A comparison of ofﬂine and
online friendship qualities at
different stages of relationship
Darius K.-S. Chan & Grand H.-L. Cheng
The Chinese University of Hong Kong
The present study was designed to compare ofﬂine and
online friendship qualities at different stages of relationship
development. The sample consisted of 162 Hong Kong
Internet users. They were asked to think of two friends, one
they knew through face-to-face interactions and one they
knew through the Internet, and then describe the qualities of
their ofﬂine and online friendships. Results revealed that
ofﬂine friendships involved more interdependence, breadth,
depth, code change, understanding, commitment, and net-
work convergence than online friendships. However,
although the qualities of both online and ofﬂine friendships
improved as the duration of the relationship increased, the
differences between the two types of friendships diminished
over time. Furthermore, contrary to the evidence typically
found for ofﬂine friendships, the qualities of cross-sex online
friendships were higher than that of same-sex online friend-
ship. These results suggest that the inﬂuence of the structural
and normative constraints typically found in face-to-face
interaction may be different in the online setting.
KEY WORDS: friendship qualities • Hong Kong • Internet users •
online versus ofﬂine friendships
Friendship is one of the most common types of interpersonal relationship.
It has long been a research focus for various disciplines and different
deﬁnitions of friendship can be found in the literature. For instance, Hartup
(1975) deﬁned friends as those ‘who spontaneously seek the company of
one another; furthermore, they seek proximity in the absence of strong
Journal of Social and Personal Relationships Copyright © 2004 SAGE Publications
(www.sagepublications.com), Vol. 21(3): 305–320. DOI: 10.1177/0265407504042834
All correspondence concerning this article should be addressed to Darius Chan, Psychology
Department, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, China [e-mail:
email@example.com]. Sandra Metts was the Action Editor for this article.
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 305
social pressures to do so’ (p. 11). According to Wright (1984), friendship is
‘a relationship involving voluntary or unconstrained interaction in which
participants respond to one another personally . . .’ (p. 119). Hays (1988)
deﬁned friendship as a ‘voluntary interdependence between two persons
over time, that is intended to facilitate social–emotional goals of the partici-
pants, and may involve varying types and degrees of companionship,
intimacy, affection, and mutual assistance’ (p. 395). All these deﬁnitions
imply that friendship evolves through voluntary interaction between two
persons over time. It typically develops through intimate, face-to-face
Recently, a new form of friendship, which we call online friendship, has
emerged. This type of friendship initiates and develops through computer-
mediated communication (CMC) in online social settings, such as chat-
rooms, newsgroups, and websites. Although the use of CMC has become
one of the most popular means of communication, only a handful of studies
have examined the characteristics of online relationships (see Wood &
Duck, 1995). In the present study, we focused on friendship, which was
found to be the most common type of online interpersonal relationship
(Parks & Roberts, 1998), and examined whether the qualities of online
friendships are comparable to ofﬂine friendships at different stages of
development. In addition, we also studied the differences, if any, between
same-sex and cross-sex online friendships, with respect to their ofﬂine
Current perspectives on online relationships
According to social presence theory (see Short, Williams, & Christie, 1976),
social presence refers to an individual’s feeling that other people engage in
personal communication and interaction. The degree of social presence
depends on the number of channels or codes available within the communi-
cation medium. CMC is typically characterized as low in social presence
because it typically allows little exchange of nonverbal cues. Similarly,
Sproull and Kiesler’s (1986) lack of social context cues hypothesis asserts
that CMC has a narrower bandwidth and less information richness than
face-to-face (FtF) interaction does.
In general, these two so-called ‘cues-ﬁltered-out’ approaches note that
the major difference between FtF interaction and CMC lies in the avail-
ability of social context cues, which are crucial to relationship development.
CMC is characterized as lacking spatial features, personal appearance, and
actors’ dynamic nonverbal cues such as facial expression, posture and
gesture (Kiesler, Siegel, & McGuire, 1984). In addition, it is likely to be
constrained by lower physical availability and frequency of exposure, which
are considered crucial for relationship development (Rawlins, 1994;
Wilmot, 1994). According to these arguments, close relationships can
hardly develop online (see also Lea & Spears, 1995). For instance, in a
study by Parks and Roberts (1998), respondents were requested to recall
and rate the qualities of one online and one ofﬂine relationship, respec-
tively. Their results showed that the overall friendship quality ratings for
306 Journal of Social and Personal Relationships 21(3)
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 306
online relationships were signiﬁcantly lower than those for ofﬂine relation-
ships in measures such as interdependence, understanding, commitment,
and network convergence.
However, the two ‘cues-ﬁltered-out’ approaches seem to ignore the
impact of the duration of relationship development on friendship quality.
Taking the social information processing perspective, Walther (1992)
suggested that the critical difference between FtF and CMC is the rate, but
not the capability. Although the limited bandwidth of CMC offers less total
information per exchange than does FtF exchange, relational communi-
cation of CMC can grow and become comparable to that of FtF, given
sufﬁcient time and message exchanges. Walther (1995) carried out an
experimental observation on task-related groups of college students and
compared the communication qualities of FtF and CMC groups. Speciﬁc-
ally, the author was interested in the message dimensions that indicate how
the participants deﬁned and regarded their relationships with team
members. The dimensions include immediacy/affection, similarity/depth,
composure/relaxation, formality, dominance, receptivity/trust, and task/
social orientation. Compared to FtF groups, in no case did the participants
interacting through CMC express less intimacy at the third meeting, which
took place in the ﬁfth week of the study.
Friendship development over time
While Walther’s (1995) study provided evidence supporting the social
information processing theory, we could not conclude from this study that
online friendships are comparable to ofﬂine friendships over time because
this study focused on measuring how the participants communicated
through CMC versus FtF interaction, but not friendship qualities. To
compare the development of online and ofﬂine friendships, we need to ask
Internet users to report their online and ofﬂine friendships and take
duration of friendship development into consideration.
Friendship development has received a considerable amount of attention
in the area of interpersonal relationships. Numerous conceptual and
empirical studies have been conducted to examine how friendship pro-
gresses. For instance, Knapp (1984) identiﬁed ﬁve stages of relationship
development, namely, initiating, experimenting, intensifying, integrating,
and bonding. Using a network analysis, Milardo (1982) found that college
students’ networks of mutual friends changed as a function of the degree
of intimacy across stages of friendship development. Hays (1985) examined
factors affecting same-sex friendship development of college students.
Using a longitudinal design, he found that those dyads that became close
friends after a 3-month period differed behaviorally and attitudinally from
dyads that did not progress. Notice that all these studies focus on friend-
ships developed through FtF interactions.
In the present study, we were interested in whether the qualities of
ofﬂine and online friendships are comparable at different stages of
development. We used the personal relationship scale developed by
Parks and Floyd (1996) to measure friendship qualities. Based on social
Chan & Cheng: Online friendships 307
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 307
psychological theories such as Altman and Taylor’s (1973) social
penetration theory and Kelley et al.’s work (1983) on close relationships,
Parks and Floyd generated a list of items for measuring seven dimensions
of interpersonal relationships: interdependence, breadth, depth, code
change, understanding, commitment, and network convergence.
According to Parks (1997; Parks & Floyd, 1996), breadth, depth, and
code change are related to the quality of communication between the two
persons. The variety of topics, activities, and communication channels
increases when a relationship develops over time. In addition, people tend
to reveal more important and personal information when their relationship
progresses. Relationship development is also characterized by communi-
cation code changes. For instance, partners may evolve specialized ways of
communicating such as personal idioms, which allow them to express
themselves more efﬁciently and reinforce their relational identity.
Understanding, interdependence, commitment, and network conver-
gence all refer to how the two parties perceive the relationship. In most
cases, a relationship develops over time as both parties come to depend on
each other more deeply, become more committed to the relationship, and
have a better understanding of the interaction (Parks, 1997). Also, network
convergence occurs as the two sides develop a common social circle. That
is, the partners introduce one another to each other’s friends and family.
These seven dimensions also capture the common qualities in relationships
that researchers typically deﬁne as friendship.
Based on the earlier discussion on friendship quality and duration of
development, we made the following predictions. First, according to Parks
and Roberts’s (1998) study, we expected that the overall qualities of ofﬂine
friendships would be higher than those of online friendships. Second, the
qualities of both online and ofﬂine friendships would improve over time.
Third, according to the social information processing approach, the differ-
ences between ofﬂine and online friendships were expected to become
smaller as the relationship progressed.
In terms of measuring friendship duration, previous longitudinal studies
on changes in friendship quality varied in the way of categorizing friend-
ship duration. For instance, Hays (1985) adopted a 3-week interval design
to capture friendship development over a semester, whereas Grifﬁn and
Sparks (1990) carried out a 4-year longitudinal study to explore the
development of friendship. In a recent longitudinal study on friendship
development conducted by Yamanaka (1998), participants were asked to
describe their friendship four times during a 3-month period. It is likely that
these differences in categorization were due to constraints faced by
individual researchers. It is also obvious that there is no consensus on how
to quantify duration in friendship development. In the present study, we
used a cross-sectional approach and divided our sample into three groups
of duration, based on the intent to form comparable group sizes.
308 Journal of Social and Personal Relationships 21(3)
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 308
Same-sex and cross-sex friendships
The issue of gender has long been a focus in the area of interpersonal
relationships. Empirical research has been conducted to examine the differ-
ences between same-sex and cross-sex friendships in the ofﬂine cases (e.g.,
Grifﬁn & Sparks, 1990; Roy, Benenson, & Lilly, 2000). It is generally
believed that, compared to same-sex cases, cross-sex ofﬂine friendships are
more difﬁcult to develop because they are constrained by both structural
and normative factors (Booth & Hess, 1974). Structural factors refer to the
lack of opportunities for males and females to meet and interact continu-
ously. For instance, cross-sex friendships in the workplace are difﬁcult to
develop because the concern for status is typically signiﬁcant. Normative
constraints refer to factors such as the social disapproval of the develop-
ment of intimate cross-sex relationships for married individuals. Along a
similar vein, O’Meara (1989) described four challenges to cross-sex friend-
ships, including confronting the issue of sexuality, determining the type of
emotional bond the pair will have, dealing with gender inequality in a
relationship that values equality, and presenting the relationship to the
It would be interesting to see if such differences in the ofﬂine cases can
be applied to online friendships. Parker and de Vries (1993) found that
both males and females reported almost twice as many same-sex ofﬂine as
cross-sex ofﬂine friendships. In contrast, Parks and Roberts (1998)
reported that most online friendships were cross-sex. Parks and Roberts
argued that the structural and normative constraints that inhibit the
growth of cross-sex ofﬂine friendships are limited in the online setting.
However, in their study, the potential differences in quality between online
same-sex and cross-sex relationships were not addressed. The second
objective of the current study was, thus, to investigate same-sex and cross-
sex online friendships, with reference to the ofﬂine cases. In addition to
the three predictions mentioned earlier, our fourth prediction was that the
differences in quality between same-sex and cross-sex friendships that
have been found in the ofﬂine cases would be less pronounced for the
Thirty-eight newsgroups, under the newsgroup name ‘hk,’ were randomly
selected. In each newsgroup, it was stated that a questionnaire studying online
friendships had been posted at a website, which was maintained by the authors.
Responses from a total of 162 Hong Kong Internet users, 71 males and 91
females, were collected. Their ages ranged from 16 to 29 years (M = 20.67,
SD = 2.12), with 89% of them between 18 and 23 years old.
According to the frequency distribution, we classiﬁed the 162 respondents
into three groups of friendship duration, 1–4 months, 5–12 months, and over 1
year, to form comparable group sizes. Table 1 provides a breakdown of the
cases. There were 52, 40, and 70 respondents in the three duration groups,
Chan & Cheng: Online friendships 309
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 309
respectively. In addition, 62 of the 162 cases were same-sex and 100 cases were
cross-sex for both ofﬂine and online friendships (see Table 1).
On the ﬁrst page of the website, we stated that participants must be Hong Kong
residents who had friendships that were made online. For online friendships,
we stated that their interaction with the online friend should solely take place
in virtual settings. Also, respondents were told to ﬁll out the questionnaire once
only. The questionnaire was put on the second page of the website. In addition
to their own age and sex, each respondent was asked to think of two friend-
ships they had – one was an ofﬂine ‘physical life’ or ‘real-life’ friendship, and
the other one initiating and developing online. The criteria were that these two
relationships had to be developed for similar durations. Respondents were also
asked to indicate the sex of their friends.
The short form of the questionnaire used by Parks and Floyd (1996) and Parks
and Roberts (1998) was used to measure friendship quality (see Table 2). Two
to three items were used to capture each of the factors of friendship quality:
1 Interdependence, which refers to the feeling of mutual dependence and
the degree to which the two parties inﬂuence one another, was captured
by items such as ‘The two of us depend on each other.’
2 Breadth, which refers to the variety of conversational topics shared
between the two parties, was captured by items such as ‘Our communi-
cation ranges over a wide variety of topics.’
3 Depth refers to the degree of self-disclosure and is the process-oriented
and communicative aspect of intimacy. This dimension was measured by
items such as ‘I feel I could conﬁde in this person about almost anything.’
4 Code change, which refers to the change in linguistic forms and cultural
codes used in a relationship, was measured by items such as ‘The two of
us use private signals that communicate in ways outsiders would not
5 Understanding refers to the agreement about which behaviors are
acceptable, the understanding of how each person’s actions can contri-
bute to the relationship, and the expectation about which responses each
is likely to have. A sample item is ‘I can accurately predict what this
person’s attitudes are.’
310 Journal of Social and Personal Relationships 21(3)
Number of same-sex and cross-sex friendships in each duration category
Duration of development Total
1–4 months 5–12 months Over 1 year
Same-sex 20 16 26 62
Cross-sex 32 24 44 100
Total 52 40 70 162
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 310
6 Commitment, which refers to the expectation that a relationship will
continue and the feeling that a relationship ought to continue, was
measured by items such as ‘This relationship is very important to me.’
7 Network convergence, which refers to the overlapping of social networks
of the two parties, was captured by items such as ‘We have introduced
each other to members of each other’s circle of friends and family.’
Chan & Cheng: Online friendships 311
Items included in the questionnaire (a short form of the questionnaire used by
Parks & Floyd, 1996)
Interdependence The two of us depend on each other
= .65; We often inﬂuence each other’s feelings toward the issues we’re
=.63) dealing with
The two of us have little inﬂuence on each other’s thoughts (R)
Breadth Our communication is limited to just a few speciﬁc topics (R)
= .76; Our communication ranges over a wide variety of topics
Depth I usually tell this person exactly how I feel
= .65; I feel I could conﬁde in this person about almost anything
=.77) I would never tell this person anything intimate or personal
about myself (R)
Code Change We have developed the ability to ‘read between the lines’ of
= .63; each other’s messages to ﬁgure out what is really on each
=.79) other’s mind
The two of us use private signals that communicate in ways
outsiders would not understand
We have special nicknames that we just use with each other
Understanding I can accurately predict what this person’s attitudes are
= .68; I do not know this person very well (R)
Commitment This relationship is very important to me
= .73; I would make a great effort to maintain my relationship with
=.72) this person
I do not expect this relationship to last very long (R)
Network Convergence We have introduced each other to members of each other’s
= .63; circle of friends and family *
=.61) This person and I do not know any of the same people (R)
represent the Cronbach’s alpha of the measures for ofﬂine and online
friendships, respectively. (R) indicates that the score was reversed.
* This item was originally worded, ‘We have introduced (face-to-face or otherwise) each other
to members of each other’s circle of friends and family’, in Parks and Floyd (1996). Because our
respondents were asked to recall an online friendship in which their interaction with the online
friend should take place solely in the online setting, we deleted the phrase ‘face-to-face or
otherwise’ in our questionnaire to avoid causing any confusion.
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 311
A total of 18 items were used. They were all presented in Chinese, anchored
by seven-point scales in which 1 represents strongly disagree and 7 represents
strongly agree. Reliabilities, measured by Cronbach’s alpha, ranged between
.61 and .83 (Table 2).
A 2 (Friendship Type: ofﬂine vs. online) 3 (Duration: 1–4 months vs. 5–12
months vs. over 1 year) 2 (Gender Composition: same-sex vs. cross-sex
friendship) MANOVA test was performed. Friendship Type was the within-
subjects factor, and Duration and Gender Composition were between-subject
factors. To recap, the Duration factor reﬂects how long the relationship had
developed, and Gender Composition examines the differences between same-
sex and cross-sex friendships. Measures of interdependence, breadth, depth,
code change, understanding, commitment, and network convergence were the
With the use of Wilks’ lambda criterion, the MANOVA test showed signiﬁcant
main effects for Friendship Type, = .37, F(7,150) = 37.10, p <.001; Duration,
= .72, F(14,300) = 3.86, p <.001; and Gender Composition, =.86,
F(7,150) = 3.42, p < .01. However, all these main effects were qualiﬁed by
signiﬁcant interaction effects between Friendship Type by Duration, =.81,
F(14,300) = 2.43, p <.01, Friendship Type by Gender Composition, =.75,
F(7,150) = 7.02, p <.001, and Duration by Gender Composition, =.82,
F(14,300) = 2.30, p < .01.
In terms of the Friendship Type main effect, univariate test results show that
the quality of ofﬂine friendships was higher than that of online friendships on
all seven dependent measures (see Table 3): Interdependence, F(1,156) = 19.84,
p < .001; Breadth, F(1,156) = 35.49, p < .001; Depth, F(1,156) = 94.13, p < .001;
Code Change, F(1,156) = 75.23, p < .001; Understanding, F(1,156) = 35.25,
p <.001; Commitment, F(1,156) = 234.15, p < .001; and Network Convergence,
F(1,156) = 88.29, p <.001. These results support our ﬁrst prediction.
312 Journal of Social and Personal Relationships 21(3)
Friendship Quality Ratings for Ofﬂine and Online Friendships
Dimension Ofﬂine Online
Interdependence*** 4.09 1.39 3.60 1.30
Breadth*** 5.31 1.44 4.75 1.61
Depth*** 5.21 1.08 4.51 1.30
Code Change*** 4.00 1.22 3.22 1.43
Understanding*** 4.55 1.30 4.01 1.35
Commitment*** 5.66 1.21 4.16 1.46
Network Convergence*** 4.27 1.61 3.04 1.66
*** p < .001.
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 312
Univariate test results also showed a signiﬁcant main effect of Duration on
Breadth, F(2,156) = 8.06, p < .001; Depth, F(2,156) = 4.05, p < .05; Code
Change, F(2,156) = 19.84, p < .001; Understanding, F(2,156) = 7.97, p <.01;
Commitment, F(2,156) = 6.73, p < .01; and Network Convergence,
F(2,156) = 8.35, p < .001 (see Table 4). Consistent with our second hypothesis,
friendship qualities improved as the relationship developed over time.
In terms of the Gender Composition main effect, same-sex friendships were
generally of higher quality than cross-sex friendships (see Table 5). Univariate
test results revealed a signiﬁcant effect on Interdependence, F(1,156) = 5.76,
p < .05; Breadth, F(1,156) = 6.89, p < .05; and Network Convergence
F(1,156) = 9.23, p <.01.
However, these main effects were qualiﬁed by three sets of interaction
effects. A signiﬁcant interaction between Friendship Type and Duration was
Chan & Cheng: Online friendships 313
Friendship quality ratings as a function of duration of development
Duration of Development
1–4 months 5–12 months Over 1 year
——————— ——————— ———————
MSD MSD MSD
Interdependence 3.68 1.17 3.74 1.33 4.04 .96
Breadth*** 4.45 1.56 4.96 .99 5.50 1.05
Depth* 4.61 1.15 4.71 1.17 5.13 .68
Code Change*** 3.01 .74 3.49 .85 4.12 1.18
Understanding** 3.82 1.41 4.16 .92 4.69 .89
Commitment** 4.54 1.16 4.73 1.07 5.29 .99
Network Convergence*** 3.13 1.21 3.64 1.28 4.06 1.37
* p < .05; ** p < .01; *** p < .001.
Friendship quality ratings as a function of gender composition
Interdependence* 4.13 .99 3.68 1.19
Breadth* 5.32 .86 4.86 1.48
Depth 4.89 .72 4.84 1.15
Code Change 3.57 .90 3.64 1.19
Understanding 4.39 .90 4.21 1.28
Commitment 4.97 .78 4.88 1.28
Network Convergence** 4.08 1.12 3.40 1.42
* p < .05; ** p < .01.
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 313
found on Breadth, F(2,156) = 4.42, p < .05; Depth, F(2,156) = 4.32, p < .05; Code
Change, F(2,156) = 4.24, p <.05; Understanding, F(2,156) = 4.60, p < .05;
Commitment F(2,156) = 6.99, p < .01; and Network Convergence
F(2,156) = 4.23, p < .05. The patterns of interaction were similar across the six
314 Journal of Social and Personal Relationships 21(3)
Means on breadth, depth, and code change for ofﬂine and online friendships
Means on understanding, commitment, and network convergence for ofﬂine
and online friendships across duration.
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 314
measures (see Figures 1 and 2). Speciﬁcally, although both ofﬂine and online
friendships developed over time, the differences between the two types of
friendships increased from the beginning to about a year of relationship
development, and then diminished. In other words, these two types of
Chan & Cheng: Online friendships 315
Means on breadth and code change for same-sex and cross-sex ofﬂine and
Means on depth and commitment for same-sex and cross-sex ofﬂine and
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 315
friendship qualities tend to converge over time. Consistent with Walther
(1995), these results provide support for our third prediction.
Regarding the effects of Gender Composition, a signiﬁcant interaction
between Friendship Type and Gender Composition was found on Breadth,
F(1,156) = 10.13, p < .01; Depth, F(1,156) = 39.12, p < .001; Code Change,
F(1,156) = 22.21, p < .001; and Commitment, F(1,156) = 21.72, p < .001. The
interaction effects were similar across the four measures (see Figures 3 and 4).
Speciﬁcally, the quality of cross-sex friendships was higher than that of same-
sex friendships in the online cases, whereas same-sex ofﬂine friendships were
of higher quality than cross-sex ofﬂine friendships. This ﬁnding is in line with
our fourth prediction.
Univariate test results also revealed one signiﬁcant Duration by Gender
Composition interaction on one dependent variable, Breadth, F(2,156) = 9.02,
p < .001. At the early stage of relationship development, communication within
same-sex friendships involved more breadth than that within cross-sex dyads.
However, this difference diminished as the relationship developed.
Consistent with the general patterns of results reported by Parks and
Roberts (1998), the signiﬁcant main effect of friendship type indicates that
our respondents also rated the quality of ofﬂine friendships as higher than
that of their online friendships. Higher degrees of breadth, depth, code
change, understanding, interdependence, commitment, and network
convergence were found in ofﬂine friendships than online friendships. This
piece of evidence seems to provide some support for the generalizability of
this phenomenon to a Chinese sample.
The ﬁnding that the quality of ofﬂine friendships was higher than that of
online friendships seems to conﬁrm the prediction of the ‘cues-ﬁltered-out’
approaches. For instance, both social presence theory and the lack of social
context cues hypothesis suggest that the limitations of communication
channels and the lack in social context cues in CMC make close relation-
ships more difﬁcult to develop online. However, these perspectives do not
consider the duration of relationship development.
Development of online friendship over time
Our results elucidate that the difference in quality between online and
ofﬂine friendships is moderated by the duration of the relationship. Ofﬂine
and online friendships seem to grow in different ways. According to our
results, the quality of ofﬂine and online friendships diverged from the
beginning to about a year of relationship development and then converged
for relationships that lasted for more than a year. Speciﬁcally, the differ-
ence in quality between online and ofﬂine friendships was mainly found for
relationships that lasted for up to about one year. For relationships that
lasted for more than a year, the differences between the two types of friend-
ship were minimal.
According to Knapp (1984), supportiveness, positiveness, and equality
316 Journal of Social and Personal Relationships 21(3)
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 316
are limited at the beginning of relationship development. Because people
do not know each other well, they may not be able to empathize or orient
to the needs of the other. In addition, the relationship is probably viewed
as too temporary to be worth exerting signiﬁcant effort to develop and
maintain. With narrower bandwidth and less information richness, the
quality of online friendships is even lower than that of ofﬂine friendships
in the ﬁrst few months of relationship development.
During the development of relationships, individuals think over what has
happened and consider the desirability of continuing the relationship
(Wood, 1982). As mentioned earlier, compared to ofﬂine friendships,
online friendships are more likely to be constrained by factors such as lower
frequency of exposure (Wilmot, 1994). Online friendships may thus be
relatively more vulnerable and develop more slowly at the initial stage.
Therefore, the differences between the two types of friendships were found
to increase until about a year of relationship development. After passing a
critical period (about 6 months to 1 year in our sample), online friendships
grow quickly and the difference between these two types of relationships
decrease after about a year of development. Perhaps this critical period is
what Wood (1982) called the revising stage of relationship development.
The general trend that both online and ofﬂine friendships develop and
their qualities converge over time provides support for the social infor-
mation processing theory. Speciﬁcally, it suggests that while CMC offers
less total information per exchange than does FtF exchange, this problem
can be alleviated by allowing longer and more frequent interactions over
time (Walther, 1992). ‘Given sufﬁcient time and message exchanges for
interpersonal impression formation and relational development to accrue,
and all other things being equal, relational (communication) in later
periods of CMC and FtF communication will be the same’ (p. 69). In other
words, relationships developed online can also become personal, if given
time, and relational partners can feel as intimate as they do in FtF inter-
Same-sex and cross-sex online friendships
With respect to the effect of gender composition, our results indicate that
the quality of same-sex ofﬂine friendships was higher than that of cross-sex
ofﬂine friendships. In contrast, cross-sex online friendships were of higher
quality than same-sex online friendships.
While the results for the ofﬂine cases are consistent with past ﬁndings
that cross-sex ofﬂine friendships are difﬁcult to develop (e.g., Hacker, 1981;
Monsour, Harris, & Kurtzweil, 1994; Rose, 1985), our results also reveal
that, in online settings, cross-sex friendships seem to be less difﬁcult to
develop. As mentioned before, the Internet removes the structural and
normative constraints existing in ofﬂine settings that impede the develop-
ment of cross-sex ofﬂine friendships (Booth & Hess, 1974; see also Parks
& Roberts, 1998). On the contrary, status differentials, such as age, race,
and social class, which may inhibit initial interaction, are less readily
apparent in CMC. Social cues, such as physical appearance and nonverbal
Chan & Cheng: Online friendships 317
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 317
information, are also limited in online interaction. In addition, the Internet
allows physical absence and anonymity and thus reduces the perception of
risk that exists in cross-sex FtF interaction, such as unwanted sexual
advances. It seems that the Internet provides alternative social venues for
men and women to interact continuously, and may be perceived as a safe
environment for people to initiate and develop cross-sex friendships with
a lower degree of reticence and reservation.
Limitations and further directions
As Parks and Roberts (1998) pointed out, individuals involved in online
relationships might go on to use other communication media as the
relationship develops, such as telephone, letters, or even face-to-face
communication. One alternative explanation of the duration effect is that
our respondents might in fact engage in face-to-face interactions with their
online friends, resulting in the comparison between online and ofﬂine
friendships of more than a year being meaningless. Even though our
instructions stated explicitly that the online friendships recalled should
refer to those taking place solely in the online settings, we obviously did
not have any control over the modes of communication of our respondents’
online friendships. Future studies may consider using an experimental
Our ﬁndings also provide some implications for cross-cultural communi-
cation research. Numerous studies have been conducted to examine the
impact of culture on interpersonal communication. Our results, however,
seem to provide evidence of the generalizability of ﬁndings in the West to
a Chinese sample. Speciﬁcally, the friendship type main effect and the
patterns of online friendship development were all consistent with those
found in the West. This partial replication seems to suggest that the cultural
differences in communication typically found in the ofﬂine setting, such as
Ting-Toomey’s (1988) research on face work, may be less pronounced in
the online setting. In other words, relationships developed through the
Internet may in fact be less likely to be subject to the cultural inﬂuences
reported in the literature, as these cultural differences are predominantly
manifested in ofﬂine, face-to-face interactions. Cross-cultural studies
should be conducted to examine this speculation.
Altman, I., & Taylor, D. (1973). Social penetration: The development of interpersonal relation-
ships. New York: Holt, Rinehart, and Winston.
Booth, A., & Hess, E. (1974). Cross-sex friendship. Journal of Marriage and the Family, 36,
Grifﬁn, E., & Sparks, G. G. (1990). Friends forever: A longitudinal exploration of intimacy in
same-sex friends and platonic pairs. Journal of Social and Personal Relationships, 7, 29–46.
Hacker, H. (1981). Blabbermouths and clams: Sex differences in self-disclosure in same-sex
and cross-sex friendship dyads. Psychology of Women Quarterly, 5, 385–401.
318 Journal of Social and Personal Relationships 21(3)
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 318
Hartup, W. W. (1975). The origins of friendships. In M. Lewis & L. A. Rosenblum (Eds.),
Friendship and peer relations (pp. 11–26). New York: Wiley.
Hays, R. B. (1985). A longitudinal study of friendship development. Journal of Personality and
Social Psychology, 48, 909–924.
Hays, R. B. (1988). Friendship. In S. W. Duck (Ed.), Handbook of personal relationships:
Theory, research, and interventions (pp. 391–408). New York: Wiley.
Kelley, H. H., Berscheid, E., Christensen, A. C., Harvey, J. H., Huston, T. L., Levinger, G.,
McClintock, E., Peplau, L. A., & Peterson, D. R. (1983). Close relationships. New York: W.
Knapp, M. L. (1984). Interpersonal communication and human relationships (3rd ed.). Boston:
Allyn and Bacon.
Kiesler, S., Siegel, J., & McGuire, T. W. (1984). Social psychological aspects of computer-
mediated communication. American Psychologist, 39, 1123–1134.
Lea, M., & Spears, R. (1995). Love at ﬁrst byte? Building personal relationships over
computer networks. In J. T. Wood & S. W. Duck (Eds.), Understudied relationships: Off the
beaten track (pp. 197–233). Newbury Park, CA: Sage.
Milardo, R. M. (1982). Friendship networks in developing relationships: Converging and
diverging social environments. Social Psychology Quarterly, 45, 162–172.
Monsour, M., Harris, B., & Kurtzweil, N. (1994). Challenges confronting cross-sex friendships:
‘Much ado about nothing?’ Sex Roles, 31, 55–77.
O’Meara, J. D. (1989). Cross-sex friendship: Four basic challenges of an ignored relationship.
Sex Roles, 21, 525–543.
Parker, S., & de Vries, B. (1993). Patterns of friendship for women and men in same and cross-
sex relationships. Journal of Social and Personal Relationships, 10, 617–626.
Parks, M. R. (1997). Communication networks and relationship life cycles. In S. Duck, R. S.
L. Mills, W. Ickes, K. Dindia, R. M. Milardo, & B. R. Sarason (Eds.), Handbook of personal
relationships (2nd ed., pp. 351–372). Chichester, UK: Wiley.
Parks, M. R., & Floyd, K. (1996). Making friends in cyberspace. Journal of Communication,
Parks, M. R., & Roberts, L. D. (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.
Rawlins, W. K. (1994). Being there and growing apart: Sustaining friendships during
adulthood. In D. J. Canary & L. Stafford (Eds.), Communication and relational maintenance
(pp. 275–294). San Diego, CA: Academic Press.
Rose, S. M. (1985). Same- and cross-sex friendships and the psychology of homosociality. Sex
Roles, 12, 63–74.
Roy, R., Benenson, J. F., & Lilly, F. (2000). Beyond intimacy: Conceptualizing sex differences
in same-sex friendships. Journal of Psychology, 134(1), 93–101.
Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecommunication.
Sproull, L., & Kiesler, S. (1986). Reducing social context cues: Electronic mail in organiz-
ational communication. Management Science, 32, 1492–1512.
Ting-Toomey, S. (1988). Inter-cultural conﬂict style: A face-negotiation theory. In Y. Y. Kim
& W. B. Gudykunst (Eds.), Theory in intercultural communication (pp. 213–235). Newbury
Park, CA: Sage.
Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: A relational
perspective. Communication Research, 19, 52–90.
Walther, J. B. (1995). Relational aspects of computer-mediated communication: Experimental
observations over time. Organizational Science, 6(2), 186–203.
Wilmot, W. W. (1994). Relationship rejuvenation. In D. J. Canary & L. Stafford (Eds.),
Communication and relational maintenance (pp. 255–273). San Diego, CA: Academic
Chan & Cheng: Online friendships 319
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 319
Wood, J. T. (1982). Human communication: A symbolic interactionist perspective. New York:
Holt, Rinehart, and Winston.
Wood, J. T., & Duck, S. (Eds.). (1995). Under-studied relationships: Off the beaten track.
Thousand Oaks, CA: Sage.
Wright, P. H. (1984). Self-referent motivation and the intrinsic quality of friendship. Journal
of Social and Personal Relationships, 1, 115–130.
Yamanaka, K. (1998). A case study of friendship development among college students.
Japanese Journal of Social Psychology, 13(2), 93–102.
320 Journal of Social and Personal Relationships 21(3)
02 Chan 042834 (dm/d) 19/4/04 9:58 am Page 320