ArticlePDF Available

Longitudinal Effects of Internet Uses on Depressive Affect: A Social Resources Approach

Longitudinal Effects of Internet Uses on Depressive Affect:
A Social Resources Approach
Katherine Bessière, Sara Kiesler, Robert Kraut, and Bonka Boneva
Carnegie Mellon University
Acknowledgements: This study was supported by NSF grant #IIS-0208900.
Key words: Depressive affect, longitudinal study, Internet uses, social support, extraversion,
interpersonal interaction, social resources.
Internet use
Using the Internet could augment people’s social resources, displace everyday
communication, compensate for resource gaps, or enhance mood. Using a longitudinal U. S.
survey of changes in participants’ depressive affect, we tested augmentation, displacement,
compensation, and mood enhancement hypotheses. Dominant uses of the Internet—
communicating with family and friends and searching for information—had no impact on
depressive affect. Using the Internet to meet people was associated with increased depressive
affect overall all and especially among those with high initial social resources, but with
reduced depressive affect among those with low initial social resources. Using the Internet
for entertainment also was associated with reduced depressive affect. We suggest that
individual differences in social resources and choice of Internet uses may account for widely
varying reports of Internet social effects.
Internet use
Longitudinal Effects of Internet Uses on Depressive Affect:
A Social Resources Approach
In this article, we examine how the sweeping communication changes embodied in
the Internet’s entry into homes during the last decade may be altering its users’ psychological
well being through changes in their everyday social resources. From the early days of
networked mainframe computers to the present, voluntary social communication has been the
technology’s most frequent use (Sproull & Kiesler, 1991). Over 90% of people who used the
Internet during a typical day in 2004 sent or received email (Pew Internet and American Life
Project, 2004), far more than used any other online application or information source. Using
email leads people to spend more time online and discourages them from dropping Internet
service (Kraut, Mukhopadhyay, Szezypula, Kiesler, Scherlis, 1999). Other Internet
communication services are increasingly popular—instant messaging, chat rooms, multi-user
games, auctions, and myriad political and charitable groups comprising “virtual social
capital” on the Internet (Putnam, 2000, pg. 170). Because most people’s use of the Internet
involves social interaction, there is reason to believe that doing so may affect people’s social
resources and psychological well-being.
Social psychologists have long explored the processes linking communication, social
resources, and psychological well being. People with more social resources—larger social
networks, close relationships, community ties, enacted and perceived social support, and
extraverted individual orientation—are likely to have better psychological functioning, lower
levels of stress, and greater happiness (e.g., Baumeister & Leary, 1995; Cohen & Wills,
1985). By contrast, those with fewer social resources--social isolation, living alone, the
absence of a close relationship, the breakdown or loss of a close relationship, low levels of
Internet use
real and perceived social support, and introversion—are likely to have poor psychological
functioning, to feel lonely, and to experience high levels of depressive affect (e.g., Bruce &
Hoff, 1994; Scheff, 2001). For example, loneliness is inversely correlated with social support
(Riggio, Watring, & Throckmorton, 1993) and positively correlated with depression
(Anderson & Arnoult, 1985). Having poor personal relationships (Burns, Sayers, & Moras,
1994; Finch & Graziano, 2001; Segrin, 1998), low social support (Finch & Graziano, 2001),
and poor social integration (Barnett & Gotlib, 1988) are associated with depressed affect.
Introversion also predicts depressive affect (Barnett & Gotlib, 1988; Finch & Graziano,
2001; Meyers & Diener, 1995). These effects can be self-reinforcing, in that people who are
lonely and depressed may reduce their social resources further by increasing their time alone
and their negative interactions with others (Hawkley et al., 2003; Joiner & Metalsky, 2001)
or by finding partners who are themselves symptomatic (Daley & Hammen, 2002).
Scholars have offered alternative arguments about the effects of Internet use on
people’s social resources, with different implications for changes in Internet users’ well
being. The social augmentation hypothesis is that social communication on the Internet
augments people’s total social resources by providing an added avenue for everyday social
interaction. Katz & Aspden (1997) conducted the first national survey of the public’s use of
the Internet in 1995, when only 8% of their sample were users. They reported that Internet
users had more total contact with family members than non-users did and made more new
friends that year, including those they talked with or met on the Internet. Many Internet users
believe that using the Internet has improved their lives in this way, even providing an
essential link to others (D'Amico, 1998; Hoffman, Novak, & Venkatesh, 2004; Isaacs,
Walendowski, Whittaker, Schiano, & Kamm, 2002; Lenhart, Rainie, & Lewis, 2001).
Internet use
Surveys of Internet users and non-users suggest that Internet users have higher levels of
community and political involvement, more everyday social interaction (Wellman, 2001),
and higher levels of generalized trust and larger social networks (Cole et al., 2000; Robinson,
Kestnbaum, Neustadtl, & Alvarez, 2000a, Uslaner, 2000). In a recent time diary study,
Internet users reported spending three times more time attending social events than non-users
(Neustadtl & Robinson, 2002). Although these findings are consistent with the augmentation
hypothesis, they may be explained by pre-existing differences between those who do and do
not use computers and the Internet. Most of these studies controlled for demographic
differences between users and nonusers, but none controlled for pre-existing differences in
social resources.
The social displacement hypothesis offers a bleaker assessment—that social
communication on the Internet displaces valuable everyday social interaction with family and
friends, with negative implications for users’ psychological well being. The first longitudinal
study of Internet use gave some support to a displacement argument. Kraut et al. (1998)
introduced a sample of people to the Internet for the first time. After a year, those who used
the Internet more were spending less time with family members, had less offline world social
contact, and felt increased loneliness and depressive affect. Weiser’s (2001) longitudinal
study of college students also showed a significant negative effect of spending more hours
using the Internet use for personal purposes on a composite scale measuring loneliness,
depression, and life satisfaction.
Kraut et al. speculated that their sample may have had comparatively superficial
online interactions with acquaintances or strangers that could have displaced time spent with
family and friends. Others’ data also suggests that Internet communications are associated
Internet use
with reduced time spent in social activities (Nie & Hillygus, 2001), less time visiting with
relatives and friends (Gershuny, 2000), reductions in the size of people’s social circle
(Mesch, 2001), fewer phone calls to friends and family (Shklovski, Kraut, & Rainie (in
press), and, for adolescents, poorer relationships with family and friends (Sanders et al.,
2000). There also is some evidence that social interactions online are not psychologically
interchangeable with social interactions offline. Although looking for sex or romantic
partners online is popular (Bargh, McKenna, & Fitzsimons, 2002), national sample studies
show that online relationships rarely lead to offline life relationships (Wolak, Mitchell, &
Finkelhor, 2003) and are more superficial and less likely to last than their offline counterparts
(Parks & Roberts, 1998; Cornwall & Lundgren, 2002). Internet communications are less
likely to support feelings of closeness with others than are face to face or telephone
conversations (Cummings, Butler, & Kraut, 2002; Moody, 2001).
Some studies, however, have found results inconsistent with displacement. In a
longitudinal study of new television and computer purchasers, Kraut et al. (2002) found that
higher Internet use predicted increased local and distant social networks, involvement in
community activities, and sense of trust, reduced loneliness, and no changes in depressive
affect one year later. One difference between the earlier and later Kraut et al. studies is the
gap of 5 years between the two surveys (1995/6 vs. 1998/9). By the time of the second study,
millions more had computers and Internet access at home. In that time also, browsers became
considerably easier to use and the Web was populated with commercial websites, games, and
many other options for information and entertainment. The second study sample also was
composed of more expert users (LaRose, Eastin, & Gregg, 2001). Sample participants may
have spent less time surfing and talking with strangers, and more time getting useful
Internet use
information connected with work, school, or hobbies (Weiser, 2001) or talking online with
friends and family who, by then, had a higher likelihood of Internet access. Such experiences
would be expected to mitigate or reverse displacement effects.
Mixed results from studies examining people’s total hours online have prompted
many investigators to wonder if the ways that Internet users spend their time on the
computer are as important to their well being as the time they spend online (e.g., Bargh &
McKenna, 2004; Caplan, 2003; Kraut et al., 2002; Shaw & Gant, 2002). The Internet, today,
serves a wide range of purposes. People can turn to the Internet for information,
communication, entertainment, or commerce. Online activities that are more utilitarian and
better integrated with people’s school-, work-, or home-life, and that support relationships
with family and friends may augment or stabilize people’s social resources rather than
displace them. For example, email among family and friends could encourage more
socializing with them offline (e.g., making plans for family reunion), increase exchanges of
concrete social support (e.g., asking grandma to babysit; obtaining homework assignments
from a friend), and increase competence and self esteem (e.g., making Web pages for work
colleagues). These online activities could increase closeness and the sense of belonging to
strong ties (Baumeister & Leary, 1985).
Some researchers have argued that whether using the Internet for different purposes
has augmentation or displacement effects may depend on a person’s initial social resources.
Kraut et al. (2002) observed that extraverts were somewhat more likely to use the Internet to
communicate with family and friends than were introverts, and they found some support for
the notion that using the Internet had augmentation effects for the extraverts in their sample.
La Rose, Eastin, and Gregg (2001) found that those with high self-efficacy and those who did
Internet use
not expect to encounter stressful events on the Internet were less likely to suffer ill effects of
being online (see also Wastlund, Norlander, & Archer, 2001).
McKenna & Bargh (1998, 2000) have developed and found some support for a social
compensation hypothesis--that using the Internet to meet new people and to participate in
online groups can have augmentative effects for those with initially impoverished offline
world social resources. New relationships and groups online could help compensate for the
social resources these people lack in the offline world. For instance, those with stigmatized
attributes who lack compatible social groups with whom to identify can find such groups
online (McKenna & Bargh, 1998). By giving such individuals a chance to meet new people
and groups online, the Internet provides these individuals with access to additional social
support and sources of social identification. The authors argue that the Internet gives people
an opportunity to meet people like themselves, and to express themselves openly.
Participants in an experiment said they were better able to express their true selves online
than offline, and they tended to project ideal qualities onto their online partners (Bargh,
McKenna, & Fitzsimmons, 2002; McKenna, Green, & Gleason, 2002).
A fourth hypothesis is implied by studies that suggest people may use the Internet to
regulate affect, particularly, to feel better (Whang, Lee, & Chang, 2003; Weiser, 2001).
Increasingly, the Internet offers pleasurable and mood-altering services and products, ranging
from pornography to multi-player online games and online chat rooms. Research on affect
regulation (for reviews see Gross, 1998; Russell, 2003) indicates that people are motivated to
regulate their emotions—those who are stressed seek escape, those who are bored seek
exciting activities, and those who are depressed seek pleasure. For example, people aggress
when they are feeling angry (Bushman, Baumeister, & Phillips, 2001), watch TV when they
Internet use
are depressed (Kubey & Csikszentmihalyi, 1990), and overeat when they are distressed
(Tice, Bratslavsky, Baumeister, 2001), especially when their bad mood implicates their self
esteem (Heatherton, Striepe, & Wittenberg, 1998). Social and nonsocial leisure activities
improve people’s moods. For example, watching television, going to church, and engaging in
sports and exercise improved positive feelings in a study of people 18 to 82 years old (Hills
& Argyle, 1998). Leisure activities such as watching TV also distract people from ruminating
and reduce self- vs. ideal-self discrepancies (Finn & Gorr, 2001).
One mood enhancement hypothesis as applied to the effects of using the Internet is
that people with high depressive affect will be more likely than others to use the Internet for
nonutilitarian, leisure activity that, in turn, elevates their mood. We propose that such
activities include accessing entertainment such as online games and music and otherwise
killing time or seeking escape online. Using the Internet to be entertained also may alleviate
depressed mood in those with poor social resources because poor resources reduce some
options for seeking pleasure—sex, romance, travel, exciting activities, and companionship.
Thus far, few studies bear on mood regulation motives for using the Internet. Whang,
Lee, & Chang (2003) surveyed a sample of self-described Internet “overusers,” who reported
accessing the Internet particularly when they were stressed by work or depressed. Weiser
(2001) examined college students’ uses of the Internet other than communicating by email.
He found activities such as chatting online and playing games to increase rather than reduce
depressive affect. However, his measure was a mix of Internet uses, and he did not examine
the impact of the students’ initial mood or social resources (i.e., interaction effects).
In sum, the existing evidence suggests that mere hours on the Internet does not have
consistent effects on well being. Rather, different uses of the Internet may have opposite
Internet use
effects through their influence on people’s social resources. Previous researchers, however,
have not compared all uses of the Internet and controlled for initial levels of social resources
and well being. Figure 1 depicts four categories of uses of the Internet that may have
different implications for people’s social resources and well being (Swickert et al., 2002,
Weiser, 2001).
Insert Figure 1 about here
We conducted a longitudinal study using state and trait measures of participants’
initial social resources and disaggregated measures of their use of the Internet to test
alternative hypotheses about the effects on well being of using the Internet use for different
purposes. The longitudinal design allowed us to test augmentation, displacement,
compensation, and mood enhancement hypotheses. Our measure of well being in this study
was depressive affect, a measure predictive of life outcomes and one employed in nearly all
studies of Internet use and well being.
We developed predictions from the augmentation, displacement, compensation, and
mood enhancement frameworks for the effects on depressive affect of four components of
Internet use shown in Figure 1: (1) communicating with family and friends (e.g., email with
relatives), (2) communicating to meet people online, (3) finding and using information (such
as reading news online), and (4) being entertained (e.g., playing online games; killing time).
The augmentation hypothesis suggests that social uses of the Internet that are well integrated
with people’s initial social resources, and especially with their strong ties, will help maintain
or augment people’s total social resources and reduce depressive affect more than other uses
Internet use
of the Internet. The displacement hypothesis suggests that social uses of the Internet,
especially communicating with strangers, can displace people’s offline social resources and
increase their depressive affect more than nonsocial uses. Note that both hypotheses can be
valid because they point to different social uses of the Internet. For example, imagine that a
teenager spending an hour online chooses between hanging out in chat rooms with people
who are not integrated with his daily life or exchanging instant messages with buddies from
school. If he chooses the chat room, he is using the Internet socially but in a manner
disconnected from everyday life could threaten the teen’s valuable social resources. If he
exchanges instant messages with his friends, he is using the Internet to support his everyday
relationships. We hypothesized that using the Internet to communicate with friends and
family would have augmentative effects (that is, would reduce depressive affect) whereas
using it to meet people online would have displacement effects (that is, would increase
depressive affect). Augmentation and displacement hypotheses imply that effects should be
comparatively small when people’s use of the Internet is mainly nonsocial (information and
The social compensation hypothesis predicts an interaction effect depending on
people’s prior social resources. Social uses of the Internet to meet people should reduce
depressive affect more than nonsocial uses for those who are less well socially integrated. If
those who are less well socially integrated lack social groups, friends, and/or support from
family (Bargh & McKenna, 2004), communicating on the Internet to meet people could
increase their social resources and reduce their depressive affect. Meeting people online
could offer a salve to loneliness, the possibility of new contacts and relationships, and a
Internet use
means to improve a currently unpleasant social situation for those with poor offline world
social resources (McKenna & Bargh, 2000).
The mood enhancement hypothesis suggests that those who are feeling depressed,
bored, or anxious may seek out pleasurable experiences on the Internet that have a high
likelihood of increasing their mood by satisfying impulses and needs, distracting them from
everyday life, reducing stress, and increasing their fun. Thus, those with high initial
depressive affect should both seek out and be helped by using the Internet for entertainment.
In summary, this national sample longitudinal study was designed to improve our
understanding of how the Internet is changing people’s well being. We examined whether
different ways of using the Internet predicted changes in depressive affect over time, and
whether these changes were affected by people’s existing social integration and depressive
affect. We statistically controlled participants’ demographic characteristics and their pre-
existing depressive affect to allow for tests of augmentation, displacement, compensation,
and mood enhancement hypotheses.
A national sample of U. S. households was contacted using random digit dialing.
Those answering were asked to list members of the household, and, if so, they were solicited
for a university study. Forty-three percent of those randomly contacted by phone agreed to
participate. This group was sent a cover letter, a consent form, and the survey; 45%
completed the survey, leaving a sample at time 1 of 1,222. Those with Internet access were
oversampled; 74% of the participants at time 1 had Internet access. Six months later, a follow
Internet use
up survey was conducted among those who answered the first survey. Of the 1,222 in the
first survey sample, 82.8% completed the second survey; 72.3% had Internet access.
Data collection was completed via two modes – a paper and pencil questionnaire for
those participants without access to the Internet or who preferred paper, and an online web
survey for those participants with access to the Internet. Participant ages ranged from 13 to
94 (mean = 50.9 years among who completed the paper and pencil survey; 40.1 years among
those who completed the online survey); 85% were adults (19 years or older). Forty-three
percent were men (40% paper; 45% online). Eighty-nine percent were Caucasian (91%
paper; 87% online), and 61% were married (57% paper; 63% online). Thirty percent had a
household income of $30,000 or less; 44% had a household income between $30,000-
$70,000; and 26% had a household income of $70,000 or more. The mean income for the
paper survey participants was between $20,000 and $30,000 and the mean income for the
online survey participants was between $40,000 and $50,000. Thus, Internet users in this
sample were younger and wealthier than non-users, mirroring national trends.
The survey was conducted between June 2000 and March 2002. Participants
completed the questionnaire at Time 1, starting in June 2001 and again six to eight months
later at Time 2, via mail or on the Internet. Sixty percent of the participants completed the
surveys online.
Control variables
Participants were asked to indicate their gender, age, marital status, level of education, race,
and income on the surveys.
Depressive Affect
Internet use
Depressive affect was measured twice using a 12-item version of the CES-D
(Radloff, 1991). This scale is used to measure dysphoria in the general population. High
scores are likely to represent depressive affect rather than clinical depression. Participants
reported how frequently in the past week they had experienced several symptoms of
depressive affect including “I felt that everything I did was an effort,” “My sleep was
restless,” and “I had trouble keeping my mind on what I was doing.” This measure is highly
reliable (Cronbach’s alpha = .89).
Social Resources
We used nine measures of participants’ social resources.
Interpersonal activity. Participants were asked to indicate the frequency of
participation in informal, interpersonal activities with friends and family (Kessler et al.,
1992). Nine items included frequency of spending time with friends, going out to dinner with
others, and eating with all family members. The reliability of this measure assessed by
Cronbach’s alpha was .84.
Group memberships. Participants were asked to indicate the frequency of their
participation in group and community activities (Kessler et al., 1992). Seven items included
playing a team sport, volunteering, and attending religious services. The reliability of this
measure assessed by Cronbach’s alpha was .80.
Social network size. Participants were asked a series of four questions to determine
the size of their social network (Fischer, 1987). These questions asked participants to indicate
the number of friends and number of relatives within an hour’s drive and more than an hour’s
drive away. These four items were summed to estimate social network size.
Internet use
Community involvement. Participants were asked questions to assess their feeling of
belonging to their community. 4 items included knowledge of the neighborhood, feeling part
of the community, and working with others in the local community. The reliability of this
measure assessed by Cronbach’s alpha was .76.
Lives with other adult(s). Information regarding the number of other adults residing
in the residence of the participant was recoded to indicate whether participants lived alone (0)
or lived with one or more other adults (1) (Hughes & Gove, 1981; Wilmoth & Chen).
Employment: Whether the participant had a job (either fulltime or part-time) or no
job was recoded into a binary variable (0=no job; 1=has a job; Dooley, Catalano, Wilson,
Perceived social support. We measured perceived social support (Cohen & Wills,
1985; Kessler et al., 1992) using the ISEL-12 (Cohen & Hoberman, 1983). This self-report
scale measures participants’ perceptions of the availability of various types of social support
such as practical help (“If I had to go out of town for a few weeks, it would be difficult to
find someone who would look after my house or apartment”), advice (“When I need
suggestions on how to deal with a personal problem, I know someone I can turn to”), and
companionship (“If I decide one afternoon that I would like to go to a movie that evening, I
could easily find someone to go with me”). The reliability of this measure assessed by
Cronbach’s alpha was .88.
Extraversion. We measured individual differences in extraversion (Costa & McCrae,
1980) using 8 items from The Big Five Inventory (John, Donahue, & Kentle, 1991).
Participants were asked to agreed or disagree with items such as, “I am talkative,” “I have an
Internet use
assertive personality,” and “I am outgoing or sociable.” The reliability of this measure
assessed by Cronbach’s alpha was .83.
Shyness. Shyness (Anderson & Arnoult, 1985) was measured (only at Time 1) using
a 14-time scale from Cheek and Buss (1981). Participants indicated the extent to which they
agreed with items such as, “I often feel nervous, even in casual get-togethers,” “I would be
nervous if I was being interviewed for a job,” and “In general, I am a shy person.” The
reliability of this measure assessed by Cronbach’s alpha was .87.
These variables should reflect differences in people’s social resources but do not
measure the same concepts, and we did not necessarily expect to see high correlations among
them. People with jobs are not necessarily more extraverted than retired or unemployed
people. Shy people may still feel connected to their local communities. Despite this diversity,
participants who reported more interpersonal contact, more membership in groups, more
connection to their communities, that they lived with another, that they had a job, more
perceived social support, more extraversion, and less shyness, were expected to have access
to more social resources in their daily lives for meeting psychological needs and maintaining
well being.
Internet Uses
A major independent variable for this research was the extent to which participants
used the Internet for different purposes. All measures of this variable were based on
participants’ estimates of the frequency with which they used a computer or the Internet at
home for 27 different purposes in the previous six months, such as “communicating with
friends,” “getting the news online,” or “playing games.” Participants responded using 7-
point, logarithmic-like Likert-scales, with response categories ranging from “several times a
Internet use
day,” “about once a day,” “3-5 days per week,” “1-2 days per week,” “every few weeks,” and
“less often” to “never.” We computed an index of overall Internet use taking the mean of
these 27 items.
In preliminary work, exploratory factor analysis of a similar list of 28 online activities
collected in a sample of 446 participants suggested 5 components of Internet use:
communication with friends and family, communication to meet people, information uses,
commerce, and entertainment. The national survey for this article used a modified set of
items: We added eleven new items, slightly changed the wording of five items, and excluded
nine items that we thought did not reflect typical Internet use at the time of the national
survey. Exploratory factor analysis confirmed the logic of the previous 5 components of
Internet use and suggested a 6th health-related component involving Web searches for health
information and talking in health related support groups.
We conducted confirmatory factor analysis to test whether a multiple-factor model
better explained the data than a single-factor one. The single-factor model represents the
hypothesis that Internet use is best measured by a single index that taps the frequency with
which participants use the Internet, regardless of their type of use. The input data consisted of
the average of a participant’s use of the Internet for each function across the two surveys
(i.e., 922 participants with Internet access by 27 function matrix). We compared the single-
factor model to several multi-factor solutions. The single-factor model, in which all items are
presumed to be caused by a single latent variable, was a very poor fit to the data (Bentler-
Bonett Normed Fit Index=.48; CFI = .49). By contrast, a six-factor model was a significantly
better fit to the data (Bentler-Bonett Normed Fit Index=.80; Comparative Fit Index (CFI) =
.90). It represents the hypothesis that one can distinguish six distinct ways of using the
Internet use
Internet: communicating with friends and family, communicating in online groups and to
meet people, retrieving and using information, seeking entertainment, shopping, and
acquiring health information or talking about health.
We tested our hypotheses using items from four of the factors that, in turn, reflected
the four components of Internet use shown in Figure 1.
Communicating with family and friends. Items included “communicating with
someone in your local area;” “keeping in touch with someone far away,” “communicating
with friends, ” “communicating with relatives” (Chronbach’s alpha = .95).
Communicating to meet people. Items were “meeting new people for social
purposes,” “participating in an online group” (Chronbach’s alpha = .81.
Information. Items were “getting the news online,” “getting information about local
events,” “finding information about national or international events,” “getting information
about movies, books, or other leisure activities,” “getting information for a hobby,” “getting
information for work or school” (Chronbach’s alpha = .95).
Entertainment. Items were “killing time,” ” “releasing tension,” “overcoming
loneliness,” ”being entertained,” “playing games,” ”listening to music” (Chronbach’s alpha =
We omitted the commerce component from analyses because it is more ambiguously
related to our hypotheses; most purchases on the Internet at the time of the national survey
were software, books, pornography, and gifts. We also omitted the health-related component
because it was even more ambiguously related to our hypotheses; items referred to both
information retrieval and communication in online groups. Analyses including these
components in the models did not change our results.
Internet use
Descriptive statistics and correlations among the control variables, social
integration variables, the component Internet use variables, and depressive affect are
described in Tables 1 and 2. The variable indicating usage of the Internet for the
purpose of meeting new people was highly skewed, so we used the log of this
variable in analyses.
Insert Table 1 and Table 2 about here
Data Analysis Strategy
We used hierarchical linear growth models to estimate how participants’ use of the
Internet influenced changes in their depressive affect (Bryk & Raudenbush, 1987, 1992;
Singer & Willett, 2003a). The outcome of interest, depressive affect, was measured at two
time periods, providing two records per participant, collected six months apart. The predictor
variables included stable characteristics of the participant (e.g., gender, race, extraversion),
time of survey, and characteristics of participants that may have varied over time (e.g., the
amount of interpersonal interaction reported by participants). Because some of the predictor
variables measuring social resources and component Internet uses could have changed over
time, to reduce ambiguity about causal direction we used only the scores on these variables at
time 1 for each individual in the equations to predict change in depressive affect between
time 1 and time 2 (Singer & Willertt, 2003b). We also modeled reverse causation to test
whether depressive affect at time 1 caused changes in social resources and component
Internet uses from time 1 to time 2.
OLS regression techniques assume that the errors are independent, normally
distributed, and have constant variance. In contrast, hierarchical linear modeling recognizes
Internet use
that responses from the same participant are not independent of each other. This analysis
separates out the error variance associated with the participant from the error variance
associated with the questionnaire nested within participant. It calculates the correct degrees
of freedom associated with each level of the analysis (participant or questionnaire) and
provides more appropriate estimates of the standard errors than does OLS regression.
The time of survey variable in the model provides an estimate of change in the
dependent variable, depressive affect, over time (i.e., whether participants’ depressive affect
increased or decreased from time 1 to time 2). The statistical interaction of time of survey
with other variables indicates whether these variables moderated changes in depressive affect
from time 1 to time 2. In this research we were especially interested in the interaction of
components of Internet use and time of survey to test augmentation, displacement, and mood
enhancement hypotheses. We expected a significant negative interaction of using the Internet
to communicate with family and friends with time of survey, indicating reduced depressive
affect (augmentation) for those who used the Internet in this way. We expected a significant
positive interaction of using the Internet to meet people with time of survey, indicating
increased depressive affect (displacement) for those who used the Internet in this way. We
also were interested in the three-way interaction effects of social resources measures,
components of Internet use, and time of survey, to test the social compensation hypothesis. In
particular, our prediction was that those who had low levels of social resources at time 1
would experience reduced depressive affect if they used the Internet to meet people online.
The latter prediction implicates initial social resources in changes in depressive affect
scores, raising the possibility that those effects could be explained by regression to the mean.
We therefore conducted analyses that account for regression to the mean, following
Internet use
procedures recommended in Singer (1998). These procedures involve treating time of survey
as a random variable and testing the covariance between the mean & slopes in the random
We conducted preliminary analyses to determine the extent of participants’ different
uses of the Internet and of the association of participants’ characteristics with uses of the
Internet. As shown in the first column of Table 1, communicating with family and friends
and getting information were participants’ dominant uses of the Internet (over 80% of
Internet users used the Internet for these purposes at least every few weeks). Over 60% used
the Internet for entertainment at least every few weeks, and a minority, just over 20%, used
the Internet to meet people at least every few weeks. We conducted a separate regression
analysis, using demographic controls and measures of social resources to predict the four
different components of Internet use. When predicting use of the Internet for a specific
purpose (e.g., entertainment), we included all the other component uses of the Internet (i.e.,
information, communicating with friends and family, communicating to meet people) in the
equations to control for overall propensity to use the Internet (see Table 3).
Insert Table 3 about here
The analyses in Table 3 show that demographic differences in gender, age, marital
status, and income predicted different component uses of the Internet, controlling for other
components. For example, women and older people were more likely to use the Internet to
communicate with family and friends, whereas men and younger people were more likely to
use the Internet for meeting people, information, and entertainment. The results mirror
Internet use
findings in polling studies (Pew, 2004). Better social resources (except for community
involvement) predicted using the Internet to communicate with family and friends. Social
resources did not predict use of the Internet to meet people, with the exception that those who
did so had smaller social networks. Social resources had mixed associations with using the
Internet for entertainment. Those who used the Internet for entertainment scored significantly
higher than others in their level of interpersonal activity and extraversion, but, consistent
with the mood enhancement hypothesis, they were more likely than others to report low
group memberships, low perceived social support, unemployment, and high levels of
Predicting Depressive Affect
We first modeled reverse causation to test whether depressive affect at time 1
predicted social resources and uses of the Internet, with demographic characteristics as
controls. None of these models showed reverse causation effects. That is, depressive affect at
time 1 did not predict any changes in either Internet use or social resources between time 1
and time 2 of the survey. None of the time X depressive affect interactions were significant.
Table 4 consists of three hierarchical linear models predicting depressive affect from
demographic characteristics (gender, age, race, marital status, and income), time of survey,
and overall Internet use and its components. To determine if use of the Internet has different
effects for people differing in initial in social resources, we included interactions with
perceived social support in the last model in Table 4.
Insert Table 4 about here
Internet use
The first model in Table 4 utilizes the index of overall Internet use in an analysis that
includes demographic variables and time of survey. By comparing this model to subsequent
ones in which Internet use is decomposed into its components, we can determine whether one
can detect the effects of an aggregate measure of Internet use or whether one needs to
decompose Internet uses into its components. As showed in Table 4, participants’
demographic characteristics predict their depressive affect scores. In particular, women
reported more depressive affect than men, young people reported more depressive affect than
older people, and poorer people reported more depressive affect than wealthier people.
Because these results are consistent with the prior literature (e.g., on gender, see Mirowsky &
Ross, 1995; on age, see Mirowsky & Ross, 1992; on income, see Conger et al., 1999) and
because they recur in each of the other models, we will not discuss them further in this
The main effect of overall Internet use had a marginally significant positive
relationship with depressive affect. That is, compared to people who did not use the Internet
at time 1 or used it infrequently, people who used the Internet for a wider variety of purposes
reported more depressive affect at time 1 and time 2. The non-significant coefficient for time
of survey shows that on average participants felt no more depressed at time 2 than 6 months
earlier, at time 1. Finally, the non-significant interaction of overall Internet use and time of
survey shows that overall Internet use at time 1 did not predict changes in depressive affect.
This result suggests either that Internet use does not cause changes in depressive affect or
that using a composite measure of Internet use is insufficient for showing the effects of
Internet use on depressive affect.
Internet use
Effects of Components of Internet Use
Our second step in the analysis, reflected in Model 2 in Table 4, was to decompose
Internet use into its four components: communication with friends and family,
communication to meet people, information, and entertainment. An examination of Model 2
shows that there is a differential association of Internet use with depressive affect, depending
upon type of use. The main effect of using the Internet for communication with friends and
family at time 1 shows a weak negative association with depressive affect. That is, people
who used the Internet for communicating with family and friends reported marginally less
depressive affect than others averaged over both administrations of the survey. The
association of using the Internet for information and for meeting new people online, although
negative, is not significant. The main effect of entertainment on depressive affect is very
significantly positive. That is, compared to people who did not use the Internet at time 1 for
entertainment or who used it infrequently for this purpose, those who used the Internet more
frequently for entertainment reported more depressive affect averaged across both surveys.
The interactions of Internet use and time of survey show whether different Internet
uses predicted changes in depressive affect. Although we had expected that use of the
Internet for communication with friends and family would be associated with declines in
depressive affect (augmentation hypothesis), we did not find this to be the case. The non-
significant interaction between time of survey and use of the Internet to communicate with
friends and family indicates that this use was not associated with changes in depressive
In support of the displacement hypothesis, the significant positive interaction between
time of survey and Internet use for meeting people shows that people who used the Internet
Internet use
more for these purposes at time 1 reported increases in depressive affect as compared with
those who did not use the Internet for this purpose. On average, those who used the Internet
for this purpose got worse, as shown in Figure 2.
Insert Figure 2 about here
Use of the Internet for information was not associated either with depressive affect or
with changes in depressive affect, suggesting that this use of the Internet has few social
psychological consequences.
The significant negative interaction between time of survey and Internet use for
entertainment shows that people who used the Internet for entertainment at time 1 reported
declines in depressive affect compared to those who did not use it for this purpose. These
effects of Internet use on changes in depressive affect are directly opposite to the main effect
of entertainment use, indicating the importance of examining the longitudinal results to
determine the direction of effect. Consistent with the mood enhancement hypothesis, those
who used the Internet for entertainment were more depressed initially than those who did not,
but using the Internet in this manner reduced their depressive symptoms. Figure 3 illustrates
this result. The other uses of the Internet we examined did not have this effect. To test for
regression to the mean, we reran the analysis modeling time as a random effect (Singer,
1998). The significant effect remains. Further, the covariance between the mean level of
depressive affect and changes in depressive affect was not significant. This null result is
inconsistent with a regression to the mean explanation.
Insert Figure 3 about here
Internet use
Moderating effects of initial social resources. From the social compensation
hypothesis, we predicted that individuals’ levels of social resources would moderate the
effects of social Internet uses on depressive affect. To test this hypothesis, we added the main
effect of perceived social support and interactions of perceived social support with the time
of survey and use variables from Model 2. Model 3 in Table 4 shows the results.
Perceived social support main effects show that, as expected, perceived support has a
strong negative association with depressive affect. Participants with less perceived social
support at time 1 had significantly higher depressive affect scores averaged over both
surveys. The interaction between perceived social support and using the Internet for meeting
people is significant. The differences in level of depressive affect between those with low
and high perceived social support was accentuated among those who used the Internet to
meet people. In particular, those with low perceived social support who used the Internet in
this way were feeling especially depressed (see Figure 4). We did not predict this cross-
sectional effect.
Insert Figure 4 about here
The three-way interactions of social support, time of survey, and components of
Internet use test the social compensation hypothesis. That is, the displacement hypothesis is
that using the Internet to meet people will lead to increased depressive affect, but the social
compensation hypothesis is that perceived support moderates changes in depressive affect
associated with using the Internet to meet people. Those with fewer social resources should
benefit from using the Internet to meet people. The social compensation hypothesis was
supported (see Figure 5). As previously discussed, those who used the Internet for meeting
Internet use
people felt more depressed on average. However, the positive 3-way interaction with
perceived social support shows that this increase was true only of those who initially reported
more perceived support. By contrast, those who initially reported less perceived support and
who used to the Internet to meet people showed declines in their depressive affect. The
results were not changed when we included effects of regression to the mean in the models
(Singer, 1998).
Insert Figure 5 about here
We conducted similar analyses using other measures of social resources. Table 5
shows these analyses using all 7 other measures of social resources: interpersonal activity,
group membership, community involvement, other adults in household, employed,
extraversion, and shyness. The pattern of results was similar across most measures of social
resources. In particular, across all the measures of social resources, those who were better
integrated reported less depressive affect. This negative cross-sectional association with
depressive affect was significant for all the social resources measures except living with
others and having a job. Extraversion, but not the other measures of social resources, was
associated with increases in depressive affect from time 1 to time 2. For all the measures of
social resources but living with others, lack of social resources was associated with more
depressive affect among people who used the Internet to meet people. Finally, for four of the
measures of social resources—perceived social support, interpersonal activity, group
membership, and extraversion—greater use of the Internet to meet people was associated
with increases in depressive affect among those who initially reported the highest social
support, but was associated with declines in depressive affect among those who initially
Internet use
reported the least social support. These effects were the same when we modeled regression to
the mean (Singer, 1998), and no such effects were found for other uses of the Internet.
Insert Table 5 about here
The Internet consumes time and attention. It also offers connections to others and
convenient, sometimes unique, information and entertainment. We argued that the social
effects of using this technology depend on people’s social resources and their ways of using
the Internet. Our longitudinal analyses of participants’ changes in depressive affect support
this general argument. Participants’ overall use of the Internet did not predict any changes in
participants’ well being, nor did their using the Internet for its most usual purposes—to
communicate with family and friends and to acquire information. By contrast, their using the
Internet for two less common purposes was associated with changes in depressive affect and
may have caused these changes. The first of these purposes was using the Internet to meet
people. Doing so predicted increases in depressive affect, except among those with poor
social resources. Among the latter group, we observed reduced depressive affect, controlling
for regression to the mean. Another influential purpose of using the Internet was for
entertainment. People who used the Internet for this purpose experienced subsequent declines
in depressive affect.
Our study provided tests of four hypotheses related to social resources and the social
impact of the Internet. The social augmentation hypothesis led us to expect those who
communicate with friends and family online to experience increased total social resources
Internet use
and reduced depressive affect but we found no support for this hypothesis. The displacement
hypothesis led us to expect that Internet users who use the Internet to meet people would be
distracted from maintaining their everyday close relationships with friends and family or
perhaps would substitute Internet socializing for more valuable offline activities with friends
and family. The results support this hypothesis. On average, and especially for those with
high levels of social resources, use of the Internet to meet people increased depressive affect.
The social compensation hypothesis (McKenna & Bargh, 1998) led us to expect that
people who used the Internet to meet people online who also had poor offline social
resources would benefit from this use. Our results supported this hypothesis. In our study,
those who had less initial perceived social support, lower interpersonal activity, fewer group
memberships, and who were more introverted experienced declines in depressive affect when
they used the Internet to meet people. This pattern was not true of using the Internet for other
The social compensation results merit further investigation. One might ask what
“meeting people” online really meant: were participants looking for romance, chatting
online, or participating in online communities about a common interest? Because only two
items in our study formed the “meet people” scale (meet new people, communicate in online
groups) we cannot get a good sense of what the scale reflected from this survey. However, in
a subsequent national study of people who have moved their residence, we added Internet use
items and identified a “meet people” factor with more items: “meeting new people for social
purposes,” “participating in online discussion groups,” “communicating with people you first
met online,” “accessing an online community website,” “participating in an online group,”
“posting factual information on a website or group communication system,” and “expressing
Internet use
ideas and opinions on a website or group communication system” (Cronbach’s alpha = .87).
Generally, the behaviors described in these items fit the social compensation assumption that
meeting people online involves spending time online with individuals or groups who were
not initially part of one’s offline world.
It would be reasonable to ask whether those with fewer social resources who used the
Internet to meet people formed new social ties that increased their social resources over time.
We conducted follow-up analyses to investigate if using the Internet in different ways
changed participants’ number or quality of social resources. In these analyses, the
components of Internet use, depressive affect at time 1, and time of survey were treated as
independent variables, and social resources was treated as the dependent variable. This
analysis was repeated for all of the measures of social resources. These analyses showed that
participants with higher depressive affect at time 1 who used the Internet for entertainment
significantly reduced the size of their social network over time (the 3-way interaction of
depressive affect X entertainment X time of survey; beta = -2.2, p = .02) whereas those with
higher depressive affect at time 1 who used the Internet to meet people experienced a
statistically significant increase in the size of their social network (beta = 6.1, p = .04). Using
the Internet to meet people also predicted somewhat less involvement in community (beta = -
.24, p = .004) and organized groups such as church and clubs (beta =-.14, p = .02). No other
measures of social resources were significantly changed. These analyses do not provide
strong evidence that meeting people online improved participants’ social resources.
Social identity processes (McKenna & Bargh, 2000; 2002) provide another
explanation that we did not examine in this study. Using the Internet for entertainment
(especially online games) and to meet people could contribute to new outlets for self
Internet use
expression and identity formation. For example, shy people can be bold, people with a
budding interest in skiing can join an active ski discussion list, and those who would be
warriors can play warrior roles in online games. Meaningful identities enacted through social
interaction and imagination could improve well being. On the other hand, those who already
have a multitude of social identities may be overloading themselves when they spend a lot of
time online, and may become “saturated” (Gergen, 2000). To sort among the alternative
explanations for our data will require more examination of the processes that ensue when
people use the Internet for different purposes.
Our results also provide support for the mood enhancement hypothesis. Those who
used the Internet for entertainment, including games, music, and “killing time” subsequently
experienced significant declines in their depressive affect, controlling for regression to the
mean. The results suggest that using the Internet for this purpose may have contributed to
mood elevation. The observed pattern of results is consistent with other research on leisure,
for example, findings suggesting that people are more likely to engage in bouts of heavy TV
watching when they are in dysphoric states (Kubey & Csikszentmihalyi, 1990), and that
distraction tends to disrupt the cycle of self-rumination and thus lessen the duration of
depressive moods (Lyubomirsky, Caldwell, & Nolen-Hoiksema, 1998; Nolen-Hoeksema,
1991; nolen-Hoeksema & Morrow, 1993). If this logic applies to the use of the Internet, it
would suggest that people who are feeling bad are using online entertainment as a form of
self-medication. When they are in a period of depression, they go online to distract
themselves and escape from these feelings. The longitudinal data suggest that this strategy
for dealing with dysphoric moods works to some degree.
Internet use
One would need more temporally fined-grained data to test this explanation in more
detail. Our research used the CES-D depression scale, which measures depressive affect as a
relatively stable trait. We waited approximately six months between measurements. To test a
self-medication mood-enhancement explanation, one would want measures of relatively
fleeting dysphoric moods and more frequent measurements. The hypothesis is that Internet
users would go online for entertainment when they were in dark moods, and that doing so
would alleviate their mood. Our data suggest that more studies of entertainment uses of the
Internet may be fruitful.
We cannot insure causality based on the longitudinal statistical analyses we used in
this study. Inferring causation depends upon accepting several strong assumptions. However,
we believe these longitudinal analyses provide clearer evidence of causation than do cross-
sectional analyses using the same variables (Singer & Willet, 2003a). Most of the claims,
positive and negative, about the impact of the Internet are based on evidence from cross-
sectional surveys, comparing individuals who have Internet access to those who do not have
it, comparing heavier users of the Internet with lighter users, or comparing earlier adopters
with later users. Most of this work also controlled only for demographic variables that
themselves are indirect causes of depressive affect, social resources, or other outcomes of
interest (e.g., Robinson, Kestenbaum, Neustadtl, & Alvaraz, 2000b). In our analyses, we
controlled for measures of social resources that might be associated with both Internet use
and depressive affect. In addition, when testing for the effect of any particular type of
Internet use, we controlled for other Internet uses, thus controlling for participants’ general
propensity to use the Internet. Even with these precautions, however, cross-sectional analyses
invariably under-control for potentially confounding variables. Because of errors in
Internet use
measurement, they under-control for variables included in the statistical models and
invariably exclude some potentially relevant variables. Longitudinal analyses are less subject
to these biases from uncontrolled third variables. Because the same individuals are measured
multiple times, individuals’ stable characteristics, such as demographic characteristics and
stable personality traits, are automatically controlled when assessing change in an outcome.
As a result, it is primarily variables that change with time that remain as threats to inferring
Inferring causation from cross-sectional data is also based on an assumption that if a
causal relationship exists between two variables, it runs in the direction specified by the
investigator. The assumption of one-way causation in most cross-sectional research on the
social impact of the Internet, however, is generally untenable. The positive cross-sectional
correlation between entertainment-oriented Internet use and depressive affect found in this
study does not necessarily imply that Internet use drives depressive affect. It seems more
plausible that the causation runs in the other direction, in that people who were in a poor
mood used the Internet to elevate their mood. This is the causal path that Kubey and
Csikszentmihalyi (1990) and many others have identified between dysphoric mood and
The longitudinal analyses used in the current research partially solve these problems,
by controlling for both prior Internet use and prior depressive affect. In predicting changes in
depression, we include predictor variables, such as Internet use and social resources, as they
were measured at the initial time period. This specification of the model eliminates the
reverse causal path between depression and subsequent Internet use. Depressive affect at
time 2 cannot cause prior Internet use. In addition, these repeated-measures analyses control
Internet use
for co-variation between the error in participants’ initial level of depressive affect and
changes in affect, in, for example, regression towards the mean. It provides separate
estimates of the cross-sectional co-variation between Internet use and depressive affect and
for the co-variation between Internet use and changes in depressive affect.
Longitudinal analyses, of course, depend upon assumptions that can be challenged.
Hierarchical linear growth models take into account individual differences in Internet use and
depression at the initial time period and the co-variation between these individual differences
and change. However, the use of hierarchical growth models to assess causation rests upon
an assumption that all the relevant variables have been measured. It is still possible that some
unmeasured variable that co-varies with time or changes in Internet use may explain why
people who used the Internet in particular ways had larger changes in depression than those
who did not use the Internet in these ways.
Our study should serve as a starting point for further research on the effects of
individual differences and Internet usage patterns on psychological well-being. We have
shown that the effects of using the Internet depends upon how it is used and that personal
characteristics affect the relationship between Internet use and depressive affect. This is an
important step in the research in this area, and could serve to explain the widely disparate
results in previous research.
Although the discussion to this point has focused on the substantive contribution of
this work, there are methodological contributions as well. This research demonstrates the
importance of conducting longitudinal panel research when examining the impact of new
technology. As we have shown here, conclusions are substantially different depending upon
Internet use
whether one examines the cross-section associations of Internet use and depressive affect or
the longitudinal association of Internet use and changes in depressive affect. Moreover, this
research demonstrates the value of decomposing Internet use into its components. The
Internet is a composite technology with a wide range of uses, sharing some features of
television, the newspaper, and the telephone. When looked at as an aggregate, overall
Internet was not associated with changes in depressive affect, but the different ways people
used the Internet made a difference in their outcomes. Our method at once avoids
technological determinism and includes consideration of baserates. Finally, our study shows
the importance of accounting for individual differences in studies of the social impact of
technology. Our results demonstrate that people’s social resources not only influenced their
well being apart from their use of the Internet but also systematically interacted with their
choices of how to use the Internet and with its effects. In that respect, our study shows how
changes in the technologies people use in everyday life can be integrated with research inm
personality and individual differences.
Internet use
Anderson, C. A., & Arnoult, L. H. (1985). Attributional style and everyday problems in
living: Depression, loneliness, and shyness. Social Cognition, 3, 16-35.
Bargh, J. & McKenna, K.Y.A. (2004). The Internet and social life. Annual Review of
Psychology, 55, 573-590.
Bargh, J., McKenna, K. Y. A., & Fitzsimons, G. M. (2002). Can you see the real me?
Activation and expression of the “true self" on the Internet. Journal of Social Issues,
58, 33-48.
Barnett, P. A., & Gotlib, I. H. (1988). Psychosocial functioning and depression:
Distinguishing among antecedents, concomitants, and consequences. Psychological
Bulletin, 104, 97-126.
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal
attachments as a fundamental human motivation. Psychology Bulletin, 117, 497-529.
Bruce, M. L. & Hoff, R.A. (1994). Social and physical health risk factors for first-onset
major depressive disorder in a community sample. Social Psychiatry & Psychiatric
Epidemiology, 29, 165-171.
Bryk, A. S., & Raudenbush, S. W. (1987). Application of hierarchical linear models to
assessing change. Psychological Bulletin, 101, 147-158.
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data
analysis methods. Thousand Oaks, CA, US: Sage Publications, Inc.
Burns, D.D., Sayers, S.L., & Moras, K (1994). Intimate relationships and depression: Is there
a causal connection? Journal of Consulting and Clinical Psychology, 62, 1033-43.
Internet use
Bushman, B. J., Baumeister, R. F., & Phillips, C. M. (2001). Do people aggress to improve
their mood? Catharsis beliefs, affect regulation opportunity, and aggressive
responding. Journal of Personality and Social Psychology, 81, 17-32.
Caplan, S. E. (2003). Preference for online social interaction: A theory of problematic
Internet use and psychosocial well-being. Communication Research. 30, 625-648.
Cheek, J. M., & Buss, A. H. (1981). Shyness and sociability. Journal of Personality and
Social Psychology, 22, 290-307.
Cohen, S., & Hoberman, H. (1983). Positive events and social supports of buffers of life
change stress. Journal of Applied Social Psychology, 13, 99-125.
Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98, 310-357.
Cole, J., Suman, M., Bel, D.V., Lunn, B., Maguire, P., Hanson, K., et al. (2000). The UCLA
Internet report: Surveying the digital future. Los Angeles: UCLA. Retrieved
October 18, 2004, from the World Wide Web:
Conger, R. D., Jewsbury Conger, K., Matthews, L. S., & Elder, G. H., Jr. (1999). Pathways
of economic influence on adolescent adjustment. American Journal of Community
Psychology, 27, 519-541.
Cornwell, B. & Lundgren, D.C. (2001). Love on the Internet: Involvement and
misrepresentation in romantic relationships in cyberspace vs. realspace. Computers
in Human Behavior, 17, 197-211.
Internet use
Costa, P. T., & McCrae, R. R. (1980). Influence of extraversion and neuroticism on
subjective well-being: Happy and unhappy people. Journal of Personality and Social
Psychology, 38, 668-678.
Cummings, J., Butler, B., & Kraut, R. (2002). The quality of online social relationships.
Communications of the ACM, 45, 103-108.
D'Amico, M. L. (1998). Internet has become a necessity, U.S. poll shows. CNNInteractive.
Retrieved October 18, 2004 from the World Wide Web:
Daley, S. E. & Hammen, C. (2002). Depressive symptoms and close relationships during the
transition to adulthood: Perspectives from dysphoric women, their best friends, and
their romantic partners. Journal of Consulting and Clinical Psychology, 70, 129-141.
Dooley, D., Catalano, R., & Wilson, G. (1994). Depression and unemployment: Panel
findings from the Epidemiologic Catchment Area Study. American Journal of
Community Psychology, 22, 745-765.
Finch, J. F. & Graziano, W. G. (2001). Predicting depression from temperament, personality,
and patterns of social relations. Journal of Personality, 69, 27-55.
Finn, S. & Gorr, M.B. (2001). Social isolation and social support as
correlates of television viewing motivation. Communications Research, 15,
Gergen, K. (2000). The saturated self: Dilemmas of identity in contemporary life. New York:
Basic Books.
Gershuny, J. (2000). Changing times: Work and leisure in postindustrial society. New York:
Oxford University Press.
Internet use
Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review
of General Psychology, 2, 271-299.
Hawkley, L. C., Burleson, M. H., Bernston, G. G., Cacioppo, J. T. (2003). Loneliness in
everyday life: Cardiovascular activity, psychosocial context, and health behaviors.
Journal of Personality and Social Psychology, 85, 105-120.
Heatherton, T.F., Striepe, M., & Wittenberg, L. (1998). Emotional distress and disinhibited
eating: The role of self. Personality and Social Psychology Bulletin, 24, 301-313.
Hills, P., & Argyle, M. (1998). Positive moods derived from leisure and their relationships to
happiness and personality. Personality and Individual Differences, 25, 523-535.
Hoffman, D. L., Novak, T. P., & Venkatesh, A. (2004). Has the Internet become
indispensable? Communications of the ACM, 47, 37 - 42.
Hughes, M., & Gove, W. R. (1981). Living alone, social integration, and mental health.
American Journal of Sociology, 87, 48-74.
Isaacs, E., Walendowski, A., Whittaker, S., Schiano, D., & Kamm, C. (2002). The character,
functions, and styles of instant messaging in the workplace. In Proceedings of the
ACM Conference on Computer Supported Cooperative Work. New York: ACM
Press, pp. 11-20.
John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The big five inventory: Versions 4a and
54. Berkeley, CA: The Institute of Personality and Social Research, University of
Joiner, T. E. Jr., and Metalsky, G. I. (2001). Excessive reassurance seeking: Delineating a
risk factor involved in the development of depressive symptoms. Psychological
Science, 12, 371-378.
Internet use
Katz, J. E., & Aspden, P. (1997). A nation of strangers? Communications of the ACM, 40,
Kessler, R.C., Kendler, K. S., Heath, A., Neale, M. C., & Eaves, L. J. (1992). Social support,
depressed mood, and adjustment to stress: A genetic epidemiologic investigation.
Journal of Personality and Social Psychology, 62, 257-272.
Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., & Crawford, A. (2002a).
Internet paradox revisited. Journal of Social Issues, 58, 49-74.
Kraut, R., Lundmark, V., Patterson, M., Kiesler, S., Mukopadhyay, T., & Scherlis, W.
(1998). Internet paradox: A social technology that reduces social involvement and
psychological well-being? American Psychologist, 53, 1017–1031.
Kraut, R., Mukhopadhyay, T., Szczypula, J., Kiesler, S., & Scherlis, W. (1999).
Communication and information: Alternative uses of the internet in households.
Information Systems Research, 10, 287-303.
Kubey, R. W., & Csikszentmihalyi, M. (1990). Television as escape: Subjective experience
before an evening of heavy viewing. Communication Reports, 3, 92-100.
LaRose, R., Eastin, M. S., Gregg, J. (2001). Reformulating the Internet paradox: Social
cognitive explanations of Internet use and depression. Journal of Online Behavior, 1.
Retrieved October 18, 2004 from the World Wide Web:
Lenhart, A., Rainie, L., & Lewis, O. (2001). Teenage life online: the rise of the instant-
message generation and the Internet's impact on friendships and family relationships.
Washington, D. C.: Pew Internet and American Life Project. Retrieved October 18,
Internet use
2004 from the World Wide Web:
Lyubomirsky, S., Caldwell, N. D., & Nolen-Hoeksema, S. (1998). Effects of ruminative and
distracting responses to depressed mood on retrieval of autobiographical memories.
Journal of Personality & Social Psychology, 75, 166-177.
McKenna, K. Y. A., & Bargh, J. (1998). Coming out in the age of the Internet: Identity
"demarginalization" through virtual group participation. Journal of Personality and
Social Psychology, 75, 681-694.
McKenna, K. Y. A., & Bargh, J. (2000). Plan 9 from cyberspace: The implications of the
Internet for personality and social psychology. Personality and Social Psychology
Review, 4, 57-75.
McKenna, K. Y. A., Green, A. S., & Gleason, M. E. J. (2002). Relationship formation on the
Internet: What's the big attraction? Journal of Social Issues, 58, 659-671.
Mesch, G. S. (2001). Social relationships and Internet use among adolescents in Israel. Social
Science Quarterly, 82, 329-340.
Meyers, D. & Diener, E. (1995). Who is happy?. Psychological Science, 6, 10-19.
Mirowsky, J. & Ross, C. E. (1992). Age and depression. Journal of Health & Social
Behavior, 33, 187-205.
Mirowsky, J. & Ross, C. E. (1995). Sex differences in distress: Real or artifact? American
Sociological Review, 60, 449-468.
Moody, E. J. (2001). Internet use and its relationship to loneliness. CyberPsychology &
Behavior, 4, 393-401.
Internet use
Neustadtl, A., & Robinson, J. P. (2002). Social contact differences between Internet users
and nonusers in the General Social Survey. IT & Society, 1, 73-102.
Nie, N.H., & Hillygus, S. (2001). Education and democratic citizenship. In Making good
citizens: Education and civil society, Diane Ravitch and Joseph Viteritti (eds.). New
Haven: Yale Univ. Press.
Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of
depressive episodes. Journal of Abnormal Psychology, 100, 569-582.
Nolen-Hoeksema, S., & Morrow, J. (1993). Effects of rumination and distraction on naturally
occurring depressed mood. Cognition & Emotion, 7, 561-570.
Parks, M., & Roberts, L. (1998). Making MOOsic: The development of personal
relationships on line and a comparison to their off-line counterparts. Journal of Social
and Personal Relationships, 15, 517-537.
Pew Internet and American Life Project. (2004). The Internet and daily life. Retrieved
November 26, 2004, from the World Wide Web:
Putnam, R. (2000). Bowling alone: The collapse and revival of American community. New
York: Simon & Schuster.
Radloff, L. S. (1991). The use of the center for epidemiologic studies depression scale in
adolescents and young adults. Journal of Youth and Adolescence, 20, 149-166.
Riggio, R. E., Watring, K. P., & Throckmorton, B. (1993). Social skills, social support, and
psychosocial adjustment. Personality and Individual Differences, 15, 275–280.
Robinson, J. P., Kestnbaum, M., Neustadtl, A., & Alvarez, A. (2000a). Information
technology, the Internet and time displacement. Paper presented at the Annual
Meetings of Am. Assoc. Public Opinion Research, Portland, OR.
Internet use
Robinson, J. P., Kestnbaum, M., Neustadtl, A., & Alvarez, A. (2000b). Mass media use and
social life among Internet users. Social Science Computer Review, 18, 490-501.
Russell, J. A. (2003). Core affect and the psychological construction of emotion.
Psychological Review, 110, 145-172.
Sanders, C. E., Field, T. M., Diego, M., & Kaplan, M. (2000). The relationship of internet
use to depression and social isolation among adolescents. Adolescence, 35, 237-242.
Scheff, T. J. (2001). Shame and community: social components in depression. Psychiatry,
64, 212-224.
Shklovski, I., Kraut, R., & Rainie, L. (in press). The Internet and social participation:
Contrasting cross-sectional and longitudinal analyses. Journal of Computer Mediated
Communication, 10.
Segrin, C. (1998). Interpersonal communication problems associated with depression and
loneliness. In P. A. Andersen & L. K. Guerrero (Eds.), Handbook of communication
and emotion: Research, theory, applications, and contexts (pp. 215-242). San Diego,
CA: Academic Press.
Shaw, L.H. & Gant, L.M. (2002). In defense of the Internet: The relationship between
Internet communication and depression, loneliness, self-esteem, and perceived social
support. CyberPsychology & Behavior, 5, 157-171.
Singer, J.D. (1998). Using SAS PROC MIXED to fit multilevel models, hierarchical models,
and individual growth models. Journal of Educational and Behavioral Statistics, 23,
Singer, J. D., & Willett, J. B. (2003a). Applied longitudinal data analysis. New York: Oxford
University Press.
Internet use
Singer, J. D., & Willett, J. B. (2003b). Applied longitudinal data analysis: Modeling change
and event occurrence. New York: Oxford University Press.
Sproull, L. & Kiesler, S. (1991). Connections: New ways of working in the networked
organization. Cambridge: The MIT Press.
Swickert, R. J., Hittner, J. B., Harris, J. L., & Herring, J. A. (2002). Relationships among
Internet use, personality, and social support. Computers in Human Behavior, 18, 437-
Tice, D. M., Bratslavsky, E., & Baumeister, R. F. (2001). Emotional distress regulation takes
precedence over impulse control: If you feel bad, do it! Journal of Personality and
Social Psychology, 80, 53-67.
Uslaner, E. (2000). Internet and social capital. Proceedings of the ACM, 43, 60 - 64.
Wastlund, E., Norlander, T., & Archer, T. (2001). Internet blues revisited: Replication and
extension of an Internet paradox study. CyberPsychology & Behavior, 4, 385-391.
Weiser, E. B. (2001). The functions of Internet use and their social and psychological
consequences. CyberPsychology & Behavior, 4, 723-743.
Wellman, B. (2001). Physical place and CyberPlace: The rise of personalized networking.
International Journal of Urban and Regional Research, 25, 227-252.
Whang, L. S.; Lee, S.; Chang, G. (2003). Internet over-users' psychological profiles: A
behavior sampling analysis on internet addiction. CyberPsychology & Behavior, 6,
Wilmoth, J. & Chen, P.C. (2003). Immigrant status, living arrangements, and depressive
symptoms among middle-aged and older adults.” Journal of Gerontology: Social
Sciences, 53: S303-S315.
Internet use
Wolak, J., Mitchell, K., & Finkelhor, D. (2003). Escaping or connecting? Characteristics of
youth who form close online relationships. Journal of Adolescence, 26, 105-119.
Table 1. Descriptive statistics for variables in study.
Time 1
Time 2
At least
Control Variables
Depressive Affect
Social Resources
Interpersonal Activity
Group Memberships
Social Network Size
Community Involvement
Lives with Other Adult(s)
Perceived Social Support
Internet Components
Internet: Family & Friends
Internet: Meet People
Internet: Information
Internet: Entertainment
*Percent respondents who used the Internet who said they used the Internet for this purpose every few
weeks or more often.
Table 2. Correlations among variables.
Note._ Correlations are computed on variables after averaging across survey administrations.
No Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Control Variables
1 Male 1.00
2 Age 0.00 1.00
3 White 0.08 0.10 1.00
4 Married 0.07 0.24 0.15 1.00
5 Income 0.07 0.05 0.13 0.39 1.00
6 Depressive Affect -0.08 -0.15 -0.03 -0.13 -0.18 1.00
Social Resources
7 Interpersonal Activity -0.06 -0.32 -0.08 -0.21 -0.06 -0.07 1.00
8 Group Memberships 0.00 -0.10 -0.02 -0.09 0.04 -0.17 0.44 1.00
9 Social Network Size 0.01 -0.06 0.01 -0.08 -0.12 -0.12 0.29 0.30 1.00
10 Community Involvment -0.04 0.25 0.07 0.14 0.14 -0.36 0.21 0.44 0.20 1.00
11 Lives with Other Adult(s) 0.06 -0.15 0.10 0.46 0.20 -0.03 0.00 0.01 0.08 0.06 1.00
12 Employment 0.13 -0.16 0.03 0.15 0.21 -0.07 -0.08 -0.09 -0.05 -0.02 0.01 1.00
13 Perceived Social Support -0.06 -0.13 0.01 0.08 0.14 -0.34 0.37 0.20 0.23 0.33 0.12 0.11 1.00
14 Extraversion -0.05 -0.12 -0.02 0.00 0.08 -0.16 0.34 0.28 0.18 0.23 0.07 0.03 0.31 1.00
15 Shyness -0.06 -0.10 0.01 -0.08 -0.12 0.29 -0.18 -0.18 -0.12 -0.27 -0.02 -0.07 -0.30 -0.63 1.00
Internet Components
Internet: Friends & Family -0.03 -0.27 -0.27 -0.09 0.22 0.00 0.28 0.21 0.14 0.00 0.01 0.06 0.14 0.15 -0.07 1.00
Internet: Meet People 0.04 -0.27 -0.06 -0.16 -0.04 0.14 0.22 0.11 0.01 -0.11 -0.03 -0.03 -0.02 0.08 0.03 0.42 1.00
18 Internet: Information 0.16 -0.30 0.00 -0.02 0.23 0.04 0.20 0.15 0.04 -0.05 0.00 0.16 0.08 0.06 -0.01 0.64 0.34 1.00
19 Internet: Entertainment 0.09 -0.38 -0.03 -0.14 0.05 0.20 0.25 0.09 0.03 -0.12 0.01 0.00 0.00 0.05 0.08 0.59 0.48 0.62
Internet use
Table 3. Predicting different uses of the Internet from participant demographics and social
Variable Beta SE t p Beta SE t p Beta SE t p Beta SE tp
Intercept 0.672 0.127 5.27 *** 0.697 0.085 8.21 *** 0.772 0.108 7.16 *** 0.244 0.120 2.03 *
Control Variables
Male -0.396 0.057 -6.91 *** 0.083 0.039 2.12 * 0.337 0.049 6.91 *** 0.147 0.054 2.69 **
Age 0.004 0.002 2.05 * -0.004 0.001 -2.63 ** -0.009 0.002 -4.89 *** -0.011 0.002 -5.86 ***
White 0.174 0.093 1.88 -0.076 0.062 -1.22 -0.086 0.079 -1.09 0.024 0.087 0.28
Married -0.263 0.078 -3.36 *** -0.054 0.053 -1.01 0.198 0.067 2.96 ** -0.052 0.074 -0.70
Income 0.092 0.013 6.98 *** -0.028 0.009 -3.15 ** 0.061 0.011 5.41 *** -0.026 0.013 -2.11 *
Social Resources
Interpersonal Activity 0.129 0.048 2.71 ** 0.031 0.032 0.98 -0.070 0.040 -1.72 0.135 0.045 3.04 **
Group Memberships 0.097 0.044 2.21 * 0.002 0.029 0.06 0.147 0.037 3.96 *** -0.103 0.041 -2.52 **
Social Network Size 0.004 0.002 2.20 * -0.003 0.001 -2.58 ** -0.003 0.001 -1.80 0.002 0.002 1.11
Community Involvement -0.113 0.038 -2.99 ** 0.020 0.025 0.80 -0.040 0.032 -1.23 -0.007 0.035 -0.21
Lives with Other Adult(s) 0.097 0.087 1.12 -0.023 0.058 -0.39 -0.293 0.073 -3.99 *** 0.101 0.081 1.25
Employment -0.046 0.062 -0.74 -0.020 0.042 -0.47 0.184 0.052 3.51 *** -0.161 0.058 -2.78 **
Perceived Social Support 0.174 0.047 3.68 *** -0.054 0.032 -1.68 0.020 0.040 0.49 -0.234 0.044 -5.30 ***
Extraversion -0.018 0.049 -0.37 0.049 0.033 1.47 -0.057 0.042 -1.35 0.090 0.046 1.95 *
Shyness -0.143 0.052 -2.76 ** 0.026 0.035 0.75 -0.029 0.044 -0.66 0.165 0.049 3.38 ***
Internet Components
Internet: Family &
0.126 0.016 7.79 *** 0.353 0.019 18.71 *** 0.246 0.022 11.15 ***
Internet: Meet People 0.277 0.036 7.79 *** -0.025 0.031 -0.82 0.330 0.033 9.99 ***
Internet: Information 0.488 0.026 18.71 *** -0.016 0.019 -0.82 0.391 0.025 15.52 ***
Internet: Entertainment 0.280 0.025 11.15 *** 0.170 0.017 9.99 *** 0.321 0.021 15.52 ***
Internet: Information
Internet: Entertainment
Internet: Family & Friends
Internet: Meet People
Internet use
Table 4. Hierarchical linear models predicting depressive affect from perceived social support,
components of Internet use, time of survey, and their interactions.
Variable Beta SE t p Beta SE t p Beta SE t p
Intercept 1.67 0.06 29.2 *** 1.68 0.06 29.83 *** 1.68 0.05 31.07 ***
Male (0=female; 1=male) -0.07 0.03 -2.26 * -0.10 0.03 -3.2 ** -0.11 0.03 -3.75 ***
Age 0.00 0.00 -3.85 *** 0.00 0.00 -2.92 ** 0.00 0.00 -4.48 ***
White (0=minority; 1=white) 0.06 0.05 1.26 0.08 0.05 1.52 0.06 0.05 1.37
Married (0=not married; 1=currently
-0.02 0.03 -0.43 -0.02 0.03 -0.62
0.01 0.03 0.43
Income -0.04 0.01 -5.79 *** -0.03 0.01 -4.12 *** -0.03 0.01 -3.94 ***
Perceived Social Support -0.34 0.04 -8 ***
Year (0=2000; 1=2001) 0.02 0.02 1.39 0.03 0.02 1.45 0.02 0.02 1.27
Internet: Overall Use 0.05 0.03 1.69 t
Year X Overall Internet Use 0.00 0.02 0.27
Internet: Friends & Family -0.04 0.03 -1.77 t 0.00 0.03 -0.04
Internet: Meet People -0.06 0.09 -0.72 -0.09 0.09 -1.04
Internet: Information -0.03 0.03 -1.1 -0.03 0.03 -1.06
Internet: Entertainment 0.13 0.03 4.96 *** 0.10 0.03 3.65 ***
Year X Internet: Friends & Family -0.01 0.01 -0.76 -0.02 0.02 -1.31
Year X Internet: Meet People 0.14 0.05 2.55 ** 0.14 0.05 2.6 **
Year X Internet: Information 0.03 0.02 1.49 0.03 0.02 1.52
Year X Internet: Entertainment -0.04 0.02 -2.38 * -0.03 0.02 -2.06 *
Perceived Support X Internet: Friends &
-0.01 0.04 -0.39
Perceived Support X Internet: Meet
-0.25 0.13 -1.92 *
Perceived Support X Internet:
-0.02 0.04 -0.49
Perceived Support X Internet:
0.05 0.04 1.32
Perceived Support X Year 0.07 0.03 2.56 **
Perceived Support X Year X Internet:
Friends & Family
0.01 0.02 0.53
Perceived Support X Year X Internet:
Meet People
0.19 0.08 2.38 *
Perceived Support X Year X Internet:
-0.02 0.03 -0.8
Perceived Support X Year X Internet:
-0.03 0.02 -1.19
Overall Internet Use
Components of Internet Use
Adding Perceived Social Support
Note: t p< .1, * p<.05, ** p<.01, *** p<.001
Internet use
Table 5. Hierarchical linear models predicting depressive affect from components of Internet use, measures of social resources, time
of survey, and their interactions.
Independent Variables Beta t p Beta t p Beta t p Beta t p Beta t p Beta t p Beta t p Beta t p
Intercept 1.72 30.64 *** 1.70 30.45 *** 1.65 29.2 *** 1.67 30.56 *** 1.70 22.02 *** 1.72 25.24 *** 1.70 30.53 *** 1.69 30.96 ***
Male (0=female; 1=male) -.12 -3.85 *** -.10 -3.36 *** -0.10 -3.24 *** -.11 -3.80 *** -.10 -3.15 ** -.10 -3.20 ** -.11 -3.62 *** -0.08 -2.72 **
Age .00 -4.26 *** .00 -3.23 ** 0.00 -3.12 ** .00 -1.87 t .00 -2.71 ** .00 -3.21 ** .00 -3.62 *** 0.00 -2.02 *
White (0=minority;
.06 1.17 .06 1.29 0.09 1.88 t .08 1.65 .08 1.59 .08 1.59 .06 1.17 0.05 0.95
Married (0=not married;
-.05 -1.47 -.04 -1.17 -0.01 -0.39 -.01 -.35 -.03 -.66 -.02 -.60 -.02 -.50 -0.02 -0.48
Income -.03 -4.27 *** -.03 -4.09 *** -0.03 -4.23 *** -.02 -3.38 *** -.03 -4.19 *** -.03 -3.58 *** -.03 -3.94 *** -0.03 -3.68 ***
Social Resourcesa-.16 -3.93 *** -.12 -3.22 ** -0.01 -3.17 ** -.18 -5.61 *** -.03 -.38 -.07 -1.14 -.24 -6.52 *** 0.29 7.27 ***
Year (0=2000; 1= 2001) .02 1.23 .03 1.42 0.03 1.73 t .03 1.80 t .00 .12 .01 .40 .02 1.36 0.03 1.52
Internet: Friends & Family -.03 -1.31 -.04 -1.42 -0.05 -1.92 * -.03 -1.25 -.08 -1.39 -.01 -.18 -.02 -.88 -0.01 -0.52
Internet: Meet People .02 .25 -.04 -.47 -0.03 -0.35 -.09 -.97 .40 1.87 t -.12 -.85 -.01 -.09 -0.06 -0.67
Internet: Information -.03 -1.02 -.02 -.83 -0.03 -0.93 -.04 -1.21 .01 .19 -.05 -.97 -.04 -1.31 -0.04 -1.41
Internet: Entertainment .13 4.87 *** .13 4.67 *** 0.13 4.75 *** .12 4.34 *** .06 .87 .05 1.20 .12 4.54 *** 0.11 4.05 ***
Integration X Internet:
Friends & Family
.02 .76 .04 1.27 0.00 1.42 .03 1.10 .05 .71 -.06 -1.11 .03 .81 -0.05 -1.59
Integration X Internet:
Meet People
-.28 -2.67 ** -.21 -2.18 t -0.02 -3.69 *** -.10 -1.06 -.57 -2.44 ** .07 .40 -.24 -2.15 t 0.23 1.92 *
Integration X Internet:
-.04 -1.18 -.02 -.55 -0.01 -2.78 ** -.01 -.31 -.06 -.79 .04 .61 .01 .17 -0.02 -0.51
Integration X Internet:
.07 1.95 t .02 .77 0.01 3.43 *** -.04 -1.35 .08 1.14 .13 2.31 t .01 .37 0.04 1.02
Year X Internet: Friends &
-.01 -.53 -.01 -.58 -0.01 -0.34 -.01 -.93 -.01 -.43 -.06 -2.40 ** -.02 -1.34 -0.02 -1.33
Year X Internet: Meet
.09 1.50 .12 2.18 t 0.11 2.15 * .15 2.71 ** -.04 -.34 .22 2.59 ** .09 1.68 t 0.12 2.34 *
Year X Internet:
.02 1.27 .02 1.28 0.02 1.17 .03 1.53 .01 .24 .06 1.86 t .03 1.66 t 0.03 1.66 t
Year X Internet:
-.04 -2.32 t -.04 -2.34 ** -0.04 -2.22 * -.04 -2.21 * .00 .04 .00 -.01 -.03 -2.02 t -0.03 -1.86 t
Year X Integration .01 .46 -.01 -.35 0.00 1.07 .01 .70 .03 .67 .02 .59 .08 3.60 *** -0.07 -2.83 **
Integration X Year X
Internet: Friends & Family
-.02 -1.01 -.02 -1.05 0.00 -1.11 -.01 -.77 .00 .11 .07 2.33 ** -.01 -.30 0.00 0.18
Integration X Year X
Internet: Meet People
.17 2.55 ** .17 2.97 ** 0.01 3.76 *** .07 1.25 .22 1.56 -.12 -1.10 .22 3.19 ** -0.19 -2.5 **
Integration X Year X
Internet: Information
.03 1.37 .01 .56 0.00 2.47 ** .01 .33 .02 .53 -.06 -1.49 .00 -.17 0.02 0.73
Integration X Year X
Internet: Entertainment
-.03 -1.49 -.02 -.99 0.00 -3.03 ** .02 .89 -.05 -1.08 -.06 -1.86 t -.02 -1.14 0.00 -0.06
Lives with Others
Has Job
Measure of Social Resources
Interpersonal Activity
Group Membership
Community Involvement
Social Network Size
Internet use
Figure 1. Predictions of changes in depressive affect from hypotheses in the literature on Internet
social effects.
Most Frequent Uses of the Internet
Less Frequent Uses of the Internet
Social Uses
of the
with family and friends
==> Augmentation hypothesis
predicts reduced depressive affect
Communication to meet people
==> Displacement hypothesis
predicts increased depressive affect
==> Compensation hypothesis
predicts reduced depressive affect
for those with fewer social resources
Social Uses
of the
==> Mood enhancement hypothesis
predicts reduced depressive affect
Internet use
Figure 2. Changes in depressive affect predicted by use of the Internet to meet people.
Note._ Interaction graph calculated from model tests (Tables 4, 5).
Internet use
Figure 3. Changes in depressive affect predicted by use of the Internet for entertainment
Time of Survey
Depressive Affect
High Entertainment Use
Low Entertainment Use
Note._ Interaction graph calculated from model tests (Tables 4, 5).
Internet use
Figure 4. The cross-sectional relationship between depressive affect, level of perceived social
support, and using the Internet to meet people.
Low High
Perceived Social Support
Depressive Affect
Low Meet People
High Meet People
Internet use
Figure 5. Changes in depressive affect predicted by level of perceived social support and use of
the Internet to meet people.
Note._ Interaction graphs derived from the model tests in Table 4. See text for discussion of regression to the mean.
... This difference is key, as these shared activities contribute to the formation of long-lasting friendship bonds with sustainable levels of self-disclosure and intimacy not traditionally found in other mediated spaces (Cole & Griffiths, 2007; Hsu, Wen, & Wu, 2009). The formation of such bonds could contribute to a preference for online interaction that is potentially greater than other mediated outlets, and, over time, lead to a variety of negative consequences for the user, such as declines in the size and quality of one's offline social circles (Bessiere, Kiesler, Kraut, & Boneva, 2012; Kraut et al., 1998; Shen & Williams, 2010; Williams, 2006) and increased loneliness (Lemmens, Valkenburg, & Peter, 2011; Morahan-Martin & Schumacher, 2003). Online games have also likely received particular attention due to the increased amounts of time that are being spent in these spaces. ...
... As the online gaming industry continues to flourish, the concern over the possible social impact of prolonged interactions within online gaming environments also continues to rise, particularly in relation to its potential long-term impact on a user's social ability , or social skills. Because increased OVG involvement has been shown to negatively impact one's level of offline social engagement (Bessiere et al., 2012; Hussain & Griffiths, 2009; Kim, Namkoong, Ku, & Kim, 2008; Kolo & Baur, 2004; Lo, Wang, & Fang, 2005; Shen & Williams, 2010; Smyth, 2007; Williams, 2006), and having and maintaining face-to-face relationships is integral to developing effective social skills and learning socially appropriate behavior (Engles, Finkenauer, Meeus, & Dekovic, 2001), becoming socially disengaged or isolated from one's offline contacts due to OVG play could substantially hinder the development, or stimulate the deterioration , of effective ''offline'' social skills, for instance the ability to verbally engage others or manage one's social self-presentation in real-time (Cole & Griffiths, 2007; Hussain & Griffiths, 2009; Shen & Williams, 2010). As outlined by Kim et al. (2008), ''the [use of] online games is associated with a decline in participants' communication . . . ...
... Another mechanism through which online gaming may affect gaming disorder is suggested by the displacement hypothesis (35). Because of the "inelasticity of time" (36), playing online games takes away time from face-to-face interactions with one's offline ties (37), which can lead to the displacement of offline social contacts for online ties (38). Therefore, gamers who are absorbed with in-game social interaction may have an overall smaller and weaker offline social circle as a result of excessive online gaming (39). ...
Full-text available
Background and Aims: Social interaction in the online games has been found to predict gaming disorder, but little research has examined the mechanism of this association. Drawing on the social capital theory, the present study investigated the mediating role of online social capital on the relationship between in-game social interaction and gaming disorder and the moderating role of alienation on the relationship between online social capital and gaming disorder. Methods: A sample of 457 Chinese massively multiplayer online role-playing game gamers was recruited to complete the In-game Social Interaction Questionnaire, Online Social Capital Scale, Alienation Scale, and Pathological Gaming Scale. Results: The results showed that online social capital was a mediator in the relationship between in-game social interaction and gaming disorder. Moreover, for individuals with low alienation, the effect of online social capital on gaming disorder was weaker than for those with high alienation. Conclusions: The present study provides new insight into the complex processes involved in the effect of in-game social interaction on gaming disorder, and the results have important theoretical and practical implications.
... Based on Young (44), Kubey and Csikszentmihalyi (47) individuals used the Internet for dealing with negative mood and real-life difficulties. Bessiere et al. (48) emphasized that individuals who have negative and inappropriate feelings used online entertainment as therapist. Caplan and High (49) believed that individuals share online massages for compensating lack of real life. ...
Full-text available
Background Self-control on Internet usage would be lost in an individual due to exposure to Internet addiction and the individual would experience negative outputs in life. The purpose of the present study is to investigate the relationship between anxiety, physical, and mental symptoms amongst students exposed to and without Internet addiction. Methods A cross-sectional descriptive study, utilizing self-report questionnaires, was conducted 2016. The statistical community consisted of 17,000 students. The sample consists of 304 students who were selected by applying the cluster sampling method. In this research, two standard instruments Young’s Internet addiction test and Beck’s anxiety questionnaire were used, where the reliability of them were accepted by 10 specialists and the validity of them were calculated by Cronbach Alpha as 0.71 and 0.92. At first, the addiction test was done in order to identify the students exposed to and without Internet addiction. After getting the scores and identifying 2 groups, an anxiety questionnaire was completed by samples. Data analysis was done by use of an independent t-test and ANOVA by SPSSv.22. Results The results showed that high total grades of anxiety and its subscales (panic, autonomic, neurophysiological, and mental) amongst students who exposed Internet addiction in comparison to other groups (P = 0.001). Furthermore, master students showed high scores in Internet addiction compared to bachelor students (P = 0.003). While there is no significant association between general anxiety, gender, and Internet addition (P > 0.001), age also does not show a meaningful relationship with Internet addiction and anxiety (P > 0.001). Conclusions The findings showed that there is a significant difference between students exposed to Internet addiction with students without Internet addiction. In reality, the students who exposed Internet addiction reported high grades of anxiety and its subscales (panic, autonomic, neurophysiological and mental). In general, Internet addiction students experienced high level of anxiety and its symptoms.
... For almost two decades, research on PIU has consistently documented associations between maladaptive online behavior and a host of serious psychological and social problems (for reviews see Aboujaoude, 2010;Caplan & High, 2011;Spada, 2014). Despite disagreements about how to best conceptualize PIU, researchers from a variety of theoretical perspectives have proposed that mood regulation motivates compulsive online activity (Bessière, Kiesler, Kraut, & Boneva, 2004;Caplan, 2010;Gámez-Guadix, Calvete, Orue, & Las Hayas, 2014;LaRose, Lin, & Eastin, 2003). Early speculation by Young (1998) suggested that Internet use has a propensity to alleviate dysphoric moods and may therefore be used to cope with real life problems. ...
The present study hypothesized that two specific positive metacognitions about Internet use (i.e. the belief that Internet use is useful in regulating negative emotions and the belief that it affords greater controllability) mediate the association between emotional dysregulation and problematic Internet use (PIU). A total of 293 undergraduate university students (male 48.4%; mean age=21.73+2.17) participated in the study. The assessed structural model produced adequate fit to the data (χ(2)=203.76; df=81; p<.001; RMSEA [90% CI]=.07 [.06-.08]; CFI=.95; SRMR=.08). Variables accounted for 46% of the variance in PIU levels. A partial mediation model in which emotional dysregulation predicted PIU levels through positive metacognitions associated with Internet use was found. The presence of a direct relationship between emotional dysregulation and PIU was also detected. Moreover, the study found that emotional dysregulation might drive symptoms of PIU to a greater extent than high negative emotionality.
... However, these generalizations often fail to recognize that online games are highly social and playful spaces that can foster a strong sense of connectedness between users (Hussain & Griffiths, 2009;Kowert, 2015;Yee, 2007). In fact, the use of online games has also been found to promote social and psychological well-being, such as reduced depression (Bessiere, Kiesler, Kraut, & Boneva, 2012). To help clarify some of the similarities and differences across the different available CMC channels, the following section will provide a broad overview of several of the more popular Internet-based social services available today. ...
... This position assumes that there is a substantial trade-off between online and offline friendships, relationships, and interactions, and that offline interactions are more socially valuable than online ones in terms of positively contributing to one's psychosocial well-being. Therefore, associations between lower psychosocial well-being and online video game involvement are attributed to the exchange, or displacement, of offline for online social contacts (Bessiere, Kiesler, Kraut, & Boneva, 2012;Blais, Craig, Pepler, & Connolly, 2008;Caplan et al., 2009;Chiu, Lee, & Huang, 2004;Morahan-Martin & Schumacher, 2003;Nie & Erbing, 2002;Williams, 2006aWilliams, , 2007. ...
... Communicating online with strong ties can lead to improvements in well-being (e.g., declines in depression symptoms) typically associated with emotional support, while communicating with strangers online does not deliver this benefit (Bessière et al., 2008;Bessière, Kiesler, Kraut, & Boneva, 2004). Adolescents who communicate with strangers online have lower levels of life satisfaction while those who talk to strong ties are much happier (Valkenburg & Peter, 2007). ...
Full-text available
The current paper presents the results from some research on the relationship between In-ternet addiction, cyberchondria, and different aspects of well-being. The information available on the Internet, which is not necessarily truthful and accurate, can unreasonably amplify users health concerns. Problematic Internet use, health anxiety aroused by online searches for health information and escalation of health concerns as an indicator of cyberchondria, are all associated with a decrease in subjective and eudaimonic well-being as well as in self-esteem. The analyses indicate positive relationships between depressive symptoms on the one hand, and Internet ad-diction and health anxiety, on the other. A conclusion regarding the existence of a relationship between Internet addiction, cyberchondria and decreased levels of well-being could be drawn from the research. Furthermore, the results suggest that self-esteem and eudaimonic well-being correlate positively with the number of people with whom users communicate online.
Full-text available
The current paper presents the results from some research on the relationship between Internet addiction, cyberchondria, and different aspects of well-being. The information available on the Internet, which is not necessarily truthful and accurate, can unreasonably amplify users health concerns. Problematic Internet use, health anxiety aroused by online searches for health information and escalation of health concerns as an indicator of cyberchondria, are all associated with a decrease in subjective and eudaimonic well-being as well as in self-esteem. The analyses indicate positive relationships between depressive symptoms on the one hand, and Internet addiction and health anxiety, on the other. A conclusion regarding the existence of a relationship between Internet addiction, cyberchondria and decreased levels of well-being could be drawn from the research. Furthermore, the results suggest that self-esteem and eudaimonic well-being correlate positively with the number of people with whom users communicate online.
The study used a five-wave latent growth model to compare two sets of competing hypotheses about the influence of overall computer and Internet use on mental health. The secondary data set from the Korean Youth Panel Survey (KYPS) with a valid sample size of 2,003 respondents was used for the analysis. The mediating role of friendship–closeness was examined to compare the social augmentation hypothesis with the social displacement hypothesis. The mediating role of academic stress was tested to compare the mood enhancement hypothesis with the problematic Internet use (PIU) model. The study was consistent with the social displacement hypothesis, where a higher initial computer and Internet use predicted a lower initial friendship–closeness, which, in turn, resulted in a faster decrease in mental health. The study provided reconciliation for the mood enhancement hypothesis and the PIU model. Although a higher initial computer and Internet use predicted lower initial mental health levels through academic stress, schoolchildren and adolescents with more computer and Internet use experienced a slower decrease in mental health because of the mediating role of academic stress.
Full-text available
A large-scale questionnaire study was conducted to test several aspects of different attributional models of everyday problems in living. College students completed scales assessing depression, loneliness, and shyness. In addition, they completed a questionnaire that measured attributional style on five causal dimensions (locus, stability, controllability, globality, and intentionality) for four types of situations (interpersonal success and failure, noninterpersonal success and failure. The results of a series of regression and correlation analyses led to the following major conclusions: (1) Globality, intentionality, and stability may be dropped from attributional models of depression, loneliness, and shyness with little loss of predictive power; (2) controllability is the single most important dimension in predicting a person's level of depression, loneliness, or shyness; (3) locus adds to the prediction of these symptoms only w hen assessed by failure items; and (4) attributional style predicts these ...
Full-text available
Women report greater distress than men, but do women genuinely experience greater distress, suggesting a heavier burden of hardship and constraint? Or do they merely report the feelings in standard indexes more frequently? Perhaps women discuss their emotions more freely. Or perhaps the indexes tap "feminine" emotions such as depression rather than "masculine" ones such as anger. This study analyzes data from a 1990 U.S. sample of 1,282 women and 749 men. Results show that men keep emotions to themselves more than women, and that women express emotions more freely than men. However, these factors do not explain the effect of sex on reported levels of distress--an effect that remains significant with adjustment for these factors. Our results also contradict the idea that the sex difference in distress would diminish if the indexes of distress contained more items that tap anger. Adjusting for emotional reserve and expressiveness, women experience anger more often than men, as they do sadness, anxiety, malaise, and aches. In fact, being female has twice the effect on the frequency of anger that it has on the frequency of sadness. Women report feeling happy as often as men, but adjusting for emotional expressiveness reveals a negative effect of being female on happiness. Overall, women experience distress about 30 percent more often than men. We discuss the possibility that drug abuse and heavy drinking mask male distress, but find little evidence that those behaviors ameliorate distress. We conclude that women genuinely suffer more distress than men.
At the heart of emotion, mood, and any other emotionally charged event are states experienced as simply feeling good or bad, energized or enervated. These states - called core affect - influence reflexes, perception, cognition, and behavior and are influenced by many causes internal and external, but people have no direct access to these causal connections. Core affect can therefore be experienced as free-floating (mood) or can be attributed to some cause (and thereby begin an emotional episode). These basic processes spawn a broad framework that includes perception of the core-affect-altering properties of stimuli, motives, empathy, emotional meta-experience, and affect versus emotion regulation; it accounts for prototypical emotional episodes, such as fear and anger, as core affect attributed to something plus various nonemotional processes.
An important part of a science aimed at the prevention of human dysfunction involves the development of empirically based models that identify processes of risk for or protection from emotional distress or behavioral problems over time. The present study developed and evaluated such a model that proposed two pathways through which family economic pressure was expected to influence change in adolescent internalizing symptoms (depression and anxiety) during the period from the eighth to the tenth grades. A total of 377 rural families in a midwestern state provided data for the analyses. The results were generally consistent with the conceptual model in that family economic pressure increased adolescent perceptions of family economic hardship, which, in turn, reduced the adolescent's sense of control or mastery over time. Lowered mastery was associated with increases in emotional distress. Also consistent with the model, prior levels of mastery appeared to reduce the magnitude of economic stress experienced by the adolescent, whereas prior emotional distress intensified the economic stress process. Although gender differences were found in these processes, the overall pattern of results suggests that girls and boys are both at risk for internalizing problems when families experience economic pressure. Implicatios of the findings for the dvelopment of effective preventive interventions with financially stressed families in rural areas are discussed.
Data from three computer-administered panel studies at the University of North Carolina were analyzed to explore relationships between motivations for television viewing and six individual-differences measures, including shyness, loneliness, self-esteem, and three measures of social support. Viewing motivations were supposed to be related to needs arising from two distinct sources: (a) social compensation, which included companionship, pass time, habit, and escape motivations, and (b) mood management, which included relaxation, entertainment, arousal, and information motivations. The results based on a sample of 290 undergraduates revealed that self-esteem and the three other social-support variables correlated positively with the mood-management viewing motive and negatively with the social-compensation viewing motive. Also as hypothesized, shyness and loneliness correlated positively with the social-compensation viewing motive.
The mass media use and social life of heavy, light, and nonusers of the Internet and personal computers are compared based on a fall 1998 survey of 3,993 nationally representative respondents age 18 and older. As in previous surveys, no significant or consistent evidence of time displacement of such media or social activities was found. Indeed, Internet users showed signs of more active social lives than non- users. These results reinforce the conclusion that personal computer/Internet use may have more in common with time-enhancing home appliances such as the telephone than they do with the time- displacing technology of television.
Just as with most other communication breakthroughs before it, the initial media and popular reaction to the Internet has been largely negative, if not apocalyptic. For example, it has been described as “awash in pornography”, and more recently as making people “sad and lonely.” Yet, counter to the initial and widely publi cized claim that Internet use causes depression and social isolation, the body of ev idence (even in the initial study on which the claim was based) is mainly to the con trary. More than this, however, it is argued that like the telephone and television before it, the Internet by itself is not a main effect cause of anything, and that psy chology must move beyond this notion to an informed analysis of how social iden tity, social interaction, and relationship formation may be different on the Internet than in real life. Four major differences and their implications for self and identity, social interaction, and relationships are identified: one's greater anonymity, the greatly reduced importance of physical appearance and physical distance as “gating features” to relationship development, and one's greater control over the time and pace of interactions. Existing research is reviewed along these lines and some promising directions for future research are described.