contexts" according to Adelman et al. (1987, p. 130). Here our research asks:

The strength of weak ties has been examined directly in relation to computer networks contexts other than social support. A network analytic study by Feldman (1987) established that an electronic network increased communication among both familiar and unfamiliar members within an organization. Focusing on two primary characteristics of such networks--asynchrony and multiple addressability--she notes that electronic message exchange has a much lower cost in time, effort, and resources, than other means of communication. A byproduct of this affordance is the easy emergence of mailing lists focused on specific interests, especially to those unknown directly to one another. Feldman argues that these lists are driven by interests, some of which are not business-related and therefore are less likely to emerge in typical organizational conversations constrained by place, time, and purpose. Because these extra-organizational interests are as likely to be shared by people who do not share place, tasks, or other topical interests, they lend themselves to weakly tied electronic links, yielding "messages between people who do not know one another, who only communicate electronically, who would not have exchanged information if there were no electronic mail and who are neither organizationally nor spatially proximate" (Feldman, 1987, pp. 88-89), the crux of her argument that electronic networks offer weak ties.

Feldman’s analysis of organizational email messages lends support to her contention. Although most messages originated from members at least slightly known to one another, such messages were predominantly (64%) related to work topics. Among the smaller set of weak tie messages, however--messages from users not at all known to one another with no other link than computers--the proportion was reversed, with 65% of messages related to non-work topics, including personal issues, political, recreational, or humorous remarks. These messages were communicated almost exclusively by distribution list rather than e-mail addressed to identifiable recipients. That is, among these potentially weakly-linked participants, CMC played a significant role: The new electronic networks connect people who do not know one another, and who are spatially and organizationally distant. Distribution list messages--messages sent to "topics" rather than "persons"--predominate these interactions.

Constant, Sproull, and Kiesler (1996) also conducted research on the utility of weak versus strong ties available through e-mail distribution lists, in a large multinational corporation. Their research was limited to work-related information exchange, and the technical questions and answers commonly sent via lists crossing departmental and geographic distances. Many questions were of the type, "does anyone know...?", and most replies--particularly the most helpful ones--were returned by users who had no personal acquaintance with the question-asker. Constant et al.’s research helps identify the way computer networks expand one’s sources of expertise within an organization. But, by analyzing only those messages that were posted to the distribution lists, the study does not therefore inform the question of whether users actually preferred closer ties to weaker ones. No effort was made to determine whether users had in fact exhausted their closer associates’ availability or expertise before resorting to electronically-linked sources. In contrast, in relation to social support and the dynamics of weak ties articulated above, we would expect some distinct advantages of weak ties to make on-line support preferable to asking one’s closest associates even if they did have potential answers.

Overall, CMC is a conduit, a channel to people with the beneficial characteristics of the interaction embedded in who the targets are, rather than in the manner of the interaction per se.

Nevertheless, different aspects of the CMC support process may affect the temporal and content aspects of communication. Some simple technical properties of CMC use are identified which may also facilitate social support exchange. Moreover, the hyperpersonal perspective of CMC may help to identify more complex CMC benefits. These aspects, taken together, suggest dimensions of CMC exchange that may make on-line support attractive and beneficial to its users.

Access. Easy access to electronic support appears to be the first advantage articulated by many commentators (e.g. Braithwaite et al., 1999; Egdorf & Rahoi, 1994; Rheingold, 1993). As in the case of many other uses of CMC, there are aspects of the medium which overcome temporal and geographic barriers usually encountered in FtF communication, or communication via other personal media. Temporally, Internet-supported support spaces, as well as those on most commercial electronic services, are available by modem connection twenty-four hours a day, seven days a week. Potential support providers are never awakened or disturbed by a message placed on such a discussion board, since users often use the system to post queries or responses asynchronously, without expectation for real time interaction. Support providers, in turn, reply when they are best able to. While these are not instantaneous conversations, they build over time, and such interaction may gradually achieve the same interpersonal intimacy and depth as FtF meeting provides (Walther & Burgoon, 1992). Geographically, interactants half way across the globe may build conversations, and one’s own location is irrelevant (and in many cases unknown) as long as one can log in (see Mitchell, 1995).

This differs substantially from FtF support, which is not available anyplace, any time. In dyadic, off-line social support, a potential provider may not be available at the precise time when a support seeker needs help (Leatham & Duck, 1990). The tradeoffs between temporal and geographic barriers may be too complex off-line: Someone who is temporally available to talk about problems may not be in a reasonably-accessed place, off-line. A critical mass of experienced or expert sources may not exist within one’s normal geography. Scheerhorn et al. (1995), for instance, detail the barriers to FtF support faced by hemophiliacs and their families in small rural Ohio towns, who may be the only such members of their geographic community and have no access to social support without electronic assistance, leading to the establishment of the HIGHNet—the Hemophiliacs in Good Health Network. And when sources of support do exist within geographic proximity, temporal issues may interfere with access. As a participant in the Usenet alt.sexual-abuse.recovery group wrote about trying to access FtF support,

The Internet, on the other hand, is not limited by capacity. CMC interactions offer "a rolling present," (Hiltz, 1992) in which time and location are discretionary. As CMC groupware stores messages for perusal at the discretion of the user at least for some period of time, and participants thus may come and go when they choose and work at their own pace. Conversational coherence (McGrath & Hollingshead, 1993), is built through the use of threads, and quoting previous messages (sometimes automatically) to which subsequent participants wish to add comments (Black, Levin, Mehan, & Quinn, 1983; Walther & Tidwell, 1995). In this way meaningful dialogues emerge.

While most electronic support interaction is asynchronous, people all over the network, or the world, may be available with help at a moment’s notice (as illustrated in Rheingold, 1993, chapter 1), or one may peruse what others have said on related topics regardless of whether one actively queries or not, by either of two methods. One method is by reading answers to "Frequently Asked Questions," or FAQ lists, which are posted periodically in a variety of Usenet groups for discretionary reference (Hersch, 1996). Another is by "lurking."

Lurkers. A relatively unique aspect of on-line support has to do with "lurking" (Egdorf & Rahoi, 1994). Lurking (and lurkers who do it) refers to reading messages posted by others on electronic spaces, without also posting one’s own messages or in any way signaling one’s vicarious observation. Some estimations place the number of lurkers to active posters as a hefty majority of users who frequent a particular group. It is not only acceptable in this context not to "speak" when one has nothing to say, it is encouraged that those who are new to a particular group spend a considerable period observing the interaction of established members (Moraes, 1996). This observation period allows two things: to learn the norms and "netiquette" of the group, and to see if answers to "frequently asked questions" are already available in some electronic file which is periodically posted. It is in a sense very efficient to see if anyone is discussing one’s concerns already (with perhaps a week’s worth of replies already posted) without the additional effort of writing them up, posting, and returning over some period of time to see what responses are have been generated.

In FtF personal relationships, of course, it is almost inconceivable to listen only to a support exchange, or to eavesdrop surreptitiously on others’ intimate self-disclosures or discussions of serious concerns. Yet according to Mickelson (1997, p. 172), online lurkers "can obtain comparison information or vicarious support without having to disclose anything about themselves…(and) obtain validation for their feelings of stigma without having to communicate those feelings to others." On-line it is acceptable, and preferable in some ways to seeing many people raise the same or similar question over and over (McLaughlin, Osborne, & Smith, 1995).

Production blocking. Research in the domain of electronic group decision-making has pointed out another advantage to on-line interaction: the ability of such systems to reduce "production blocking" relative to FtF communication (Diehl & Stroebe, 1987; Valacich, Dennis, & Nunamaker, 1992). In CMC, participants may type their comments at any time, regardless of what any other participant may be doing. Unlike FtF groups and dyads, there is no need to take turns or manage conversations in the usual manner. As a result, participants need not suppress or forget their comments as they might during a delay that could even preclude a FtF utterance. There is no limit on the time a group has for a meeting in CMC; everyone gets a chance to speak. Thus the research question:

Hyperpersonal aspects. While much early research on CMC characterized it as "cold" relative to FtF interaction, recent research has found it quite tenable for establishing impressions and interpersonal relations (Walther, 1993; Walther & Burgoon, 1992). Indeed recent work finds that previously unknown interactants are able to achieve greater intimacy via CMC than they do in parallel FtF interaction (Walther, 1996, 1997). While this outcome is not universal, there are certain affordances of CMC and some combinations of social context factors that are known to affect relational communication and interpersonal perceptions in extraordinarily positive directions.

The hyperpersonal model of CMC suggests that extraordinary intimacy levels depend on activation of psychological processes related to receiving, sending, channel, and feedback in ways that CMC enhances (Walther, 1996). When message receivers assume that they are part of a group of similars, messages stimulate idealization of group and interpersonal perceptions (Spears & Lea, 1992). CMC hides from one another differences that would be more easily detected and more salient through FtF interaction. As senders, CMC users engage in selective self-presentation, imbuing in their messages highly preferred personal and relational cues. As a written medium, these cues are more controllable than would be the case in the more spontaneous FtF interaction. The impressions these may garner may be especially friendly, knowledgeable, pathetic, or otherwise, as the sender may prefer; they are arguably more deliberative and effective than one might manage off-line.

Another component of the model addresses the affordances of the asynchronous channel in the above processes. This has particular importance for electronic social support. As CMC does not demand real-time utterance, participants may craft their messages and their self-presentations with greater care than may be possible in spontaneous FtF settings. They may stop and think, edit, rewrite, even abort and re-start a message that is not to their liking. Additionally, Walther (1996) argues, CMC users need not maintain any other expressive system than writing (i.e. they need not monitor their gestures, facial expressions, voice, or physical appearance). They may in turn re-allocate greater cognitive resources to the articulation of their desired message. These possibilities, along with the relative solitude accompanying CMC use, have been found to heighten users’ self-awareness (Matheson & Zanna, 1988). These advantages allow CMC users to present whatever degree of vulnerability or wisdom that their verbal skills are able to muster. In this way it is easy to imagine that requests for support, and its provision--especially for emotional support--may be enhanced.

By contrast, it is easy to imagine that FtF conversants may not have at their disposal relevant references, addresses, or contact information during a specific, impromptu support conversation. On-line, however, since the "conversation" does not continue until each participant is ready, there is no apparent interruption while one searches for external information (M. Altman, Walther, & Edelson, 1996). Hence, the asynchronous and electronic nature of CMC allows time to craft a good request or a good response. Thus, the actual quality of support may be increased. This, in turn, may enhance interpersonal perceptions; as Rice (1987) has suggested, one's image in cyberspace is likely to be a function of the quality of one's information, and the character one reveals through one's writing. Based on these aspects of the hyperpersonal model we ask:

Anonymity. CMC, over the Internet to Usenet news groups, can offer absolute, objective anonymity, which many social support participants use in discussing their situations. Soliciting support or reflecting common experience in the provision of social support often reflects a personal concern, that many might consider to be a problem, inadequacy, maladaptation, weakness, or other stigma, which would not do well to be attributable to their off-line personal or professional identity. Participants often disclose extremely personal details on-line in seeking social support, and others sometimes respond with equally intimate details as they relate similar experiences or traumas of their own. Anonymity, while infrequently used, may be one of the most dramatic things about CMC in the realm of social support.

Anonymity may be achieved by several means. Some commercial on-line systems allow users to adopt "screen names" which are marginally traceable to one’s actual identity. The Internet goes a step farther: using one or more of the freely available anonymous remailer systems, one can post messages to Usenet newsgroups in such a way that one’s real username and e-mail address are not apparent and virtually untraceable (Bacard, 1995, 1996). While such systems have raised concerns about information security and a cloak of anonymity for subversive communication (see Lee, 1996), they remain available in the difficult-to-regulate Internet.

It is through low sociometric density or actual anonymity, says Rheingold (1993, p. 27), on-line venues are "a place that people often end up revealing themselves far more intimately than they would be inclined to do without the intermediation of screens and pseudonyms." Thus the research question:

An original questionnaire was developed and distributed by electronic mail to people who had posted messages to selected Usenet support groups. Based on the twelve research questions and background reviewed above, thirty-three scale items were developed reflecting hypothetical aspects of the electronic support advantages, and disadvantages of FtF support. Items were constructed with a mixture of positively- and negatively-worded items, scaled as five-interval Likert-type items.

The questionnaire was formatted for e-mail in such a way that respondents could simply execute a "reply" command to return it with their scores. They were asked to type their 1 to 5 response to the left of each item as it appeared on their screens. A key with the semantic values for each number 1 to 5 was repeated twice, as it might be at the top of a printed page in a multi-page questionnaire, so that respondents did not lose sight of the response anchors as they scrolled down the questionnaire on their computer screens. Additionally, scale items were lettered consecutively rather than numbered, to avoid confusion or bias that numbers might otherwise introduce appearing adjacent to the blanks where numerical responses were called for.

Additionally, demographic items were included, as well as single items indicating experience with FtF support groups, use of anonymous remailers, and whether respondents pay for their Internet access. Finally, respondents were invited to add their own thoughts on the value of electronic social support by means of a single, open-ended question.

Sampling procedures in organic electronic spaces raise a number of troubling issues. The first one apparent in this research is estimating the size of the potential participant population, despite readership estimates, and determining how to access it. One has very little idea how many people passively participate by lurking in these spaces. Additionally, there is no simple way to determine, of those who do post, whether they represent one person or more (as people may share email addresses), or whether more than one address corresponds to what is in actuality a single individual (since an individual may have more than one way to access Usenet). In some cases we have observed what appears to be multiple personalities of the same individual person (or at least the same address) posting messages which are related but in different "voices" or age-related styles. We also consider the possibility that some postings are exercises in creative writing and/or attempts to gain attention in which the posters may not actually manifest the symptoms or experiences about which they write (see e.g. Hellerstein, 1990). Concerns over deception on the Internet are not new (see e.g. Van Gelder, 1985), and the fluidity of one’s identity presentations, whether as a game (see Curtis, 1992), a reflection of psychological multiplicity (Stone, 1995), or as a rather normal and strategic affordance of the medium (Walther, 1996), is an issue that occupies a great deal of discussion and inquiry of late (e.g. Allen, 1996; Turkle, 1995). Regardless, in this effort we took posters at face value, particularly since we sought ways to contact them rather than analyze their messages. Acknowledging that we would miss the lurking sub-population by this strategy, we collected posters’ e-mail addresses from their postings to several Usenet groups.

Newsgroups were identified on the basis of their apparent appropriateness for the general topic of social support. All groups containing the word "support" and "recovery" in their names were selected within both the alt.* and the soc.* newsgroup hierarchies (with the exception of alt.support.jockstrap, due to its lack of activity). On the basis of other pointers such as support-related web pages and FAQs, we were able to locate other pertinent groups without "support" in their names (e.g. alt.sexual-abuse.recovery) which we verified through observation as involving social support exchange.

Based on typical response rates to mail surveys and a similar methodology for on-line questionnaires developed by Parks and Floyd (1996), we projected a thirty-three per cent response rate to the questionnaire distribution. In order to achieve some stability for subsequent factor analyses, we aimed for at least ten subjects per questionnaire item. Thus 1,000 addresses were used. In order to get both a diversity of responses as well as to be sensitive to similarities within groups, a randomized block sampling method was adopted to obtain these addresses. Once identified as pertinent, the groups’ names were arrayed in a list. Every fifth group was identified, and within these groups, we recorded up to fifty unique username/addresses from the messages posted to that group. If there were less than fifty showing at the time we examined, more addresses were added from the next consecutive group on the list, until fifty were gathered. Then the next group, of the original every fifth group, was examined similarly. A list of all groups sampled appears in Table 2. The same method was used to gather additional addresses for each of the original set that we learned were not valid. The sampling method makes the groups unequal at the stage where we solicited responses, and imperfect response rates skew the representation of various groups in the final sample even more. However, the randomized approach to subject recruitment should bolster the generalizability of the results.

As we do not have knowledge of the base of potential subjects, either because they did not happen to post messages during a time that would preserve their response for our detection, or they do not post but rather read only, this sample differs from the parallel of a census block, which has a known population from which a sample is an estimable proportion. Despite the uncertainties of the sampling method, however, response rate was very typical compared to other electronic and non-electronic research. Of the 1,000 questionnaires sent out, 340 were returned sufficiently completed for most analyses.

Initial analyses revealed an interesting profile of our respondents. Their ages were extremely diverse, ranging from 16 to 79. Five per cent of the sample was below 20 years of age; 23% was in their twenties; 33% in their thirties; 22% in their forties; 8% in their fifties; and the remaining 2% in their sixties or seventies. The mean age was 36.29 years. There were slightly more females (55%) than males (44%) in this sample, contrary to folk knowledge of the Internet in general as sporting more male users (as many as 2:1 ratio by some recent estimates), yet the distribution was similar to that obtained in Mickelson's study of online support. Two subjects identified themselves as transgendered.

Many of these subjects (65%) paid for their Internet access, as opposed to getting it free through their employment or educational institutions. This may indicate a serious user population, for whom network access is personally important. Finally, 73% of these respondents reported having participated in some form of organized FtF support groups at some point. This suggests a very well-informed sample from which to inform us of relative advantages of on-line support. In general the descriptive demographic data reveal a diverse and knowledgeable population for purposes of this research.

Scores were subjected to principal components factor analysis with orthogonal rotation, in order to determine the most parsimonious set of dimensions of attraction to electronic social support. Criteria for the acceptability of a solution included the following: eigenvalues of 1.5 or greater for as many factors as would be accepted; items having a primary loading of .50 or greater on the primary factor, and less than .30 on any other factor; that any dimension has a minimum of three items meeting the loading criteria, and that factors be readily and meaningfully interpretable. Eigenvalues indicated as many as 8 factors, but initial statistics indicated that no substantial increments in variance would be accounted for beyond four or five. A final four-factor solution met all analytic criteria and accounted for 51% of the variance. The items and factor loadings are presented in Table 3. In presenting the final factors and their interpretation, illustrative comments from the open-ended question are offered.

The first factor, which accounted for the greatest variance, appears to reflect a notion of social distance, and clusters together concepts from several areas reviewed for this research. The items include advantageous aspects of the strength of weak ties, as well as disadvantageous aspects related to autonomy and dependence issues from FtF support relationships. In particular, items related to stigma management are included, dealing with the embarrassment one might experience if off-line acquaintances were to know of the users' concerns. Dependency issues surfaced in items dealing with dependency, giving up control, and feeling obligated in FtF social support. Items reflecting CMC partners' gentler and more expert treatment than that provided FtF were also included. Overall, this dimension suggests a more distant rather than close relationship offers certain benefits and the detriments of unmediated support and the benefits of mediated support are indeed complementary. Respondent's comments reflected this dimension. Concurrence with the notion of reduced obligation/dependency may be seen in the comments of a 17 year old male from alt.support.shyness:

A 33 year old female from soc.support.transgendered seems to concur: "You get a wider spectrum of people. People from more diverse backgrounds and experiences. You get a larger *number* of people as well, increasing the likelihood of hitting it off with someone, or getting the answer you need (provided that you’re capable of weeding through the chaff)." A 27 year old female participant from soc.support.depression.crisis stated, "Usenet has given me the opportunity to communicate with people I would not ordinarily meet. And yes, sometimes it is a great help to be able to bare your soul to someone who has some *understanding* of what you’re going through, which many of my actual friends do not."

The second factor related to anonymity. Items mentioned that no one knows who one is on-line, and that people can say personal things without knowing one another. One item in this cluster implies that anonymity, too, precludes embarrassment. Given that the anonymity factor factored separately from the social distance factor--which also contained items about embarassment and stigma-related issues--it appears that respondents saw differently the social distance of weak ties than the technological distance anonymity may provide.

Factor three appears to focus on interaction management. Most of these items reflect the hyperpersonal perspective's notions about the medium's relative advantage in terms of message composition. FtF does not, while CMC does, allow users to take the time and put into words what they really desire, is reflected here. An additional item states "I am not expected to continue my interaction/relationship after I receive online support" broadens this dimension beyond message composition to interaction management somewhat more generally.

Another respondent's comment illustrates the hyperpersonal notion of message management reflected in this factor; a 47 year old male contacted through alt.support.loneliness observed:

The fourth and final factor relates to access. It is somewhat surprising that this factor was not more prominent, given the salience of the access issue in discussions about the benefits of on-line support and virtual communities. The item ratings on the overall factor did show a generally high appreciation for this dimension. That on-line support is available any time, and that non-electronic support is not, is the essence of this dimension.

The present research attempted to ascertain some dimensions of participants’ attraction to social support transacted through CMC. The numerous potential benefits of online social support were suggested by the social support literature, both from disadvantages of close personal relationship for support exchange, and from the benefits of weak-tie interaction as affect social support. CMC research and theory also offered several potential benefits, as on-line support may offer connections to different kinds of people, through different communication behaviors, than is traditionally associated with support exchange. This background yielded twelve research questions identifying potential attractions of electronic social support: candor (both less harsh and more forthright responses to problems), less negative judgement, reduced obligation to reciprocate support, less relational dependency, more immediate ability to seek support, greater expertise in the network, stigma management, intimacy, access, uninterrupted composition, more expressive communication, and anonymity. The findings from an online survey of Usenet support group users indicate that these specific attributes combine to form four overall factors of attraction, some of them pure and some combining attributes of CMC in unanticipated ways. Findings suggest that users recognize several benefits of on-line social support, which resulted in four dimensions: (1) they are aware of the kinds of people and relationship in which they transact their concerns, as seen in social distance; this distance provided greater expertise, stigma management, and more candor. (2) The level of anonymity with which their actual identity may be associated. (3) The affordances of the medium in optimizing their expressiveness, turns, and ongoing obligations, in interaction management; and (4) the ability to seek or provide support anytime, in terms of access.

Judging by the variance in the statistical analysis, the users who responded to our research appeared to vary in the degree to which they valued each of these dimensions. At the same time, specific support groups may have particular rewards that, like their own norms of conduct, evolve and develop over time among their own user communities. While general research on social support or electronic interaction would not necessarily benefit through attempts to delineate specific groups’ specific attractions, such work may be useful in "site-specific" concerns. That is, practitioners, therapists, system designers, and those in need of support would be well-advised to attempt to understand the communication patterns and dynamics of specific groups which may indeed vary between one and another. While such an approach may address certain applied questions, it remains to be seen whether our knowledge about the support process online will be advanced by empirically examining these diverse online support activities without basing these studies on alternative or varied theoretical perspectives.

There are other aspects of electronic social support which deserve exploration. One aspect is the manner in which many online participants seek social support, in ways that are not considered in the social support literature. A common online support-seeking strategy, disclosing a very personal narrative and/or revelation of feelings and concluding with a question if anyone else has similar experiences. Does not seem common in typological studies of support seeking (see, e.g., Barbee & Cunningham, 1995),³ perhaps because this is an uncommon strategy in FtF support encounters. On the other hand, it is quite common to send a narrative "into the air, so to speak" (as one respondent put it), in electronic support groups. This is also the case in other domains of CMC such as organizational electronic mailing lists or Usenet newsgroups on other topics, such as technical issues to hobbies, where users frequently send messages along the lines, "does anybody know/have any experience with..." whatever issue (e.g. Constant et al., 1996). While social support information requests thus offered tend to be met with less success than those prefaced with an indication of familiarity with the group, according to Galagher et al., 1998), this remains an interesting support seeking strategy. Whether it is relatively undocumented in FtF encounters, or something relatively unique to on-line support, remains to be seen.

While we have argued that there may be many advantages to exchanging social support in electronic networks, we are hesitant to speculate whether this kind of support alone may be sufficient for the kinds of known benefits that are associated with traditional social support. Even if information and shared experience exchanged among similar participants may be useful, it may not satisfy other needs often met through FtF support exchange. In regard to off-line interactions, Schwarzer and Leppin (1991) have demonstrated that support relationships may provide both cognitive social support as well as behavioral social support. Wellman and Wortley (1990) found that females tend to exchange verbal and nonverbal messages of emotional support, while males tend to give support by way of doing instrumental activities to assist others. Such emotional support and instrumental activities may potentially be less available through online discussions among relative strangers, no matter what else the benefits might be.

The benefits of these types of support activities to psychological and physical health and well-being remains unknown. A direct implication for future research is to assess the use of on-line social support to the extent that it correlates with an "individual's subjective sense of being supported by others as well as the appraisal of adequacy...of that support" (Albrecht at al., 1994, p. 423), as implied by any of several existing measures (see Albrecht et al. 1994; Sarason, Sarason, & Pierce, 1990). This type of research will be difficult, however: it is not clear whether the majority of posters to online social support groups are seeking or providing support. If the latter, sampling addresses of message posters, as this study did, may skew results toward exaggerated levels of social adjustment. On the other hand if, once feeling supported, online support seekers tend not to use the system, polling recent posters may skew results to reflect people disproportionately experiencing problems.

In a recent field study by Kraut et al. (1998), despite negative effects of Internet use on depression and loneliness, there were no negative effects obtained on an assessment of self-reported social support access online. They found that previous non-users showed a slight increase in depression and loneliness (R² = .0225) in association with hours of Internet use per week; and that their interaction in geographically close social circles and families was reduced, while their conversations in larger social circles went up. These results were not anticipated, and the authors conjectured that the replacement of strong relationships with weak ones--paralleling the sociometric definition of links but, like Egdorf and Rahoi, taking them to connote qualitative values--leads to the negative psychological reactions. The report suggests that weakly linked, virtual support partners do not deliver adequate support when one is ill, typifying the superficiality of online relations that eventually disappoint or depress the user. This conjecture was offered despite no detection of a relationship sought between Internet use and declined social support, however. It is thus unclear what, if anything, caused users to become more depressed with greater Internet use. Our own research suggests that online social support may make up for, or supplement, such a gap in offline life.

While we share the concern over online support's ability to transact tangible aid or instrumental support, we have doubts about Kraut et al.'s suggestion that these online relations lack meaning for their participants or that they are emotionally shallow compared to "close tie" relationships. There are several bases for our skepticism. First, empirical work by Parks and Floyd (1996) found that most Usenet users "make friends" online and that many of these friendships are comparable to offline friendships on a variety of relationship assessment dimensions (a notable exception being how much time partners spend together; it seems to remain the case that we spend more time in the company of those with whom we see than those with whom we type). Furthermore, just as FtF social support relationships may become multiplex friendships (Burleson, 1990; Burleson et al. 1994), many Usenet-based friendships come to use alternative media, and as many as 60% meet face-to-face (Parks & Floyd, 1996). It is not clear the extent to which subjects in the Kraut et al. study used Usenet per se--the 1998 research looked primarily at e-mail and Web use, while a 1995 report (The HomeNet Project, 1995) focused on Usenet but in relation to other issues. We would be surprised if these contrary findings are that channel-specific, and the comments of our respondents, as well as other research about online relationships, does not reconcile easily the depiction of a "sad, lonely world" of Internet use (see Harmon, 1998) and displacement from one's family that Kraut et al. suggest. We maintain that discussions with one's family may not be the best place, for example, for a woman to ask how can one maintain a healthy sexual relationship with her husband; or one's circle of close friends may not be a comfortable venue in which to ask for advice about a possible divorce and custody fight. The Internet affords certain benefits in such cases.

Indeed, the present research demonstrates how new technology can be appropriated to address needs that may go unmet through the use of traditional communication channels. It offers some evidence to Egdorf and Rahoi's (1994) suggestion that electronic social support helps to satisfy needs unmet needs in non-mediated relationships. Whether it's because one doesn't have access to a critical mass of similar others or experts, or one's close relations might denigrate or obfuscate in response to a request for help, or one needs respite from the double-edged sword of dependency and obligation that accompanies support in close relationships, accessing social support via CMC and electronic networks offer a potent alternative. It will be useful to bear this in mind as new developments in multimedia telecommunications are increasingly resembling mediated interaction to FtF communication. Such innovations as "desktop video" would undermine the affordance of anonymity which, to some electronic support participants, is highly valuable (but see O'Sullivan & Levine, 1999).

In their comprehensive review of support communication, Albrecht at al. (1994, p. 435) suggest that researchers "should be especially interested in the features of situations, messages, interactions, relationships, and networks that can alleviate or reduce dilemmas." That is, the dilemmas of meeting support needs while avoiding the negative aspects of FtF personal social support. Internet and CMC users have found this interesting as well, as their frequent usage and self-reflective comments about on-line social support demonstrate. The notion is ironic, to say the least, but worthwhile nonetheless: Sometimes FtF communication is not what is wanted or best suited to a situation, even when or especially so when the topic is personal, delicate, requiring support. In such cases, different aspects of new electronic communication venues can be an attractive alternative.

ACLU v. Reno, No. 96-963, 1996 U.S. Dist. (E.D. Pa. June 11, 1996). Retrieved June 11, 1999 from the World Wide Web: http://www.aclu.org/court/cdadec.html

Adelman, M. B., Parks, M. R., & Albrecht, T. L. (1987). Beyond close relationships: Support in weak ties. In T. L. Albrecht & M. B. Adelman (Eds.), Communicating social support (pp. 126-147). Newbury Park, CA: Sage.

Albrecht, T. L. & Adelman, M. B. (1987). Communication networks as structures of social support. In T. L. Albrecht & M. B. Adelman (Eds.), Communicating social support (pp. 40-63). Newbury Park, CA: Sage.

Albrecht, T. L., Burleson, B. R., & Goldsmith, D. (1994). Supportive communication. In M. L. Knapp & G. R. Miller (Eds.), Handbook of interpersonal communication (2nd ed., pp. 419-449). Thousand Oaks, CA: Sage.

Allen, C. L. (1996). Virtual identities: The social construction of cybered selves. Unpublished doctoral dissertation, Northwestern University.

Altman, I., Vinsel, A., & Brown, B. B. (1981). Dialectic conceptions in social psychology: An application to social penetration and privacy regulation. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 14, pp. 107-160). New York: Academic Press.

Altman, M., Walther, J. B., & Edelson, D. C. (1997). Electronic collaboration and problem-based learning. Paper presented to the Computers in Healthcare Education Symposium, Health Sciences Libraries Consortium, Philadelphia.

Baccard, A. (1995). The computer privacy handbook: A practical guide to e-mail encryption, data protection, and PGP privacy software. Berkeley, CA: Peachpit.

Baccard, A. (1999, May 13). Anonymous remailer FAQ. Retrieved June 11, 1999 from the World Wide Web: http://www.well.com/user/abacard/remail.html

Barbee, A. P., & Cunningham, M. R. (1995). An experimental approach to social support communications: Interactive coping in close relationships. In B. R. Burleson (Ed.), Communication yearbook 18 (pp. 381-413). Thousand Oaks, CA: Sage.

Barnes, M. K., & Duck, S. (1994). Everyday communicative contexts for social support. In B. Burleson, T. Albrecht, & I. G. Sarason (Eds.), Communication of social support: Messages, interactions, relationships and community (pp. 175-194). Thousand Oaks, CA: Sage.

Black, S. D., Levin, J. A., Mehan, H., & Quinn, C. N. (1983). Real and non-real time interaction: Unraveling multiple threads of discourse. Discourse Processes, 6, 59-75.

Braithwaite, D. O., Waldron, V. R., & Finn, J. (1999). Communication of social support in computer-mediated groups for persons with disabilities. Health Communication, 11, 123-151.

Burleson, B. R. (1990). Comforting as social support: Relational consequences of supportive behaviors. In S. Duck (Ed.), Personal relationships and social support (pp. 66-82). London: Sage.

Burleson, B. R., Albrecht, T. L., Goldsmith, D. J., & Sarason, I. G. (1994). Introduction: The communication of social support. In B. R. Burleson, T. L. Albrecht, & I. G. Sarason (Eds.), Communication of social support: Messages, interactions, relationships, and community (pp. xi-xxx). Thousand Oaks, CA: Sage.

Burgoon, J. K., Parrott, R., Le Poire, B., Kelley, D., Walther, J., & Perry, D. (1989). Maintaining and restoring privacy through communication in different types of relationships. Journal of Social and Personal Relationships, 6, 131-158.

Constant, D., Sproull, L., & Kiesler, S. (1996). The kindness of strangers: The usefulness of electronic weak ties for technical advice. Organization Science, 7, 119-136.

Critical Art Ensemble (1996). Utopian promises -- net realities. Retrieved June 11, 1999 from the World Wide Web: http://mailer.fsu.edu/~sbarnes/lectures/interface.html

Curtis, P. (1992). Mudding: Social phenomena in text-based virtual realities. In D. Schuler (Ed.), DIAC-92: Directions and implications of advanced computing (pp. 48-68). Palo Alto, CA: Computer Professionals for Social Responsibility.

Cutrona, C. E., & Suhr, J. A. (1992). Controllability of stressful events and satisfaction with spouse support behaviors. Communication Research, 19, 154-174.

Diehl, M., & Stroebe, W. (1987). Productivity loss in brainstorming groups: Toward the solution of a riddle. Journal of Personality and Social Psychology, 53, 497-509.

Egdorf, K., & Rahoi, R. L. (1994, November). Finding a Place Where "We All Want To Hear It": E-Mail as a Source of Social Support. Paper presented at the annual meeting of the Speech Communication Association, New Orleans.

Feldman, M. S. (1987). Electronic mail and weak ties in organizations. Office: Technology and People, 3, 83-101.

Fernback, J. & Thompson, B. (1995, May). Computer-mediated communication and the American collectivity: The dimensions of community within cyberspace. Paper presented at the annual meeting of the International Communication Association, Albuquerque, NM. Rpt. as Virtual communities: Abort, retry, failure?, retrieved June 11, 1999 from the World Wide Web: http://www.well.com/user/hlr/texts/VCcivil.html

Ford, L. A., Babrow, A. S., & Stohl, C. (1996). Social support messages and the management of uncertainty in the experience of breast cancer: An application of problematic integration theory. Communication Monographs, 63, 189-207.

Fowler, F. J. (1988). Survey research methods (rev. ed.). Newbury Park, CA: Sage.

Froomkin, A. M. (1995). Anonymity and its enmities. Journal of Online Law. Retrieved June 11, 1999 from the World WideWeb: http://www.law.cornell.edu/jol/froomkin.htm

Galagher, J., Sproull, L., & Kiesler, S. (1998). Legitimacy, authority, and community in electronic support groups. Written Communication, 15, 493-530.

Goldsmith, D. (1992). Managing conflicting goals in supportive interaction: An integrative theoretical framework. Communication Research, 19, 264-286.

Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360-1380.

Goldsmith, D., & Parks, M. R. (1990). Communicative strategies for managing the risks of seeking social support. In S. W. Duck & R. C. Silver (Eds.), Personal relationships and social support (pp. 30-45). Newbury Park, CA: Sage.

Harmon, A. (1998, August 30). Researchers find sad, lonely world in cyberspace. New York Times on the Web. Retrieved August 30, 1998 from the World Wide Web: http://www.nytimes.com/library/tech/98/08/biztech/articles/30depression.html

Harris, S. (1999, February 27). Emotional support on the Internet V1.23. Retrieved June 11, 1999 from the World Wide Web: http://www.compulink.co.uk/~net-services/care/list.htm

Hellerstein, L. (1990). Electronic advice columns: Humanizing the machine. In G. Gumpert & S. L. Fish (Eds.), Talking to strangers: Mediated therapeutic communication (pp. 112-127). Norwood, NJ: Ablex.

Hiltz, S. R. (1992). Constructing and evaluating a virtual classroom. In M. Lea (Ed.), Contexts of computer-mediated communication (pp. 188-208). London: Harvester Wheatsheaf.

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

King, S. (1994, July). Analysis of electronic support groups for recovering addicts. Interpersonal Computing and Technology: An Electronic Journal for the 21st Century, 2 (3), pp. 47-56. Retrieved June 11, 1999 from the World Wide Web: http://www.helsinki.fi/science/optek/1994/n3/king.txt

Kraut, R., Lundmark, V., Patterson, M., Kiesler, S., Mukopadhyay, T., & Scherlis, W. (1998). Internet paradox: A social technology that reduces social involvement and psychological well-being? American Psychologist, 53, 1017-1031.

La Gaipa, J. J. (1990). The negative effects of informal social support systems. In S. Duck & R. C. Silver (Eds.), Personal relationships and social support (pp. 122-139). London: Sage.

Leatham, G., & Duck, S. (1990). Conversations with friends and the dynamics of social support. In S. Duck (Ed.), Personal relationships and social support (pp. 1-29). London: Sage.

Lee, G. B. (1996). Addressing anonymous messages in cyberspace. Journal of Computer-Mediated Communication, 2 (1). Retrieved June 11, 1999 from the World Wide Web: http://www.usc.edu/dept/annenberg/vol2/issue1/anon.html

Matheson, K., & Zanna, M. P. (1988). The impact of computer-mediated communication on self-awareness. Computers in Human Behavior, 4, 221-233.

McGrath, J. E., and 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 and J. Valacich (Eds.) Group support systems: New perspectives (pp. 78-96). New York: Macmillan.

McLaughlin, M. L., Osborne, K. K., & Smith, C. B. (1995). Standards of conduct on Usenet. In S. G. Jones (Ed.), Cybersociety: Computer-mediated communication and community (pp. 90-111). Thousand Oaks, CA: Sage.

Mickelson, K. D. (1997). Seeking social support: Parents in electronic support groups. In S. Kiesler (Ed.), Culture of the Intertnet (pp. 157-178). Mahwah, NJ: Erlbaum.

Mickelson, K. D., Helgeson, V. S., & Weiner, E. (1995). Gender effects on social support provision and receipt. Personal Relationships, 2, 211-224.

Millar, F. E., & Rogers, L. E. (1976). A relational approach to interpersonal communication. In G. R. Miller (Ed.), Explorations in interpersonal communication (pp. 87-104). Beverly Hills, CA: Sage.

Mitchell, W. J. (1995). City of bits: Space, place, and the infobahn. Cambridge, MA: MIT.

Moraes, M (1996). A primer on how to work with the Usenet community. monthly posting to news.announce.newusers

Moursund, J. (1997). SANCTUARY: Social support on the Internet. In J. E. Behar (Ed.), Mapping cyberspace: Social research on the electronic frontier (pp. 53-78). Oakdale, NY: Dowling College.

Myers, D. (1987). Anonymity is part of the magic. Qualitative Sociology, 10, 251-266.

Navaro, E., Gureli, A., Gursel, O., & Starnella, K. (1996). Internet access in the world. Unpublished paper, Northwestern University, Evanston IL.

O'Sullivan, P. B., & Levine, K. J (1999, May). Mediated social support groups: A bona fide group perspective. Paper presented at the annual meeting of the International Communication Association, San Francisco.

Parks, M. R., & Floyd, K. (1996). Making friends in cyberspace. Journal of Communication, 46, 80-97.

Rafaeli, S. (1986). The electronic bulletin board: A computer-driven mass medium. Computers and the Social Sciences, 2, 123-136.

Rapaport, M. (1991). Computer mediated communications: Bulletin boards, computer conferencing, electronic mail, information retrieval. New York: John Wiley & Sons.

Rheingold, H. (1993). The virtual community: Homesteading on the electronic frontier. Reading, MA: Addison-Wesley.

Rice, R. E. (1987). Communication technologies, human communication networks and social structure in the information society. In J. Schement & L. Lievrouw (Eds.), Competing visions, complex realities: Social aspects of the information society (pp. 107-120). Norwood, NJ: Ablex.

Robinson, D. (1988). Self-help groups. In R. S. Cathcart & L. A. Samovar (Eds.), Small group communication: A reader (5th ed., pp. 117-129). Dubuque, IA: Wm. C. Brown.

Sarason, B. R., Sarason, I. G., & Pierce, G. R. (1990). Traditional views of social support and their impact on assessment. In B. R. Sarason, I. G. Sarason, & G. R. Pierce (Eds.), Social support: An interactional view (pp. 9-25). New York: Wiley.

Scheerhorn, D., Warisse, J., & McNeilis, K. (1995). Computer-based telecommunication among an illness-related community. Health Communication, 7, 301-325.

Schwarzer, R., & Leppin, A. (1991). Social support and health: A theoretical and empirical overview. Journal of Social and Personal Relationships, 8, 99-127.

Shade, L. R. (1996). Is there free speech on the net? Censorship in the global information infrastructure. In R. Shields (Ed.), Cultures of Internet (pp. 11-32). London: Sage Publications.

Shields, R. (1996). Introduction: Virtual spaces, real histories and living bodies. In R. Shields (Ed.), Cultures of Internet (pp. 1-10). London: Sage Publications.

Spears, R., & Lea, M. (1992). Social influence and the influence of the "social" in computer-mediated communication. In M. Lea (Ed.), Contexts of computer-mediated communication. London: Harvester-Wheatsheaf.

Stone, A. R. (1995). The war of desire and technology at the close of the mechanical age. Cambridge, MA: MIT.

Thompson, T. L., & Seibold, D. R. (1978). Stigma management in normal-stigmatized interactions: Test of the disclosure hypothesis and a model of stigma acceptance. Human Communication Research, 4, 231-242.

Valacich, J. S., Dennis, A. R., & Nunamaker, J. F. (1992). Group size and anonymity effects on computer-mediated idea generation. Small Group Research, 23, 49-73.

Van Gelder, L. (1985, October). The strange case of the electronic lover. Ms. Magazine. Rpt in C. Dunlop & R. Kling (Eds.), Computerization and controversy: Value conflicts and social choices (pp. 364-375). Boston: Academic Press.

Walther, J. B. (1993). Impression development in computer-mediated interaction. Western Journal of Communication, 57, 381-398.

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

Walther, J. B. (1997). Group and interpersonal effects in international computer-mediated collaboration. Human Communication Research, 23, 342-369.

Walther, J. B., & Burgoon, J. K. (1992). Relational communication in computer-mediated interaction. Human Communication Research, 19, 50-88.

Walther, J. B., & Tidwell, L. C. (1995). Nonverbal cues in computer-mediated communication, and the effect of chronemics on relational communication. Journal of Organizational Computing, 5, 355-378.

Wellman, B. & Gulia, M. (1999). Net surfers don't ride alone: Virtual communities as communities. In M. A. Smith & P. Kollock (Eds.), Communities in cyberspace (pp. 167-194). London: Routledge.

Wellman, B., & Wortley, S. (1990). Different strokes from different folks: Community ties and social support. American Journal of Sociology, 96, 558-588.

Usenet Newsgroup	Estimated Monthly Readers Worldwide

alt.support.depression	22,000

alt.dads-rights	21,000

alt.support	21,000

alt.support.diet	20,000

alt.support.cancer	18,000

soc.support.youth.gay-lesbian.bi	16,000

alt.support.shyness	11,000

alt.support.eating-disord	11,000

alt.support.divorce	11,000

alt.support.arthritis	8,100

Subscale Alpha Reliability, Means, Standard Deviations, and Rotated Factor Matrix.

_______________________________________________________________________________________________________________________________

Note: * indicates items reflecting the greatest reliability, retained for dimensional analyses to be reported elsewhere.

1. Indeed, a vehement objection to attempts by the Prodigy network to censor certain words, and to the potential of the Communications Decency Act enacted by the US government in 1995, was that it could as equally prohibit discussion of "breasts" as related either to breast cancer as well as in lascivious discussions. Just such issues over the censorship of language with the potential for prosocial or lascivious discussion were part of the U.S. District Court's reasoning in striking down the CDA (ACLU v. Reno, 1996, Sec. III, Dalzell, C.):

2. [sic]. In this and other direct quotations from respondents or on-line sources, all text is reproduced as originally appeared without corrections to spelling, punctuation, and grammar. Permission obtained my e-mail through anonymous remailer.

3. The closest we have noted is the support-seeking strategy Barbee and Cunningham (1995) call "tell details," which suggests some diclosure as a seeker describes a problem: A seeker "describes entire situation; explains the reasons for the problem; talks openly and states factual evidence; sticks to the facts..." (p. 394). Yet this approach seems overly rational compared to the intimate, personal, subjective narratives we have frequently observed informally as calls for response in many support newsgroups.

			Mean	St Dev	I	II	III	IV
	Factor I: Social Distance	a = .88	2.69	0.82
*	I'm afraid to talk to my close, personal, face-to-face contacts about this topic.		2.37	1.23	.73	.24	.05	-.01
*	My face-to-face acquaintances might think less of me if we talked about this issue		2.56	1.24	.72	.11	.17	-.01
*	I would feel embarrassed if people I know off-line knew I had these concerns.		2.59	1.33	.69	.32	-.03	-.06
*	My face-to-face relationships hold things back to spare my feelings		2.99	1.19	.65	.33	-.03	-.06
*	In face-to-face support relationships, you get too dependent on others.		2.55	1.15	.64	.09	.05	-.01
*	People I know face-to-face are too harsh about this topic.		2.45	1.17	.64	.10	.02	.08
*	In seeking support in my face-to-face relationships, I give up a measure of control over my life.		2.49	1.12	.63	.09	.31	.11
*	I wouldn't want some of my friends and family to be aware of this issue.		2.80	1.40	.63	.36	.03	-.05
*	My on-line acquaintances are less likely to judge me negatively		3.02	1.17	.54	.29	.05	.22
*	After I receive face-to-face social support I feel obligated to return the support even though I may not want to.		2.68	1.14	.52	.01	.23	-.01
*	The advice I receive from my face-to-face relationships is not as objective		2.76	1.15	.50	.22	.24	.23
	The advice I receive from people I may not know is more objective than from people I know well.		3.12	1.15	.42	.25	.01	.29
	My face-to-face social support network does not have enough expertise in this area		3.26	1.24	.35	.03	.08	.24
	Factor II: Anonymity	a = .75	3.28	0.83
*	On-line, I can say personal things without others knowing who I really am.		3.37	1.17	.21	.76	.09	.04
*	I can say things anonymously on-line.		3.27	1.24	.16	.66	.26	.03
	Others can say personal things because we don't know who they are on-line.		3.50	1.14	.40	.59	.01	.15
*	There is less embarrassment being anonymous on-line.		3.25	1.13	.29	.54	.27	.08
*	I always get an opportunity to express myself on-line.		3.87	0.96	-.10	.52	.26	.32
*	No one knows my true identity on-line.		2.65	1.28	.19	.52	.23	-.12
*	Factor III: Interaction Management	a = .72	3.53	0.71
*	Face-to-face interactions do not always allow me time to compose my thoughts.		3.50	1.16	.28	.09	.67	.08
*	It's hard to put my thoughts into words on the spot face-to-face.		3.12	1.19	.35	-.04	.65	-.04
*	I have time to craft a well-thought message on-line		4.00	0.92	-.05	.30	.65	-.01
*	There's plenty of time to say exactly what I mean on-line.		3.93	0.86	-.09	.28	.59	.15
*	I am not expected to continue my interaction/relationship after I receive on-line support		3.14	1.03	.14	.22	.56	.05
	I do not always get a chance to speak in face-to-face support.		3.06	1.18	.36	-.09	.41	.26
	I cannot receive anonymous support face-to-face.		3.38	1.22	.29	.26	.38	.09
	In seeking support here, I do not give up any control over my life.		3.50	1.11	.12	.35	.37	.16
	Factor IV: Access	a = .72	3.92	0.78
*	On-line support is always there when I need it.		3.89	0.99	-.04	.15	.02	.70
*	Face-to-face social support is not available 24 hours a day		3.72	1.17	.19	-.02	.16	.66
*	It is available any time of day or night.		4.23	0.92	-.06	.05	.21	.65
*	I cannot always receive face-to-face social support when I want it		3.84	1.19	-.01	.05	-.12	.57
	This on-line group has enough expertise to be helpful in this area		3.75	1.02	.43	-.08	.14	.56
	Compared to face-to-face relationships, you don't have to get to know people as long before you can ask for help.		3.75	1.07	.24	.34	.16	.37