Content uploaded by David C Giles
Author content
All content in this area was uploaded by David C Giles on Dec 01, 2015
Content may be subject to copyright.
Content uploaded by David C Giles
Author content
All content in this area was uploaded by David C Giles on May 27, 2015
Content may be subject to copyright.
Requests for reprints should be sent to David C. Giles, School of Health and Social
Sciences, Coventry University, Coventry CV1 5FB, UK. E-mail:
D.Giles@coventry.ac.uk
Parasocial Interaction: A Review
of the Literature and a Model
for Future Research
David C. Giles
School of Health & Social Sciences
Coventry University, UK
This paper considers the phenomenon of parasocial interaction (PSI) used by
media researchers to describe the relationship between media users and
media figures (from celebrities to fictional characters). Although the concept
has been used consistently across the past two decades in media research, it
is argued here that it has not been sufficiently developed at a theoretical level
to be taken up by psychologists. A number of key issues have not been
addressed: firstly, how PSI might, as its originators put it, be “integrated into
the matrix of usual social activity” (Horton & Wohl, 1956); secondly, how PSI
might vary according to different types of media figure; and thirdly, what
processes over time and media use bind user and figure into a “parasocial
relationship.” In this paper the existing literature on PSI is extensively
reviewed, and an original model of PSI is developed for use in future social
psychological research, which places PSI within the realm of ordinary social
interaction and suggests ways in which different media use and different types
of media figure interact to produce different styles of relationship. Finally,
some applications of more detailed research into PSI are suggested.
The concept of parasocial interaction has become fairly well established in the
media and communication literature in the four decades since its first appearance
in a paper by Horton and Wohl (1956). This article discusses ways in which the
interaction between users of mass media and representations of humans
appearing in the media (“media figures,” such as presenters, actors, and
celebrities) can produce a form of parasocial relationship, to which the user
responds as though in a typical social relationship. The authors conclude by
MEDIAPSYCHOLOGY, 4, 279–305.
Copyright © 2002, Lawrence Erlbaum Associates, Inc.
THEORETICAL INTEGRATION ESSAY
2 8 0 GILES
suggesting that social psychologists “learn in detail how these parasocial
interactions are integrated into the matrix of usual social activity” (p. 225). For
various reasons, this suggestion has yet to be taken up in earnest, but in the
course of this article I suggest some ways in which parasocial interaction may be
viewed as “usual social activity.”
Although a profoundly psychological topic, the implications of parasocial
interaction have received little consideration from psychologists, and it was not
until the 1980s that Alan Rubin and other researchers began to develop the
concept extensively within the field of communication science. However, the
concept of parasocial interaction, and detailed examination of the behavioral
phenomena that it seeks to explain, have considerable potential for developing
psychological theory. In keeping with the call by Reeves and Anderson (1991)
for the study of media to enrich psychology, this article is an attempt to extract
key findings from the media and communication literature, and to construct a
model for use in psychological research.
PARASOCIAL INTERACTION (PSI)
AS A MEDIA PHENOMENON
The literature review begins by charting the development of PSI as a research
concept in mass communication literature, although its origins lie in the field of
psychiatry.
Following the Horton and Wohl (1956) paper, there was little significant
interest in PSI until the advent of the uses and gratifications approach to mass
communication research in the early 1970s (McQuail, Blumler, & Brown, 1972;
Rosengren & Windahl, 1972). In a study of a British television audience, McQuail
et al. (1972) found many of the phenomena described by Horton and Wo h l
appearing in viewer responses to early soap opera. Following a car crash in
Coronation Street, a viewer commented, “You feel as if they had been in a real
road accident and you’d like to do something for them” (p. 157). The authors
identified two essential functions of PSI: companionship and personal identity.
Soap characters frequently reminded viewers of people they knew, and viewers
used characters’situations and behavior as ways of understanding their own lives.
In a typology of audience–media figure relations, Rosengren and Wi n d a h l
(1972) argued that PSI could be identified when a viewer interacted with a media
figure, but did not identify with the figure. This is an important distinction, because
identification has a longer history than PSI, deriving initially from psychoanalytic
t h e o r y. For Rosengren and Windahl, like McQuail et al., PSI’s most important
function was as a source of alternative companionship, resulting from “deficiencies”
in social life and dependency on television (i.e., as compensation for loneliness).
Subsequent Scandinavian research in the 1970s argued in favor of broadening
audience–media figure research to include PSI, identification and “capture”
(where the viewer both interacts and identifies with the figure), integrating these
functions into more general “media interaction” (Nordlund, 1978). The idea of
PSI as a functional alternative to social interaction was not supported by data
collected from Swedish adolescents, marking something of a hiatus in
Scandinavian PSI research (Rosengren, Windahl, Hakansson, & Johnsson-
Smaragdi, 1976).
The next major development in PSI research took place in North America,
with Mark Levy’s (1979) important study of older adults and local television
news. Levy conducted a number of focus group interviews concerning, among
other things, viewers’PSI with newscasters, and used this data to construct a 42-
item psychometric scale to measure strength of PSI with local newscasters. This
scale was correlated with a number of demographic variables in a sample of
viewers in a broader age band. Among the items most strongly agreed with were,
“I compare my own ideas with those of newscasters” and “When the newscasters
joke around with each other it makes the programme easier to watch.” Of the
demographic variables, education was strongly negatively correlated with PSI (r
= –.51), but other variables were not significantly related when education was
taken into account.
Most subsequent PSI research has been conducted in the psychometric
tradition of uses and gratifications research, where PSI has been entered alongside
other behavioral variables into models predicting media use. Most of these studies
have operationally defined PSI by using variations on a scale devised by A. M.
Rubin, Perse, and Powell (1985), henceforth referred to as the PSI scale. T h e
original 29-item instrument was constructed using a number of Levy’s (1979)
items and other items based on previous PSI research. Following its use with an
adult sample (n= 329), nine redundant items were eliminated, resulting in
acceptable internal reliability (α= .93). Asingle factor solution best described the
data, with the factor accounting for 45.7% of the variance. This study continued
the theme of measuring PSI with newscasters and replicated the earlier
Scandinavian findings that PSI did not seem to be associated with loneliness.
Later studies have used variations on the 20-item scale to measure PSI with soap
characters (A. M. Rubin & Perse, 1987), comedians (Auter, 1992), T V s h o p p i n g
hosts (Grant, Guthrie, & Ball-Rokeach, 1991), and favorite television personalities
of any type (R. B. Rubin & McHugh, 1987; Tu r n e r, 1993). A. M. Rubin and Perse
(1987) reduced the scale further, to 10 items, and the short version was found to
have high internal reliability, and high correlation with the 20-item scale. The 10-
item version has been used in a number of subsequent studies, notably by Perse and
R. B. Rubin (1989) and Conway and A. M. Rubin (1991). The latter study found
PA R A S O C I A LI N T E R A C T I O N 2 8 1
2 8 2 GILES
that PSI was associated with most viewing motives, and was a better predictor of
television use than many other behavioral measures; indeed, the authors argued that
PSI may be a more important viewing motivation than program content itself.
Studies using the PSI scale have found that perceived realism, and attraction
to the media figure, were highly correlated with the measure (A. M. Rubin et al.,
1985; A. M. Rubin & Perse, 1987, R. B. Rubin & McHugh, 1987), which
suggests that media users evaluate media figures along similar criteria to people
they encounter in the flesh. Television dependency and the amount of time spent
watching television have also been found to correlate with PSI (Gleich, 1997;
Grant, Guthrie, & Ball-Rokeach, 1991; A. M. Rubin et al., 1985; Vorderer,
1996;), which is important from a uses and gratifications perspective, and also
for cultivation theory (Cohen, 1999). However for many variables, particularly
demographic variables, there is less consistency across studies. For example, PSI
scores have not always correlated with viewing alone, and there are mixed
findings with respect to age, gender, and education level.
In the majority of studies the data is best described by a single factor solution,
often accounting for over 30% of total variance, but an interesting exception is
the study by Gleich (1997), who argued that PSI may not be a unitary concept.
This study was carried out using a German sample (the majority of these studies
have been conducted in North America), and three factors seem to account for
most of the variance.
F i r s t l y , companionship describes many of the items: These are perhaps those
aspects of PSI that gratify a need for social interaction, for example, “I feel as if I
am part of their group.” The second factor, person–program interaction, concerns
items that are directly related to programme content, for example, “If X appeared
on another program I would watch it.” The third factor, empathetic interaction,
refers to items that imply some degree of behavioral or affective response, such
as verbally addressing the media figure, or feeling embarrassed when they make
a mistake. Gleich found that the scores for these items differentiated the German
sample from the U.S. sample of A. M. Rubin et al. (1985). This may be a
reflection of broader cultural differences between these audiences, or variations in
broadcasting tradition, with U.S. presenters and anchors perhaps addressing the
camera in ways more likely to promote affective viewer responses.
Another German study (Vo r d e r e r, 1996) also found that a three-factor solution
best described the data obtained from a modified version of the PSI. The item
loadings in this analysis were similar to the Gleich study, except that the three
factors were interpreted, respectively, as “quasisocial relationship” (items relating
to interaction that was similar to social interaction), unique media relationships
(items peculiar to mediated interaction) and “star relationships” (items that were
more indicative of relationships with celebrities, such as “I find X attractive”).
PA R A S O C I A LI N T E R A C T I O N 2 8 3
Amore recent challenge to the notion that PSI represents a unitary concept
has been made by Auter and Palmgreen (2000), who have developed a
multidimensional measure of PSI, the Audience–Persona Interaction (API)
Scale. Forty seven initial items were derived from an open-ended questionnaire,
and these were reduced to 22, with acceptable internal reliability (α= .84). Data
collected from a high school sample were best described by a four-factor
solution, the factors being interpreted as identification with a favorite character;
interest in a favorite character; interaction with a group of favorite characters
(e.g., a sitcom family); and a favorite character’s problem-solving abilities. This
measure differs notably from previous PSI scales in considering “group
identification” as a characteristic of PSI, measured by items such as “The
characters’interactions are similar to mine with my friends.” Overall API scores
correlated with total viewing time, dependency on television, and perception of
TV as reality.
In addition to Likert-type scales, other PSI studies have used semantic
differential scales (Auter & Davis, 1991; Koenig & Lessan, 1985), and repertory
grids (Gleich, 1997). Finally, a small number of studies have used different
qualitative approaches (Alperstein, 1991; Papa et al., 2000; Sood & Rogers,
2000).
PARASOCIAL INTERACTION
AS A PSYCHOLOGICAL PHENOMENON
So far, the literature in this review has been drawn almost exclusively from
media and communication research, the one striking exception being the original
Horton and Wohl (1956) paper. So far, psychologists have shown little interest in
the concept, yet it raises many important questions about social psychology, and
the nature of relationships that are problematic for existing theories in those
fields. Furthermore, there is an increasing awareness of the importance of media
in psychological research (Kirschner & Kirschner, 1997; Livingstone, 1998).
The PSI research to date, although not conducted under the explicit banner of
psychology, contains many issues of considerable interest to psychologists, both
those interested in the psychology of the media, and those working in the broader
fields of social psychology and relationship theory.
Similarities Between PSI and Social Interaction
Anumber of studies have addressed a key psychological issue for PSI; namely,
how similar are parasocial relations to ordinary social relations? It seems likely
that, once we have made a person judgement about a media figure, or attributed
2 8 4 GILES
person characteristics to that figure (e.g., an anthropomorphized cartoon animal),
then we will subsequently respond to that figure “as if” it occupies our physical
space, thereby becoming incorporated into our social network. If this is the case,
then we might expect to identify similar psychological processes underpinning
the course of parasocial relationships to those found in face-to-face relationships.
R. B. Rubin and McHugh (1987) examined the importance of social, task, and
physical attraction for the development of parasocial relationships with favorite
television performers in an undergraduate sample. They found that social
attraction (i.e., the media figure could be a friend) was a more important
motivating factor in developing a parasocial relationship than physical attraction.
Perse and R. B. Rubin (1989) applied uncertainty-reduction theory to the
development of parasocial relationships and found that higher levels of
attributional confidence were associated with greater “parasocial complexity”
measured by descriptions of favorite and disliked soap opera characters. Turner
(1993) examined PSI from the perspective of homophily and found that
similarity was an important factor in the strength of the parasocial relationship,
particularly in relation to attitudes, appearance, and background. However these
effects varied according to the type of media figure.
More recently, Gleich (1996) has compared ratings of relationship quality
made toward media figures with those made toward friends and neighbors. On
most dimensions (confidence, proximity, idealism, strength of character),
respondents’best friends were evaluated considerably higher than favorite media
figures, but the ratings for a “guter nachbar” (good neighbor) and favorite media
figure were much closer than those for friends. Indeed, on some dimensions
(passion and sociability), favorite media figures were rated more highly than
good neighbors.
Generally, it seems that many attributes of PSI are similar to those of social
interaction. This is not surprising; as A. M .Rubin and Perse (1987) argued, PSI
may arise from an altruistic human instinct to form attachments with others, at
no matter how remote a distance. From an evolutionary perspective, Reeves and
Nass (1996) argued that this is an example of the “media equation” where social
responses are automatically elicited by any cues that are related to human
characteristics (such as the appearance of a human face on a screen).
Despite the noted parallels between social and parasocial interaction, the
status of parasocial relationships as relationships is doubtful if using long-
established definitions of relationships, such as Hinde (1979), who argued that “a
relationship exists only when the probable course of future interactions between
the participants differs from that between strangers” (p. 16). In a parasocial
relationship the media user is a “stranger” throughout. In order to develop
research into parasocial relationships, therefore, it may be necessary to redefine
relationships in general.
S u r p r i s i n g l y, key social psychological theories of interaction and
relationships have only been adapted slowly to account for technological
influences on social behavior. Some research has been conducted on computer-
mediated relationships (McKenna & Bargh, 2000; Parks & Floyd, 1996; Parks &
Roberts, 1998). However, this work has been less concerned with challenging
fundamental notions of relationships than with exploring the internet as a means
of facilitating traditional relationships (e.g., use of internet dating sites). Other
researchers have praised the internet for its “liberating” potential (e.g., Turkle,
1995), where one can try out different identities, freed from physical constraints
of age, gender, or ethnicity, although this cuts both ways, and we can also be
misled and exploited by other users similarly masquerading under “false”
identities (Wallace, 1999). The most significant challenge to the study of
relationships as face-to-face phenomena to date has come from Rohlfing (1995)
and Lea and Spears (1995), whose work on long-distance relationships and
cyberfriends raised questions about the need for proximity to form emotionally
close relationships, and about the role of fantasy in relationship development.
Can on-line interaction be regarded as functionally equivalent to face-to-face
social interaction? Papachrissi and A. M. Rubin (2000) investigated the
possibility that individuals who find face-to-face interaction unrewarding use the
internet as a functional alternative. They found some support for this hypothesis,
in that those internet users who found interpersonal communication most
rewarding were more likely to use the net for information or entertainment,
whereas those who were most socially anxious were more likely to use on-line
newsgroups and chat rooms.
This finding would seem to conflict with some of the earlier research that failed
to find support for the argument that PSI was a functional alternative to social
interaction (Rosengren et al., 1976; A. M. Rubin et al., 1985). However, on-line
communication and parasocial interaction are only alike in that the interactants are
distant from one another. As Wellman (1996) argued, an internet user group is
e ffectively a social network, especially when communication is synchronous (e.g.,
chat rooms). In such groups, interaction is reciprocal, at least until the group
reaches a particular size (Haythornthwaite, Wellman, & Garton, 1998).
In a study of listeners to talk radio, Armstrong and A. M. Rubin (1989) found
that those listeners who actually made calls to the show also found face-to-face
interaction less rewarding than those who merely listened. This finding suggests
that mediated social interaction can be regarded as a functional alternative to
face-to-face interaction, but only when it shares the principle of reciprocity (a
PA R A S O C I A LI N T E R A C T I O N 2 8 5
talk radio caller actually engages in synchronous communication with the show’s
host and/or guests), and—as with on-line socializing—this is qualitatively
different from PSI.
Although Lea and Spears (1995) argued that the existence of cyber-
relationships implies that all relationships need to be “socially situated,” the
concept of parasocial relationships seems to imply the opposite. Here, the
imaginary relationship alone is sufficient. This is clearly not the case in all
parasocial relationships, however; much of the time, there is always a remote
possibility of meeting the media figure, and the function of fan clubs and
organizations may provide direct access.
For these reasons, it is important to consider the relative functions of different
types of media figure. A fully comprehensive theory of PSI needs to distinguish
between media figures who are direct representations of real people (such as
newscasters), and fictional creations, whether dramatic characters played by
actors, or fantasy figures with “low modality” such as cartoon characters (Hodge
& Tripp, 1986). It is always possible for the media user to encounter some of
these figures in the physical realm, whereas others must necessarily remain
imaginary. Until such contact has been established, however, the nature of the
relationship remains parasocial for the media user.
At one level, this is a philosophical issue regarding the authenticity of
mediated experience. A number of authors (notably Picirillo, 1986) argued that,
for the viewer, the mediated experience is a real experience. The implication that
PSI is “imaginary,” or “pseudo-social,” pathologizes viewers who form strong
parasocial attachments. Nevertheless, the “imaginary” position is often held by
viewers themselves, who may fervently deny the reality of their attachment,
possibly as a consequence of the “third-person effect” (Davison, 1983). A study
of responses to the deaths of Princess Diana and the British TV presenter Jill
Dando found several instances of profound confusion from fans who had never
appreciated the extent of their parasocial relationships (Giles & Naylor, 2000).
In summary, there appears to be a strong case for citing PSI in the domain of
social psychology (or, to be precise, relationship psychology), albeit at the
individual (social cognitive) level. However, such a position runs the risk of creating
a psychopathological dimension to PSI. To avoid this happening, it is now necessary
to consider how PSI might be treated as an extension of normal social activity.
PSI as a Dimension of Normal Social Behavior
As Perse and R. B. Rubin (1989) pointed out, people use fundamentally the same
cognitive processes in both interpersonal and mediated communication
2 8 6 GILES
(specifically, in their study, personal constructs and uncertainty reduction).
Indeed, Horton and Strauss (1957) suggested that a form of PSI exists even in
face-to-face social situations where there are large audiences (at a show or a
lecture), where there is a large gap in status between performer and audience.
This is an important consideration for the study of PSI at a psychological
level, because, as suggested earlier, there are many situations in which the
user–figure relation is ambiguous. In normal social interaction there may be a
degree of PSI (e.g., a schoolboy develops a “crush” on a classmate with whom
he has never directly interacted). In mediated interaction there are elements of
direct interaction (for example, talking to a presenter or celebrity guest on a
phone-in show). If these intermediate positions suggest some sort of continuum
of social interaction, with full face-to-face interaction at one end, and PSI with a
cartoon character or a fictional protagonist at the other end, then this may
constitute a new way of theorizing about social interaction in general.
Afurther consideration is the study of relationships at the individual level.
Social cognitive approaches are, by definition, the study of social interaction as
individual cognitive activity. It is normally accepted that this approach is
inadequate by itself for the study of relationships (Duck, 1994). However, there is
a small but growing literature on the role of imagination in social interaction
( C a u g h e y , 1984; Honeycutt, 1993). These studies suggest that individuals plan, and
rehearse, interaction with others, and that this imaginative activity may be an
influential factor in the outcome of real social interaction. Such cognitive activity
may originate in the imaginary friends of childhood, which are often created to
satisfy a need for companionship, usually in the absence of other siblings (Gleason,
Sebanc, & Hartup, 2000; Ta y l o r, 1999). However, no data exists to support the
hypothesis that television has led to a decline in imaginary companions.
What seems certain, in fact, is that PSI is not exclusively a modern-day
phenomenon. The work of James Caughey (1984) is important in this respect; he
has catalogued a variety of instances, both across cultures and through history,
where individuals form strong relationships with imaginary figures. These
include relationships between readers and fictional protagonists (e.g., characters
in a novel); between citizens and major political figures and monarchs; and even
between individuals and gods or spirits. Clearly, the status of these figures, and
reverence accorded to them, varies enormously; nevertheless, it is argued that the
individual requires similar powers of imagination in order to successfully enter
into a relationship with the figure.
This section has looked at ways in which PSI can be regarded as an extension
of normal social cognition, specifically in terms of the use of the imagination.
Drawing on the PSI literature as well, it can be argued that the psychological
PA R A S O C I A LI N T E R A C T I O N 2 8 7
processes acting at the individual level parallel those used in ordinary social
activity and relationship building.
Developmental Aspects of PSI
In addition to considering how PSI might evolve historically and culturally, it is
necessary to consider how it develops across the lifespan. PSI in childhood is a
seriously neglected topic in the communication literature, and the influence of
media in childhood has received little attention from developmental
psychologists, a situation that seems absurd given the sheer quantity of media
consumed by young children. Furthermore, media is similarly neglected in much
of the adolescence literature. These are fields to which a broader model of PSI
might make a very useful contribution.
Anumber of studies have examined young children’s perception of television
characters. Reeves and Greenberg (1977) and Reeves and Lometti (1979) used
multidimensional scaling to show how children from 7 to 11 years of age
evaluate characters on the basis of typical human personality dimensions.
Hoffner (1996) was the first researcher to apply the concept of PSI to children’s
choice of favorite characters. Using an oral modification of the PSI scale with 7-
to 12-year-olds, she found that sex-role stereotyping was an important predictive
factor, with boys overwhelmingly choosing male characters, whereas girls
preferred equivalent numbers of male and female characters. Using
semistructured interviews, Giles and Long (1998) collected some qualitative data
that suggest that identification is a more important influence on 5- to 6-year-olds’
choice of character (particularly for boys), whereas by 10–11 years, children’s
rationale for selecting characters is influenced more by PSI, with characters
being described as similar to friends.
The influence of media figures in adolescence is perhaps less widely studied,
although the use of celebrities and other media figures as role models has long
been a topic of popular concern (Pleiss & Feldhusen, 1995). There has been
particular anxiety over the influence of thin celebrities in adolescent eating
disorders (Harrison, 1997; Heilman, 1998). Most studies in this area draw on
identification as the key psychological process involved; however, little research
has investigated how adolescents develop parasocial relationships with media
figures. Cohen’s (1999) study of Arab and Jewish teenagers’PSI with characters
from an Israeli soap opera suggests that, in adolescence, favorite characters are
likely to be related to as pseudofriends rather than figures who are idolized and
imitated. Nevertheless, recent research on PSI and talk radio indicates that PSI
with a radio host may itself be enough to influence listeners’ attitudes and
2 8 8 GILES
behavior (A. M. Rubin & Step, 2000). Therefore the psychological processes
involved in attitudinal influence and modelling may be highly complex.
Clearly, more research is needed on developmental aspects of PSI and media
use in general. However the little data that exist suggest that children begin to use
personality traits to evaluate media figures from an early age, and that the
characteristics of PSI have begun to appear by early adolescence. But the relative
importance of PSI and identification processes is far from clear, and it may be
that these processes vary according to different types of media figure.
OTHER THEORETICAL ISSUES
The literature reviewed so far supports the contention that PSI is a topic worthy
of closer scrutiny by psychologists, as well as contributing usefully to
communication research. In this next section I outline two important
considerations that should be central to any future model of PSI. The first
concerns the nature of the parasocial relationship itself; how it might be defined
in behavioral, cognitive, or discursive terms; the second, closely related to the
first, concerns the nature of the media figure and how different types of figure
may determine different manifestations of PSI.
Types of User–Figure Relationship
Asuccessful typology of media user–media figure encounter needs to begin by
drawing a clear distinction between PSI and identification, as proposed by
Rosengren and Windahl (1972). However, identification itself produces different
types of user–figure interaction. Feilitzen and Linne (1975) distinguished
between similarity identification, where the television viewer identifies with a
character because they share certain salient characteristics, and wishful
identification, where the viewer desires to emulate the character. Cohen’s (1999)
description of four discrete types of user–figure relationship encapsulates this
distinction and draws in affinity as a discrete category.
PSI. Cohen defined the parasocial relationship as one in which “the viewer
is engaged in a role relationship with a television persona” (p. 329). However, it
is probably more helpful in this model to regard it as a user response to a figure
as if s/he was a personal acquaintance. This response consists of both behavioral
responses (e.g., greeting a newscaster out loud), and cognitive responses (e.g.,
making psychological inferences about a figure’s behavior). From a media
perspective, this role is dictated to some extent by programming conventions.
PA R A S O C I A LI N T E R A C T I O N 2 8 9
For example, as Scannell (1996) described, UK radio broadcasters made a
deliberate change during the 1940s from a clipped, formal style to an informal,
“fireside” approach, in which the listener was addressed as though a personal
acquaintance. It is important to bear in mind, however, that PSI is more than just
a media-related phenomenon, as argued earlier in the article; imaginary social
relationships are characteristic of most societies through history.
Identification. This category represents Feilitzen and Linne’s (1975) concept of
similarity identification, whose key feature is the sharing of perspective. W h a t
distinguishes this form of identification from PSI is that for identification a user
needs to recognize some salient characteristic in the figure that is shared by
themselves. However the user may still engage in PSI without sharing any
perspective; this enables us to interact with media figures whom we actively dislike.
Wishful Identification. In this form of identification we desire to emulate the
figure with which we identify, either in general terms (e.g., as a role model for future
action, or identity development), or in specific terms (e.g., imitating a particular
behavior). Again it is important to distinguish this relationship from a parasocial
one, particularly in terms of adolescent attachments to media figures (i.e., PSI does
not necessarily imply a wish to emulate the figure, as often feared by parents).
Affinity. This covers all instances where a media user displays a liking for a
media figure, without identifying with them, or forming a parasocial relationship.
Affinity has occasionally been used as a complementary measure to PSI; for
example, A. M. Rubin (1981) found a correlation between affinity and the
“companionship” aspect of television, and it correlated positively with PSI in the
A. M. Rubin et al. (1985) study. Affinity may be most useful in relation to
fictional and fantasy characters, where PSI might seem too strong a description.
In addition to Cohen’s four basic types of relationship, we should also
consider instances of PSI that result in the user actually meeting the figure,
whether in person or “in character,” or perhaps in a remote sense (e.g., via a
phone-in show). An important consideration is the moment at which PSI
becomes normal social interaction. This “gray area” raises all manner of issues
about media experience and the presentation of self by media figures themselves.
Types of Media Figure
One of the main ideas behind Cohen’s typology is that types of user–figure
interaction vary with different types of media figure. He argued that PSI is most
2 9 0 GILES
appropriate for describing media figures who directly address the viewer, such as
newscasters and presenters. Some evidence comes from studies that show that,
when a comedian “breaks the fourth wall” and engages the audience directly,
higher levels of PSI are reported (Auter, 1992; Auter & Davis, 1991). Cohen also
argued that identification is more appropriate for protagonists, such as central
characters in a film, and that affinity is a better term to describe relationships
with characters in serials (e.g., soap operas).
In some respects, this distinction is constrained by the notion of PSI as
momentary activity within a single encounter, rather than to the cultivation of
relationships that persist after the encounter. Behavioral aspects of PSI are most
likely with forms of direct address by media figures, although strong PSI is also
exhibited with soap characters (Perse & A. M. Rubin, 1987) and cartoon
characters (Hoff n e r, 1996), suggesting that direct address is not a prerequisite for
PSI. Indeed, when we consider how parasocial relationships might form over
time, both types of identification listed earlier might enter into the mix as well.
Thus a parasocial relationship may exist with many different types of figure,
regardless of the overt interaction between user and figure in any given encounter.
Nevertheless, three important characteristics of different types of media
figure can be identified that may determine the nature of the parasocial
relationship that can evolve beyond the immediate encounter.
Authenticity/Realism. Akey element in the user–figure relationship is the
extent to which users are able to make person, or character, judgements about the
figure. Therefore, it is necessary for the figure to present a credible persona—one
of the essential ingredients for the appreciation of soap opera (Geraghty, 1991).
Anumber of studies have found “perceived realism” to be a significant predictor
of PSI (Alperstein, 1991; A. M. Rubin et al., 1985; A. M. Rubin & Perse, 1987).
However, this may well depend on the nature of the media figure. Clearly,
television personalities such as newscasters, presenters and celebrities need to
appear authentic, because this is part of their appeal. A popular celebrity can
easily harm his or her reputation by “faking it,” for example, endorsing an
inappropriate product (Alperstein, 1991). Perceived realism is also important for
PSI with soap characters (A. M. Rubin & Perse, 1987).
Nevertheless, intense parasocial attachments may be formed with figures who
are clearly not “authentic,” such as a pop star who takes an obvious pseudonym
and is only known through his or her “act,” a fantasy character such as Mickey
Mouse, or a cartoon character such as Homer Simpson. What psychological
explanations lie behind these types of attachments? Clearly research is necessary
PA R A S O C I A LI N T E R A C T I O N 2 9 1
on this topic. For the time being, however, it may be possible to conceive of
authenticity as a dimension within the overall model of PSI.
Representation Across Different Media Outlets. One quality of media
figures often overlooked in the PSI literature is the continuity of their
representation across different media outlets. Newscasters, the genre of figure
that inspired the PSI scale research paradigm, are somewhat unusual among
media figures in that they rarely appear in media other than in a specific
television genre. Apop star or film star, on the other hand, may appear on several
different television and radio programs, as performer or chat show guest, give
interviews to dozens of magazines, and be the object of discussion by daily
newspapers. Furthermore, the repeated viewing of a video cassette featuring the
star will intensify visual aspects of PSI with that star. Newscasters, on the other
hand, although inviting affective and behavioral responses during the bulletin,
are unlikely in most cases to engage media users far beyond the viewing episode.
User Contexts. Another feature of most PSI research is that the media user
is typically characterized as a television viewer, often a solitary figure, whose
PSI gratifies a need for social interaction. Although findings generally fail to
support the notion of PSI as compensation for a lack of social outlets (e.g., A. M.
Rubin et al., 1985), the emphasis has been exclusively on individual responses to
media figures. What implications does co-viewing have for PSI? If co-viewers
reinforce initial responses to a media figure, this may strengthen the parasocial
relationship; indeed, co-viewing—and subsequent discussion of the figure—may
strongly influence the development of the parasocial relationship.
Future research in PSI needs to address the above issues. The different types of
user–figure interaction can be addressed by conceptualizing PSI as an extension of
ordinary social interaction, and by examining features of social encounters that are
significant for parasocial encounters. The different types of media figure can
inform this conceptualization, but in addition, theories of PSI need to distinguish
between PSI as momentary activity and longer term interaction—the development
of parasocial relationships as conceived by R. B. Rubin and McHugh (1987).
These issues are now discussed in relation to the formulation of a model of PSI.
DEVELOPMENT OF A MODEL FOR PSI RESEARCH
In this section a model for future research in the area of PSI is described, drawing
on the existing literature and on the issues raised in the previous section.
2 9 2 GILES
McQuail and Windahl (1993) have outlined three functions of communication
models that this model respects. Firstly, it has an organizing function: It draws
together the themes from the literature in an attempt to elucidate the processes in
the formation of parasocial relationships; secondly, it has a heuristic function: It
presents a theory of PSI as extended social activity in the form of a continuum;
and thirdly, it has a predictive function: By outlining the processes in parasocial
relationship formation, it anticipates possible outcomes of media use.
The model is conceived as having two basic components. The first is a
continuum of social activity based around social encounters. This enables us to
identify the nature of an encounter at any point in its duration through assessing
four qualities of the encounter, and to determine which qualities define an
encounter as parasocial. It also allows us to distinguish between parasocial
encounters according to the type of media figures involved, and the nature of the
possible relationship between the user and the figure. The second component is a
diagrammatic chart outlining the processes involved in a parasocial encounter that
may ultimately bind the media user into a parasocial relationship with a figure.
Model Component 1: Continuum of Social–Parasocial Encounters
The first part of the model addresses Horton and Wohl’s (1956) call for PSI to be
incorporated into “the matrix of usual social activity” (p. 225). It describes four
qualities of social encounters that can be used to locate a given encounter on a
continuum that stretches from an unambiguous social encounter (a face-to-face
dyad) at the top end to a genuinely parasocial encounter at the bottom. The four
qualities are (a) number of persons involved (broadly, whether a dyadic
encounter or a group encounter); (b) physical distance between interactants
(either proximate, i.e. face-to-face, or distant, i.e., separated by an amount of
space that renders face-to-face communication impossible); (c) social
conventions (whether formal or informal; this distinction is based on ritual
features of the interaction as well as power relations); (d) potential relationship
between the interactants (in most face-to-face communication this is unlimited,
but will differ across levels of PSI).
At the “social” end of the continuum, encounters are ranked largely in terms
of group size, from dyads to large groups. The rationale here is that, the more
other people are involved in that encounter, the weaker the quality of interaction
with a specific person becomes. Up to this point, the encounters have not been
defined according to the individual characteristics of the persons involved.
Midway along the continuum, however, the emphasis switches from ordinary
social groups to encounters with media figures. The most “social” of these
PA R A S O C I A LI N T E R A C T I O N 2 9 3
2 9 4 GILES
encounters is a one-to-one meeting with a media figure, in which the relationship
possibilities are potentially the same as with any social encounter, albeit
restrained by social conventions.
As we approach the “parasocial” end of the continuum, the interaction
becomes weaker according to the authenticity or realism of the representation of
the person. For parasocial encounters, distance and formality become redundant
qualities of the interaction; all PSI inevitably takes place across a distance and is
entirely constrained by the social (or communicative) conventions. Therefore
parasocial encounters are distinguished according to the nature of the media
figure, and by the relationship that is possible with that figure (determined by the
figure’s authenticity/realism). These can be broadly identified across three levels
of PSI.
First-order PSI refers to the type of encounter defined by Cohen (1999) as
parasocial, where the media figure addresses the user directly, for example a talk
show host facing the camera and greeting the viewer.
Second-order PSI concerns encounters where the media figure is to some
degree inauthentic. Typical of this level is a soap character portrayed by an actor.
Here there is no doubt as to the fictional nature of the representation, yet the
character’s physical counterpart is responded to as a real person. Ultimately, a
user might make face-to-face contact with the figure, but would only be able to
enter into a social relationship with the actor and not the character to whom s/he
has established a parasocial relationship. Thus user–figure meetings may be
problematic; for example, there are numerous anecdotes concerning soap actors
who play unpopular roles receiving abuse from viewers based on their
character’s behavior.
Third-order PSI concerns encounters with fantasy or cartoon figures who
have no real-life counterpart. These are distinguished from first- and second-
order encounters in that a social relationship with the figure is impossible.
This continuum is outlined in Table 1, along with examples at each level.
These examples are not intended to be definitive by any means but simply
illustrate the principles involved (as it stands they are mostly culture specific). In
the “potential relationship” column, the examples are intended to refer to
interaction with a single person in each setting. So, for “party,” the potential
relationship is defined by the probabilities at the outset; for example, you walk
into the party and see a stranger who, over the course of the evening, could
develop into a friend. The potential relationships set up by this informal large
group setting are significantly different from those in a third-order PSI (where
friendship is impossible) and other encounters progressively down the
continuum (where friendship is an increasingly remote prospect).
PA R A S O C I A LI N T E R A C T I O N 2 9 5
Clearly there is plenty of potential slippage between levels, although this is
most likely across levels of formality and informality at the “social” end of the
continuum (a colleague has every chance of becoming a best friend, but this is not
true of all formal dyadic relationships). The further toward the “parasocial” end
one goes, the less likelihood there is of formal relations evolving into informal
relations. Thus formality is one of the essentially binding features of PSI.
This component of the model, then, addresses some of the features of PSI that
have hitherto been overlooked in the literature: varying degrees of PSI based on
the varying nature of media figures. Perhaps most importantly, it considers the
possibilities for social contact with media figures. These may play an important
role in determining the nature of the parasocial relationship, particularly in the
case of fans of specific individuals, and in extreme cases where individuals have
contacted celebrities in a threatening manner or even been accused of stalking
celebrities (Dietz et al., 1991; Giles, 2000).
Model Component 2: Stages in the Development
of the Parasocial Relationship
The second component of the model is aimed at identifying the processes
through which an encounter with a media figure may evolve into a parasocial
relationship. It draws on the continuum presented earlier in that the nature of the
parasocial relationship is determined partly by the nature of the media figure. It
also considers the effects of successive encounters both within the same medium
(e.g., successive episodes of a drama serial) and across several media (e.g., a
celebrity appearing on different television shows, radio broadcasts,
advertisements, and in magazines). It also considers the function of other people,
particularly coviewers of television, in shaping the outcome of PSI.
Figure 1 contains a flowchart that displays the behavioral outcomes of a
single viewing episode (or any single episode of interaction with a media figure).
The cognitive activity during the episode is contained in the oval at the top of the
chart, and consists mainly of making judgements about the media figure (person)
based on knowledge of that person. For example, physical knowledge may give
rise to physical attraction, or the expression of an opinion may chime with the
opinion of the user and create a positive judgment based on attitude homophily.
Within the viewing episode, a number of outcomes are likely, and these are
based largely on Cohen’s (1999) types of relationship. The user may express
affinity for the figure, for example, liking a character in a drama, but the
interaction goes no further than the viewing episode. Alternatively, he or she may
identify with the figure. If this is the case, then it is possible that the user’s
2 9 6 GILES
behavior beyond the viewing episode may be influenced by the encounter, in the
shape of imitative behavior or modelling. This path is identified in the flowchart
by a box and one-way arrow. Future encounters with the figure may also
reinforce (or disrupt) the imitative behavior.
Figure 1. Stages in the development of a parasocial relationship.
Where PSI takes place, the development of a relationship can only occur if
there are further media encounters, so this connection is marked by a one-way
arrow. An important link in the model is the role of other media users, who can
influence the PSI process in two ways; firstly, they can affect the relationship
development in the light of other encounters (for example, by revealing some
gossip about a celebrity from a different source). This is particularly true where
the opinions of other users do not agree with those of the individual user, for
example a user’s friends disapprove of a particular soap character of whom the
user had initially formed a positive opinion. If the user is highly influenced by
peers, the discussion may substantially color the person judgments made in the
next viewing episode. The second way that other users may influence PSI occurs
PA R A S O C I A LI N T E R A C T I O N 2 9 7
Person
judgement
Person
knowledge
Other
encounters
Imitative
behaviour
(modelling?)
Imagined
interaction
Discuss
figure with
others
The viewing episode
Make attempt to
contact
Person
judgment
Imitative
behaviour
(modeling?)
Discuss
figure with
others
The viewing episode
Other
encounters
Person
knowledge
during the viewing episode itself, which is why the arrow for this connection is
bi-directional. Person judgements may be continually updated during the
viewing episode in the light of comments of other viewers.
The behavioral outcome of PSI is dependent on a number of encounters with
the media figure. At some point the user may consider whether or not to make
contact with the figure. Where this is the case, a period of imagined interaction
may intervene, because the user will need to weigh the costs and benefits of
making such an approach. Most of the time PSI fails to reach this stage, but
simply forms part of the user’s cultural and social life.
In summary, this model of PSI goes well beyond the conceptualization of PSI
presented in the literature to date. It presents PSI as an extension of normal social
activity by considering shared and different qualities of social and parasocial
encounters. It considers psychological and other issues that have not been
addressed in the PSI literature: the distinction between PSI as momentary
activity and the longer term development of parasocial relationships; the
importance of different types of media figure, and different types of medium, on
the development of such relationships; the role of co-viewers and discussants in
shaping PSI; and the dual function of PSI and identification.
APPLICATIONS OF THE MODEL
The most important reason for devising this model of PSI is to provide a
theoretical framework that might inform future research on aspects of PSI that
are underdeveloped in the existing literature.
In relation to the psychometric research of the uses and gratifications
tradition, the model expands the notion of PSI as a behavior that might vary in
nature with regard to different types of media figure. An initial test of the model
would be to compare PSI with, say, newsreaders, soap characters, and cartoon
characters, perhaps by entering values for all three types as predictors in models
of different aspects of media use. It may also be that parasocial relationships at
different levels develop along different lines; here, R. B. Rubin and McHugh’s
(1987) model of parasocial relationship development could be examined across
different types of media figure.
Other aspects of the model require different methodologies to those that have
been used to study PSI thus far. For example, the influence of coviewing on PSI,
and the subsequent integration of information in the progress of a parasocial
relationship might require longitudinal designs, or even ethnographic research,
particularly as a means of examining the meaning of parasocial relationships for
media users. In this respect, it will be necessary to apply qualitative methodologies
2 9 8 GILES
to the study of PSI in order to gain a richer understanding of the psychological
implications of PSI.
Perhaps the most important area for future PSI research lies in the
developmental psychology field. Firstly, the creation of adolescent heroes and
role models is a concern for many parents and professionals working with young
people. Much of the existing research in this area draws on social learning theory
in suggesting that media figures function largely as “models” that are imitated by
impressionable youngsters (e.g., Austin & Meili, 1994; Harrison, 1997).
However, the literature on PSI suggests that media users develop parasocial
relationships with media figures that are more complex than modelling theory,
based largely on the concept of identification.
The two components of the model might inform future research on this topic.
Firstly, the continuum, with its consideration of different types of relationships
determined by different types of media figure, offers an explanatory framework
that might be used to explore the nature of attachments to a variety of media
figures. Secondly, the model might be used to chart the evolution of a parasocial
relationship in adolescence, to investigate how successive exposures to a media
figure bind the user into a relationship, and, significantly, the contribution of
other users (both co-viewers and discussants) in shaping the course of the
relationship. The conceptual distinction—made in the model—between PSI and
identification will be a critical factor in further research in adolescent media use.
An important application of PSI research is in the clinical field, where
interesting research has recently been conducted into “special media interests”
with a population of learning disabled adults (Whomsley, 2000). Media figures
were found to have considerable influence over the lives of the participants in
this study, with favorite characters acting as models for planned behavior in
social situations, or providing a cathartic or escapist function. The effects of
special media interests included both positive and negative features; again the
model might provide a clear theoretical framework for investigating the
processes involved in the development of parasocial relationships.
Asecond clinical application of the model might be in the area of physical
d i s a b i l i t y , exploring the value and functions of PSI in the lives of people with
restricted mobility. This would include older adults, who formed the basis of much
early PSI research examining the association between PSI and social isolation.
Here, the ideas underpinning the continuum might provide a useful theoretical base
for research. Television producers and program makers might also have particular
interest in the factors promoting strong PSI in these populations.
I n e v i t a b l y, in some cases the effects of PSI are psychopathological in nature.
Failure to discriminate between PSI and ordinary social activity may lead to
PA R A S O C I A LI N T E R A C T I O N 2 9 9
situations where individuals believe that their PSI has been reciprocated by the
media figure. A classic example of this is the case of John Hinckley Jr., who shot
former U. S. president Ronald Reagan in an attempt to demonstrate his strength of
feeling for actress Jodie Foster (Giles, 2000). Like many aspects of PSI, this
phenomenon, known as erotomania in the clinical literature, has a history that dates
back long before mass communication. To d a y, we are likely to associate such
behavior with stalking, a matter of deep concern for celebrities and other media
figures. Perhaps a final application of the model would be to identify the points of
departure between ordinary fandom and delusional behavior, and at what point in
the development of the parasocial relationship such points become salient.
REFERENCES
Alperstein, N. (1991). Imaginary social relationships with celebrities appearing
in television commercials. Journal of Broadcasting and Electronic Media,
35, 43–58.
Armstrong, C. B., & Rubin, A. M. (1989). Talk radio as interpersonal
communication. Journal of Communication, 39, 84–94.
Austin, E. W., & Meili, H. K. (1994). Effects of interpretations of televised
alcohol portrayals on children’s alcohol beliefs. Journal of Broadcasting and
Electronic Media, 38, 417–435.
Auter, P. J. (1992). TVthat talks back: An experimental validation of a parasocial
interaction scale. Journal of Broadcasting and Electronic Media, 36,
173–181.
Auter, P. J., & Davis, D. M. (1991). When characters speak directly to viewers:
Breaking the fourth wall in television. Journalism Quarterly, 68, 165–171.
Auter, P. J., & Palmgreen, P. (2000). Development and validation of a parasocial
interaction measure: The audience–persona interaction scale.
Communication Research Reports, 17, 79–89.
Caughey, J. L. (1984). Imaginary social worlds: A cultural approach. Lincoln:
University of Nebraska Press.
Cohen, J. (1999). Favorite characters of teenage viewers of Israeli serials.
Journal of Broadcasting and Electronic Media, 43, 327–345.
Conway, J. C., & Rubin, A. M. (1991). Psychological predictors of television
viewing motivation. Communication Research, 18, 443–463.
Davison, W. P. (1983). The third-person effect in communication. Public
Opinion Quarterly, 47, 1–15.
3 0 0 GILES
Dietz, P. E., Matthews, D. B., Van Duyne, C., Martell, D. A., Parry, C. D. H.,
Stewart, T., Warren, J., & Crowder, J. D. (1991). Threatening and otherwise
inappropriate letters to Hollywood celebrities. Journal of Forensic Sciences,
36, 185–209.
Duck, S. (1994). Meaningful relationships: Talking, sense, and re l a t i n g .
Thousand Oaks, CA: Sage.
Feilitzen, C,. & Linne, O. (1975). Identifying with television characters. Journal
of Communication, 25, 51–55.
Geraghty, C. (1991). Women and soap opera. Cambridge: Polity Press.
Giles, D. C. (2000). Illusions of immortality: A psychology of fame and celebrity.
Basingstoke: Macmillan.
Giles, D. C., & Long, L. (1998, September). Expanding the concept and
m e a s u rement of parasocial interaction. Paper presented at the IXth
European Conference on Developmental Psychology, Spetses, Greece.
Giles, D. C., & Naylor, G. C. Z. (2000, December). The construction of intimacy
with media figures in a parasocial context. Paper presented at the BPS
London Conference, Institute of Education.
Gleason, T. R., Sebanc, A. M., & Hartup, W. W. (2000). Imaginary companions
of preschool children. Developmental Psychology, 36, 419–428.
Gleich, U. (1996). Sind Fernsehpersonen die “Freunde” des Zuschauers? Ein
Vergleich zwischen parasozialen und realen sozialen Beziehungen [Are
media figures “friends” of the viewers? A comparison between parasocial
and real social relationships]. In P. Vorderer (Ed.) Fernsehen als
“Beziehingskiste”: Parasoziale beziehungen und interaktionen mit TV-
personen (pp. 113–144). Opladen: Westdeutscher Verlag.
Gleich, U. (1997). Parasocial interaction with people on the screen. In P.
Winterhoff-Spurk & T. H. A. Van der Voort (Eds.). New horizons in media
psychology: Research co-operation and projects in Europe (pp. 35–55).
Opladen, Germany: Westdeutscher Verlag.
Grant, A., Guthrie, K., & Ball-Rokeach, S. J. (1991). Television shopping: A
media dependency perspective. Communication Research, 18, 773–798.
Harrison, K. (1997). Does interpersonal attraction to thin media personalities
promote eating disorders? Journal of Broadcasting and Electronic Media,
41, 478–500.
Haythornthwaite, C., Wellman, B., & Garton, L. (1998). Work and community
via computer-mediated communication. In J. Gackenbach (Ed.) Psychology
and the internet (pp. 199–226). San Diego, CA: Academic.
PA R A S O C I A LI N T E R A C T I O N 3 0 1
Heilman, E.E. (1998). The struggle for self: Power and identity in adolescent
girls. Youth and Society, 30, 182–208.
Hinde, R. A. (1979). Towards understanding relationships. London: Academic.
Hodge, B., & Tripp, D. (1986). Children and television: A semiotic approach.
Cambridge: Blackwell.
Hoffner, C. (1996). Children’s wishful identification and parasocial interaction
with favorite television characters. Journal of Broadcasting and Electronic
Media, 40, 389–402.
Honeycutt, J. M. (1993). Memory structures for the rise and fall of personal
relationships. In S. Duck (Ed.) Individuals in relationships (pp. 60–86).
Newbury Park, CA: Sage.
Horton, D., & Strauss, A. (1957). Interaction in audience-participation shows.
American Journal of Sociology, 62, 579–587.
Horton, D., & Wohl, R. R. (1956). Mass communication and para-social
interaction. Psychiatry, 19, 215–229.
Kirschner, S., & Kirschner, D. A. (1997). (Eds.) Perspectives on psychology and
the media. Washington, DC: American Psychological Association.
Koenig, F., & Lessan, G. (1985). Viewers’relationship to television personalities.
Psychological Reports, 57, 263–266.
Lea, M., & Spears, R. (1995). “Love at first byte?” Building personal
relationships over computer networks. In J. T. Wood & S. Duck (Eds.)
Under-studied relationships: Off the beaten track (pp. 197–233). Thousand
Oaks, CA: Sage.
Levy, M. R. (1979). Watching TV news as para-social interaction. Journal of
Broadcasting, 23, 69–80.
Livingstone, S. (1998). Making sense of television: The psychology of audience
interpretation (2nd ed.). London: Routledge.
McKenna., K. Y. A., & Bargh, J. A. (2000). Plan 9 from cyberspace: The
implications of the internet for personality and social psychology.
Personality and Social Psychology Review, 4, 57–75.
McQuail, D., Blumler, J. G., & Brown, J. R. (1972). The television audience: A
revised perspective. In D. McQuail (Ed.) Sociology of mass
communications: Selected readings (pp. 135–165). Harmondsworth:
Penguin.
McQuail, D., & Windahl, S. (1993). Communication models for the study of
mass communication (2nd ed.). Harlow: Longman.
3 0 2 GILES
Nordlund, J. (1978). Media interaction. Communication Research, 5, 150–175.
Papa, M. J., Singhal, A., Law, S., Pant, S., Sood, S., Rogers, E. M., & Shefner-
Rogers, C. L. (2000). Entertainment-education and social change: An
analysis of parasocial interaction, social learning, collective efficacy, and
paradoxical communication. Journal of Communication, 50, 31–55.
Papachrissi, Z., & Rubin, A. M. (2000). Predictors of internet use. Journal of
Broadcasting and Electronic Media, 44, 175–196.
Parks, M. R., & Floyd, K. (1996). Making friends in cyberspace. Journal of
Communication, 46, 80–97.
Parks, M. R., & Roberts, L. D. (1998). “Making MOOsic”: The development of
personal relationships on line and a comparison to their off - l i n e
counterparts. Journal of Social and Personal Relationships, 15, 517–537.
Perse, E. M., & Rubin, R. B. (1989). Attribution in social and parasocial
relationships. Communication Research, 16, 59–77.
Picirillo, M. S. (1986). On the authenticity of televisual experience: A critical
exploration of para-social closure. Critical Studies in Mass Communication,
3, 337–355.
Pleiss, M. K., & Feldhusen, J. F. (1995). Mentors, role models, and heroes in the
lives of gifted children. Educational Psychologist, 30, 159–169.
Reeves, B., & Anderson, D. R. (1991). Media studies and psychology.
Communication Research, 18, 597–600.
Reeves, B., & Greenberg, B. (1977). Children’s perceptions of television
characters. Human Communication Research, 3, 113–117.
Reeves, B., & Lometti, G. (1979). The dimensional structure of children’s
perceptions of television characters: A replication. Human Communication
Research, 5, 247–256.
Reeves, B., & Nass, C. (1996). The media equation: How people tre a t
computers, television and new media like real people and places. Stanford
University: Cambridge University Press.
Rohlfing, M. E. (1995). “Doesn’t anybody stay in one place anymore?” An
exploration of the under-studied phenomenon of long-distance relationships.
In J. T. Wood & S. Duck (Eds.) Under-studied relationships: Off the beaten
track (pp. 173–196). Thousand Oaks, CA: Sage.
Rosengren, K. E., & Windahl, S. (1972). Mass media consumption as a functional
alternative. In D. McQuail (Ed.), Sociology of mass communications: Selected
readings (pp. 119–134). Harmondsworth: Penguin.
PA R A S O C I A LI N T E R A C T I O N 3 0 3
Rosengren, K. E., Windahl, S., Hakansson, P., & Johnsson-Smaragdi, U. (1976).
Adolescents’ TV relations: Three scales. Communication Research, 3,
347–366.
Rubin, A. M. (1981). An examination of television viewing motivations.
Communication Research, 8, 141–165.
Rubin, A. M., & Perse, E. M. (1987). Audience activity and soap opera
involvement: A uses and effects investigation. Human Communication
Research, 14, 246–268.
Rubin, A. M., Perse, E. M., & Powell, R.A. (1985). Loneliness, parasocial
interaction, and local television news viewing. Human Communication
Research, 12, 155–80.
Rubin, A. M., & Step, M. M. (2000). Impact of motivation, attraction, and
parasocial interaction of talk radio listening. Journal of Broadcasting and
Electronic Media, 44, 635–654.
Rubin, R. B., & McHugh, M. (1987). Development of parasocial interaction
relationships. Journal of Broadcasting and Electronic Media, 31, 279–292.
Scannell, P. (1996). Radio, television and modern life. Oxford: Blackwell.
Sood, S., & Rogers, E. (2000). Dimensions of parasocial interaction by letter-
writers to a popular entertainment-education soap opera in India. Journal of
Broadcasting and Electronic Media, 44, 386–414.
Taylor, M. (1999). Imaginary companions and the children who create them.
New York: Oxford University Press.
Turkle, S. (1995). Life on the screen: Identity in the age of the internet. New
York: Simon and Schuster.
Turner, J. R. (1993). Interpersonal and psychological predictors of parasocial
interaction with different television performers. Communication Quarterly,
41, 443–453.
Vo r d e r e r, P. (1996). Picard, Brinkmann, Derrick & Co. als freunde der zuschauer:
Eine explorative studie uber parasoziale beziehungen zu serienfiguren.
[Picard, Brinkmann, Derrick & Co. as friends of the viewer: An exploratory
study of parasocial relationships with T V characters]. In P. Vorderer (Ed).
Fernsehen als “Beziehingskiste”: Parasoziale beziehungen und interaktionen
mit TV-personen (pp. 153–172). Opladen: Westdeutscher Ve r l a g .
Wallace, P. (1999). The psychology of the internet. Cambridge: Cambridge
University Press.
3 0 4 GILES
Wellman, B. (1996). An electronic group is virtually a social network. In S.
Kiesler (Ed). Culture of the internet (pp. 179–205). Mahwah, NJ: Lawrence
Erlbaum Associates.
Whomsley, S. R. C. (2000). Identity search: Special media interest in a clinical
learning disabilities population. Unpublished thesis submitted in partial
completion of a doctorate in Clinical Psychology, University of Leicester.
PA R A S O C I A LI N T E R A C T I O N 3 0 5
