November 9, 2007
Internet Use, Abuse, and Dependence Among Students at a Southeastern Regional University
By Fortson, Beverly L Scotti, Joseph R; Chen, Yi-Chuen; Malone, Judith; Ben, Kevin S Del
Abstract. Objective: To assess Internet use, abuse, and dependence. Participants: 411 undergraduate students. Results: Ninety percent of participants reported daily Internet use. Approximately half of the sample met criteria for Internet abuse, and one-quarter met criteria for Internet dependence. Men and women did not differ on the mean amount of time accessing the Internet each day; however, the reasons for accessing the Internet differed between the 2 groups. Depression was correlated with more frequent use of the Internet to meet people, socially experiment, and participate in chat rooms, and with less frequent face-to-face socialization. In addition, individuals meeting criteria for Internet abuse and dependence endorsed more depressive symptoms, more time online, and less face-to-face socialization than did those not meeting the criteria. Conclusions: Mental health and student affairs professionals should be alert to the problems associated with Internet overuse, especially as computers become an integral part of college life. Keywords: dependence, depression, Internet abuse
Sex Differences in Internet Use
Although researchers have shown little difference in the amount of time men and women spend online, they have consistently found that men and women differ in their reasons for accessing the Internet. Weiser7 found that men were more likely than were women to use the Internet for purposes related to entertainment and leisure, whereas females used it primarily for interpersonal communication and educational assistance. Odell et al8 similarly found that men reported greater use of the Internet for visiting sex sites, researching purchases, checking the news, playing games, and listening to and copying music, whereas more women used the Internet for e-mail and school-related research. Researchers9-11 have obtained many of these same results cross-culturally; however, results by Joiner et al12 were not entirely consistent with these previous studies. Joiner et al found that men were more likely than were women to use the Internet for leisure activities (eg, downloading material from the Internet, using game Web sites), but women did not use the Internet for communication more than men did.
Problems Related to Internet Use
A small percentage of college student Internet users develop problematic behaviors, such as cravings, sleep disturbance, depression, and withdrawal symptoms, as a result of their Internet use.13,14 From a sample of 531 college students, Scherer4 found that 13% met criteria for Internet dependency and, as such, believed their Internet usage interfered with their daily functioning. The Internet-dependent students were predominantly male and reported more leisure-time Internet use than did nondependent students. Approximately 9% of the college students in Anderson's5 study endorsed dependence on the Internet. Morahan-Martin and Schumacher15 also found that 8.1% of college students met their criteria for pathological Internet use. Again, most of the pathological users were male and were more likely to use online games and technologically sophisticated sites (eg, file-transfer protocols, remote support communication software, virtual reality). Morahan- Martin and Schumacher15 also found that pathological Internet users reported being generally lonelier than others and more socially disinhibited online.
In the research on problematic Internet use, experts have typically defined abuse and dependence using criteria similar to that for pathological gambling, suggesting that Internet addiction is considered a behavioral addiction. Such a definition is controversial, with opponents of the use of these criteria holding that Internet addiction (as well as sex or food addictions) is not based on empirical research, as is pathological gambling.13 In addition, pathological gambling involves more serious financial issues (eg, loss of large sums of money, illegal activities to repay losses, heavy borrowing from legal and illegal sources) than does pathological Internet use. In response to such issues, Anderson5 used criteria modeled after the substance-related disorders from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition text revision (DSM-IV-TR16) to evaluate pathological Internet use.
In the present study, we examined Internet use among college students using a questionnaire that we constructed primarily on the basis of the work of Scherer.4 Like Scherer, we assessed social styles and preferences for the types of therapy one would choose if one were to seek professional help for a problem, such as Internet dependence. Unlike Scherer, we modeled our Internet abuse and dependence criteria after the DSM-IV-TR16 criteria for substance abuse and dependence. We considered participants to abuse the Internet if they endorsed 1 or more symptoms supporting a maladaptive pattern of behavior that resulted in significant impairment or distress (eg, failure to fulfill major roles, legal problems related to Internet use, continued use despite problems). We also considered them Internet dependent if they endorsed 3 or more symptoms of dependency (eg, tolerance; withdrawal; being online for periods longer than intended; impairment in social, occupational, or recreational activities because of Internet use). In a departure from the previously cited studies, we gathered data via paper surveys and over a restrictedaccess Internet site.
We recruited 485 (55% female) undergraduate students enrolled in an introductory psychology course at a large southeastern regional university. Of these, we deleted 74 from the final data set because of incomplete questionnaires; these students did not differ from the final sample on any available demographic variables.
The final sample of 411 participants was 56% female, 91% Caucasian (4% African American, 2% Asian American, and 2% Hispanic), and primarily (50%) from West Virginia (Pennsylvania, 20%; Maryland, New Jersey, and Virginia, 6%-7% each; and the remaining 11% from 13 other states and 3 countries). Of the participants, 63.7% were freshmen; 22.6%, sophomores; 9.8%, juniors, and 3.9%, seniors. On average, participants were aged 20.4 years (SD = 3.2, range = 18- 56). Men were slightly older (M = 20.8 years, SD = 3.5) than were women (M = 20.1 years, SD = 2.9), t(406) = 2.0, p
We included questions pertaining to sex, race and ethnicity, year of birth, year in college, and current state or country of permanent residence to describe the sample. Participants also answered questions about social style and preferences for therapy. The first item on social style was about perceived sociability (1 = very sociable, 2 = sociable, 3 = sociable, but shy or introverted, 4 = not really sociable; somewhat of a loner); the second item was about contexts for social interaction (1 = more face-to-face than on the Internet, 2 = equally face-to-face and on the Internet, 3 = more on the Internet than face-to-face, 4 = seldom, I do not socialize much face-to-face or online). We also assessed preferences for 7 therapy formats (from face-to-face with an individual to online with a group) if one were to ever seek psychological treatment (eg, for Internet abuse or dependence). Tables 1, 2, and 3 show all items and related rating scales.
The demographics questionnaire also contained 9 items to evaluate whether participants felt that their use of the Internet interfered with their daily functioning; that is, did they meet criteria for Internet abuse and dependence (on the basis of DSM-IV-TR criteria, as previously discussed).
The Internet Usage Questionnaire consisted of 17 items to determine how often the participants accessed the Internet and for what purposes they did so. We constructed these items on the basis of Scherer's4 work and scored them from 0 to 4 (0 = no; 1 = yes, at least once per year; 2 = yes, at least once per month; 3 = yes, at least once per week; 4 = yes, at least once per day). Table 4 shows the questionnaire items. We obtained a Cronbach's alpha of .62, indicating an acceptable level of internal consistency for a short research survey with nonhomogenous items of this kind.17,18 Depression
The Center for Epidemiological Studies Depression Scale (CES-D) is a 20-item questionnaire that we used to identify the presence and severity of depressive symptomatology. Higher scores on the measure indicate a higher frequency of occurrence of the symptoms, with a score of 16 or more suggesting clinical cases of depression. The CES- D has high internal consistency, moderate test-retest reliability, and concurrent and construct validity.19
We conducted this study, which the university's institutional review board approved, as a portion of a larger project20 comparing the results of psychometric surveys completed via the Internet with those completed on paper. We entered participants into 1 of 2 conditions (paper or Internet) on the basis of the data collection session they attended; participants were unaware of the conditions prior to arrival. All participants completed informed consent agreements. Those in the Paper Condition completed the survey on paper immediately; we gave those assigned to the Internet Condition a slip of paper with the Web address to the restricted study site (along with a user name and password), asked them to complete the questionnaire within the next 24 hours, and dismissed them. Of the 411 participants in the final sample, 211 completed the survey on paper (51%), and 200 completed it via the Internet (49%). Participants in these 2 conditions did not differ on key demographic variables nor on the time they typically spent accessing the Internet each day. After survey completion, participants received extra credit for a course. (This research project was one of multiple opportunities for students to earn extra credit in their courses.) All participants also received a list of referrals in the event that answering the questions led to psychological difficulties or if an individual wanted to follow up with psychological services after participating in the research.
Paper Versus Internet Survey Completion
We conducted multivariate analyses of variance (MANOVAs) to examine whether differences existed in responding between those who completed the survey on paper versus the Internet. We analyzed each major variable category (eg, frequency of accessing the Internet for different reasons and services, sociability, symptoms of Internet abuse and dependence, therapy preferences, and depressive symptoms) in separate MANOVAs. We found no differences between Internet and paper respondents for most categories and their component variables. The one exception was a significant effect for survey version: Internet and paper respondents differed in their reasons for accessing the Internet, Hotelling's T = .04, F(6, 406) = 3.0, p
Depression, Sociability, and Therapy Options
On the CES-D, we obtained a mean score of 13.9 (SD = 8.9), with 33% of the sample exceeding the clinical cutoff score of 16 or above, suggesting the presence of clinical depression. Table 1 shows ratings of participants' levels of sociability and their preferences for modes of socialization. Most students described themselves as very sociable (34%) or sociable (46%) and reported that they socialized more face-to-face (74%) than they did by other means. Table 2 shows preferences for therapy options. If seeking treatment for psychological problems, most participants reported that they would prefer treatment to be face-to-face with an individual (M = 4.0; SD = 1.1).
On the Internet Usage Questionnaire, 90% of the participants reported daily use of the Internet for some activity (9.5% weekly use, 0.5% monthly use), such as e-mail, Web access, chat rooms, and shopping. Time accessing the Internet ranged from less than 30 minutes a day (20%), to 30-60 minutes (31%), 1-4 hours (37%), 4-8 hours (9%), 8-12 hours (1%), and 12-24 hours (1%). Table 4 provides descriptive data on reasons for using the Internet and the services accessed.
Internet Abuse and Dependence
We calculated Internet abuse and dependence by using a set of liberal (ie, ratings at the midpoint or higher on each relevant item; sometimes to very frequently or somewhat to yes definitely) and conservative (ie, ratings only at the high point on each relevant item; very frequently or yes definitely) criteria. More than half (57.2%, n = 235) of the sample reported a pattern of behavior sufficient to meet criteria for Internet abuse under the liberal criteria; 21.9% (n = 90) met the definition for abuse using the conservative criteria. In both cases, more than 95% of those meeting abuse criteria endorsed continued Internet use despite current social problems, and less than 5% reported legal problems related to Internet use.
One-quarter (26.3%, n = 108) of the sample reported a pattern of behavior sufficient to meet criteria for Internet dependence under the liberal criteria (60.2% of these indicating tolerance and/or withdrawal); 1.2% (n = 5) met the definition for dependence using the conservative criteria (80% of these indicating tolerance and/or withdrawal). Participants endorsed all the individual criteria for dependence at similarly high rates. Table 3 contains the statistics for each of the individual abuse and dependence items.
Sex Differences in Internet Use
A MANOVA indicated a sex difference in reasons for accessing the Internet, Hotelling's T = .39, F(6, 404) = 26.3, p
A MANOVA indicated a sex difference in frequency of accessing different Internet services, Hotelling's T = .14, F(11, 399) = 5.0, p
We did not find sex differences on perceived sociability and social behavior or on the mean amount of time accessing the Internet each day. In addition, there were no sex differences on individual questions regarding symptoms of Internet abuse and dependence or on mean CES-D symptoms. There were no sex differences in meeting criteria (liberal or conservative) for Internet abuse or dependence, except that men (62.8%) were more likely than were women (52.8%) to meet the liberal criteria for Internet abuse, chi^sup 2^(1, N = 411) = 4.1, p
Relations of Internet Use and Psychological Symptoms
We conducted chi-square analyses to examine whether participants meeting the liberal criteria for Internet abuse or dependence reported different reasons for accessing the Internet or used different services than did those participants not meeting the liberal criteria. (Analyses using the conservative criteria were highly similar or could not be calculated because of the small sample sizes.) Participants meeting, versus not meeting, the liberal criteria for Internet abuse were more likely to report accessing the Internet to maintain relationships, meet people, socially experiment, and seek illegal/immoral material, as well as to use the Web, conduct searches, use chat rooms, and play single- and multi- user games, minimum chi^sup 2^(4, N = 411) = 11.2, p
Last, depression also was negatively correlated with the likelihood of using therapy in a face-to-face situation, either individually, r(411) = -.11, p
Our study consisted of a sample of frequent Internet users, with 90% of the participants using the Internet daily. The majority (68%) of the participants reported using the Internet between 30 minutes and 4 hours daily. These figures are similar to-if not somewhat higher than-those obtained in previous research.2,4-6 Men and women did not differ on the mean amount of time accessing the Internet each day; however, as with past research, we found differences between men and women for reasons for accessing the Internet and services used. Past researchers7,8,12 generally have found that men are more likely than are women to use the Internet for purposes related to entertainment and leisure, whereas women use it primarily for interpersonal communication and educational assistance.7,8,10,21- 24 In our study, men were significantly more likely to use the Internet to meet new people, seek sexual material, and seek illegal or immoral material. In addition, men were significantly more likely than were women to use the Internet to (1) surf the Web, (2) participate in newsgroups, chat rooms, and bulletin boards, and (3) play games (both single- and multiuser games).
It is interesting to note, however, that men and women in our study, unlike in past studies,7,8 did not differ on their use of the Internet for educational or academic assistance (eg, library services, course access)-41% of the participants used the Internet daily for academic purposes. These results may have been skewed, however, particularly with regard to male use, because courses on this particular university campus require that students access course materials over the Internet, especially in the course from which we drew this sample. We also found similar rates of Internet use for men and women in shopping and e-mailing. Although several past researchers7,8 have shown differences between men and women in these activities, Joiner et al,12 in addition to us, found no differences between men and women in these activities. These results suggest, therefore, that it is likely that as Internet access has become more commonplace, especially on college campuses, there are certain activities (eg, shopping and e-mailing) that also have become more common. For example, 80% of the participants in our study reported daily use of e-mail.
Most participants (80%) described themselves as sociable, whereas the remaining 20% described themselves as shy but sociable (19%) or as not sociable/loners (1%). Socialization reported by these participants occurred more often face-to-face (74%) or equally face- to-face and on the Internet (23%). A small percentage (3%) stated that they socialized more via the Internet, whereas the remaining 1% seldom socialized. Scherer4 used such socialization patterns to determine whether dependent Internet users fit the stereotype of the socially introverted computer geek. Our results support her contention (and the results found in other studies) in that those participants meeting the liberal criteria for Internet abuse and dependence had higher depression scores and reported less face-to- face interaction.
Prior researchers4,5,15 have suggested that between 8% and 13% of all college students meet the criteria for Internet dependence. About half of the students in our study met the liberal criteria for Internet abuse, and one quarter met the liberal Internet dependence criteria (22% and 1.2% using the conservative criteria, respectively). These numbers differ from prior findings likely because most researchers have defined Internet abuse and dependence using the DSM-IVTR criteria for pathological gambling, whereas we used the criteria for substance abuse and dependence and used rating scales rather than yes/no responses. Anderson5 used substance abuse criteria similar to ours, but with yes/no responses, and found that approximately 9% of students met criteria for dependence. To further investigate the prevalence of Internet abuse and dependence, researchers will need to agree on specific diagnostic criteria and behaviors, as well as how to evaluate the presence, absence, or severity of symptoms. This will help future researchers and clinicians to more fully appreciate the extent to which Internet overuse may interfere with lives and thus constitute a behavioral health problem.
There was no relation between total time online and depressive symptoms; however, depression was correlated with more frequent use of the Internet to meet people, socially experiment, and participate in chat rooms and with less frequent socialization. In addition, individuals meeting the criteria for Internet abuse and dependence endorsed more depressive symptoms and time online and less face-to- face socialization than did those not meeting the criteria. Although it is tempting to suggest that individuals who are depressed may prefer less face-to-face interaction and thus spend more time online, thereby becoming abusive or dependent on the Internet, these data cannot speak to the directionality of the relations. Understanding such pathways will prove an important research direction and provide guidance to clinicians who may be addressing Internet abuse and dependence in their clientele.
The information we have highlighted has several implications for mental health and student affairs professionals. Mental health professionals should be alert to the problems associated with excessive Internet use, including depression, social withdrawal, a failure to fulfill major responsibilities, and behaviors that resemble the patterns seen in tolerance and withdrawal in substance dependence. As the Internet becomes a more integral component of college life, student affairs professionals may need to expend greater effort alerting students and faculty to the potential difficulties that may arise from significant Internet overuse, including personal difficulties and interference with school- related work and assignments. The modern work environment also is largely computer dependent, making this issue relevant to employee assistance programs. In addition, mental health professionals may need to explore the Internet behaviors of clients, particularly those who are depressed or socially introverted. However, providing therapy resources over the Internet appears to be a somewhat acceptable therapeutic modality, although still less preferred than face-to-face therapy.
As the Internet becomes a more popular method of data collection for research purposes, experts should develop standardized survey instruments, particularly with regard to determining the amount of time an individual spends online, reasons for accessing the Internet, and the services used, so that surveys are more comparable and less idiosyncratic. Investigators also should use available psychometrically sound measures of pathology (eg, the CES-D) rather than basing constructs on a small number of untested items (eg, sociability, as defined in this and most prior research). Last, if evaluation of Internet abuse and dependence is to be a viable area of clinical research, then experts need to agree on specific criteria (eg, is Internet abuse more similar to pathological gambling or substance abuse), and use standardized measures (eg, modifying existing substance abuse measures).
We did not find differences in survey results when participants responded via paper or on the Internet20; however, we identified our participants beforehand and assigned them to these conditions. Known respondents on the Internet are likely to produce different results than are unknown respondents who serendipitously come upon the Internet survey site, as seen in several prior studies.25,26 Thus, researchers will need to clearly define their samples and means of survey access. For clinical research focusing on the potential problem of Internet abuse and dependence, known samples specifically invited to participate in the research would be most appropriate. The increasing availability of Internet courseware on college campuses would make this a practical method of participant recruitment. Because most of the researchers have investigated college student Internet abuse and dependency, future researchers should move to the general population and investigate a wider range of factors that could contribute to abuse and dependency and to determine the extent to which these may constitute a new behavioral health problem. ACKNOWLEDGMENT
The authors thank Cara O'Connell, Jennifer Guriel, Serena Gibson, Amisha Dean, and Tara Parsons for the time they devoted to the project.
At the time of the study, all authors were with West Virginia University.
Copyright (c) 2007 Heldref Publications
1. Perry T, Perry LA, Hosack-Curlin K. Internet use by university students: an interdisciplinary study on three campuses. Internet Res Electron Netw Appl Policy. 1998;8:136-141.
2. United States Department of State. College students outpace general US population in Internet use. United States Department of State. Available at: http://usinfo.stae.gov. Accessed November 20, 2002.
3. Attitudes and characteristics of freshmen at 4-year colleges. Chron Higher Educ. 2001:48:26.
4. Scherer K. College life online: healthy and unhealthy Internet use. J Coll Stud Dev. 1997;38:655-665.
5. Anderson KJ. Internet use among college students: an exploratory study. J Am Coll Health. 2001;50:21-26.
6. Rotunda RJ, Kass SJ, Sutton MA, Leon DT. Internet use and misuse: preliminary findings from a new assessment instrument. Behav Modif. 2003;27:484-504.
7. Weiser EB. Gender differences in Internet use patterns and Internet application preferences: a two-sample comparison. Cyberpsychol Behav. 2000;2:167-177.
8. Odell PM, Korgen KO, Schumacher P, Delucchi M. Internet use among female and male college students. Cyberpsychol Behav. 2000;3:855-862.
9. Colley A. Gender differences in adolescent perceptions of the best and worst aspects of computing. Comput Human Behav. 2003;19:673- 682.
10. Teo TSH, Lim VKG. Gender differences in Internet usage and task preferences. Behav Inf Technol. 2000;19:283-295.
11. Ho SMY, Lee MC. Computer usage and its relationship with adolescent life in Hong Kong. J Adolesc Health. 2001;29:258-266.
12. Joiner R, Gavin J, Duffield J, et al. Gender, Internet identification, and Internet anxiety: correlates of Internet use. Cyberpsychol Behav. 2005;8:371-378.
13. Kubey RW, Lavin MJ, Barrows JR. Internet use and collegiate academic performance decrements: early findings. J Commun. 2001;51:366-382.
14. Young KS, Rogers RC. The relationship between depression and Internet addiction. Cyberpsychol Behav. 1998;1:25-28.
15. Morahan-Martin J, Schumacher P. Incidence and correlates of pathological Internet use among college students. Comput Human Behav. 2000;16:13-29.
16. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed., text rev. Washington, DC: American Psychiatric Association; 2000.
17. DeVellis RF. Scale Development: Theory and Applications. Newbury Park, CA: Sage; 1991.
18. Nunnally JC. Psychometric Theory. New York, NY: McGraw-Hill; 1978.
19. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1:385-401.
20. Fortson BL, Scotti JR, Del Ben K, Chen Y. Reliability and validity of an Internet traumatic stress survey with a college student sample. J Trauma Stress. 2006;19:709-720.
21. Jackson LA, Ervin KS, Gardner PD, Schmitt N. Gender and the Internet: women communicating and men searching. Sex Roles. 2001;44:363-379.
22. Miller LM, Schweingruber H, Bradenberg CL. Middle school student's technological practices and preferences: reexamining gender differences. J Educ Multimedia Hypermedia. 2001;10:125-140.
23. Morahan-Martin J. Males, females and the Internet. In: Gackenback J, ed. Psychology and the Internet. San Diego, CA: Academic Press; 1999:169-197.
24. Sherman RC, End C, Kraan E, et al. The Internet gender gap among college students: forgotten but not gone? Cyberpsychol Behav. 2000;3:885-894.
25. Buchanan T, Smith JL. Using the Internet for psychological research: personality testing on the World Wide Web. Br J Psychol. 1999;90:125-144.
26. Joinson A. Social desirability, anonymity, and Internetbased questionnaires. Behav Res Methods Instrum Comput. 1999;31:433-4.
Beverly L. Fortson, PhD; Joseph R. Scotti, PhD; Yi-Chuen Chen, PhD; Judith Malone, BS; Kevin S. Del Ben, PhD
Portions of this manuscript were submitted as the master's thesis of the first author. The West Virginia University Department of Psychology Alumni Fund at West Virginia University provided partial funding for this research.
For comments and further information, please address correspondence to Dr Beverly L. Fortson, University of South Carolina-Aiken, 471 University Parkway, Box 2, Aiken, SC 29801 (e- mail: [email protected]).
Copyright Heldref Publications Sep/Oct 2007
(c) 2007 Journal of American College Health. Provided by ProQuest Information and Learning. All rights Reserved.