An Exploratory Analyses Of The Social Nature Of Internet Addiction:
A Research Paper Submitted To The Electronic Journal Of Sociology


Al Bellamy Department of Interdisciplinary Technology School of Technology Eastern Michigan University

Ypsilanti, Michigan

Cheryl Hanewicz Department of Interdisciplinary Technology School of Technology Eastern Michigan University

Ypsilanti, Michigan


This study explored the influence that personal relations and communications within Internet chat rooms has on a user’s Internet Predisposition – a concept developed to refer to what psychologists have vicariously termed as Internet addiction. Two measures of Internet Predisposition were utilized: a quantitative measurement based upon time spent in chat rooms and the Internet, and a four item Internet Predisposition Scale (IPS) developed by the authors. Results of the study indicate that the IPS is significantly correlated with certain personal relations and communication variables. It further revealed stronger correlation between these factors and the IPS in comparison to the quantitative measurement of Internet addiction. The study also examined the moderator influence of gender, locus of control, and sociability on the relationship between Internet addiction and chat room personal relationships and communications. Gender was shown to have the strongest moderator influence upon these relationships.


The concept of Internet addiction has recently entered the social problem lexicon. Inordinate amounts of time spent engaging in various types of Internet activities such as muds, chat rooms, and discussion groups have been cited as having a negative impact on social relationships, marriages, school achievement, work performance, health, and other vital life functions (Young, 1998; King, 1996). Given the prediction that 80 percent of American households will be connected to the Internet by 2000 (Young, 1998), Internet addiction is perceived as a possible societal epidemic. To address this increasing concern, The Center for On-Line Addiction (1998) has classified Internet addiction into five specific types:

Cybersexual Addiction – Addictions to adult chat rooms or cyberporn.

Cyber-relationship Addiction – On-line friendships made in chat rooms, MUDS, or newsgroups that replace real-life friends and family.

Net Compulsions – Compulsive online gambling, online auction addiction, and obsessive online trading.

Information Overload – Compulsive web surfing or database searches

Computer Addiction – Obsessive computer game-playing or to programming aspects of computer science.

The study reported within this paper is related specifically to the second type, Cyber-relationship Addiction, and focuses primarily on attitudes toward and conduct in electronic chat room communication. This type of interchange is a group and mass communication system, in which users send and receive text-based messages. The time delay of these computer-mediated messages can be nearly instantaneous or “real time” text interchange (December, 1996). The social aspects of chat room communication make the subject of addiction amenable to sociological investigation. Currently, the vast majority of the empirical and theoretical inquiry into this phenomenon has utilized psychological frameworks.

Purpose OF Study

The purpose of this paper is to explore the extent to which social factors, delineated in terms of interpersonal communication variables, are associated with what is referred to as Internet addiction among college students at a large Midwestern University. We also want to determine if gender, locus of control, and individual sociability moderate the relationships between communication variables and Internet utilization. This type of analysis makes the current study more systematic in its approach for understanding the dynamics underlying the so-called Internet addiction phenomena. Previous studies on this topic have been more descriptive in their methodological approach and have not systematically attempted to empirically describe the manner in which the variation in Internet utilization is explained by selected social or social psychological factors. We contend that such an analysis will offer a different viewpoint of the so-called Internet addiction phenomena that has been presented by the psychological frameworks that currently predominate the discussion and analyses on this topic.

The paper will proceed with the following format:

  1. A brief overview of the psychological approaches to Internet addiction
  2. Discussion of sociological factors that may affect Internet utilization
  3. Statement of research questions that incorporate social communication variables
  4. Description of research methods
  5. Analyses of data
  6. Discussion

Psychological Inquiry Into Internet Addiction

Psychologists have labeled Internet addiction as Internet Addiction Disorder – IAD, a term first used by Goldberg (1996). According to Goldberg, IAD exists when the individual experiences “decreased occupational, academic, social, work-related, family-related, financial, psychological, or physiological functioning”. He suggests that a parallel of IAD would be pathological gambling.

In an online survey of 496 web users, Young (1998) categorized 80 percent of the respondents as being addicted to the Internet. She concludes that the most salient factors contributing to IAD are the capabilities of an individual to take on different roles, the anonymous nature of computer-mediated communications, and the prospects for developing meaningful interpersonal relationships. She used the following criteria (similar to that of compulsive gambling and alcoholism) in determining Internet addiction:

  1. Feeling preoccupied with the Internet (for example, thinking about previous online activity or anticipating the next online session).
  2. Feeling a need to use the Internet with increasing amounts of time in order to achieve satisfaction.
  3. Making repeated unsuccessful attempts to control, cut back, or stop Internet use.
  4. Feeling restless, moody, depressed, or irritable when attempting to cut down or stop Internet usage.
  5. Remaining online longer than intended.
  6. Jeopardizing the loss of a significant personal relationship, job educational, or career opportunity.
  7. Lying to family members, friends, or a therapist to conceal the extent of their involvement with the Internet.
  8. Using the Internet as a means for escaping from problems (Young, 1998, pp. 4-5).

In another online survey conducted by Brenner (1996) of 185 Internet users it was reported that 17 percent may be addicted to the Internet, which included spending 40 or more hours per week online. The survey indicated that nearly half of the respondents experienced adverse effects in their work as a result of online usage. Ten percent reported problems with employment and school due to online activity.

Cojac, (1996) reports that there is an inverse relationship between the number of hours spent per day on the Internet and the average number of sleep hours per night. It appears from this particular research that Internet utilization affects sleep patterns, an outcome that is common to other addictions such as alcohol and drug addictions.

A common idea that is reported within the majority of these studies is that Internet addiction is affected by social interactions. However, none of the studies have empirically examined the ways in which IAD correlates with social factors. Furthermore, none of the studies have utilized conceptual frameworks that would explain the sociological nature of their findings.

Our objective within the next section of this paper is to discuss some sociological aspects of computer mediated communication that could serve as a context for exploring heavy Internet utilization from a sociological perspective.

The Sociological Context OF Internet Addiction

Contrary to earlier propositions that the cueless structure of the Internet, and chat rooms in particular, would create a normless context in which individuals would be less constrained to exhibit deviant type behavior, and where interpersonal relations would be tenuous and superficial (Dubrovsky, Kiesler, and Sethna, 1991; Siegel, Dubrovsky, Kiesler and McGuire, 1986; Rice, 1993; Rice and Love, 1987; Short, Williams, and Christie, 1976; Sproull and Kiesler, 1991), recent research has shown that electronic communication mediums such as chat rooms are organized according to social standards (Bellamy and Hanewicz, 1998) and that altruistic friendship relations are common (Parks, 1996). Indeed, it is the social characteristics of chat rooms that may be the primary antecedents and reinforcements generating the so-called addictive behavior (Young, 1996; Suler 1996; King, 1998).

Attributes of Internet communications that enhance its social appeal are:

  1. Individuals have easy and flexible access to communication 24 hours a day.
  2. The Internet increases the range of potential social networks, and this diversity is appealing to many individuals (Wellman, 1996).
  3. Individuals have almost total control of self-presentation. Individuals can create and maintain aspects of themselves that would be difficult to present in face-to-face (ftf) situations. Electronic communications ameliorates significant characteristics of the individual (i.e., gender, ethnicity, age, handicap, etc.) that might adversely affect social relationship formation in ftf encounters. It is what King (1998) states as the “hyperpersonal aspect” of Internet communications, a “way to be more selective about how one presents one’s self. This promotes a sense of group membership, one that is solely dependent on the perceptions of the receiver.”
  4. The anonymous identity of individuals makes it possible for them to send and receive communications that they would be reluctant to transmit in conventional ftf situations.
  5. The Internet allows individuals to leap over geographical boundaries and “..expand the ability of people with common interests to share ideas important to them” (King, 1999).

These as well as other factors may make Internet communication very attractive for some individuals, individuals who have been labeled as Internet Addicts within current psychological literature. The Internet, where conventional information symbols are absent, represents a radically different situational platform for interpersonal electronic communication. It allows the individual to more freely engage in reciprocal communication in comparison to traditional face-to-face environments.

Sensitizing Conceptual Platform

Given the nascent state of empirical and theoretical knowledge on Internet communications within sociology, it is indeed helpful to refer to extant theoretical frameworks that can be utilized to engender a sociological imagination towards Internet research. While the major objective of the present paper is to explore the correlational structure of Internet communication factors, we will reference tenets of symbolic interaction theory (Mead, 1934) and Stryker’s identity theory (1980). and utilize them as “sensitizing” concepts” (Giddens, 1996) while analyzing our data.

From a sociological perspective, identity emerges from the recursive interplay between taking the role of others in a given situation (where the process of taking-the-role of the other is influenced by the symbols present within a given social situation) and the social interactions that occur from

a situational context that has been symbolically defined by the individual. (Mead, 1934). The Internet, where conventional information symbols are absent, represents a radically different situational platform for self-presentation, which in turn may have an influence on identity formation.

Within sociology, identity is considered a subset of “self”. That is, self consists of several identities, wherein the manifestation of a specific identity is contingent upon the exigencies of a social situation. Stryker, (1980) suggests that identities are arranged according to a “salience hierarchy”, whereas the dynamics of a social situation determine the identity or identities that will be evoked. “The salience hierarchy determines those identities that respond as people orchestrate their roles and interpret the role behaviors of others……when an interaction situation is isolated from structural constraints, or these structural constraints are ambiguous, individuals will have more options in their choice of an identity” (Turner, 1998).

The point that we are attempting to make here with Stryker’s concept of the saliency hierarchy of identity is this: When simultaneously taking into consideration the factors that we have delineated that enhances the appeal of the Internet, and the idea that the Internet is a situation having few

structural constraints as described by Stryker, the Internet can be conceived as a viable context in which identity can be more readily manipulated by the individual (This may be particularly relevant for presenting one’s “ideal” self). Fewer symbolic props are needed in order to present self (Goffman, 1959) to others in this type of social environment than in FTF conditions. This is a very important and significant sociological characteristic of electronic communication. The individual may effectively present not just one or two identities, but as many that are contained within one’s compendium of self. In other words, “self” can be more systematically and comprehensively presented within this environment than in FTF situations. Furthermore, the saliency attached to each identity is more in the control of the individual, rather than the social context, which may give the person a heightened sense of “self control” and /or personal life satisfaction. However, the processes underlying these factors are social, for the identities that are being projected by the individual are yet still attached and nourished by the expectations of others (with similar self objectives and methods of defining the situation) within the electronic communication environment. This idea is supported by Walther (1996) in his statement regarding communication feedback on the Internet:

“Another component of the model, feedback, suggests that these heightened self-presentations and idealized perceptions magnify each other to a superordinal level, as users reciprocate each other’s partial and selective presentations.”

Once again, the theoretical ideas presented above are not directly tested within this literature.

However, it does provide conceptual legitimacy to a study of Internet addiction from a sociological perspective.

The Moderating Influence of Gender

Men and women have different cultural orientations that may influence the way in which they define electronic communication situations. “Gender characteristics are a primary means by which we sort and define self and others. Sex attributes provide basic information about how to conduct interactions with others and how to organize social reality” (Obrien, 1999). Gender, as a “social” construct has not been systematically examined within Internet addiction research. Our objective is to measure the extent to which gender influences the connection between so-called Internet addiction and social communication factors.

This idea is supported by Walther (1996) in his statement regarding communication feedback on the Internet. “Another component of the model, feedback, suggests that these heightened self-presentations and idealized perceptions magnify each other to a superordinal level, as users reciprocate each other’s partial and selective presentations.”

Taken altogether, these structural characteristics of chat rooms may function as very powerful enticements for human beings. From a sociological perspective, heavy chat room users may be people who attach a higher degree of saliency (Stryker and Serpe, 1982) to communication related factors such as establishing friendships in chat rooms than individuals who utilize it less frequently. They may be more oriented towards defining chat rooms as places where feelings can be expressed and where personal relationships are important, where people are good listeners and so on. Since the appeal of the Internet has a lot to do with interpersonal communication, we would expect heavy users to be relatively sociable people. We will test this proposition within the current report by testing for the moderator influence of an individual’s sociability orientation on the relationships between our Internet Predisposition scale (a term that will be used in place of Internet Addiction) and various factors related to social communications.

The freedom that the Internet allows for self-presentation would also suggest that a heavy user would possibly be a person of moderate to high self-esteem and confidence. This idea will be examined by reviewing the impact that locus of control (Rotter, 1966) has on the correlation between Internet Predisposition and various communication variables.

Research Questions

Based upon the ideas presented within the theoretical discussion, this paper will attempt to examine the correlation between chat room and Internet utilization, and a scale that measures Internet predisposition and the following attributes of individuals:

  1. Perceptions regarding personal relationships formed on the Internet.
  2. Perceptions of chat room communications processes.
  3. Locus of control (which is a surrogate measurement of self esteem and confidence).


The extent to which there are differences between individuals categorized as High Internet Predisposition (HIP) and Low Internet Predisposition (LIP) chat room users in terms of why they visit chat rooms.

Men and woman have different cultural orientations that may influence the way in which they define electronic communication environments. Such differences in turn may affect the nature of the correlation between the measurements of Internet addiction and the personal relationship and communication factors delineated above. This paper will examine the extent in which gender moderates these relationships, which has not been conducted in previous Internet addiction studies.


Data for this study was collected from 114 undergraduate and graduate students in a relatively large university in Southeast Michigan during the months of April – June 1998. The undergraduate students were enrolled in a technology and society-type class that satisfies a basic studies requirement at the university. Subsequently, the sample population represents a wide spectrum of undergraduate degree programs and career orientations within the university. A full sample was taken among students who identified themselves as chat room users (for both undergraduate and graduate students). The graduate students were all enrolled in an interdisciplinary technology program. Each student completed a 104-item questionnaire (during class time) that measured a variety of Internet and chat room utilization factors.

The following summarizes the demographic structure of the study:

Demographic Structure of Study



Internet Predisposition (Ip)

An underlying theme of this paper is that the term Internet Addiction as delineated by psychologists is a construct that has not been empirically validated. Subsequently, this paper will deploy the term High Internet Predisposition (HIP) to describe individuals who score high on our IP scale. The following four (4) items were used to measure Internet Predisposition:

  1. I spend less time doing the things that I used to do now that I use the Internet.
  2. Other people (i.e., friends and relatives) have complained that I spend too much time on the Internet.
  3. Spending time on the Internet has interfered with my relationships with other people.
  4. Spending time on the Internet has affected my academic and/or work activities

Each of the above items utilized a five-point, Likert-type scale with response categories ranging from Agree to Disagree. These four items were summed together to form an Internet Predisposition Scale (IPS). The alpha reliability of the scale is .74. The range of scores for this sample is 4 – 19 with a mean average of 7.18 and a median of 6.00.

For the purposes of this study, HIP has been operationalized as being a score of above 11: n=18 or 16% of the sample population. Based upon the concepts presented within the literature on what constitutes Internet Addiction, this scale appears to be high on content validity. In order do conduct a comparison between “addictive” and “non-addictive” chat room users, the IP scale was recoded into two categories. Scores ranging from 4 thru 11 denote low Internet predisposition (LIP). This category contains 96 cases or approximately 84% of the study population. Scores ranging from 12 thru 19 constitute high Internet predisposition. Eighteen (18) cases or approximately 16% of the study population are included within this range of scores.

Another method for measuring IP is the number of hours per week that people spend in chat rooms and the Internet in general. We included both of these variables within our survey. The correlation between the two variables is .75 (p=.00). This very highly correlation indicates that both variables are measuring the same construct. – time spent on the Internet. Nevertheless, this paper will present data for each of these variables. The correlation between time spent in chat rooms and the Internet with the IP scale is .225 (p=.00) and .342 (p=.00) respectively.

Personal Relationships

The following items were used to measure perceptions regarding personal relations in chat rooms:

  • The personal relationships that I have formed in chat rooms are just as important as those that I have formed outside of chat rooms. (Perform)
  • It is easy for me to form personal relationships in chat rooms. (Easyform)
  • I feel that I could confide in the people that I form personal relationships with in chat rooms. (Confide)
  • I feel quite close to the people I have formed personal relations with in chat rooms. (Feelclose)
  • I am very committed to maintaining my chat room relationships. (Commitment)
  • I form personal relationships in chat rooms almost immediately. (Formper)

Perceptions of Chat Room Communications

  • I feel comfortable communicating in chat rooms. (Comfortable)
  • It is easy to express your feelings in a chat room. (Feelings)
  • People in chat rooms are better listeners than people outside of chat rooms. (Listeners)
  • It is easier for me to communicate with people in chat rooms rather than face-to-face. (Easycom)
  • People that I chat with on the Internet are more accepting of who “I am” than people outside of the Internet. (Accepting)
  • Interacting with people in chat rooms has given me a different perspective of myself. (Self)
  • People in chat rooms are just as friendly as people I meet face-to-face. (Friendly)

Sociability (Soc)

Sociability was measured by a 7-item, Likert-type scale consisting of anchor points ranging from Agree to Disagree created by Hanewicz and Bellamy (1998). Scores ranged from 7 to 50 with a mean and median of 26.2 and 27 respectively. The alpha reliability for this scale is .72. (See Appendix A for entire scale.)

Locus of Control (Loc)

Locus of Control (Rotter, 1966) is a personality orientation variable which delineates how individuals attribute outcomes related to their actions. People who see themselves as being able to “control” events of their actions are referred to as internals. Those who are more oriented towards believing that events are outside of their control are characterized as being externals. Locus of control was measured by a 10-item scale developed by Burger (1986) which consisted of 7 scale points ranging from strongly agree to strongly disagree. Scores ranged from 24 to 80 with a mean and median of 49.1 and 51 respectively. Higher scores indicate an internal orientation. The alpha reliability coefficient for this scale is .62. (See Appendix A for the entire scale.)


The correlation between the IP scale and these variables is a way to determine whether individuals with higher levels of IP are attaching higher saliency towards personal relationships. The data revealed in Table 1 indicates that the IP scale is significantly correlated with five out of the seven personal relationship variables and the direction of each of these is positive. Overall, this finding appears to be suggesting that Internet users who report that time spent within chat rooms are negatively affecting their social and personal lives, are more likely to report that they form personal relationships in chat rooms, that they form relationships easily, that they confide in individuals with whom they have formed personal relations, and that such relations are characterized with a degree of closeness and commitment. However, these correlations are somewhat small, thus giving limited (yet, statistically significant) support to our idea that HIP individuals are attaching greater saliency to personal relationships.

The results in Table 1 illustrate that the IP scale is a much better predictor of interpersonal perceptions than either of the quantitative measurements of Internet addiction (at least for this sample). What is interesting is that the IA scale is significantly correlated with both the chathour and nethour variables. This appears to allude to the possibility that our social psychological measurement of IP is measuring a different construct than the quantitative measurements.

If Internet addiction is similar to other addictions such as gambling and alcoholism one would expect that it would be significantly and negatively correlated with variables that measure aspects of self esteem such as Locus of Control. That is, the higher the addiction, the lower the self esteem. Since locus of control is a surrogate measurement of self-esteem and confidence, it is expected that there would be a high correlation between IP and LOC. However, none of the addiction measurements are significantly associated with locus of control. Instead, significant correlations appear between the IP factor and social factors, giving additional support to the idea that much of what has been labeled Internet Addiction might not be, in fact, an addiction. This is not to say that true Internet addiction does not exist, only that it may not take the form of other, more “traditional” addictions, and it may be hasty to apply current addiction doctrine to a new form of social communication.

Demographic Structure of Study

Table 1: Zero Order Correlations Between Internet Predisposition, Sociability, Locus of Control, and Personal Relationship Variables For Entire Population