SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author

Literature suggests 4 hypotheses to explain social outcomes of online communication among adolescents: displacement, increase, rich-get-richer, and social-compensation hypotheses. The present study examines which hypothesis is supported, considering differences in social ties (time vs. quality of social relationships; parent-child relationships; friendships; school connectedness). This study's sample was 1,312 adolescents ages 12 to 18. Displacement hypothesis predicted negative associations between time in online communication and time with parents, but time with friends was not displaced. Examination of relationships among earlier sociability, online communication, and cohesive friendships supported the rich-get-richer hypothesis. That is, adolescents who already had strong social relationships at earlier ages were more likely to use online communication, which in turn predicted more cohesive friendships and better connectedness to school.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author

Among American households with 8- to 18-year-olds, 85 percent have personal computers. The youth uses home computers more often for recreational activities than for educational purposes. Their most common recreational activities are playing games and communicating through instant messaging (Roberts, Foehr, & Rideout, 2005). Even for preschool children under 6, almost half have some experience surfing online looking at websites for kids (Calvert, Rideout, Woolard, Barr, & Strouse, 2005). Historically, the introduction of a new medium has always aroused similar promises and concerns that old media had about its influences on children and adolescents. Research has been shaped by public concerns, including worries about learning, socialization, emotions, sleep patterns, and moral development (Wartella & Reeves, 1985). Particularly for the Internet, which has two contradictory features such as a personalized/individualized medium vs. an interactive/connecting medium, the issue regarding its influence on adolescents' social isolation or social interaction may be more prominent than for other media.

The observation that a computer is placed in an individual's room rather than a family room and that a child uses a computer alone without any other family members' presence amplifies concerns about social isolation and harmful influences on children's social development (Roberts, Foehr, Rideout, & Brodie, 1999). A few empirical studies identify that time spent using the Internet displaces time spent in face-to-face interaction with family members and friends (Nie, Hillygus, & Erbring, 2002). On the other hand, the statistical figures regarding the use of e-mail and instant messaging to communicate with families and friends support the argument that people are using the Internet as a social tool. Adolescents use the Internet for making connections and maintaining relationships with other people (Lenhart, Madden, Macgill, & Smith, 2007). From a developmental perspective, adolescents try to establish their identity and self-esteem in the relations with friends. They develop and practice advanced social skills within their peer groups, while being involved in many activities, particularly something new and popular among their friends. That is, peer relationships are a social context influencing adolescents' achievement of developmental tasks (Buhrmester, 1990; Hartup, 1983). Adolescents utilize the potentials of interactive media facilitating connections with social contexts, in order to achieve their developmental tasks. Even electronic games and information online can be a source of conversation and interaction among peer groups (Livingstone & Bovill, 2001; Orleans & Laney, 2000; Suoninen, 2001; Valentine & Holloway, 2002).

While positive evaluations and concerns over the Internet coexist, research has tried to elaborate the hypotheses regarding the impact of Internet use on social interaction, social relationships, and social capital.1 The present study examines the following hypotheses to identify whether or not adolescents' use of online communication contributes to maintaining their existing social ties. The hypotheses include (1) the displacement hypothesis, (2) the increase hypothesis, (3) the rich-get-richer hypothesis, and (4) the social compensation hypothesis.

Displacement Hypothesis

The Internet may be harmful to young people's social development because online time displaces time they would spend interacting with families and friends, and weak ties formed online displace strong ties offline. The displacement hypothesis has been one of the most dominant hypotheses explaining the negative effects of media on child development. This argument is based on a zero-sum assumption of time use. That is, people have limited time, thus time spent in one activity interferes with time that would be spent in another activity. While all activities are not displaced by media use, some studies empirically found that social interaction is one valuable activity that is displaced by Internet use (e.g., Kraut, Patterson, Lundmark, Kiesler, Mukopadhyay, & Scherlis, 1998; Mesch, 2003; Nie, Hillygus, & Erbring, 2002).

One of the earlier studies prompting the displacement hypothesis of Internet use and social interaction is Kraut et al.'s (1998) study, called the HomeNet study. This study used longitudinal data to examine the relationships between Internet use, social involvement, and psychological consequences. They found that greater use of the Internet was associated with declines in family communication and size of the local social circle, as well as increases in loneliness and depression. Despite some limitations, such as a small sample size in one city, panel attrition, the possibility of exposure to factors like history and maturation, and no control group, their study has been frequently cited in the press and in other academic papers, which have amplified public concerns about the harmful effects of Internet use. Among recent studies, Mesch (2003, 2006) who examined Israeli adolescent Internet use found that Internet use was negatively related to family closeness, and positively related to family conflicts. Nie et al. (2002) also support the displacement between Internet use and social time. They utilized the time diary method, criticizing inaccurate measures of time through global estimates. Based on time diary data, they found that greater Internet use was related to less time spent with family and less time with friends.

Among studies arguing the displacement hypothesis, there are a few that focus on the quality of online communication (Cummings, Butler, & Kraut, 2002; Gross, Juvonen, & Gable, 2002; Parks, 1996). For instance, Cummings et al.(2002) argue that people perceive e-mail as a less useful means for developing and maintaining close social relationships than face-to-face contact and telephone conversations, and the listserv as less valuable than offline small groups for establishing a sense of belonging and for gaining social support. The perception of inferiority of online communication may result from a lack of social cues and absence of physical proximity.

In sum, the displacement hypothesis has been supported by findings that time online is negatively related to time in face-to-face interaction, that online communication is less useful or valuable than face-to-face communication, and that Internet use is negatively related to existing intimate relationships such as families. However, these findings should be interpreted with caution: First, time displacement is not enough to account for the negative influence of Internet use on the quality of relationships. While time displacement examines changes in the amount of time, good relationships may be explained by the quality of time rather than the quantity of time. Second, even though online communication is inferior to face-to-face communication, online communication can serve as an additional, not a tradeoff, to face-to-face communication. Particularly in certain situations such as distant relationships, online interaction can be an alternative to face-to-face interaction.

Increase Hypothesis

The increase hypothesis suggests that Internet use increases social interaction, the size of social networks, and closeness with others, as a means of maintaining social ties and creating new ones. This positive argument is partially based on the potential of the Internet as an interactive medium that can connect people to people while overcoming the barriers of time and place. In addition, the characteristics of the Internet such as anonymity and lack of social cues may facilitate users initiating new relationships. However, whether people utilize the technological potential or not is another question, which can be explained by the uses and gratifications perspective. People use the media based on their needs and motives, and their media use is reinforced by the obtained gratifications. This perspective has identified social interaction is one of people's common reasons for media use (Blumler & Katz, 1974; McQuail, 1987). Particularly for adolescents who begin to expand their interests outward, and have strong needs to be connected with their friends, social interaction will be one of the most primary motives for online communication. The theoretical approach of uses and gratifications to the interactive services of the Internet provides the basis for the increase hypothesis. The increase hypothesis of the impact of the Internet on social relationships and social networks is supported by the following empirical studies.

Unlike the findings of their earlier study in 1998, Kraut, Kiesler, Boneva, Cummings, Helgeson, and Crawford's follow-up study with the original sample in 2002 found that the negative relationships between Internet use and family communication and size of social network were no longer significant. In addition, the analysis of a new sample consisting of recent purchasers of computers and television presented that those who used the Internet had larger increases in face-to-face interaction with friends and family, in the size of their local social circles, and in the size of their distant social circles. Particularly for teens, their frequent use of the Internet increased family communication and social support. They noted that the shift in findings between an early study and a later study might result from maturation of participants, changes in the way participants use the Internet, and changes in the Internet environment and services.

Meanwhile, Wellman, Hasse, Witte, and Hampton (2001) argue that the Internet serves as a supplementary means of offline interaction. That is, the Internet is more helpful for maintaining existing social networks than for creating new ones. Offline interaction stimulates, rather than is stimulated by, online interaction. They found that frequent Internet users were more likely to contact friends and relatives via e-mail than were less frequent users, while contact via e-mail did not reduce the frequency of face-to-face contact and phone calls. Shklovski, Kraut, and Rainie (2004) also found that visiting a family member increased the frequency of e-mailing that person, but e-mailing neither increased nor decreased the frequency of communicating by phone or in person.

Studies focusing on adolescents' Internet use consistently present the idea that instant messaging is used as an additional communication tool rather than displacing the telephone (Gross et al., 2002; Howard, Rainie, & Jones, 2001; Lenhart, Madden, & Hitlin, 2005; Lenhart, Rainie, & Lewis, 2001). Lee and Kuo (2002) found that the amount of time spent on the Internet increased time interacting with friends. Furthermore, Valkenburg and Peter (2005, 2007) found that Internet use, particularly online communication enhances the quality of friendships directly and indirectly through increased time with friends. According to Gross (2004), most instant messaging partners are friends or best friends from school. Their online interactions occur in a private setting, with friends who are part of their daily offline lives, with ordinary yet intimate topics. The communication of intimate topics strengthens their closeness with friends.

The increase hypothesis is further extended into the rich-get-richer hypothesis and the social compensation hypothesis. These hypotheses are interested in who, a socially integrated person or a socially isolated person, uses the Internet more frequently, particularly focusing on online communication, and accordingly they discuss who benefits more from Internet use.

Rich-Get-Richer Hypothesis

The rich-get-richer hypothesis, which was suggested by Kraut et al. (2002), proposes that those who already have strong social networks and social skills benefit the most from the Internet. That is, initial social connection or competence functions as a moderator based on the interaction effect of Internet use with extroversion. They found that Internet use was associated with better outcomes for extroverts and worse outcomes for introverts. For extroverts, using the Internet was related to increases in well-being, including increases in self-esteem, and decreases in loneliness, and negative affect. In contrast, introverts showed declines in well-being associated with these same variables.

It is also argued that existing social connection or competence may be an antecedent variable of Internet use. For instance, Gross et al. (2002) found that teens with strong connections to school-based peers use the Internet to seek out additional opportunities to interact with them, while teens who felt lonely and socially anxious tried online communication with strangers. The Pew Internet survey also found that buddy list size was directly related to the intensity and duration of instant messaging use (Lenhart et al., 2005). Bryant, Sanders-Jackson, and Smallwood (2006) presented that adolescents who had more friends were more likely to use instant messaging than those who had fewer friends. Valkenburg and Peter (2005) performed path analysis to examine which hypothesis, between the rich-get-richer hypothesis and the social compensation hypothesis, is appropriate in explaining adolescents' Internet use and friendships. They found that social anxiety was negatively related to online communication, which in turn was positively related to closeness to peers. That is, their path analysis supports the rich-get-richer hypothesis. In sum, they argue that online communication contributes to the solidarity of peer group networks for sociable adolescents, because sociable adolescents are more likely than socially anxious adolescents to use online communication more frequently. However, because of cross-sectional data, we cannot identify causal relationships between initial social resources or sociability and Internet use.

Social Compensation Hypothesis

The social compensation hypothesis, proposed as the alternate model to the rich-get-richer hypothesis in Kraut et al.'s (2002) study, states that the Internet is more beneficial for socially anxious and isolated people. The Internet may compensate for lack of a social network offline because socially anxious people may feel more at an advantage in developing intimate relationships online. McKenna and Bargh's (1999) conceptual framework well explains how online interaction compensates for lack of sociability and social network. Stigmatized identity, constrained identity, social anxiety, and loneliness serve as motivators for online interaction. The characteristics of the online environment such as text-based communication, lack of visual and auditory cues, and anonymity facilitate disclosure of a true or idealized self, gaining intimacy with others through self-disclosure, and leading to formation of new relationships. As a consequence, social networks increase, and loneliness and depression decrease.

McKenna and Bargh's theoretical arguments are partly supported by several empirical studies which examine introverts' attitudes toward online communication, and the association between self-disclosure and formation of online relationships. For instance, introverts more strongly agreed that online modes of communication offer greater freedom of expression, and they were more likely than extroverts to choose online communication for interactions with friends (Goby, 2006). Shy people had much lower levels of shyness, lower levels of rejection sensitivity, and higher levels of interpersonal competence in initiating relationships online than offline (Stritzke, Nguyen, & Durkin, 2004). Better self-disclosure over the Internet was related to formation of close online relationships, and moved to face-to-face relationships (McKenna, Green, & Gleason, 2002). Meanwhile, Bessiere, Kiesler, Kraut, and Boneva (2008) found that overall Internet use was not related to users' well-being, but using the Internet for meeting people reduced depressive affect for people with low initial social resources. This finding suggests that the possibility for socially isolated persons to get benefits from the Internet depends on their goals or motives behind the use of the Internet. Thus, we need to examine if socially isolated persons have some motives for online communication. As the uses and gratifications approach argues, different users have different needs or motives for a particular media. If those who are anxious about face-to-face communication and who does not find satisfaction from face-to-face communication consider online communication a functional alternative (Papacharissi & Rubin, 2000), they could compensate for their weak ties through online interaction.

Summary

The literature suggests two contrary hypotheses in explaining social outcomes of Internet use among adolescents. While the displacement hypothesis predicts that the Internet weakens adolescents' existing relationships with families and friends, the increase hypothesis argues that Internet use contributes to maintaining those social relationships, particularly friendships, and developing new social connections. The increase hypothesis is extended into the rich-get-richer and social compensation hypotheses which suggest the consideration of initial social relationships or sociability. However, existing studies have some sampling and measurement issues which may limit full understanding of social impact of Internet use among adolescents. The present study adds to the existing literature in the following ways. First, the data utilized in this study come from the Child Development Supplement to the Panel Study of Income Dynamics, a nationally representative sample of U.S. adolescents. Second, considering different impact of Internet use by the characteristics of social ties, the present study examines the associations of Internet use with family relationships and peer relationships separately. Third, each type of social relationship as an outcome is measured in two ways, that is, time-based measurement (e.g. time with friends) and quality-based measurement (e.g. closeness with friends), and analyzed separately. Fourth, for the displacement hypothesis of online time and social time, the present study uses time diary data (as part of the CDS data) which is considered more reliable than global estimates (Vandewater & Lee, in press). Fifth, for the rich-get-richer and social compensation hypotheses, the longitudinal relationship between initial sociability and online communication is examined. Figure 1 and Figure 2 present the conceptual models tested in the present study.

image

Figure 1. Time Displacement Hypothesis.

Download figure to PowerPoint

image

Figure 2. Links Among Earlier Sociability, Online Communication, and Cohesive Social Relationships (Rich-Get-Richer Hypothesis vs. Social Compensation Hypothesis).

Download figure to PowerPoint

Method

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author

Participants

The data for this study came from the Panel Study of Income Dynamics (PSID)—Child Development Supplement (CDS). The PSID is a longitudinal study of a representative sample of U.S. individuals and their families. In 1997 the first wave of data (CDS I) was collected on 0- to 12-year-old children and their parents (3,563 children), focusing on children's developmental outcomes within the context of family, neighborhood, and school environments. In 2002-2003 (CDS II), they recontacted the families and children, and interviewed 2,907 children and adolescents aged 5 to 18 years (http://psidonline.isr.umich.edu/CDS/cdsii_userGd.pdf). Appropriately weighted, these data provide nationally representative estimates (For a detailed discussion, see http://psidonline.isr.umich.edu/CDS/weightsdoc.html).

The present study focuses on adolescence, because this developmental stage is widely accepted as a time in which individuals spend a great deal of energy creating and maintaining social relationships. Among the sample in the second wave of data, 1,312 adolescents ages 12 to 18 were utilized for the present study. They participated in the child interview, which provides data regarding the frequency of Internet use for specific purposes and social interactions. In addition to the child interview, interviews with a primary caregiver about the target child and time diaries were used as the methods of data collection.

Measures

The data utilized in the present study were collected through the primary caregiver interview, the child interview, and the time diary. The primary caregiver interview includes a primary caregiver's responses of the target child's general quality of social relationships and internalizing behavior problems in Wave 1. The time diary data provide the information about the amount of time they spent on a computer, interacting with parents, and interacting with friends at Wave 2. The child interview provides information about a child's Internet use for specific purposes, family relationships, friendships, and connectedness to school at Wave 2.

Earlier Sociability

Earlier sociability is measured by two different variables, the quality of social relationships and internalizing behavior problems. These two variables were collected through interviews with a primary caregiver at Wave 1. The quality of social relationships includes five items assessing a child's relationships with friends, with the primary caregiver, with the other parent, with siblings, and with a teacher. The scale ranged from 1 (poor) to 4 (excellent). The measure for internalizing behavior problems is a subscale of the Behavior Problem Index (BPI) which was originally developed by James Peterson and Nicholas Zill (1986) from the Achenbach Behavior Problems Checklist. Among 14 items of the internalizing behavior problem measure, the five items with highest factor loadings were used as indicators of the internalizing behavior problem construct. The items include “withdrawn,”“trouble getting along,”“feeling others are out to get him,”“unhappy and sad,” and “feeling worthless.” A primary caregiver was asked whether these behaviors were not true, sometimes true, or often true of the target child.

Time Spent on a Computer

The time diary captures computer use for communication. Each child completed two time diaries, one for a weekday and one for a weekend day. These days were randomly selected when the interviewer completed the initial contact for the household. On the diary, every minute of the two 24-hour periods was accounted for with a primary activity and, if applicable, a secondary activity. Children were also asked who was doing the activity with them and who was there but not participating directly in the activity. Children's computer use was calculated by summing the minutes spent using a computer on a weekday and a weekend day. Computer use was separately coded by specific purposes: (1) computer communication, (2) computer use for study and school work, (3) computer games, and (4) other recreational computer activities. Computer communication includes e-mail, computer/video/speaker phone, Internet phone, teleconferencing, chat rooms, instant messaging, and e-cards. Computer use for study includes using a computer for homework, studying, research, and reading related to classes. Other recreational computer activities included surfing the net, downloading pictures, music, movies, burning CDs, watching DVDs on the computer, and creating/programming.

Time Interacting with Parents

Time with parents was measured by using the time diaries to indicate the amount of time the child spent with either parent while doing an activity together on one weekday and one weekend day. The time that the child used the computer with either parent was subtracted from the total amount of time they spent with parents to prevent overlap between independent and dependent variables.

Time Interacting with Friends

Time with friends was measured by summing the amount of time the child spent with friends while doing an activity together on one weekday and one weekend day. The amount of time that the child used a computer with friends was subtracted from the total amount of time they spent with friends.

Frequency of Online Communication

Information on Internet use is drawn from the child interview. Adolescents were asked to report how often they used the Internet in the last month: (1) to use e-mail, (2) to use a chat room or instant messenger. The scale ranged from 1 to 6 (1 = never , 2 = once or twice in the last month, 3 = about once a week, 4 = two or three times a week, 5 = almost every day, 6 = every day).

Cohesive Parent-Child Relationships

The construct of cohesive parent-child relationships consists of three scales: conversations, closeness, and support. Intimate conversation with parents: As one of scales to indicate cohesive parent-child relationships, this was measured by three items assessing the frequency of a child's conversation with his/her mother about his/her social life, such as “in the last month, how often did you talk with your mother/stepmother about things that are going on with your friends; about your plans for the future; about problems you are having in school?” Rather than measuring all the topics about which an adolescent could talk with parents, it was limited to the topics about which an adolescent having good relationships with parents could talk with their parents. The questions about frequency in conversation with the father were asked as well. However, 40% of the children reported they did not live with their father, and as a result 40% of the sample did not respond to conversations with father items. Thus, conversation with parents included only three items asking about conversations with their mother in order to maintain the sample size. The scale ranged from 1 (never) to everyday (6), and was created by the mean score of the three items. Closeness to parents: This variable was measured by asking a child, “How close do you feel towards your mother?” The scale ranged from 1 (not very close) to 4 (extremely close). Support to parents: Adolescents were asked to report in the last 6 months how often they have helped their parents with things they had to get done such as chores and running errands; and how often they have provided emotional support to their parents such as making them feel better when they were sad. The scale ranged from 1 (almost never) to 7 (everyday), and was created by the mean score of the two items.

Cohesive Friendships

The construct of cohesive friendships also consists of three scales: conversations, closeness, and support. Intimate conversation with friends: This was measured by three items assessing the frequency of a child's conversation with his/her friends about his/her social life. The questions were “in the last month, how often did you talk with your friends about things that are going on with your friends; about your plans for the future; about problems you are having in school?” The scale ranged from 1 (never) to 6 (everyday), and was created by the mean score of the three items. Closeness to friends: This variable was measured by asking a child, “How close do you feel towards your friends?” The scale ranged from 1 (not very close) to 4 (extremely close). Support to friends: Adolescents were asked to report in the last 6 months, how often they have helped their friends with things they had to get done, such as homework or chores, and provided emotional support to their friends such as giving them advice on a problem or making them feel better. The scale ranged from 1 (almost never) to 7 (everyday), and was created by the mean score of the two items.

School Connectedness

Connectedness to school was measured by three items assessing the degree of inclusiveness, closeness, and happiness at school. The questions included: “In the last month, how often did you feel like you were part of your school; close to people at your school; and happy to be at your school?” The scale ranged from 1 (never) to 6 (everyday).

Analysis Plan

Overarching Analytical Strategy

The structure of the CDS data demands specific analytical attention. First, the CDS data have a sampling weight issue. The CDS data are based on an oversample of low income families, mostly African Americans. In addition, as a panel study, CDS II data has a sample attrition issue. Thus, all the analyses were performed using a CDS II sample weight, which was created by the inverse of the probability of sample selection and attrition adjustment factor. Using this weight, the CDS data present a nationally representative sample. Second, the CDS data has a nested nature, because the sample includes up to two children from one family. The siblings have same information in the family level data. That is, siblings in one family have same values for the family variables such as family income to needs ratio, parental education level and neighborhood quality. This data feature violates the assumption of independence stating that the error associated with each data point is independent of every other error value, thus resulting in an increase in the type 1 error rate. Thus, nonindependence in the CDS sample should be corrected in the analysis. Third, the missing values on variables are handled by full information maximum likelihood estimation (FIML). Thus the present study could maintain the same sample size in all analyses. MLR, a maximum likelihood estimator with robust standard errors, can handle biases resulting from these characteristics of CDS data. The estimates by MLR are robust to nonnormality and nonindependence of observations (Muthen & Muthen, 2006). The following analyses were performed in Mplus 4.2.

Regression Analysis

The time displacement hypothesis, which suggests time spent using a computer and the Internet displaces time with families or friends, is examined by using time use variables, because the reliable measurement of time and full count of 24 hours are crucial. Regression analyses are performed to examine whether time spent on a computer displaces time with parents and time with friends, controlling for family income to needs ratio, household head's education level, race, age, gender and earlier sociability.

Structural Equation Modeling (SEM)

Structural equation modeling technique is used to identify whether online communication displaces or enhances cohesive parent-child relationships, friendships and school connectedness, and whether earlier sociability serves as an antecedent of online communication. The quality of social relationships and internalizing behavior problems measured in CDS I were entered as exogenous variables to indicate earlier sociability. Three structural equation models where cohesive parent-child relationships, cohesive friendships, and school connectedness were separately entered as an endogenous variable were tested, controlling for family income to needs ratio, household head's education level, child race, age, and gender.

Underlying assumptions of SEM were validated: multivariate normality, outliers, linearity, multicollinearity, and missing data. Multivariate normality can be detected by skew and kurtosis of univariate distributions. Skewness and kurtosis of all the variables except for items related to internalizing behavior problems were within the range of ± 1.96, which indicates acceptable boundaries of normality. Given that maximum likelihood estimation (ML), which is the common method in SEM for estimating path coefficients, requires multivariate normality, particularly normality of endogenous variables, skew and kurtosis of items related to internalizing behavior problems—one of the exogenous variables—is acceptable. Moreover, MLR (robust ML) used in this study is maximum likelihood estimation with standard errors that are robust to nonnormality and nonindependence of weighted and cluster data.

The model is evaluated by chi-square (χ2), the comparative fit index (CFI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). A good fit is denoted by a nonsignificant chi-square, the RMSEA less than .05, the SRMR less than .10, and the CFI more than .90 (Klein, 2005; Marsh, Hau, & Wen, 2004; Zhu, Walter, Rosenbaum, Russell, & Raina, 2006).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author

Time Spent on a Computer and Social Time

Regression analyses were performed to test whether time spent on a computer displaces social time or not (Table 2). Overall, the findings supported the displacement hypothesis between computer time and social time, but interestingly, time spent on computer communication was not related to time spent interacting with friends. Unstandardized coefficients indicate that an increase of 1 minute in computer-mediated communication explains a decrease of 0.4 minute in time spent interacting with parents. In terms of hours, an increase of 1 hour in computer-mediated communication results in a decrease of 24 minutes in time with parents. Computer use for recreation also displaced time with parents. Even though computer use for study and computer games had negative coefficients, their relationships were not statistically significant.

Table 2.  Unstandardized Regression Coefficients for Time With Parents and for Time With Friends
 Time interacting with ParentsTime interacting with Friends
  1. ***p < .001, *p < .05

Computer Communication−.40***.00
Computer for study−.20−.51*
Computer Games−.13−.26
Computer for Recreation−.36***−.21*
R2.15.08

Earlier Sociability, Online Communication, and Social Outcomes

For the measurement model, the factor loadings of indicators for each latent construct are presented in Table 3. Standardized path coefficients from all covariates to constructs in the model are given in Table 4. For the structural model, the path coefficients among latent constructs and the variance explained for each endogenous variable are presented in Figures 3 and 4.

Table 3.  Standardized Factor Loadings in the Measurement Model
Latent construct and Observed IndicatorsFactor loading
Social Relationships 
  With friends.61
  With a primary caregiver.58
  With other caregiver.54
  With siblings.56
  With a teacher.59
Internalizing behaviors 
  Withdrawn.45
  Trouble getting alone.54
  Feeling other are out to get him.60
  Unhappy, sad, depressed.65
  Feeling worthless.75
Online communication 
  Email.89
  IM/Chat.73
Cohesive Parent-Child Relationships 
  Conversation.69
  Closeness.53
  Supports.52
Cohesive friendships 
  Conversation.74
  Closeness.49
  Supports.63
Connectedness to school 
  Inclusiveness to school.66
  Closeness to school.74
  Happiness at school.65
Table 4.  Standardized Path Coefficients From Covariates to Constructs in the Structural Model
 Covariates
Latent Endogenous ConstructsIncome to needs ratioParent EducationRaceaGenderbAge
  1. aWhite = 1; Non-White = 0

  2. bGirls = 1; Boys = 0

  3. ***p < .001,

  4. **p < .01,

  5. *p < .05

Online Communication.12**.15**.17***.21***.18***
Cohesive parent-child relationships.05−.03−.04.20**.08
Cohesive friendships−.01−.11.00.33*.07
Connectedness to school.00.16**−.06−.16**−.19***
image

Figure 3. Result of Hypothesized SEM on Cohesive Friendships.

Download figure to PowerPoint

image

Figure 4. Result of Hypothesized SEM on Connectedness to School.

Download figure to PowerPoint

Cohesive Parent-Child Relationships

The associations among earlier sociability, online communication, and cohesive parent-child relationships were tested. Among fit indices, RMSEA and SRMR suggested that the estimated model fit the data, while CFI indicates unacceptable model fit (χ2(141) = 436.31, p < .000; CFI = 0.87; RMSEA = .04; SRMR = .05). In addition, the path coefficient from online communication and parent-child relationships was not statistically significant (β = .07, n.s).

Cohesive Friendships

The associations among earlier sociability, online communication, and cohesive friendships were tested. Figure 3 shows the model fit indices, the standardized path coefficients and the variance explained for each endogenous factor. The tests of model fit suggested that the hypothesized model was acceptable (χ2(141) = 379.94, p < .000; CFI = 0.90; RMSEA = .04; SRMR = .04). The chi-square test was significant, which means that the estimate model was significantly different from the observed data, in other words, the estimate model did not fit the data. However, a chi-square is sensitive to sample size, and usually leads to model rejection. Thus, given that a chi-square/degree of freedom ratio, “normed chi-square” that does not exceed five indicates reasonable model fit (Klein, 2005), the normed chi-square in this model was 2.69. CFI, RMSEA, and SRMR indicate acceptable model fit.

Twenty-nine percent of the total variance in friendships was explained. Initial social relationships were positively related to online communication, which was in turn positively related to cohesive friendships. That is, adolescents who had better earlier social relationships more frequently used online communication, which in turn was related to the outcome of better friendships. However, earlier internalizing behavior problems were not significantly related to online communication.

According to Holmbeck's recommendation (Holmbeck, 1997), the present study compared chi-squares of the nested model and the alternative model with the direct paths between the quality of social relationships and cohesive friendships, and between internalizing behavior problems and cohesive friendships. The alternative model fit the data as well (χ2(139) = 373.95.18, p < .000; CFI = 0.90; RMSEA = .04; SRMR = .04). To compare the nested model and the alternative model, the Satorra-Bentler scaled chi-square difference test (TRd), which is used in estimations using MLR, was performed (Muthen & Muthen, 2005). Using this test, the chi-square was not significantly improved in the alternative model (Trd = 2.71, that is, χ2(2) = 2.71, n.s). In addition, the relationship between earlier social relationships and cohesive friendship (β = .23, p < .001) was reduced to .15 (p < .05), when a mediator of online communication is considered. Thus, based on model parsimony principle (i.e. if two models have similar model fit, the simpler one is to be preferred), the nested model presented in Figure 3 was considered the final model in this study.

In sum, the findings supported the rich-get-richer hypothesis, indicating initial social relationships were positively related to frequent use of online communication, which in turn was positively related to cohesive friendships. Even though earlier relationships were directly related to later cohesive friendships, online communication partially mediated the association between earlier social relationships and cohesive friendships.

Connectedness to School

The SEM with the five latent variables, quality of social relationships, internalizing behavior problems, online communication, cohesive friendships, and connectedness to school, was tested. Figure 4 shows the model fit indices, the standardized path coefficients and the variance explained for each endogenous factor. The tests of model fit suggested that the hypothesized model was acceptable (χ2(193) = 447.11, p < .000; CFI = 0.91; RMSEA = .03; SRMR = .05). The normed chi-square in this model was 2.32. CFI, RMSEA, and SRMR indicated that the model was acceptable. The direct effect of online communication on school connectedness was reduced to nonsignificance after adding cohesive friendships. Online communication was indirectly related to connectedness to school through cohesive friendships (β = .19, p < .001).

The model was compared with the alternative model with four direct paths between quality of social relationships and cohesive friendships, between internalizing behaviors and cohesive friendships, between quality of social relationships and connectedness to school, and between internalizing behaviors and connectedness to school. The alternative model also fit the data (χ2(189) = 426.96, p < .000; CFI = 0.92; RMSEA = .03; SRMR = .04). However, the Satorra-Bentler scaled chi-square difference test indicated no significant differences between the nested model and the alternative model (Trd = 1.84, that is, χ2(4) = 1.84, n.s). Thus, this study decided the nested model presented in Figure 4 as the final model.

Discussion and Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author

The present study examined how Internet use, particularly online communication is related to adolescent social ties, parent-child relationships, friendships, and school connectedness, considering their earlier sociability. According to the findings, for parent-child relationships, time using a computer for communication was negatively related to time interacting with parents. However, frequent use of online communication did not weaken or strengthen the quality of parent-child relationships, even though one of the model fit indices indicated that the hypothesized model did not fit the data very well. For friendships, time using a computer for study and recreation was negatively related to time with friends, whereas time in online communication was not related to time with friends. Rather, frequent online communication was associated with cohesive friendships. As Wellman et al. (2001) and Shklovski et al. (2004) argue, these findings also suggest that adolescents use online communication as an additional communication modality to enhance the quality of friendships, while time in face-to-face interaction with friends is not decreased or increased. For school connectedness, adolescents who more frequently used online communication were more likely than less frequent users to feel connectedness to school, by having friendships that are more cohesive.

The findings that online communication displaces time with parents, but it does not displace time with friends can be explained by the principles of displacement such as physical and psychological proximity, and marginal fringe activities (Neuman, 1991). That is, computer activity and family interaction usually occur in the same place at home, and online communication may provide as much as or more psychological satisfaction than interaction with parents. Additionally, for adolescents family interaction is a lower priority and less attractive than peer interaction. Thus, time spent with parents is more likely to be displaced by online communication than time spent with friends. Given that family time is a crucial resource for children's healthy development, the finding of displacement of time interacting with parents may support public concerns of children's media use. Particularly for young children, interactive activities with parents promote cognitive and physical development as well as social development. Even though this study found that online communication was not cross-sectionally associated with the quality of parent-child relationship, family time displaced by online communication might have negative implications for children's overall development in the long term.

Examination of the associations among initial sociability, online communication, and cohesive friendships supported the rich-get-richer hypothesis. The hypothesized model fit the data. It is also noteworthy that the model held in the face of a variety of covariates including family income to needs ratio, household head's education level, child race, age, and gender. Adolescents who already had strong social relationships in their earlier age were more likely to use online communication, which in turn predicted more cohesive friendships, and furthermore better connectedness to school. Contradictory to the social compensation hypothesis, adolescents with poor social relationships tended to have less needs for e-mail and instant messaging than those who had strong social relationships. A few prior studies suggest that the social compensation hypothesis may predict the relationships of sociability and online communication under certain conditions. Specifically, Bessiere et al. (2008) found that when using the Internet for the purpose to meet people, depressive affect decreased for those who had poor social resources. Valkenburg and Peter (2005) found that the extent of self-disclosure was positively related to online communication. Taken together, if socially anxious and shy individuals who expect to meet new people through online communication are willing to self-disclose using online communication, then they would gain some social benefits which may compensate for their poor sociability offline.

The finding supporting the rich-get-richer hypothesis may be explained by the uses and gratifications approach. Users have different needs for Internet use and different gratifications. While socially integrated persons may have motives for online communication in order to maintain their existing social relationships, socially isolated persons may have needs for initiating and building new social relationships online. If e-mail, instant messaging, and online chatting are more helpful for maintaining existing relationships than for forming new relationships, socially integrated persons will have more gratification through e-mail and instant messaging, as a result, they will have more frequently use them than socially isolated persons. Thus we need to further examine socially isolated and anxious individuals' motives for online communication, the extent of their obtained gratifications, willingness to self-disclosure, and some features of online communication applications or services which can enhance their gratifications.

Users of social network sites such as Facebook also have more motives for keeping in touch with old friends and checking out someone they met offline than for meeting new people or strangers (Ellison, Steinfield, & Lampe, 2007). According to Ellison et al. (2007), undergraduate students' intensive use of Facebook was related to bridging and bonding social capital, and particularly for users with low self-esteem, intensive use of Facebook was related to greater bridging social capital than for users with high self-esteem. This is partly because social network sites have some features helping users make visible their “latent” social ties (Boyd & Ellison, 2007). Users of social network sites can present their profiles, build their own mini pages, and see other members' profiles and homepages. Through the profiles, they can develop latent ties or weak ties, which started offline, such as friends of friends or someone sharing a class at school. They can also join in virtual communities based on their common interests and hobbies. That is, social network sites have more room than e-mail and instant messaging for socially anxious persons to experiment their identities and articulate their invisible social networks. Thus, the social compensation hypothesis may be well examined through adolescents' use of social network sites rather than e-mail and instant messaging.

As a follow-up study, structural equation modeling for subgroups was conducted separately: boys vs. girls, and White adolescents vs. Black adolescents. Even though the models tested in the study had controlled for the influences of the demographic variables on online communication, the study conducted a follow-up analysis in order to examine if the model is applicable across groups. As a result, all the models for subgroups fit the data very well. The directions of path coefficients were consistent with the model for the whole sample, while the sizes of coefficients were a little bit different, but almost same. For example, for the group of girls the path coefficient of online communication on friendships was a little bit larger than the coefficient for the group of boys (β: .40 vs. .38). Thus, we can say that even though boys and girls, and White and Black adolescents show differences in the extent of use of online communication, the impact of online communication on friendships and school connectedness was the same across subgroups. Regardless of their gender and race, the rich-get-richer hypothesis predicts the associations among earlier quality of social relationships, online communication, and cohesive friendships, and furthermore, connectedness to school.

The present study is benefited by the use of PSID-CDS I and CDS II, which has a representative sample, rich indicators of social outcomes, and reliable time diary data. However, the longitudinal data with only two time points have some limitations in developing longitudinal structural equation models. That is, while the present study identified the longitudinal relationship between earlier sociability at time 1 and Internet use at time 2, which is crucial for examination of the rich-get-richer and social compensation hypotheses, the associations between Internet use and cohesive social relationships were examined by the data collected at the same time. Thus, social outcomes of Internet use cannot imply causal relationships.

Conclusively, a computer and the Internet are integrated into young people's daily lives. With the prevalence of new technology in their lives, it is time to identify its social impact and who benefits more. The present study examined competing hypotheses explaining social outcomes of Internet use. The hypotheses supported by the data depended on the type of online activity and the type of social outcome examined. This suggests that discussion regarding social impact of Internet use should be specified by the type of online activity and the type of social networks. The findings supporting the rich-get-richer hypothesis imply that Internet use may lead to the digital divide in social capital. As economically rich persons have more chance to access the Internet, socially rich persons with strong ties more frequently use online communication, and as a result, they can build or maintain more cohesive friendships and connectedness to school than persons with lower sociability. The quality of social relations such as cohesiveness and connectedness is an important element to facilitate the transfer from alters' resources to an ego actor. While the findings of this study support the rich-get-richer hypothesis, future research should investigate how we can motivate socially isolated adolescents to get such benefits of Internet use as cohesive peer relationships and connectedness.

Notes
  • 1

    The definition of social capital has some variation among scholars. For instance, Bourdieu defines social capital as “the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance and recognition, in other words, to membership of a group” (Bourdieu, 1986, p. 248). He noted an actor's social capital depends on the size of network, and the volume of capital/resources by each of those to whom he is connected. Portes defined social capital as “the ability of actors to secure benefits by virtue of membership in social networks or other social structures” (Portes, 1998, p.3). They emphasize different elements of social capital. For instance, Bourdieu emphasize the resources embedded in the networks or possessed by each member linked to the networks, and Portes highlights an actor's ability to get benefits from social networks. However, they have a commonality about the mechanism by which they can benefit from social network. Boudieu noted that for an individual to get benefits from social networks, the networks should be based on solidarity, and Portes suggested internalized norms, bounded solidarity, reciprocity, and enforceable trust as sources of social capital. That is, the quality of social relations such as solidarity/cohesion, reciprocity and trust explains the process that resources embedded in networks or resources possessed by others are transferred to an actor. Thus, the quality of social relations is one of the most significant components generating an actor's social capital. The present study focuses on the quality of social relation of an individual's existing social networks in terms of social capital.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author
  • Bessiere, K., Kiesler, S., Kraut, R., & Boneva, B. (2008). Effects of Internet use and social resources on changes in depression. Information, Communication & Society, 11(1), 4770.
  • Blumler, J. G. & Katz, E. (1974). The uses of mass communication. Newbury Park, CA: Sage.
  • Bourdieu, P. (1986). The forms of capital. In J. G.Richardson (Ed.), Handbook for Theory and Research for the Sociology of Education (pp. 241258). New York: Greenwood Press.
  • Boyd, D. M., & Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13(1).
  • Bryant, J. A., Sanders-Janckson, A., & Smallwood, A. M. K. (2006). IMing, text messaging, and adolescent social networks. Journal of Computer-Mediated Communication, 11(2).
  • Buhrmester, D. (1990). Intimacy of friendship, interpersonal competence, and adjustment during preadolescence and adolescence. Child Development, 61, 11011111.
  • Calvert, S. L., Rideout, V. J., Woolard, J. L., Barr, R. F., & Strouse, G. A. (2005). Age, ethnicity, and socioeconomic patterns in early computer use. American Behavioral Scientist, 48(5), 590607.
  • Cummings, J. N., Butler, B., & Kraut, R. (2002). The quality of online social relationships. Communication of the ACM, 45(7), 103108.
  • Ellison, N. B. Steinfield, C., & Lampe, C. (2007). The benefits of Facebook “friends:” Social capital and college students' use of online social network sites. Journal of Computer-Mediated Communication, 12(4).
  • Goby, V. P. (2006). Personality and online/offline choice: MBTI profiles and favored communication modes in a Singapore study. CyberPsychology & Behaviors, 9(1), 513.
  • Gross, E. F. (2004). Adolescent Internet use: What we expect, what teens report. Applied Developmental Psychology, 25, 633649.
  • Gross, E. F., Juvonen, J., & Gable, S. L. (2002). Internet use and well-being in adolescence. Journal of Social Issues, 58(1), 7590.
  • Hartup, W. W. (1983). Peer relations. In P.H.Mussen (Ed.) Handbook of Child Psychology: Vol. 4. Socialization, Personality, and Social Development (pp. 103196). New York, NY: John Wiley.
  • Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the study of mediators and moderators: Example from the child-clinical and pediatric psychology literatures. Journal of Consulting and Clinical Psychology, 65, 599610.
  • Howard, P. E. N., Rainie, L., & Jones, S. (2001). Days and nights on the Internet. American Behavioral Scientist, 45(3), 383404.
  • Klein, R. B. (2005). Principles and practice of structural equation modeling. New York, NY: The Guilford Press.
  • Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., & Crawford, A. (2002). Internet paradox revisited. Journal of Social Issues, 58, 4974.
  • Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., & Scherlis, W. (1998). Internet paradox: A social technology that reduces social involvement and psychological well-being? American Psychologist, 53(9), 10171031.
  • Lee, W., & Kuo, E. C. Y. (2002). Internet and displacement effects: children's media use and activities in Singapore. Journal of Computer-Mediated Communication, 7(2) .
  • Lenhart, A., Madden, M., & Hitlin, P. (2005). Teens and technology: Youth are leading the transition to a fully wired and mobile nation. Washington, D.C.: Pew Internet & American Life Project.
  • Lenhart, A., Madden, M., Macgill, A. R., & Smith, A. (2007). Teens and social media: The use of social media gains a greater foothold in teen life as they embrace the conversational nature of interactive online media. Washington, D.C.: Pew Internet & American Life Project.
  • Lenhart, A., Rainie, L., & Lewis, O. (2001). Teenage life online: The rise of the instant-message generation and the Internet's impact of friendships and family relationships. Washington, D.C.: Pew Internet & American Life Project.
  • Livingstone, S., & Bovill, M. (2001). Children and their changing media environment. Manhwah: Lawrence Erlbaum Associates.
  • Marsh, H. W., Hau, K., & Wen, Z. (2004). In search of golden rules: comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler's findings, Structural Equation Modeling, 11, 320341.
  • McKenna, K. Y. A., & Bargh, J. A. (1999). Causes and consequences of social interaction on the Internet. Media Psychology, 1, 249269.
  • 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(1), 931.
  • McQuail, D. (1987). Mass communication theory: An introduction (2nd Ed.). London: Sage.
  • Mesch, G. (2003). The family and the Internet: The Israeli case. Social Science Quarterly, 84(4), 10381050.
  • Mesch, G. (2006). Family relations and the Internet: Exploring a family boundaries approach. Journal of Family Communication, 6(2), 119138.
  • Muthen, L. K., & Muthen, B. O. (2006). Mplus User's Guide. Los Angeles, CA: Muthen & Muthen.
  • Neuman, S. (1991). Literacy in the television age. Norwood: Ablex Publishing.
  • Nie, N. H., Hillygus, D. S., & Erbring, L. (2002). Internet Use, interpersonal relations, and sociability: A time diary study. In B.Wellman & C.Haythornthwaite (Eds.), The Internet in Everyday Life (pp. 215243). Malden: Blackwell Publishers Ltd.
  • Orleans, M., & Laney, M. C. (2000). Children's computer use in the home: Isolation or sociation? Social Science Computer Review, 8(1), 5672.
  • Parks, M. R. (1996). Making friends in cyberspace. Journal of Communication, 46(1), 8097.
  • Peterson, J., & Zill, N. (1986). Marital disruption, parent-child relationships, and behavioral problems in children, Journal of Marriage and the Family, 48(2), 295307.
  • Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 124.
  • Roberts, D. F., Foehr, U. G., & Rideout, V. (2005). Generation M: Media in the lives of 8–18 year-olds. Washington, D.C.: The Henry J. Kaiser Family Foundation.
  • Roberts, D. F., Foehr, U. G., Rideout, V. J. & Brodie, M. (1999). Kids & Media @ the New Millennium. Washington, D.C.: The Henry J. Kaiser Family Foundation.
  • Shklovski, I., Kraut, R., & Rainie, L. (2004). The Internet and social participation: Contrasting cross-sectional and longitudinal analyses. Journal of Computer-Mediated Communication, 10(1).
  • Stritzke, W. Nguyen, A., & Durkin, K. (2004). Shyness and computer-mediated communication: A self-representational theory perspective. Media Psychology, 6, 122.
  • Suoninen, A. (2001). The role of media in peer group relations. In S.Livingstone & M.Bovill (Eds.), Children and changing media environment. Menhwah: Lawrence Erlbaum Associates.
  • Valentine, G., & Holloway, S. L. (2002). Cyberkids? Exploring children's identities and social networks in on-line and off-line worlds. Annals of the Association of American Geographers, 92(2), 302319.
  • Valkenburg, P., & Peter, J. (2005). Adolescents' online communication and their closeness to friends. Paper presented at the International Communication Association, New York.
  • Valkenburg, P. M., & Peter, J. (2007). Online communication and adolescent well-being: Testing the stimulation versus the displacement hypothesis. Journal of Computer-Mediated Communication, 12(4), article 2. http://jcmc.indiana.edu/vol12/issue4/valkenburg.html.
  • Vandewater, E. A. & Lee, S. J. (in press). Measuring children's media use in the digital age: Issues and challenges. American Behavioral Scientist.
  • Wartella, E., & Reeves, B. (1985). Historical trends in research on children and the media: 1900-1960. Journal of Communication, 35, 118133.
  • Wellman, B., Haase, A. Q., Witte, J., & Hampton, K. (2001). Does the Internet increase, decrease, or supplement social capital? American Behavioral Scientist, 45(3), 436455.
  • Zhu, B., Walter, S. D., Rosenbaum, P. L., Russell, D. J., & Raina, P. (2006). Structural equation and log-linear modeling: A comparison of methods in the analysis of a study on caregivers' health, BMC Medical Research Methodology, 6(49), http://www.biomedcentral.com/1471-2288/6/49.

About the Author

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion and Conclusion
  7. References
  8. About the Author

Sook-Jung Lee is a full-time lecturer in the Department of Mass Communication at Chung-Ang University. Her research focuses on social impact of Internet use, children's media use in a family context including parental mediation and developmental outcomes, and methodological issues in media research.

Address: 221 Heukseok-Dong, Dongjak-Gu, Seoul 156-756, Rep. of Korea.