WHI!WWWWIHHIHIHINIIHIIHWIHIWI HTHS THESIS \ liiiiiiifllliilml|1l|H||Hl|HH|H| IHIHIIHHI 3 1293 01712 This is to certify that the thesis entitled ? 2/75 [of anT/u'co M ‘V’ \ i’AN/n m OHD n 133%??ng Z: air/7W CiM/V’WMfia/j m presented b Mic/”lad ’5, 7/2554 has been accepted towards fulfillment of the requirements for MA degree in CommciIViCIif/On 5 3/ 9/07?) 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Mlchlgan State Unlverolty PLACE IN RErURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MTE DUE DATE DUE DATE DUE my? -' ,‘_ Eam— 1/98 chlRCIDdoDuopGG—p.“ THE IMPACT OF ANONYMITY ON DISINHINIBITIVE BEHAVIOR THROUGH COMPUTER-MEDIATED COMMUNICATION By Michael J. Tresca A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1998 ABSTRACT THE IMPACT OF ANONYMITY ON DISINHINIBITIVE BEHAVIOR THROUGH ’ COMPUTER-MEDIATED COMMUNICATION By Michael J. Tresca The purpose of this thesis was to examine effect of anonymity on disinhibitive behavior over computer-mediated communication. 484 posts from the alt.law- enforcement newsgroup were examined and coded for inflammatory and informational disinhibition. It was found that as anonymity increased, informational disinhibition increased. ii COPyright by MICHAEL J. TRESCA 1998 To Amber, whose strength and support allowed me to finish this thesis. ACKNOWLEDGMENTS A publication such as this would not be possible without the assistance, guidance, and kind words provided by many people. Their strength, dedication, and perseverance allowed me to battle my way through paperwork and long nights, news posts and blurry screens. This thesis is for them. Special thanks go to my advisor, Steven McCornack, and my committee members, Lawrence Busch and Jim Dearing for their patience and wisdom dispensed in judicious amounts to a hyperactive graduate student. Scott Preston’s advice and Gerad Middendorf’ s constructive commentary were invaluable. Thanks also to my parents who supported me even in my times of self-doubt. Finally, thanks must go most of all to my soul-mate, Amber Reifert. Our chance meeting over the Internet provided the inspiration for this thesis. TABLE OF CONTENTS LIST OF TABLES ....................................................................................................................................... vii INTRODUCTION .......................................................................................................................................... 1 CHAPTER 1 RESEARCH ABOUT COMPUTER-MEDIATED COMMUNICATION .................................................... 3 CHAPTER 2 METHOD ..................................................................................................................................................... 14 CHAPTER 3 RESULTS ..................................................................................................................................................... 19 CHAPTER 4 DISCUSSION .............................................................................................................................................. 22 APPENDICES .............................................................................................................................................. 30 LIST OF REFERENCES .............................................................................................................................. 41 vi LIST OF TABLES TABLE 1 INFLAMMATORY AND INFORMATIONAL DISINHIBITION ............................................................. 19 TABLE 2 TOTAL DISINHIBITION ............................................................................................................................ 20 TABLE 3 STANDARD DEVIATION AND MEANS ................................................................................................. 2] vii INTRODUCTION Technology changes the way we perceive each other and ourselves. Humanity has fi'equently sought mechanical means of controlling and enhancing face-to-face communication to serve the needs of security, socialization, and fantasy (Cathcart and Gumpert, 1983). Over the years, technological advancements in communication have allowed us to “reach out and touch someone” even as computers have managed to physically isolate us. Computer-mediated communication, for example, allows us to contact thousands of people within seconds without actually standing in their presence. This anonymity affects how we perceive each other and ourselves, how we interact with these perceptions of others, and the degree to which our social environment restricts us. Many users feel uninhibited and unrestrained because of a lack of social context cues and therefore exhibit more “disinhibition” in the form of insults, swearing, and hostile language (Walther, 1993) than if they were communicating in a face-to-face situation with the same people (Siegel, Kiesler and McGuire, 1984; Kielser, Dubrovsky, Siegel and Powers, 1986). The lack of social context cues can cause excited and uninhibited communication as Christian Sandvig (1995), a police officer, explained: “When I first started responding to local discussion about crime on USENET newsgroups, I quickly discovered that the Internet’s playing field is very level. My user name at the time, ‘David Police,’ prompted vicious attacks on police generally, belligerence, and questions about crime best classified as beyond the bizarre. Behavior that would not occur on the telephone, or in a letter, (or especially face to face, with me in uniform). I was stripped of my boundaries, my protections.” This issue will become even more critical as the use of networks perceived as anonymous, like the Internet, continue to increase in popularity. Even today, there are “anonymous remailers” that accept email messages and forward them to their intended recipient, stripped of any identifying information (Denning and Lin, 1994). Unfortunately, past research has focused more upon the productivity and effectiveness of computer-mediated communication rather than the human consequences of employing such communication technology, such as disinhibitive behavior (Chesebro and Bonsall, 1989). The purpose of this study is to determine if experience with computer-mediated communication will alter a computer user’s behavior and perceptions. Specifically, this study will test the effect of objective anonymity and experience upon disinhibitive behavior in computer-mediated communication. I begin by reviewing the existing literature on computer-mediated communication in the present chapter, and providing my rationale and hypotheses. In Chapter Two, I elaborate on my experimental methods. Chapter three presents the results thereof. Finally, in Chapter Four, I discuss the outcome of the thesis and implications for future research. Chapter 1 RESEARCH ABOUT COMPUTER-MEDIATED COMMUNICATION There has been no research on disinhibition in computer-mediated communication. Consequently, the purpose of this study is to examine these issues. Walther (1993) did some work on this study in an initial exploration. This exploratory study will hopefully provide incentive for future ventures into this neglected topic. In this chapter, I will first define the terminology used throughout this thesis, and then examine the conditions which require acceptance of computer-mediated commmiication. I will then examine computer-mediated communication in general to provide a better perspective on the nature of this research, and finally, my literature review will focus upon a more general review of computer-mediated communication. Computer-Mediated Communication in General Computer-mediated communication is the synchronous or asynchronous process by which people create, exchange, and perceive information using networked telecommunications systems that facilitate encoding, transmitting, and decoding messages (Kiesler and Sproull, 1992). For the purposes of this study, this definition does not include such mediums as the World Wide Web, as the Web incorporates advanced visual cues that electronic mail and bulletin boards do not normally provide. More visual cues have a higher relevance over their textual counterparts and therefore would provide “normal artifacts” which are more similar to face-to-face communication than text-based communication lacks. Computer-Mediated Society The study of anonymity and communication has its basis in societal interaction. Ulric Neisner first noted virtual societies in 1964 in his early study of programmers. He observed that in the relatively fast development atmosphere of one of the first interactive systems, the only way users were able to keep up was by informal communication within the close community that developed (Kerr and Hiltz, 1982). In order to become part of such a virtual society, a user defines himself first, and then the society helps determine or modify that definition. Computer-mediated social interactions provide the behavioral basis on which network culture is generated and reproduced. Such interactions are also the subject of behavior that is regulated by culture at large. Values are formed about what’s desirable, behavioral norms which provide the basis for judging in individual situations whether a value is being observed, and thus regulating behavior in the interest of implementing or reinforcing values and beliefs that provide history, mythology, and world views (Denning and Lin, 1994). In this manner, others help provide a reflection through which the user perceives himself. Schaeferrneyer and Sewell’s (1988) survey indicated that 17.6 percent of users felt the greatest advantage of computer-mediated communication was the advantage of reaching large and diverse groups of people. Due to this flexibility, ease of choice, and accessibility, a user has virtually no restrictions upon his interactions with others, with little fear of retaliation or personal responsibility. The user is free to pick and choose his social reality. Which social reality that user chooses, however, requires a certain degree of acceptance. Such acceptance only exists if a number of conditions hold true. Acceptance Conditions First, users must have access to an information computer network. Users who have unstable connections to the Internet experience severe impairment of message transmission (lag or other transmission interference), or who have hardware or software malfunctions will be highly unlikely to accept the medium as a valid means of communication. Second, users understand how to operate the computer-mediated communication system. To communicate via computer, a user must be able to type and know how to utilize the communications software (Belson, 1994). Unfamiliarity can lead to distrust of the motivations of others, as measured by the perception that some of the group’s members act unethically, which invariably results in lower use of computer-mediated communication. Comparatively, a user who is familiar with computers and has adequate typing skills is likely to find the medium less inhibiting and is more likely to treat the medium as a normal form of social interaction (Kerr and Hiltz, 1982). Self-expression becomes second nature, allowing users to concentrate on more abstract concepts. Without this familiarity, users may be unable to distinguish between different forms of computer-mediated communication, such as bulletin boards and email (Denning and Lin, 1994). The advent of more user-friendly technologies has reduced the learning curve to utilize computer systems significantly. As these technologies become more refined, users should feel even less inhibited in the future. Third, users must be completely free to use alternative systems for their communication activities. Users who are forced to use computer-mediated communication are unlikely to feel comfortable communicating over it. Finally, users must be motivated to use the system (Kerr and Hiltz, 1982). Older users (50+) tend to use such systems the least and have the lowest levels of subjective satisfaction over computer-mediated communication, perhaps because older users need lengthier or different kinds of training than younger users who are more likely to have previous experience with computer systems. Or, it may be that older users are less likely to accept changes in their basic communication patterns, despite any special training efforts. This is a historical measure that should change with the aging of the population. Individuals who are assertive, have high internal control, and high tolerance for ambiguity accept and use computerized communication systems more than those without such traits (Kerr and Hiltz, 1982). Definitions A user’s perceived anonymity is the ability to remain unaccountable within his or her social environment for actions and dialogue, both legally and socially. A user’s personal views are influenced by lack of visual appearance, the flexibility of a label that is different from the user’s normal persona, and relative protection from physical and social repercussions. Users may often perceive that they are anonymous when they are not, may find a deeply personal attachment to a label so that it becomes part of their own personality, or may be far less protected from repercussions than they perceive. The ability to accurately perceive one’s objective anonymity in relation to others over computer-mediated communication is directly tied to the user’s experience with that medium (Orcutt and Anderson, 1996). Since there is no way of knowing who has signed onto a specific account in the absence of voice identification or who has actually read the messages, users may not be motivated to communicate sensitive issues over computer- mediated communication. A user is never guaranteed a response to her computed-mediated communications and does not have nearly as many communication channels to resort to should the response fail to materialize. This is not necessarily the case in face-to-face scenarios, where other cues, the environment itself, and preconceived notions of social status, ability, etc., are all in force regardless of any verbal communication. A user who does not respond to her email may give the impression of stony-faced silence, but may in actuality be simply too busy to answer the message at the time, or may not even have received it. These concerns are more prevalent among new users than experienced users (Kerr and Hiltz, 1982). Because of a lack of immediate artifacts, initial messages showing attempts at asserting authority and dominance are more common in early computer-mediated interactions than similar face to face situations. Such informational disinhibition can be perceived as overcompensation for the incompleteness of computer-mediated communication (Chesebro and Bonsall, 1989). Conversely, good writers and more literate people have the same social advantage that physically attractive people have in face-to-face over computer-mediated communications (Kerr and Hiltz, 1982). Walther (1995) found that both extremes eventually become attenuated to patterns similar to face- to-face communicators over time. Disinhibitive behavior is not limited to only rude or offensive interaction. Users of computer-mediated communication systems can express informational disinhibition, becoming more connected with people who share common interests but with whom they would otherwise be unacquainted (Rice and Love, 1987). Nearly 70 percent of those polled in Schaeffrneyer and Sewell’s study (1988) used electronic mail to communicate with those of similar interests at different locations. Of that 70 percent, 63.8 percent use email for social activity. With such an accessible audience, a user is able to communicate quickly and easily with others. Anonymity All beginning users start out as anonymous. They are both anonymous to other users, and other users are anonymous to them. Users who wish to distinguish themselves from their anonymous counterparts must provide information about themselves, while users who desire to know more about whom they are communicating with must ' accumulate more information through further communication. Each of these forms of anonymity creates a different form of disinhibition. Inflammatory disinhibition can occur because without nonverbal and paralinguistic cues, a users’ attention becomes selfish, more concerned about oneself than the feelings or welfare of others. She is less likely to form impressions of other users as distinct individuals because she does not have enough information to distinguish them (Kiesler and Sproull 1992, Walther, 1992). To achieve a distinct impression of another’s psychological makeup requires a large quantity of information. This information is developed through communication with another, gaining insight of the other user’s beliefs, the rationale behind those beliefs, and the cognitive structures underlying their motivation. Users utilize knowledge-generating strategies to acquire this information, including interrogation, self-disclosure, deception detection, environmental structuring, and deviation testing. This development is inhibited in computer-mediated communication encounters because information takes longer to accumulate, with fewer verbal exchanges over time and a consistent lack of nonverbal cues (Walther, 1992). The lack of nonverbal cues about physical appearance, authority, status, and turn taking also allows users to participate more equally in computer-mediated communication systems than in many face-to-face interactions. These cues are flexible, as there is no consistent visual appearance to which a user is bound. Participants use words alone to reconstruct contexts in their own image, adding imagined actions as metadescriptions to the running dialogue, typographically set apart from words meant as straight dialogue. They also add modifiers to the strict definition of words, indicating intention, mood, or other contextual cues that would normally be lacking (Froomkin, 1995) Some investigators (Rice and Love, 1987; Walther, 1992) have hypothesized that computer-mediated communication technologies attenuate social context cues even more than does paper communication. Uncertainty reduction needs combined with accessibility and a selective channel for expression allow the user to choose his self- representation and relational behavior. In this mode, the user may plan, contemplate, and edit his comments much more than in spontaneous face-to-face interactions. While limited in scope, users can engage in disinhibitive behavior by expressing political views 10 without fear of retaliation, engaging in whistle-blowing while remaining undetected, and seeking advice about embarrassing personal problems with the knowledge that their true identity is kept secret (F roomkin, 1995). Over time, as a user’s knowledge of the medium increases, her ability to created desired impressions should increase. The absence of social context cues could result in more disinhibitive behavior and polarization of attitudes, creating more negative perceptions of group members. Because others exist only within the context of the computer medium, a user can shut the computer off and the entire virtual society ceases to influence that user (Brook and Boal, 1995). Likewise, a user usually communicates over computer-mediated communication singularly, as it does not require any other human beings to be physically present in the room at the time of the interaction. This causes a user’s perceptions to be focused more on herself than others (Chesebro and Bonsall, 1989). It can also create a feeling of great isolation despite the ease of accessibility to the computer medium. Lack of cues can also cause informational disinhibition. Anonymity can help people overcome social inhibitions, encourage communication across social or psychological boundaries, and deregulate group behavior (Kiesler and Sproull, 1992). In groups, such de-individuation involves a decrease in self-awareness and a corresponding loss of identity. This occurs when group salience is high, because the prototypical image or ideal of the group is not contaminated by viewing the other users, who, by their mere appearance or presence, may be perceived as deviating from group norms in some way (Spears, Lea, and Lee, 1990). The absence of these social context cues make it difficult for people to perceive and adapt to social roles and situational norms because static cues are derived from ll artifacts that delineate levels of power and authority (Kiesler and Sproull, 1992). Because computer-mediated communication lacks social context cues, the ability to perceive authoritarian and dominance positions is lessened. In essence, everyone looks the same on the Internet. Thus the capacity for social influence via computer-mediated communication will vary depending on the user’s perceived relation to others on the network. We would expect substantial differences in rates of disinhibition for people at different levels of technical anonymity such that people utilizing highly anonymous accounts will be more disinhibitive. Therefore, we should see a user’s inflammatory disinhibition over computer-mediated communication decrease as the user’s objective anonymity decreases (H1). Experience Much socioemotional content is corrrrnunicated through relational icons, contrived sideways faces that can be made by combinations of punctuation marks (e. g., >:)). Electronic paralanguage also enables the user to express himself, including grammatical markers, intentional misspelling, onomatopoeias, spatial arrays, absence of corrections, and capitalization. This computer-mediated paralanguage adds affective information and indicates informality (Walther, 1992). Computer users who understand the rules of “netiquette” are generally more appealing to communicate with. A speaker who has no grasp on slang terminology is at a distinct disadvantage and loses much of the conversation. The amount of paralanguage and personal content deemed appropriate 12 over computer-mediated communication systems was found to be at least partially a function of experience and familiarity (Rice and Love, 1987). Perceptions of other users as individual personalities should eventually increase, given enough time, over computer-mediated communication. Textually conveyed information will accumulate slowly. As interpersonal knowledge accumulates, communicators express more personal messages in both face-to-face and computer- mediated communication (Raefeli, 1988). This increase accounts for the presence of more personalized impressions and message transmission through ongoing asynchronous conferencing, as well as for the more task-oriented and impersonal messages in time- lirnited computer-mediated interactions among unacquainted interactants (Walther, 1992). Such interactions are relevant because a user then becomes more capable of processing information as well as reacting to it, has a multitude of communications with several users, helps to define as well as be defined by his perceived peers, and becomes a concrete individual with a contextual base to which others will react. Individuals who do not have time to formulate such interactions will have poorly-defined “virtual perceptions” of themselves and others, and will thus indicate more formality and less comfort with computer-mediated communication. Walther (1992) found that, given the lack of nonverbal cues and requirements for greater time to achieve group objectives in computer-mediated communication, forming distinct individual impressions of other users would similarly require more time in the computer-mediated environment than in similar face-to-face encounters. It would take longer to observe and decode impressions from verbal and textual cues alone than from multichannel cues common in face-to-face interaction. A single message exchange may 13 not carry as much social information as would a similar exchange in a non-mediated setting. This extended process creates short-time users who react with disinhibition over the medium because they are not accustomed to or socialized by the particular context of the computer-mediated environment. Conversely, long-time users who are fully adapted to the environment and socially accepted with other experienced users should treat computer-mediated communication with the full social status that is accorded to face-to- face interaction. Based upon the literature reviewed, we could speculate that one symptom of disinhibition would be to use more abusive language. Therefore, users who post more often in newsgroups should exhibit less inflammatory disinhibition. If this is the case, we should see a decrease in a user’s inflammatory disinhibition over computer- mediated communication the more experienced the user is (HZ). Anonymity over computer-mediated communication can be empowering. History illustrates the use of anonymous communications to protect dissidents from the scrutiny or retaliation of autocratic governments. On the other hand, if users are granted true anonymity then there can be no justice system that will right the wrongs they commit and there can be no rights that people will be able to vindicate (Denning and Lin, 1994). This confidence in one’s own “invisibility” can lead to conspiracy, electronic hate-mail, electronic stalking, libel, disclosure of trade secrets and other valuable intellectual property. Either sex can masquerade as the other, children as adults, and vice-versa (Froomkin, 1995). With no fear of retaliation, the inability to perceive or clearly delineate authoritarian positions, and the relative ease with which anonymity can be achieved, a user’s level of disinhibition is likely to be high. Chapter 2 METHOD The purpose of this experiment was to examine the coded results of computer- mediated interaction over electronic newsgroups. Although there is a lack of prior research on the topic, Walther’s work (1993) seemed to suggest that people are more likely to be disinhibitive when they are more anonymous. If this reasoning is valid we would see a direct positive relationship between users’ experience and users’ disinhibition. Pretest A pretest was run to determine the impact of “newsgroup influence.” It also tested the experience and anonymity measure. Responses were polled from 50 randomly selected Internet users. The correlation of experience and anonymity was tested by two sets of five questions. The experience questions attempted to verify the level of self- reported computer-mediated communication while the anonymity questions focused on self-perceived anonymity. On a Likert scale of 1 to 7 the average experience level of Internet users was 5.3. This indicated that the testing of newsgroups could also apply to computer-mediated communication in general. The pretest also attempted to measure the most neutral newsgroup that would be least likely to encourage or discourage disinhibition. Several newsgroups were listed, fi'om alt.fishing to alt.party. The alt.law-enforcement newsgroup indicated an average 14 15 disinhibition rating of 2.8, on a Likert-scale of 7. This was the newsgroup closest to a score of 4 that indicated it was least likely to encourage or discourage disinhibition. Both experience and anonymity were tested for. In the original drafi, the anonymity measure was measured by the email addresses of the posters, divided into three categories of educational accounts (.edu), commercial accounts (.com), and miscellaneous accounts (.net). However, because email accounts are no longer grouped into these three categories, the results were discarded. See Appendix A. Main Project Gender was not accounted for because lack of identification of gender is part of anonymity. This is a more natural environment than a simulated experiment under laboratory conditions and is a much more authentic representation of users rather than a level of forced anonymity. A random sampling of 484 users was coded. Users were assigned scores on two dependent variables, experience and anonymity. All levels of experience were traced by Deja News, a Web program that tracks the number of posts from a particular site (Appendix B). Deja News also provides the number of total posts written by a particular user based on any one e-mail, including e-mail which is obviously anonymous (Appendix C). Coding of Data Coders were trained and tested in examining newsgroup results and identifying levels of anonymity and indicators of disinhibition. Messages were each counted for the number of words, excluding titles, signatures, and technical information displayed by posting software. Anonymous posters were given a rating of 1 through 3, one being l6 completely anonymous, three being unanonymous. The number of posts were taken from Deja News for the experience measure, which provides the total number of posts previously written by the email address used to make the post examined (Appendix C). Content was then divided into two categories, inflammatory and informational disinhibition. Each word was considered for informational and inflammatory content. Informational content included the following words: me, mine, my, you, your, yours, our, they, theirs, and I. It also included proper names of self and others, and questions. Inflammatory content included expletives, exclamation points, repeated punctuation of question marks or exclamation points, and emphasized text utilizing *, “, _, or capital letters. The total number of informational and inflammatory messages was then calculated as a percentage of total number of words in a post. See Appendices D and E. The experience variable was complied by totaling the number of posts per poster and calculating the means. Inflammatory and informational data were calculated as a percentage of the total word content of a post, and then the means calculated across the three levels of anonymity. Definition of Variables Experience Users who communicate over the Internet more frequently are much more likely to have a better knowledge of “netiquette” and the social rules of the computer-mediated medium than those who have less interaction with the medium. By utilizing the number of postings by a particular user as a measure of this experience, several factors are taken into account. Users who post over the Internet are, at the minimum, users capable of 17 reading and understanding news reading programs. Users who post repeatedly indicate a desire to pursue a topic or topics, indicative of their experience with the medium. Anonymity Beyond electronic mail, posters have no other means of being identified. Legally performing an experiment wherein the users know they are being catalogued, observed, or otherwise tested (including signing a document which grants this privilege) is likely to spoil any true feelings of anonymity. Thus, the most natural form of anonymity is in posters who have chosen to be anonymous of their own accord. They post through anonymous remailers or utilize alternate methods of concealing their email address and other identifying features. Anyone obviously utilizing a nickname (Captain Meat) or simply a nonsensical electronic mail address (noemail@forme.please) is included in this category. Moderately anonymous users neither post information detailing who they are, nor make attempts to conceal their electronic mail address. However, some information can be gleaned based on a user’s e-mail using the right software. Universities most commonly have this information displayed for public perusal by students, professors, and staff, including name, address, phone number, and e-mail. Most universities only restrict this information if a user requests its removal and very ofien users are simply unaware that the information is available. Low anonymity users provide information identifying their last names, addresses, telephone numbers, or web pages. Low anonymity posters usually place this information in a signature file at the bottom of the post. Note that this is merely the user’s intent; 18 there is no way to know if the lack of anonymity is genuine without violating the user’s rights. More importantly, it is the perception of identity the user is trying to formulate that is measured. Disinhibition Disinhibition is the dependent variable. This variable consisted of inflammatory and informational disinhibition. Inflammatory indicators included expletives, sarcasm, capital letters, emoticons, exclamations, and other forms of emphasis (e.g., asterisks, underscores, or arrows immediately before and afier a word). Informational indicators included references to oneself (1, me, my, mine, we, ours), or another (you, yours), questions, and addressing another poster by name. Chapter 3 RESULTS 484 posts were examined fi'om the alt.law-enforcement newsgroup. The messages were selected at random. No author was duplicated. These messages were individually entered into Endnote 2, coded, exported into Microsoft Excel, AN OVA was run and the results examined. Table 1. Inflammatory and Informational Disinhibition ANONYMITY (N) EXPERIENCEl INFLAMMATORY2 INFORMATIONAL High (143) 300.2 0.05 0.04 Medium (102) 292.5 0.03 0.05 Low (239) 390.4 0.02 0.03 1 Average number of posts per user. 2 Average response from a range of l to 3. l9 20 Table 2. Total Disinhibition ANONYMITY (N) TOTAL DISINHIBITION High (143) 0.09 Medium (102) 0.08 Low (239) 0.05 A one-way AN OVA examining the effect of anonymity on informational disinhibition found that there was not a significant difference between high, medimn, and low levels of anonymity (F 2, 483 = 1.93, NS). A one-way AN OVA examining the effect of experience on anonymity found that there was not a significant difference between high, medium, and low levels of anonymity (F 2, 483 = .41, NS). Hypothesis 1 In order to test hypothesis 1 a one-way AN OVA examined the effect of a user’s anonymity on inflammatory disinhibition. Ex post-facto tests were conducted to assess the nature of the between group difference. Using Scheffe’s procedure it was discovered that highly anonymous individuals indicated a small but significant result from low anonymous individuals exhibiting inflammatory behavior (P< 0.05). F(2,483 = 4.21, p < 0.02, eta squared = 0.02) Hypothesis 2 21 In order to test hypothesis 2 a one-way AN OVA examined the effect of experience with computer-mediated communication on inflammatory disinhibition. The results were not significant. No other significant differences were detected. Means, standard deviations, and Ns for each group are displayed below in table 3. Table 3. Standard Deviation and Means ANONYMITY LEVEL N MEAN STANDARD DEVIATION l 143 0.0486 0.15 2 102 0.0254 0.09 3 239 0.0175 0.06 TOTAL 484 0.0283 0.10 When experience increased, disinhibition was lower between the low and high and low and medium anonymity groups. The moderately anonymous group, the group into which posters which were neither particularly anonymous nor made any effort to identify themselves, exhibited more experience than the low anonymous group. Posts of the informational nature tended to be personal experiences, often requesting advice regarding legal situations involving the law. Posts of an inflammatory nature were usually related to lawsuits or the behavior of police officers. The highest inflammatory posters tended to be those who were advertising and thus emphasizing every word in the post, or those using expletives to denigrate another poster. The least disinhibitive posters tended to keep their posts short and their answers in the third person. New users ofien treated the newsgroups as if they were writing a formal letter, and clearly identified themselves as low experience users (e. g., “This is my first time posting. . .”). Chapter 4 DISCUSSION The purpose of this study was to determine if experience with computer-mediated communication would alter a computer user’s behavior and perceptions. Specifically, the study tested the relationship between disinhibitive behavior under conditions of objective anonymity and experience with computer-mediated communication. It was argued that the results suggest highly anonymous groups exhibit inflammatory disinhibition. However, there was little support for the belief that low anonymity groups would exhibit higher informational disinhibition. And yet there is an important difference. The reason I included experience as a variable in determining the effects of anonymity is because there are clearly different forms of disinhibition. The data indicated that one form is offensive, often vulgar, argumentative, and sometimes bizarrely irrelevant to the current topic (or thread) in a newsgroup. Others are well thought-out, often abrasive responses or contests of verbal one-upmanship. Such verbal contests are a battle to attain status verbally in a social construct that does not give such cues freely. Both rich and poor, handsome and ugly, must resort to words to distinguish themselves. Although this experiment was exploratory, we can speculate as to why anonymity had an effect on inflammatory disinhibition. First, based on Walther’s work, users communicating over computer-mediated communication lacked the initial visual cues that help define first-time encounters with others. This lack of definition as a person makes a target of disinhibition much less of a person. A user may not be aware she is 22 23 inflicting disinhibitive behavior upon another because she does not recognize the other users as individuals. Second, because a user cannot suffer immediate physical harm and cannot be held legally or socially responsible for his actions, inflammatory disinhibition is a much more appealing recourse. There is no immediate accountability for any damage inflicted upon another user except verbal, and even that may be limited if a user is capable of ignoring or ending the conversation. While a face-to-face communication may be forced to continue because one person is screaming and the other cannot help but listen, a computer-mediated communication can be immediately ended. News posts can be ignored entirely because the user has much more control over communication channels. There was no significant effect for anonymity on informational disinhibition however. Users may have unintentionally provided more information about themselves. One possible reason for a user divulging such information is because users may associate a text-based medium that requires typing skills with more traditional forms of communication like letters and memos. The user then provides his name, address, and phone number in the letterhead, grouping all other users into one category and addressing the letter to them all (e.g., “Dear Sirs,”). Posts of a commercial nature may also provide informational content but intentionally be less anonymous in order to avoid retaliation. A computer-mediated community may highly resent newcomers pushing commercial advertising, and are much more likely to retaliate over computer-mediated communication rather than in person. Thus, a store wishing to advertise its wares or provide a phone number may include a 24 large amount of information but intentionally disguise its email address to avoid retaliation. There was also no significant effect for experience on inflammatory disinhibition. One possible reason might be that users who posted more often were doing so because they were more argumentative in the first place. Thus, a talkative user may exhibit more posts which do not necessarily reflect her experience level and instead be representative of her desire to communicate more often than other users. Users may also post repeatedly in order to argue their point, and thus express a gradual increase in disinhibitive behavior as the argument continues. Individuals who feel they have an identity to defend over a computer-mediated communication might argue more vehemently when attacked. A witty, invisible poster will gain little prestige if he cannot identify himself in some way. Limitations Unfortunately, it was not possible to delve too deeply into the experience of the poster beyond using the number of posts as a guide. A poster could well have repeated the same post several times, or used different computers to post each time, both of which would have skewed the number of posts for a user’s experience variable. Without the ability to carefully track posts as they are being created, it was exceptionally difficult to determine the true measure of a user’s experience with computer-mediated communication. There are plenty of computer users who have a specific medium they prefer to employ among forms of computer-mediated communication and thus may be highly experienced with one form and not the other. 25 The issue of anonymity is also problematic. We cannot measure perceived anonymity, because the presence of an observer would contaminate the construct. In addition to anonymity being difficult as a measure, anonymity masks other controls as well, such as gender and age. Historical factors also influence the experience measure because the more a user posts, the more experience he has. New users, although plentiful, are exceptionally difficult to isolate over a group over any given period of time. Because the nature of newsgroups is ever changing, what applies to the law- enforcement newsgroup today may no longer be applicable if the popular opinion on law- enforcement changes. The law-enforcement newsgroup ranged over a large number of topics, from police officers to lawyers, gun control to birth control. Finding a newsgroup that was a neutral topic to remove newsgroup influence was a challenge because only experienced users of those newsgroups can accurately determine their content. Newsgroups that scored closest to a 4 on a Likert scale were dominated by inflammatory posts and were discarded, indicating that the title of a newsgroup (alt.party) often has little to do with the content in the newsgroup. Implications These results provide several implications. Although there was a small sample of posters we can make attributions about Internet behavior from this newsgroup study. We were able to perceive a verbal difference manifested based upon a technical designation. Users who were anonymous were more likely to use expletives, exclamation points, capital letters, and emphasize their points. The more anonymous a user is, the more 26 sentences containing disinhibitive indicators will be present. A higher level of anonymity thus indicates an increased likelihood of disinhibition being exhibited. For computer-mediated forums, those wishing to avoid disinhibitive behavior should provide some means of identifying other users. By clearly holding everyone accountable for their communication by labeling their messages with full names, addresses, and emails, the likelihood of disinhibitive behavior will decrease. In a business or professional setting, this may be a very desirable precaution. Conversely, computer-mediated forums that wish to encourage disinhibitive behavior may want to ensure every user’s anonymity is secure. Making each user completely untraceable may facilitate a freer flow of communication. Newsgroups specifically devoted to argumentative or coarse discussions would thus attract more anonymous users. It is much easier to identify a single individual than to have several individuals identify themselves. It is up to the individual user to paint the appearance and expressions of the others he interacts with. Thus, without additional information to otherwise change such sculptures, narrow-minded views and opinions are allowed to fester. Technology now allows selective filters for news and commercials, allowing a user to only receive information over computer-mediated communication that she believes she would be interested in, and thus never allowing a dissenting voice or opinion to cross the user’s screen. Even the medium of computer-mediated communication is changing. It is rapidly combining and absorbing all of the other communication technologies, including radio, telephone, and television. With the addition of visual and audio cues, and complete 27 control over them, the subject of anonymity of self becomes an even greater issue. There are now virtual beings creating records in Japan and hosting game shows, given voice by mysterious people we never see. Virtual-reality chat rooms provide everyone with a face, even though it may not be their own. When dealing with such virtual personas, the identity definition defined by the graphics of the character must be questioned. Computer-mediated communication in the media and the perception of it also changes how users associate with the medium. Some commercials now show an attractive woman providing her e-mail address to someone asking her out on a date. More and more, the view of computer-mediated communication as being an intellectual playground is disappearing. This will affect the experience measure over time, creating more computer-savvy users who know much more about the individuals they are interacting with prior to using computer-mediated communication. Future Research Future research might distinguish the anonymity of self versus the anonymity of others, creating a dichotomy wherein inflammatory disinhibition can be tested for those who have little information about other users and informational disinhibition can be tested for those about whom little is known. In this way, the inflammatory disinhibitive user might be verbally aggressive to her audience because she does not recognize them as individuals, while the informational disinhibitive user might be more verbally open about herself because she thinks the other users do not recognize her as an individual. A controlled group of posters who voluntarily identified themselves prior to posting by 28 filling out a form which was not visible to other posters would preserve the sense of anonymity as well as allow tracking and identification of posts and post content. A larger cross-section of experienced and non-experienced users whose identities can be verified would greatly add to the veracity of the anonymity measure. By allowing such “tagged” individuals to treat the medium as different levels of anonymity (perhaps by informing the poster before he or she posts of his or her level of anonymity), the posters are then assured of being aware of their level of anonymity when they post. It was difficult to determine just how aware the low experience posters were of their anonymity. Just as new users have difficulty distinguishing personalities or even recognizing others as personalities over computer-mediated communication, users strive to be distinguished. While new users may be unable to appreciate their virtual community, established users are often all too aware of it, proudly displaying their ideas and ethos in a public forum. Future studies that focus on the perception of individuals over computer- mediated communication, the formation and cultivation of computer-mediated personas, and the impact of new information technology on users will help us learn more about ourselves and society. Given time, computer-mediated communication will soon follow in the footsteps of telephones, radio, and television. Like them, computer-mediated communication has been given an unqualified amount of praise for its ability to bring education and equality to the masses. Like them, computer-mediated communication is feared for that very same power it grants to the unknown user. And just as telephones, radios, and television were neither a great menace nor the perfect solution, computer-mediated communication 29 will eventually find its own niche in society, until our children’s grandchildren refer to it as just another form of communication. APPENDICES APPENDIX A APPENDIX A Pretest Experience Pretest $099397" I check my e-mail every day. I log onto the web frequently. I’ve been communicating over the Internet for a long time. I consider myself to be a highly experienced Internet user. I post to newsgroups ofien. Anonymity Pretest 1. SAP.“ I feel anonymous when I send an e-mail message with the account I’m using to answer this survey. I don’t trust this e—mail account because you never know who may read it on the other end. I avoid identifying myself over e-mail. I feel my privacy is guaranteed over e-mail. I avoid e-mailing sensitive information to anyone due to privacy concerns. 32 APPENDIX B APPENDIX B Deja News Power Search Resu/ts Microsoft Internet Exp/191 er . p __ '3 Home - Resource Directories *1 Qfimjnfih1hzlfiunhwfilmJNIrInnnllfll'flmflaflnfln ' d e b Power Search Results I w; ' ........................... Ear ......... Matches 1-25 of exactly 136 for search' [Worm & "9 {ammo} Date Set Sub ect Newsgroup Author 1. 97/11/10 036 Re: Accidents (was: Dope talk.p011tioa.dru¢s tirehawkfltiac 2. 97/11/10 035 Re: Accidents (was: #1/3 talk.politios.dru¢s firehawkfitiac 3. 97/11/03 034 Re: Alaska Curfew Termin alt.lal-entoroa-ent meflme.com 4. 97/11/01 034 Re: Pepe: Spragiand litt a1t.drugs.pot "D.G. Porter' 5. 97/10/27 032 Re: Accidents (was: #2/6 talk. politiol.drugs tirehaukfltiac G 97/111178 0’29 Da- Dn1~rr~a,hra ('r1m1ff'? I? a1L 1n-nnfnrnn-M...Frnngi Garnic'"“ V ‘ ' ’. . . e . ~ 5 7 V V W . V 7‘ . ‘ t , . - M; 4"“ (Thwfiifié‘f‘ .g'f.,§.-e.‘,~,f, . i‘ . -.ja‘.4.~?f~o‘ #3.,- 5:4,}? ~_.J4 \fihht I} r- 'J ‘1, ”1:33.11, _ I’m-A. _. ._ -.. "f'f 5““ ‘9 fl .fl.‘ , . ,.. ; "9,. . ’ ‘A' . x. ‘ - . .~ . . . . .. z . . ... . - . . , , . .. ..... . 4. . .. , . .. . -. . .. ,, . ,‘..‘.'.'.' '. ' , ., 34 APPENDIX C APPENDIX C .... Deja News- Author Pratt/e on meg e. cm - Mticrosot tntemet Exp... BE} E; 519 Edit flew 90 Favorites Help " " ' ~ » , . I v. .-..‘. AI'L.'~1~-;.t-~.'_',- ‘.--' -._-..-- 11-. (rs-9i.r\'uuc-mr1/teu.e{u- "11.7,; 'GF‘FL't‘ Home - Resorn'ce Directories W Let's see your caWarkers try to steal Jthis phane back from your desk. Author: me@me.com 0 19817 tmique articles posted. . Number of articles posted to individual newsgroups (slightly skewed by cross-postings): 0 17004 a1t.binaries.mac.games 0 315 gr_lt. sex. exhritionism . 213 altbalgpot 0 135 alt.poliucsimmigration : , 113. alt-88888833. .. . .. .A . _. ,. .. ,. , __._ .. s 36 APPENDIX D APPENDIX D Sample Post and Coding From: meat@captain.net (Captain Meat) Officer Blah from his recent Usenet posts *may* too join the ranks of hypocrites who used marijuana, but now tell others not to...even though marijuana is just a plant and has been used for thousands of years throughout the world with few adverse effects other than the zealous attempts to ban it. Marijuana has never directly killed anyone unlike alcohol which kills over 100 people *directly* each year...surely most remember the recent LSU incident where a young student died from alcohol. Over 70 million Americans have used marijuana including some major hypocrites which includes President Clinton, Al Gore, and Newt Gingrich! If marijuana is so dangerous, addictive, bad, etc. then surely there would be a marijuana epidemic considering over 70 million Americans have used it, but of course the only epidemic is the arrest of millions for possessing a plant. It's a sad commentary and as others have correctly noted, the hypocrites despite their efforts to the contrary are actually accelerating the reform process for who can really take their zero-tolerance, "just say no" message seriously when they used marijuana themselves...it should be no coincidence that drug use statistics have showed a steady increase since President Clinton's lame admission of smoking marijuana on MTV in 1992. Anonymity From: meat@captain.net (Captain Meat) Posting under a fictional name, high anonymity = 1 Inflammatory Disinhibition Emphasis (**) 2, Exclamation (l) 1, Name Calling (hypocrite) 2 TOTAL INFLAMMATORY DISINHIBITION: 5 Informational Disinhibition Proper Name of Another Poster (Officer Blah) 1, Proper Name of Others (President Clinton) 4, His/Her/Theirs (his) 1 TOTAL INFORMATIONAL DISINHIBITION: 6 TOTAL WORDS: 204 TOTAL INFLAMMATORY DISINHIBITION PERCENTAGE: 2.45% TOTAL INFORMATIONAL DISINHIBITION PERCENTAGE: 2.94% 38 APPENDIX E APPENDIX E Sample Posting and Coding II From: jsmith@ameritech.net (John Smith) Is the increased number of traffic deaths really due to the no-daytime-highway-limit, or are there other factors to consider? Is the increase perhaps a statistical anomaly? Was the increase comprised mainly of no limit-highway fatalities, or posted-limit road fatalities? Or maybe the increase is due to abuse of posted limits by drivers accustomed to no limit on the highway? Maybe law enforcement has shifted resources to the no-limit highways to keep speed in check, while secondary road speed-limit enforcement is ill- maintained? There will always be highway fatalities, and most will be blamed on 'excessive speed', which kills by rapidly decelerating the fragile human organism into hard objects. So too is the cause of death in most airline disasters this same event. Drivers and airlines are both regulated, and the regulators of both are continually etching an arbitrary line of rule enforcement. That line ofien strikes the balance of life and death for both the innocent and the culpable. How would you draw that line? What price speed? If you were a cop, where would you draw your line? I recall that even Officer Blah, who as an investigator must witness the aftermath of fatal road crashes, and who staunchly defends traffic law, will yet not set a hard and fast rule of when he will give a speeding ticket, but only say that breaking the speed limit is an indefensible offense of the law. So if you were a cop, where would you draw your line, or if you *are* a cop, where do you draw your line? Anonymity From: : jsmith@ameritech.net (John Smith) Posting under a normal name, last name used, low anonymity = 3 Inflammatory Disinhibition Emphasis C”) 1 TOTAL INF LAMMATORY DISINHIBITION: 1 Informational Disinhibition Proper Name of Another Poster (Officer Blah) 1, His/I-Ierffheirs (his) 1, Question (8), You/Y our (10), I (1), He/Him/Her (1), TOTAL INFORMATIONAL DISINHIBITION: 32 TOTAL WORDS: 258 TOTAL IN F LAMMATORY DISINHIBITION PERCENTAGE: 0.38% TOTAL INFORMATIONAL DISINHIBITION PERCENTAGE: 12.4% 40 LIST OF REFERENCES Belson, D. (1994). The Network Nation Revisited, paper presented at the Stevens Institute of Technology. Brook, James and Ian A. Boal (editors), “Resisting the Virtual Life: The Culture and Politics of Information,” City Lights: San Francisco, 1995. Cathcart, R. & Gmnpert, G. (1983). Quarterly Journal of Speech, Q, 267-277. Denning, Dorothy E. and Herbert S. Lin (editors), “Rights and Responsibilities of Participants in Networked Communities,” National Academy Press: Washington, DC, 1994. F roomkin, M. (1995). Anonymity and Its Enmities. Online 4. Kerr, E. B. and Hiltz, S. R. “Computer Mediated Communication Systems: Status and Evaluation,” Academic Press: New York, 1982. Kielser, S., Siegel, K., and McGuire, T.W. (1984). Social psychological aspects of computer-mediated communication. American Psychologist, 39, 1123-1134. Kiesler, S. & Sproull, L. (1992). Organizational Behavior and Human Decision Processes, 52, 96-123. Orcutt, J. D. and Anderson, R. E. (1993). Social Interaction, Dehumanization and the “Computerized Other”, fl, 380-397. Rafaeli, S. (1988). Interactivity: From new media to communication. In R.P. Hawkins, J .M. Wiemann, and S. Pingree (Eds), Advancing communication science: Merging mass and interpersonal process (p. 110-134). Newbury Park CA: Sage. Rice, R. E. & Love, G. (1987). Communication Research, fl, 85-108. Sandvig, C. (1995). Special Focus: Government Online, CMC Magazine, 7-10. Schaeferrneyer, M. J. & Sewell, E. H., Jr. (1988). Communicating by electronic mail. American Behavm Scientist. 32. 112-123. Siegel, J ., Dubrovsky, V., Kiesler, S., and McGuire, T.W. (1986). Group processes in computer-mediated communication. Organizational Behavior a_nd Hm Decision Processes. 37, 157-187. 41 Spears, R. & Lea, M. & Lee, S. (1990). British Journal of Social Psychology, 29, 121-134. Walther, J. B. (1992). Interpersonal Effects in Computer-Mediated Interaction. Communication Research. 12, 52-90. Walther, J. B. (1993). Impression Development in Computer-Mediated Interaction. Western Journal of Communication. 57, 381-398. Walther, J. B. (1994). Anticipated Ongoing Interaction Versus Channel Effects On Relational Communication in Computer-Mediated Interaction. Human Communication Research. 20. 473-501. Walther, J .B. (1995). Relational aspects of computer-mediated communication: Experimental observations over time. Organization Science. 6 (2), 186-203. 42 "‘titiiiiltiilt