Process of parasocial interaction in local television news watching

MISSING IMAGE

Material Information

Title:
Process of parasocial interaction in local television news watching
Physical Description:
xi, 157 leaves : ; 29 cm.
Language:
English
Creator:
Choi, Yangho, 1957-
Publication Date:

Subjects

Subjects / Keywords:
Television broadcasting of news -- Psychological aspects   ( lcsh )
Television viewers -- Psychology   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 147-156).
Statement of Responsibility:
by Yangho Choi.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001943429
notis - AKB9675
oclc - 31063985
System ID:
AA00003654:00001

Full Text









PROCESS OF PARASOCIAL INTERACTION IN
LOCAL TELEVISION NEWS WATCHING















By

YANGHO CHOI


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1993


































Copyright 1993

by

Yangho Choi












ACKNOWLEDGMENTS


First, I am grateful to my committee chairman, Dr. John

Wright, for all his advice, direction, and patience during the

course of this research. I also want to thank him for being

concerned for my personal well being.

I would like to thank the members of my committee: Dr.

Mary Ann Ferguson, Dr, Leonard Tipton, Dr. Richard Lutz, and

Dr. Kent Lancaster. Most of all, I am grateful to Dr. Mary

Ann Ferguson for many constructive suggestions and comments,

and for helping me resolve many problems that came up during

the course of this research. I thank Dr. Leonard Tipton for

his thoughtfulness, kindness and encouragement. I thank Dr.

Richard Lutz for providing many valuable insights into the

improvement of this research. I must also extend my special

thanks to Dr. Kent Lancaster for his generosity in being on my

dissertation committee.

Most important, I am grateful to my wonderful parents to

whom I dedicate this dissertation. My father instilled in me

the love of learning and pride in accomplishing a goal. My

mother always provided her warm support and love. Without

them, my long stay in the United States might have been more

difficult.


iii













TABLE OF CONTENTS

page
ACKNOWLEDGMENTS .......... ........... ................... iii

LIST OF TABLES ..................................... vii

LIST OF FIGURES ...................................... viii

ABSTRACT ............................................. x

CHAPTERS

1 INTRODUCTION ...................................... 1

2 REVIEW OF LITERATURE ............................ 7

Statement of the Problem ........................ 7
Uses and Gratifications Approach ................. 12
Development of Parasocial Interaction (PSI) ..... 18

Sociological Origins of PSI ................... 22
Psychological Origins of PSI ................... 25
Personality and PSI ...... ..................... 27
Media Use, Motives, and PSI ................... 28
Loneliness and PSI .............. ................ 34
PSI from Interpersonal Perspective ............. 37
Television as Significant Others ............... 39
Operationalization of PSI ..................... 41

Eysenck's Personality Dimension .................. 46
The Person or the Situation ..................... 49
Mass Comm. from Individual Difference Perspective 51

Extraversion, Neuroticism, Psychoticism, & TV.. 51
Sensation Seeking and Media Use ................ 54
Personality, Drugs, and TV Use ................. 55
Risk-Taking, Information Seeking, & Media GS... 56
Personality & Communication-Related Behaviors.. 57

Local Television News ............................ 60







page

3 RESEARCH DESIGN .................................. 64

Proposed Model of PSI ........................... 64

Moderator Model ............. ....... ............ 67
Mediator Model ................................. 68

Hypotheses ...................................... .. 69
Procedure ...................................... .. 72
Measures ........................................ .. 72

Local TV News Exposure ........................ 73
Local TV News Reliance ........................ 75
Parasocial Interaction ........................ 75
Loneliness .................................... .. 76
Extraversion .................................. .. 77
Neuroticism ...................................... 79
Need for Social Interaction ................... 80

4 RESULTS ......................................... 81

Relationships between Local TV
News Watching & PSI ............................ 85
Relationships between Individual Difference
Variables & PSI ................................ 90
Relationship between the Situational Variable
Loneliness & PSI .......... ................... 91
Relationships between Moderators & Local TV
News Watching ................................. .. 92
Multivariate Analysis for Proposed Model of PSI.. 99

Regression Analysis for Whole Sample (N=587)... 99
Regression Analysis for Adult Sample (N=283)... 103
Regression Analysis for Student Sample (N=280). 112

5 CONCLUSION AND DISCUSSION ....................... 123

Summary of Findings ............................. 123
Limitations of the Study ........................ 128
Conclusions ....................... ... ........... 131
Implications for the PSI Research ................ 132
Directions for Future Research ................... 133








pacte


APPENDICES

A QUESTIONNAIRE USED IN TELEVISION NEWS
WATCHING STUDY .................................. 136

B MEASURES OF KEY VARIABLE ........................ 143

REFERENCES ........................................... 147

BIOGRAPHICAL SKETCH .................................. 157












LIST OF TABLES


TABLES page

3-1 Reliability Coefficients for Key Measures.......... 74

3-2 Factor Analysis of Individual Difference Items... 78

4-1 Means, Standard Deviations & Ranges of Variables. 82

4-2 Zero-Order Correlations for TV News Variables.... 83

4-3 Comparisons of Means betw. Students & Adults for
Key Variables............ ..... .................... 86

4-4 T-Tests betw. Adults & Students for Key Variables 86

4-5 Zero Order Correlations for Key Variables(N=587). 87

4-6 Zero Order Correlations for Key Variables(N=281). 88

4-7 Zero Order Correlations for Key Variables(N=280). 89

4-8 Summary of Hypotheses Tests ...................... 122


vii












LIST OF FIGURES


FIGURES page

3-1 Proposed PSI Model ............................... 66

4-1 Regression Coefficients for PSI Model w/o
Interaction (N=587) .............................. 94

4-2 Regression Coefficients for PSI Model w/
Interaction (N=587) .............................. 94

4-3 Regression Coefficients When NSI is a Moderator
betw. TV Exposrue & PSI (N=587) .................. 95

4-4 Regression Coefficients When Extraversion
is a Moderator betw. TV Exposrue & PSI (N=587).. 95

4-5 Regression Coefficients When Neuroticism
is a Moderator betw. TV Exposrue & PSI (N=587).. 96

4-6 Regression Coefficients When Loneliness
is a Moderator betw. TV Exposrue & PSI (N=587).. 96

4-7 Regression Coefficients When NSI is a Moderator
betw. TV Reliance & PSI (N=587) .................. 97

4-8 Regression Coefficients When Extraversion
is a Moderator betw. TV Reliance & PSI (N=587).. 97

4-9 Regression Coefficients When Neuroticism
is a Moderator betw. TV Reliance & PSI (N=587).. 98

4-10 Regression Coefficients When Loneliness
is a Moderator betw. TV Reliance & PSI (N=587).. 98

4-11 Regression Coefficients for PSI Model w/o
Interaction (N=283) ............................... 104

4-12 Regression Coefficients for PSI Model w/
Interaction (N=283) ............................... 104

4-13 Regression Coefficients When NSI is a Moderator
betw. TV Exposrue & PSI (N=283) .................. 105


viii






page
4-14 Regression Coefficients When Extraversion
is a Moderator betw. TV Exposrue & PSI (N=283).. 105

4-15 Regression Coefficients When Neuroticism
is a Moderator betw. TV Exposrue & PSI (N=283).. 106

4-16 Regression Coefficients When Loneliness
is a Moderator betw. TV Exposrue & PSI (N=283).. 106

4-17 Regression Coefficients When NSI is a Moderator
betw. TV Reliance & PSI (N=283)................... 107

4-18 Regression Coefficients When Extraversion
is a Moderator betw. TV Reliance & PSI (N=283).. 107

4-19 Regression Coefficients When Neuroticism
is a Moderator betw. TV Reliance & PSI (N=283).. 108

4-20 Regression Coefficients When Loneliness
is a Moderator betw. TV Reliance & PSI (N=283).. 108

4-21 Regression Coefficients for PSI Model w/o
Interaction (N=280) .............................. 113

4-22 Regression Coefficients for PSI Model w/
Interaction (N=280) ........ ...................... 113

4-23 Regression Coefficients When NSI is a Moderator
betw. TV Exposrue & PSI (N=280)................... 114

4-24 Regression Coefficients When Extraversion
is a Moderator betw. TV Exposrue & PSI (N=280).. 114

4-25 Regression Coefficients When Neuroticism
is a Moderator betw. TV Exposrue & PSI (N=280).. 115

4-26 Regression Coefficients When Loneliness
is a Moderator betw. TV Exposrue & PSI (N=280).. 115

4-27 Regression Coefficients When NSI is a Moderator
betw. TV Reliance & PSI (N=280) .................. 116

4-28 Regression Coefficients When Extraversion
is a Moderator betw. TV Reliance & PSI (N=280).. 116

4-29 Regression Coefficients When Neuroticism
is a Moderator betw. TV Reliance & PSI (N=280).. 117

4-30 Regression Coefficients When Loneliness
is a Moderator betw. TV Reliance & PSI (N=280).. 117












Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy

PROCESS OF PARASOCIAL INTERACTION IN
LOCAL TELEVISION NEWS WATCHING


By

Yangho Choi

December 1993

Chairman: Dr. John W. Wright
Major Department: Journalism and Communications

A model was proposed explaining parasocial interaction

(PSI) in local television news watching. It was hypothesized

that the individual difference variables of Need for Social

Interaction, Extraversion, and Neuroticism as well as the

situational variable Loneliness could explain the parasocial

interaction in local TV news watching.

Questionnaires were completed by 283 adults and 280

undergraduate students. Mixed support was found for the

proposed PSI model. Extraversion was significantly correlated

with local TV news exposure only among adults. For college

students, however, there was significant association between

extraversion and local TV news reliance. Neuroticism had a

negative correlation with local television news exposure.

Loneliness, a situational variable, is significantly related






to local TV news exposure among adult members. Further,

Loneliness (situation variable) is the only variable which

remains significantly associated with PSI when controlling for

other moderation variables and local television news watching

measures. Lonely people are less likely to communicate

interpersonally. They are more likely to turn to television

when experiencing loneliness.

The multiple regression analysis suggested that local TV

news exposure, local TV news reliance, need for social

interaction, and loneliness were the only substantial,

positive contributors to the explanation of parasocial

interaction among a sample of adults. These variables

explained 27 percent of variance in parasocial interaction.

For a sample of students, on the other hand, all the

individual difference variables including local TV news

watching measures were inconsequential components of the

equation in the multiple regression. In future research, it

would be advantageous to include situational variables as well

as individual difference variables as moderators in the

relationship between local television news watching and

parasocial interaction.












CHAPTER 1
INTRODUCTION



Since Horton and Wohl (1956) introduced the concept of

parasocial interaction (PSI)--the television viewing

audience's relationship of "intimacy at a distance" with

television personae--the uses and gratifications researchers

have considered the PSI concept as deserving of investigation

(Auter & Palmgreen, 1992; Bowes & Rand, 1992; Levy, 1979;

Perse & Rubin, 1989; Rubin & McHugh, 1987; Rubin, Perse, &

Powell, 1985; Miller, 1983). But findings have been

inconsistent, showing low to moderate relationships between

parasocial interaction and other variables, including media

use and loneliness. Decreased social interaction

opportunities have also not been found to be strongly

associated with parasocial interaction (Levy, 1979; Nordlund,

1978; Rosengren & Windahl, 1972, 1977; Rosengren, Windahl,

Hakansson, & Johnsson-Smaragdi, 1976).

Parasocial interaction has been shown to be one of many

important motives in media use. A main assumption of PSI

research is that parasocial interaction relationships develop

over time in a manner similar to interpersonal relationships.

In other words, PSI relationships develop after repeated

exposure to the TV personae.






2

Another assumption underlying parasocial interaction

research is that every person has a basic need for social

interaction and that one uses the mass media to fulfill such

a need. However, the intensity of the need for social

interaction may vary from one individual to another. There

may also be several factors preventing a person from

fulfilling social interaction needs to a satisfactory extent.

For example, such needs may be difficult to fulfill due to

work environment or lack of opportunity for meeting others.

If one perceives that social interaction needs are not met to

a satisfactory degree, one will search for functional

alternatives for need fulfillment. A person may substitute

actual social interactions with parasocial interaction with TV

personalities. As Nordlund (1978) pointed out, how one seeks

to fulfill one's social needs is dependent on "relevant social

and psychological structures and the interplay between them"

(p.153). Nordlund suggested that a person "who is often media

interacting [parasocially interacting] will tend to turn to

the media, rather than to other individuals, in stressful and

otherwise difficult or frustrating situations, e.g., when one

is feeling lonely and the like" (p.154).

Parasocial interaction has been found to be an important

gratification sought or obtained by some television viewers

(Auter & Palmgreen, 1992; Palmgreen, Wenner, & Rayburn, 1980,

1981; Rubin & Perse, 1987a; Rubin, Perse, & Powell, 1985).

Parasocial interaction has been found to be correlated with






3

perceived importance and realism of a TV program, attention

level during watching, physical attraction to television

personalities, relationship development with TV personalities,

intention to watch TV programming, and affinity with TV or TV

personalities (Perse, 1990; Rubin & Perse, 1987a; Rubin &

Perse, 1987b; Rubin et al., 1985; Rubin & McHugh, 1987).

In a focus group interview, Levy (1979) found qualitative

evidence that audience members form their parasocial

relationships with television news personalities over repeated

exposures. Quantitatively, Rosengren and Windahl (1972) found

a significant relationship between television exposure and

parasocial interaction in two survey studies (r=.22, .26,

P<.001). Nordlund (1978) found an association between

television exposure and parasocial interaction -with television

entertainment programs in general (gamma=.32, p<.001). Levy

(1979) found a significant correlation between TV news

exposure and PSI (r=.22, p<.01), but they were not able to

determine the causal direction of this relationship.

On the other hand, Rubin, Perse, and Powell (1985) did

not find any relationship between local television news

exposure and parasocial interaction. They also did not find

significant relationships between loneliness and PSI for local

television news watching. But they argue that "lonely people

are watching and perceiving television news differently than

are nonlonely persons" (p. 173).






4

There appears to be a theoretical problem with a

sociologically oriented approach which looks at availability

of functional alternatives to gratify needs of social

interaction (Rubin, Perse, & Powell, 1985). Previous research

(Levy, 1979; Nordlund, 1978) found a weak association between

reduced social interaction opportunity and parasocial

interaction. Rubin, Perse, and Powell (1985) also give three

reasons for failure to find a strong relationship between

loneliness and parasocial interaction:

(1) the need for interaction is made up of both a need
for an intimate other and a need for general social
network; (2) the need for interaction is not constant
among individuals, nor is it constant throughout an
individual's lifespan; (3) individuals vary in their
abilities to obtain satisfying social interactions
(p. 157).

As Katz, Blumler, and Gurevitch (1974) suggested, many more

theoretical/empirical studies of psychological origins of

media gratifications are needed for understanding the uses

that people make of the media. By understanding the

personality traits of local television news audiences we

should be better able to explain the development of the

parasocial interaction process.

Horton and Wohl (1956) conceptualized parasocial

interaction as an imaginary, one-sided friendship that the

television audience has with a television "persona." Rubin

(1985) offered a conceptual model for the process of

parasocial interaction. Rubin's model emphasized the

psychological origins of media consumption motives and the






5

consequences of media use. In Rubin's study, the significant

contributors to the parasocial interaction were: news affinity

(beta=.48), perceived news realism (beta=.20), and the

information viewing motive (beta=.19). Unexpectedly, he found

no significant relationship between loneliness and parasocial

interaction in local television news watching. One reason was

that there was a little variance in the loneliness scores

among his sample. Another reason may be that he did not take

personality traits such as amount of extraversion into account

in his study.

Rubin, Perse, and Powell (1985) attempted to "link

psychological need to media use and its outcomes" (p. 171).

In contrast to previous research that emphasized demographic

and sociological antecedents of parasocial interaction, "it

was assumed that the psychological dimension of individual

needs could aid the explanation of parasocial interaction" (p.

172). Rubin et al. (1985) treated PSI as a consequence of

media use, but PSI can be treated as an antecedent to future

media use. They argued that "parasocial interaction is an

outcome of viewing at some point in time, but similar to the

affinity and realism concepts, it also should become an

affective orientation that influences subsequent media use and

behavior" (p. 174). Parasocial interaction has been

identified in both entertainment and news programs (Rubin,

Perse, & Powell, 1985; Perse & Rubin, 1989). In television

news studies, uses and gratifications researchers (Levy, 1979;






6

Palmgreen, Wenner, & Rayburn, 1980) identified friendliness,

companionship, and reality exploration through the TV

personality as elements of PSI. Parasocial interaction has

been found to be correlated with greater news affinity, or

perceived news importance, higher levels of perceived news

realism, attention, and involvement (Levy, 1979; Rubin &

Perse, 1987b; Rubin et al., 1985). These findings increase

understanding of the development of parasocial interaction

with local television news personalities.











CHAPTER 2
REVIEW OF LITERATURE


Statement of the Problem



Despite the spate of recent research, findings are

inconsistent about why people watch television news for

parasocial interaction. Most researchers have examined either

sociological or psychological antecedents to the development

of parasocial interaction and have found a weak relationship.

This lack of significant findings may exist because

researchers have not included in their parasocial interaction

model both sociological and psychological factors. Television

viewers are not equally active in watching television.

Individual differences as well as situational factors may

explain variability in parasocial interaction with local

television newscasters.

Conway and Rubin (1991) argue that uses and

gratifications research has been criticized for its failure to

link systematically media gratifications and their social and

psychological origins. Understanding the role of

psychological/personality variables in media dependence may

help explain parasocial interaction. Although research has

identified social characteristics of PSI, little has been done







to examine its psychological and personality origins

empirically.' This study attempts to examine the role of

personality variables in parasocial interaction.

Auter and Palmgreen (1992) give several possible reasons

for no direct relationship between TV exposure and parasocial

interaction. First, development of the 20-item scale by Rubin

et al. was not based on qualitative research. Second, the

self-report measure may not be suitable for measuring the

development of PSI over repeated exposures to TV persona.

Third, "because the survey focuses on a subject's favorite

personae, and is administered in absence of any actual

audience-message interaction, evidence of development over

time may be obscured" (pp. 14-15). Finally, the PSI scale

developed by Rubin et al. appear to measure a personality

trait rather than a development process.

Palmgreen, Wenner, and Rosengren (1984) pointed out that

there is a long-standing gap between the "psychological

origins" and the gratifications sought from the mass media.

The Eysenck's (1975) EPQ scale may provide us with some

theoretical direction for the psychological origins.2 If we

assume that media can meet social interaction needs, it would

seem that we should be most likely to find a relationship



'Examples of social characteristics are as follows: social
roles, frequency of social contact or organizational affiliation,
or work satisfaction.

Eysenck's personality dimension will be explained in detail
in the later section.





9

between media use for parasocial interaction and personality

factors. This is based on an assumption that the media,

especially television, are functional alternatives to social

interaction. From a functional perspective it makes sense

that media meet audience members' social interaction needs.

It is argued in this study that parasocial interaction in

local television news watching may be traced to personality

traits such as Eysenck's personality dimensions, including

Extraversion and Neuroticism, as well as situational factors

such as loneliness. In other words, the individual difference

perspective may, along with the sociological perspective,

contribute to the development of the parasocial interaction

concept. Local television news watching may serve the

functions of combating loneliness and substituting for face-

to-face interactions with friends and relatives.

Nordlund (1978) found positive correlations between

neuroticism and media exposure. He notes that "this could

partly be understood by the assumption that individuals with

a relatively high degree of neuroticism will experience

problems both in interacting socially with others and in

deriving subjective satisfaction from such interaction. Hence

they may try to 'interact' in a 'less demanding way' via the

mass media" (p. 161). He also found a positive relationship

between media exposure and PSI. He demonstrated some

significant relationships among leisure activity, neuroticism,

and media interaction [PSI], but he found no significant






10

relationship between media interaction and social interaction.

Thus, his findings indicate that parasocial interaction may

lead to increased media dependence, or vice versa, as well as

media exposure at times of loneliness.

Rosengren and Windahl (1977) found some evidence that

neuroticism and intraversion/extraversion were associated with

parasocial interaction with mass media content. For

neurotics, in particular, they showed some evidence for

parasocial interaction for both intraverts and extraverts.

For non-neurotics, they found that intraverts show the

strongest PSI when situational social interaction is low. But

when situational social interaction is high, non-neurotic

intraverts show the weakest PSI. They argue that "it may seem

quite natural, perhaps, that non-neurotic introverts with a

high interaction potential should have the lowest tendency for

involvement in the media content consumed. Presumably they

get enough real-life-interaction not to bother much about the

para-social interaction, etc. offered by the mass media"

(p.344). Rosengren and Windahl (1977) also found a positive

correlation between neuroticism and situational social

interaction.

Miller (1983) found a significant relationship between

parasocial interaction, amount of TV watching, and importance

of television in the lives of the isolated elderly viewers.

Most of previous parasocial interaction research has

concentrated on sociological and demographic aspects of the






11

development of the PSI process. Parasocial interaction also

has been seen among uses and gratifications researchers as an

antecedent to media use. But, in this study, PSI will be

treated as a consequence or an outcome of media use.

What factors or variables contribute to parasocial

interaction? In particular, what are some social and

psychological origins of parasocial interaction? The purpose

of this study is to empirically test a proposed model of the

process of parasocial interaction. Specifically, this study

attempts to find out what roles personality traits as well as

the situational variable loneliness might be playing in the

parasocial interaction process. In the 1980s, Gunter and his

associates found some significant relationships between

personality factors (e.g., the Eysenck's EPQ scores) and mass

communication processes. Based on the review of the

constructs of Eysenck's EPQ scale (extraversion/intraversion;

neuroticism/ stability; psychoticism/superego) in personality

psychology and theoretical perspectives of uses and

gratifications, this study empirically tests hypotheses about

the association between personality traits, loneliness, and

parasocial interaction with local television news

personalities.

To explicate the role of personality and situation in

media uses and effects, this study focused on parasocial

interaction with a local TV news personality. This study

selected local television news for several reasons. First,







12

local television news is a trusted and relied upon source of

information. Second, local news is widely viewed by people.

Third, local television news is "a mixture of information and

entertainment" (Perse, 1990b). As of late, local news

provides increased emphasis on human and local interest

stories as well as hard news stories. Finally, parasocial

interaction has received empirical support in local television

news watching (Houlberg, 1984) as well as in general news

watching (Levy, 1979; Palmgreen, Wenner, & Rayburn, 1980).

Thus, local news offers an opportunity to examine the

development of parasocial interaction.



Uses and Gratifications Approach

Within the active audience framework, uses and

gratifications researchers have examined the relationship

between the content of the mass media and audience members'

gratifications sought and obtained. The uses and

gratifications approach to mass media is based on three key

assumptions: 1) audiences are goal-directed in their behavior,

2) they are active media consumers, and 3) they are aware of

their needs and select mass media to satisfy those needs

(Katz, Blumler, & Gurevitch, 1974). In their 1974 volume,

Katz, Blumler, and Gurevitch describe the uses and

gratifications approach as one concerned with

(1) the social and psychological origins of (2) needs,
which generate (3) expectations of (4) the mass media or
other sources, which lead to (5) differential patterns of
media exposure (or engagement in other activities),





13

resulting in (6) need gratifications and (7) other
consequences, perhaps mostly unintended ones. (p. 20)

Most uses and gratifications research has concentrated on

elements 2, 3, 4, 5, and 6, paying less attention to elements

1 and 7 (social and psychological origins and media effects).

The research agenda seemed to contradict assumptions of uses

and gratifications research, as it originally developed out of

the effects research tradition and sought to utilize

gratifications as intervening variables in the study of media

use or media effects (Palmgreen, Wenner, & Rosengren, 1985).

Newton and Buck (1985) argue that "the uses and gratifications

approach has been criticized for its weak theoretical

development, for its reliance on the yet unproven assumption

that people know their needs and act rationally in meeting

them, and for ignoring social and environmental influences"

(p. 293).

Theoretically, the uses and gratifications approach is

basically a functional approach, which is based on the

assumption that the actions of the social world are

functionally interdependent. Uses and gratifications

researchers perceive television as performing various

functions for society as a whole, as well as for an individual

audience member. In other words, television news watching by

the audience members may be viewed as behavior that meets

needs generated through an interaction of one's psychological

dispositions and experience, which are a part of his or her

social situation (McQuail & Gurevitch, 1974).






14

Functionalism has received the most attention in the uses

and gratifications research literature. According to Little

(1991), functional explanations attempt to "explain a feature

of society in terms of the beneficial consequences it has for

the larger social system" (p. 91). Little (1991) uses the

term "quasi-teleology" for a functional explanation--a

tendency toward maintaining a state of equilibrium (e.g.,

social order, economic efficiency, or homogeneity of value

systems) against "the random perturbations of the

environment." Functionalism in social science regards social

systems as dynamic systems in which the subsystems play a

functional role. And the objective of the functional

explanation is to identify the role played by a particular

subsystem in relation to the social system as a whole.

Little (1991) is also concerned with the "causal history

requirement" for the functional explanation. The causal

history requirement holds that if B (Benefits) is the function

of P (Social institutions or practice) within S (Society), it

must be that: "P's disposition to produce B causes the

persistence of P in S." (Little, p. 95) In other words, "If

P ceased to have the disposition to produce B, then the random

fluctuations of social life would gradually undermine the

practice of P and P would disappear in S." (Little, pp. 95-96)

Therefore, if a functional relationship between subsystems

within a whole social system is to be considered as






15

explanatory, it is necessary to have some causal mechanisms

preserving the functional relationship.

According to Wright (1974), functions, as intended in the

theoretical paradigm presented many years ago, referred to the

consequences of certain routine, regular, and standardized

components of communications. As such, they were distinct

from the intended effects or purposes of the communicator and

from the intended use or motivation of the receiver. However,

Wright contends that satisfaction of needs may be seen as a

function of the mass media. Other mass communication

researchers seem to support this view.

The major assumption underlying functionalism is that the

actions and phenomena of the social world are functionally

interdependent (i.e., systematically related in causal chains

and circles) A more comprehensive list of the assumptions of

functionalism in relation to uses and gratifications approach

is provided by Katz, Blumler, and Gurevitch (1974, pp. 21-22):

1. Audience is active.

2. Much initiative in linking need gratifications and
media choice lies with the audience member.

3. The media compete with other sources of need
satisfaction.

4. Individuals are capable of understanding their needs
and reporting them in an intelligible manner.

5. Value judgments about the cultural significance of
mass communication should be suspended while studying
media gratifications.

6. Uses and gratifications research treats audience
requirement as intervening variables in the study of
traditional communication effects.







16

All of the articles in uses and gratifications drew

attention to the first assumption to the list--that media

behavior reflects prior interests and preferences. Blumler

(1979) agrees that the audience is active, but he argues that

this is an empirical question as well.

The development of uses and gratifications research may

be divided into four phases: (1) description phase, (2)

operationalization/typologies phase, (3) theoretical

integration (extension) phase, and (4) theory construction

phase (Palmgreen, Wenner, & Rosengren, 1985) The first phase

of development, as observed by Blumler and Katz (1974, p. 13),

offered "insightful description of audience subgroup

orientations to selected media content forms," giving rise to

conceptual and methodological problems. But the second phase

of development in media gratifications studies emphasized

"operationalization of the social and psychological variables

presumed to give rise to differentiated patterns of media

consumptions" as well as concern with typologies (e.g.,

surveillance, correlation, socialization, and entertainment).

The third phase of development, according to Blumler and

Katz (1974, p. 13) was concentrated on "attempts to use

gratifications data to provide explanation of such other

facets of the communication process with which audience

motives and expectations may be connected." In other words,

after long periods in which uses and gratifications

researchers primarily emphasized description and measurement





17

of audience uses and motives, the main concentration shifted

in the mid-1970s to building the theory. During the fourth

phase of development, many media gratification researchers

indeed concentrated on the specification and testing of

hypotheses about gratifications and media use, the

relationship between gratifications sought and obtained, the

social and psychological origins of media use, and

gratifications and media effects (Palmgreen, Wenner, &

Rosengren, 1985).

McQuail (1985) mapped out the field of theory, research,

and critical perspective, and regarded the uses and

gratifications approach as "multitheoretical" (not as

theoreticala" as some criticize). He also provided a number

of research questions deserving of investigation:

1. Media gratifications--their nature and substance: What
are gratifications or similar concepts of satisfaction,
use, motive, and so on? How should they be
differentiated and labeled? What gratifications goes
with what content? How is their expression distributed
in a given population? How are separate gratifications
interrelated?

2. Gratifications and measure: How do expectations relate
to kind and amount of media use? Do they predict use?
Are they consistent with use? Is choice active, guided
by expectations in advance? How are expectations related
to eventual satisfactions, by way of media use?

3. Social origins and media gratifications: Do expressed
gratifications vary with social background and current
circumstances? What processes underlie patterns of
interrelationship--for instance, escape, compensation,
adjustment, or chance?

4. Gratifications and effect: Do gratifications predict
measured effects on behavior, on learning or change in
opinions, attitudes, or the like? What kinds of
gratification go with what kinds of content, with what






18

kinds of effect? What are the associated conditions of
the gratification-effect relationship?

5. The sequence--originations--motivations--media use--
satisfactions--consequences/effects: Can such sequences
be empirically discovered or alternative orderings of the
same elements? What is the relative importance of
different steps and the nature of relationships between
them? (pp. 163-164)


Development of Parasocial Interaction (PSI)

The concept of parasocial interaction falls into the uses

and gratifications research tradition. However, it is a

specific type of media gratifications that has been poorly

understood. The development of the parasocial interaction

concept will be presented and discussed below.

Horton and Wohl (1956) introduced the concept of

parasocial interaction. Parasocial interaction (PSI) has been

defined as follows: "simulacrum of conversational give and

take" between television 'personae' and their audience (Horton

& Wohl, 1956, p. 215); "a perceived interpersonal relationship

on the part of a television viewer with a mass media persona"

(Perse & Rubin, 1989, p. 59); "the interaction with somebody

of the mass media world more or less as if he were present in

person, without losing even momentarily one's identity"

(Rosengren & Windahl, 1972, p. 173); "constituent elements of

an overarching, more multifaceted phenomenon of audience

involvement in certain portions of media fare" (Nordlund,

1978, p. 151); "interpersonal involvement of the media user

with what he or she consumes" (Rubin, Perse, & Powell, 1985,





19

p. 156); "a sense of affective interpersonal involvement with

media personalities" (Rubin & Perse, 1988, p. 248).

In other words, parasocial interaction is the simulated

interaction between TV "personae" and audience. Horton and

Wohl (1956) expected the audience to "benefit by his [the

personae's] wisdom, reflect on his advice, sympathize with him

in his difficulties, forgive his mistakes, buy the products

that he recommends, and keep his sponsor informed of the

esteem in which he is held" (p.219). In short, the television

audience is supposed to interact with the "personae"--the quiz

show hosts, the reporters, or the newscasters. They wrote:

the TV personality "faces the spectator [TV viewer], uses the

mode of direct address, [and] talks as if he were conversing

personally and privately [with the TV viewer]" (p. 215). On

the part of television viewers, they should try to know the

television personality just as they know their close friends,

and they try to take part in the illusion of a face-to-face

interaction with the personality rather than passive

observation. In addition, over time the television audience

would start believing that they knew the television

personalities intimately and that they understood the values

of the persona. Accordingly, the television viewers would

consider the TV personality as a close friend.

Television even offers opportunities for role playing for

which the television audiences find no opportunity in their

social environment. Horton and Wohl (1956) call this function






20

of parasocial interaction compensatory because "it provides

the socially and psychologically isolated with a chance to

enjoy the elixir [cure-all] of sociability" (p. 222).

Horton and Wohl (1956) argue that television plays roles

in satisfying audiences' social needs. In other words, the

television personality serves as a substitute friend for some

people who do not have sufficient social interaction to meet

their social needs. "The personality [TV] program," as Horton

and Wohl (1956, p. 223) wrote, "... is particularly favorable

to the formation of compensatory attachments by the socially

isolated, the socially inept, the aged and invalid, the timid

and rejected." Horton and Wohl (1956) also pointed out that

the intimacy of prolonged parasocial interaction with the

television personalities may increase the amount of television

watching. "They 'know' such a persona in somewhat the same

way they know their chosen friends: through direct observation

and interpretation of his appearance, his gestures and voice,

his conversation and conduct in a variety of situations" (p.

216).

In a later work, Horton and Strauss (1957) suggest that

television producers and personalities manipulate the

technical components of production in audience-participation

programs in order to control the studio audiences' reactions.

For example, the television personality controls types of

camera shots to maintain the illusion of intimacy. Even

though the works of Horton, Wohl, and Strauss (1956, 1957)






21

established the parasocial interaction concept, they did not

provide empirical evidence for their new concept.

Early works that provided some verification of the

parasocial interaction concept were Rosengren and others

(1972, 1976, & 1977). From social and social psychological

perspectives, Rosengren and Windahl (1972) examined

relationships between a viewer's social interaction needs and

his or her functional alternatives to satisfy these needs, and

found some evidence for the parasocial interaction concept.

Cathcart (1969) questioned what motivated television

viewers to faithfully watch a particular television

newscaster. He stated that there may be something that

attracts television viewers to the newscaster. There may be

something in television newscasters that satisfies viewer

needs and desires. Cathcart suggests that "on both the

national and local levels, the newscasting 'person' has been

replaced by today's newscasting 'personality.'" Koenig and

Lessan (1985) termed TV newscasters "quasi-friends" because

viewers felt closer to TV journalists than to acquaintances.

Based on focus group interviews with 32 adults, Houlberg

(1984) indicated the existence of a parasocial interaction

with television newscasters. In his study, he found the PSI

dimension emerged as a stronger factor than physical or

professional attributes of the newscaster. His suggestion for

future research was to study how the parasocial interaction

was formed.






22

Based on Houlberg's (1984) tripartite division of

newscaster attributes into physical (gender, age, general

attractiveness, voice and hair), professional (competence,

fairness, knowledge, accuracy and being articulate) and

parasocial (reassuring, understanding, warm, friendly and

familiar), Bowes and Rand (1992) examined general differences

between audience and professionals in evaluating television

personalities. While importance ratings were close between

the audience members and broadcast professionals for

professional attributes for TV newscasters, they differed for

the parasocial and physical attributes. Professionals put

greater emphasis on parasocial and physical attributes than

did audience members.

Sociological Origins of PSI

According to Blumler (1979, p. 21), most uses and

gratifications researchers "have always been strongly opposed

to 'mass audience' terminology as a way of labelling the

collectivities that watch TV shows, attend movies, and read

magazines and newspapers in their millions." From this

perspective, many of the media-related needs of audience

members may derive from interaction with their social

environment. Blumler postulated three primary social origins

of media gratifications: (1) normative influences, which lead

to certain expectations based on life-cycle position, social

roles, etc.; (2) socially distributed life-chances,

constituting factors facilitating a richer involvement with






23

contents of media (e.g, frequency of social contact and

organizational affiliation); (3) the subjective reaction or

adjustment of the individual to his or her situation (e.g.,

work or role satisfaction).

McQuail, Blumler, and Brown (1972), in their study of

gratifications types of the British radio serial The Dales,

found some evidence that the audience had a vicarious

relationship or substitute companionship with media

personalities. The most frequently mentioned items were: 1)

"It is good company when you're alone"; 2) "The characters

have become like close friends to me"; 3) "I like the sound of

the characters' voices in my house" (p. 157). They even

illustrated that parasocial interaction with characters from

the British program The Dales was the strongest among

"solitary" listeners: "the characters [from The Dales] may

become virtually real, knowable and cherished individuals, and

their voices are more than just a comforting background which

breaks the silence of empty house" (p.157).

Rosengren and Windahl (1972) hypothesized that the more

dependent a viewer is on mass media as "purveyors of

functional alternatives to real interaction," the higher the

degree of involvement. In turn, the higher the degree of

involvement, the larger the mass media usage. It was expected

that "effects of mass media consumption should vary not only

with amount of consumption, but also with degree of

dependence, motives for seeking the functional alternatives







24

offered by the mass media, degree of involvement and degree of

reality proximity of the content consumed" (p.176). Rosengren

and Windahl (1972) found significant relationships between PSI

(degree of involvement), interaction potential (composed of

three indices: status, means and opportunities for

interaction, and partner availability), actual interaction and

media use. They found a negative correlation between

interaction potential and PSI. In other words, a person

isolated from social interaction will be more likely to use

media for parasocial interaction. They also found that mass

media use was significantly related to parasocial interaction.

Thus, the greater one's parasocial interaction with television

personalities, the more television news one will watch. When

a person does not have satisfactory individual and

environmental possibilities to satisfy the need for social

interaction, then as a substitute for the need for interaction

one may attend to such mass media as radio and television.

Rosengren and Windahl's (1972) work is important to our

understanding of PSI for two reasons: 1) It provides a strong

conceptual framework within which the PSI concept can fit, and

2) it operationalizes the many variables related to the PSI

concept.

Rosengren, Windahl, Hakansson, and Johnsson-Smaragdi

(1976) suggested that parasocial interaction and solitary

identification, when combined, result in what they term

"capture." Their findings did not support the relationship






25

between PSI and social and psychological variables. The PSI

scale was not significantly correlated with interaction

potential (sociometric status, partners, and opportunity for

interaction), social interaction, and amount of TV watching.

But Rosengren et al. (1976) found evidence that lack of

sibling interaction leads to capture, in turn giving rise to

long-term identification with a TV personality. Unexpectedly,

they found negative relationships between TV watching and both

capture and long-term identification. For their sample

(school children from 11 to 15 years old), television watching

was not a functional alternative to social interaction. They

argued that their theoretical model originally for adults did

not seem to apply to 11- to 15-year-old adolescents.

Psychological Origins of PSI

Psychological factors may also point to origins of

parasocial interaction. For example, McGuire examined a large

number of personality theories for their usefulness in

generating hypotheses about media gratifications.

McGuire (1974) divided the eight psychological theories

of cognitive and the eight theories of affective motivations

for mass media use by initiation (active vs. passive),

orientation (internal vs. external), and stability

(preservation vs. growth), the former emphasizing "the

person's information processing and attainment of ideational

states, while the latter stressing "the person's feelings and

attainment of certain emotional states." McGuire's Expressive






26

theories (active) suggest that through parasocial interaction

with the TV characters, one can "vicariously express one's

feelings rewardingly."

Conway and Rubin (1991) empirically tested McGuire's

eight selected active psychological theories representing

cognitive and affective orientations. They assessed whether

psychological variables, based on McGuire's (1974) theoretical

paradigm, could explain motivation for using television. They

found that PSI was significantly correlated with information,

entertainment, relaxation, and pass-time motives. Consistent

with past research, parasocial interaction was positively

associated with information viewing. Rubin et al. (1985)

found that the information viewing motive predicted parasocial

interaction with local television newscasters. Conway and

Rubin (1991) indicated that psychological characteristics

mediate media exposure and parasocial interaction. But, there

is not much research examining the role of psychological

variables in media uses and effects.

Recently, Perse and Rubin (1990) pointed out that social

and psychological characteristics of audience members mediate

mass media effects. In contrast to previous assumptions

(Rosengren & Windahl, 1972; Rubin, Perse, & Powell, 1985) that

the need to compensate for psychological deficiencies results

in goal-directed media use, Perse and Rubin (1990) assume that

chronic loneliness leads to passive use of local television

news.









Personality and PSI

Rosengren and Windahl (1977) assumed that there are two

possibilities for satisfying the need for social interaction:

individual and situational. They result in four possible

combinations of degree of dependence on mass media as

functional alternatives for the social interaction need: high-

high, high-low, low-high, and low-low (a typology of

individual and situational possibilities for satisfying social

interaction need). For example, a person with low individual

and situational potentials for social interaction is more

likely to watch television for parasocial interaction than one

with high individual and situational possibilities. The

incidence of PSI can be either more or less likely for

extraverts or intraverts. Is an extravert satisfied with his

or her social interaction so as not to watch television for

parasocial interaction, or does he or she need more social

interaction? Is an intravert not satisfied with his or her

social interaction and wants more parasocial interaction than

an extravert, or does he or she need not social interaction?

Based on Eysenck and Eysenck's (1969) EPI scale, Rosengren and

Windahl divided between two types of intraverts: neurotic and

non-neurotic intraverts. Neurotic intraverts are the 'real'

intraverts, "who for some reason or another feel it difficult

to interact socially," while non-neurotic intraverts "do not

seem to care very much about the whole thing" (p. 342).






28

Nordlund (1978) attempted to test hypotheses about the

social and psychological origins of parasocial interaction.

Nordlund called PSI "media interaction." In contrast to

conceptualization of Rosengren and his colleagues (1972, 1976,

1977) in which parasocial interaction and solitary

identification are distinct variables which, when combined,

result in what they term "capture," (PSI + Solitary

identification -> Capture). Nordlund (1978) considered PSI as

"constituent elements of an overarching, more multifaceted

phenomenon of audience involvement" in media contents: "Such

involvement may range from readiness to be preoccupied with

and muse upon media characters, to talk about them with

others, to relate their circumstances to one's own, to feel as

if in some relationship with them and, at the extreme, to get

fully involved in their fate" (p.151).

Media Use, Motives, and PSI

Miller (1983) suggested that "a news personality who is

on television every day and who is greatly relied upon by

thousands or perhaps millions is going to be high on any

parasocial scale" (p. 35). Miller also suggested two levels

of parasocial interaction--cognitive and affective: "there is

a need for information and intellectual stimulation from the

media personae as well as a general likability of (or affinity

toward) the media personae" (p. 40). Miller found significant

positive relationships between parasocial interaction,






29

television exposure, and the importance of television among

the elderly.

Levy (1979) suggested that psychological and

environmental (situational) circumstances generate the

expectations and needs that a person brings to television. He

argues that needs do not arise in a social vacuum but are the

product of interactions between personality and social

environment. A television anchorperson who appears frequently

on television encourages PSI by "speaking in conversational

tones directly into the television camera, by engaging in

clever monologues which appear to require audience

reciprocity, and by interacting in a casual way with other

media communicators ('side-kicks,' confidants, antagonists,

and the like)" (p. 69). For most audiences, PSI is regarded

as complementary to social interaction. However, for people

with few or weak social ties, parasocial interaction may give

a functional alternative for inadequate interaction

opportunities. Levy argued that "to demonstrate that news

audiences interact on a para-social level with news personae,

it is necessary to show not only that viewers interpret and

evaluate the behavior or 'gestures' of the newscasters, but

that audiences also act or react based on those attributed

meanings" (p. 70).

Levy's hypotheses assume that viewers value their

parasocial interaction and that after an initial period during

which the PSI tie is formed, audience members will try to





30

maintain that valued relationship by continuing to watch the

television news at a relatively high rate. Levy (1979) found

a negative relationship between gregariousness and PSI (r=-

.12, 2<.05). Parasocial interaction was also significantly

correlated with television news watching (r=.22, p<.01).

Based on focus group interviews, Levy made a distinction

between cognitive and affective PSI. He found about one third

of his participants use the TV newscaster as a "cognitive

guide," comparing their own ideas to those of the newscaster.

As for the affective quality of the newscaster, about two-

thirds believed that the banter, jokes, and 'happy talk'

between local newscasters made the news easier to take. Levy

concludes his findings as follows:

People who watch television news engage in varying
degrees of para-social interaction with the news
personae. Those viewers who find the para-social
relationship particularly attractive or gratifying
increase their exposure in order to increase their
'contact' with the news personae. While it is
possible that there is a threshold of exposure
beyond which individuals will not increase their
viewing in order to increase their para-social
interactions, it is also likely that establishing
and maintaining para-social interaction with the
news personae is an important determinant of how
much television news some people will watch. (p.
78)

Levy (1979) also found that people with limited social

interaction, such as the elderly, perceived television

newscasters as substitute friends.

Rubin, Perse, and Powell (1985) took a psychologically

oriented approach to examine the relationship between

loneliness and parasocial interaction. They developed a





31

conceptual model explaining parasocial interaction from both

a social interaction need due to loneliness and instrumental

television news watching. They pointed out that "involvement

may take many forms including seeking guidance from a media

persona, seeing media personalities as friends, imagining

being part of a favorite program's social world, and desiring

to meet media performers" (pp. 156-157). The parasocial

relationship is nourished by a combination of factors: "degree

of reality approximation of the persona and the media,

frequency and consistency of appearance by the persona,

stylized behavior and conversational manner of the persona,

and effective use of the formal features of television" (Rubin

et al., 1985, p. 156). The factors work together to make the

TV personality a "predictable, nonthreatening, and, hence,

perfect role partner" for the audience members. Their PSI

scale contains elements of "empathy, perceived similarity, and

physical attraction." Their hierarchical multiple regression

analysis indicates that "news affinity" (beta =.48) (e.g.

perceived importance of watching TV news), "perceived news

realism" (beta=.20), and "the information viewing motive"

(beta=.19) are "the only substantial, positive contributors to

the explanation of parasocial interaction." Rubin et al.

(1985) found a significant relationship between instrumental

television use and parasocial interaction with local

television news personalities. They found that information

seeking viewing motives are associated with parasocial





32

interaction with local television newscasters. Rubin et al.

(1985) also found that perceived news realism and news

affinity explain parasocial interaction with local television

newscasters.

Finn and Gorr (1988) examined the relationship between

individual differences and motivations for television

watching. Their factor analysis of the eight TV viewing

motivations using oblique rotation supported two distinctive

dimensions for TV viewing motivations -- social compensation

(e.g., the companionship, pass time, habit, and escape

motivations) and mood management (e.g., relaxation,

entertainment, arousal, and information motivations)

(Eigenvalue=3.87, percentage of total variance=48.4 for social

compensation and 1.41 and 17.7 for mood management). Shyness

was positively correlated with social compensation motives

(r=.30, p<.001), but loneliness was not significantly related

with the motives (r=.09, p>.05). Finally, self-esteem was

negatively related to the social compensation motives (r=-.24,

p<.001), while Appraisal (r=.20, p<.001) and Belonging (r=.ll,

p<.05) correlated positively with the mood management motive.

Rubin and Perse (1988) found that viewing attention and

perceived realism of soap opera content were significantly

related to parasocial interaction. Their findings suggest

that parasocial interaction reflects attention to

realistically perceived content during the TV watching

experience. Parasocial interaction was also correlated with






33

more instrumental viewing motives, stronger attitudes, and

greater audience activity.

Auter and Palmgreen (1992) found a significant

relationship between parasocial interaction and television

exposure (r=.28, p<.0005), but they noted that it was not

clear whether exposure leads to parasocial interaction or vice

versa. They also argued that parasocial interaction is a

construct consisting of at least four dimensions: 1)

Identification with Favorite Character, 2) Interest in

Favorite Character, 3) Group Identification/ Interaction, and

4) Favorite Character's Problem Solving Ability.

Auter and Palmgreen conducted an experiment in which

"levels of exposure to a particular situation comedy are

manipulated, and post-viewing levels of PSI" are measured with

the Audience Persona Interaction scale. Their experiment may

yield causal evidence that greater exposure results in higher

levels of parasocial interaction.

They argue that the PSI processes tapped by the four

subdimensions of the API post-viewing scale are rather general

processes which may apply to other television genres, not just

comedy. Finally, Auter and Palmgreen (1992) suggested:

The differences between our scale developed for
situation comedies and the Rubin scale developed
originally for television news suggest that
different content genres may involve somewhat
different parasocial processes. On the other hand,
the similarities among the various PSI indices
developed over the years, including the API scale,
imply that the fundamental nature of parasocial
interaction with television personae is the same
regardless of content. Only continued research on






34

this important process involving a variety of
genres can provide answers to such questions" (p.
25).

Loneliness and PSI

Perlman and Peplau (1981, p. 31) conceptualized

loneliness as "the unpleasant experience that occurs when a

person's network of social relations is deficient in some

important way, either quantitatively or qualitatively."

Rubin, Perse, and Powell (1985, p. 158) defined loneliness as

"a descrepancy between the amount of interaction individuals

need and the amount that they perceive is fulfilled." Perse

and Rubin (1990) conceptualized loneliness as "reduced

personal interaction and somewhat greater media use." On the

other hand, chronic loneliness was defined as a

"psychological, apathetic state that decreases motivation to

communicate interpersonally" (pp. 47-48). A sense of

decreased quantity and quality of interpersonal interaction

accompanies chronic loneliness. Thus, people watch television

to reduce boredom.

The UCLA Loneliness Scale (Version 2 & 3) is the most

frequently used loneliness measure (Russell & Cutrona, 1988;

Russell, Peplau, & Cutrona, 1980). The original and revised

UCLA Loneliness Scales have been employed in hundreds of

loneliness studies. Version 2 of the UCLA scale has been used

in studies of loneliness and parasocial interaction (Finn &

Gorr, 1988; Rubin et al., 1985). In conducting a survey of

the elderly, Russell et al. found the elderly encountered






35

difficulties due to the high reading level required by Version

2 of the UCLA scale. They created Version 3 of the UCLA

Loneliness Scale to solve these problems. It has well-

established reliability and validity and consists of eleven

statements worded in the "lonely" direction and nine in the

"nonlonely" direction. Version 3 of the scale has been

administered to a variety of different populations, including

college students (Cronbach's alpha=.92), nurses (alpha=.94),

teachers (alpha=.89), and the elderly (alpha=.89) (Russell &

Cutrona, 1988). It should be noted that the word "loneliness"

does not appear in any of the statements and the scale is

unidimensional.

Based on the assumption from the uses and gratifications

approach that "social psychological motivations may cause

people to turn to mass media for companionship and other

gratifications," Perloff, Quarles, and Drutz (1983, p. 353)

predicted that situational variables (e.g., dating

involvement, number of friends, and age) cause dispositional

variables (e.g., dissatisfaction with relationships and

depression), leading to television exposure and

gratifications. The absence of dating behavior in college

students was found to have a strong impact on feelings of

dissatisfaction with intimate relationships (beta = -.61,

p<.05). Also, dating involvement has direct effects on TV

consumption (beta = -.15, P<.05). Dissatisfaction mediated

the relationship between dating involvement and TV consumption






36

(beta = .26, P<.05). Dissatisfaction also had a significant

effect on depression (beta = .31, R<.05), but depression was

not significantly related to TV exposure.

Perse and Rubin (1990) found that chronic loneliness was

associated with reduced interpersonal interaction, increased

television watching, and passive television use. They suggest

that loneliness and mass communication is a promising area for

examining the development of parasocial interaction.

Rubin, Perse, and Powell (1985) examined the relationship

between loneliness and parasocial interaction in local

television news watching. They did not find any significant

relationship between loneliness and parasocial interaction but

found a significant negative relationship between loneliness

and the learning motive for watching local television news.

They offered two reasons for the findings: (1) most of the

sample might not be very lonely, and (2) news watching is more

likely motivated by "need states," such as a need for

orientation, than by loneliness.

The conceptual model of Rubin, Perse, and Powell's (1985)

study was:

(1) An individual who experiences a greater degree
of loneliness will seek, in lieu of perceived or
actual functional alternatives, (2) to use local
television news in a (3) goal-directed,
instrumental manner. (4) This instrumental use will
exhibit characteristics of (5) news program
dependency, namely higher news program viewing
levels, longer duration of news viewing, and
greater affinity with and perceived realism of
local television news. This pattern of local
television news viewing will result in (6)








parasocial interaction with favorite local
television news casters (pp. 161-162).

PSI from Interpersonal Perspective

Rubin and McHugh (1987) explored the development of

parasocial interaction by applying uncertainty reduction

theory. Their findings indicated a path from "(a) social and

task attraction to (b) parasocial interaction to (c) a sense

of relationship importance" (p. 279). Rubin and McHugh's

study examined a parallel pattern of relationship development

with television personalities, applying principles from both

uses and gratifications and uncertainty reduction theories to

understand the development of the parasocial interaction

process. Rubin and McHugh (1987) argued that interpersonal

and mediated relationships follow a similar process of

relationship development. In other words, they suggest that

PSI with TV personalities can be a functional alternative to

interpersonal relationships. The path model in Rubin and

McHugh's study demonstrated that parasocial interaction is a

significant part of mediated relationship development.

Attraction leads to parasocial interaction, and perceived

relationship development importance is mediated by parasocial

interaction. They argue that relationship development with a

TV persona is not a matter of sheer TV exposure, but a

function of attraction that results in parasocial interaction.

Perse and Rubin (1989) explored parasocial interaction

from the interpersonal attribution perspectives. Their

findings demonstrated that PSI with favorite soap opera





38

characters was based on "reduction of uncertainty and the

ability to predict accurately the feelings and attitudes of

the persona" (p. 59). They suggested that we need to

understand the cognitive as well as affective foundations of

PSI development. Parasocial interaction is based on the

belief that the television newscaster is like other people in

the audience's social circle. As Perse and Rubin (1989)

noted, like social interaction, parasocial interaction

develops over a period of time and is reinforced when

television presentations are similar to interpersonal

interaction. For instance, television close-ups let audience

members read a TV persona's nonverbal gestures and allow the

personality to create an illusion of a two-sided relationship

through scripted reactions to anticipated comments of viewers.

Citing Horton and Wohl (1956), Perse and Rubin (1989) pointed

out: "some media personae establish mutual eye gaze during

conversations with the audience, confide in the audience

during close-up asides, and express their fidelity to the

audience and ask for reciprocity. Indeed, a media personality

is a perfect friend--dependable, discreet, and uncritical" (p.

61).

Perse (1990b) examined the contribution of cognitive and

emotional involvement to parasocial interaction with local

television news personalities. Perse found PSI was explained

by higher levels of news realism, less recognition of news,

and feeling happy while viewing the news. Perse argues that






39

happiness is significantly related to PSI because it is a

pleasant emotion signalling interpersonal interaction. Perse

(1990b) provides three reasons why PSI develops out of seeking

information from news.

First, viewers seeking information may also acquire
knowledge about the news personality. Increases in
interpersonal knowledge are an important antecedent
to relationship development. Second, friendship is
based, in part, on how others help us meet needs.
So, newscasters who provide information to
information-seeking viewers will be perceived more
favorably. Third, newscaster credibility is based
on interpersonal perceptions such as dynamism,
extroversion, and sociability (p. 21).

Television as Significant Others

Koenig and Lessan (1985) examined distance in semantic

space between "the self" of television viewers and friends,

acquaintances, and television personalities. They found that

a TV audience member puts his or her "self-concept" to

favorite TV personality in closeness between friends and

acquaintances. They argue that the relationship between

viewer and persona is "not true interaction" in the sense that

both are "self-other-role playing," responding to each others'

cues, behaving in terms of expected responses. But some

social realities exist as the television character interacts

with other characters or looks directly at the audience and

acts as if he or she were talking one on one. Citing Horton

and Wohl, they note that television personalities "appear

regularly in one's home, and one's whole weekly schedule can

be arranged around their electronic presence." For example,

a TV pe sona takes on identifiable features, always speaking






40

in the predictable pattern. Few people have friends counted

on to the same degree. The television personality has the

capability to fill the position of role partner in someone's

life. In addition to the personality's characteristics of

being attractive, trusted and friendly, the regularity,

predictability, and immediacy in television viewing make it

possible for television personalities to take on some

functions of friends. Consequently, their presence in homes

can reduce loneliness and boredom for the audience members.

Newton and Buck (1985) emphasized the role of television

as "significant other" and the value of "TV and me" as an

index of parasocial interaction. Based on symbolic

interactionism in which "individuals develop self-concepts

through interaction with significant others," they assume that

"television is for some viewers a type of significant other

against which--through a type of parasocial interaction

process--viewers develop, maintain, and revise their self-

concepts" (pp. 293-294). From the symbolic interactionist

perspective, Newton and Buck examined the relationships

between self-concept and television. Specifically, they

treated television as significant other by examining how close

people put themselves to television. As Newton and Buck put

it, the roles and attributions that we attribute to ourselves

are much closer to our relationship with television in a type

of parasocial interaction than with actual viewing. Thus, the

proximity of "TV and me" may be a better indicator of






41

television as a significant social and psychological influence

than is simple television exposure.3 The overall low

correlations with "TV and me" and exposure suggest that some

individuals may not watch television a great deal, but still

feel a close affinity with what is shown on television.

Newton and Buck (1985) noted that the social and psychological

significance to a viewer is more than simply the amount of

time he or she spends watching television.

Based on discourse analysis, Mancini (1988) explored a

set of "meta-discoursive" indicators, through which "it should

be possible to define and analyze the role the television

journalist assumes before his audience, and the kind of

interaction he tries to develop with the televiewer" (p. 151).

In his formulated "hypothesis of simulated interactions,"

Mancini suggested that the TV personality plays an explicitly

social role and simultaneously assumes a specific social role

for the audience: "The television presenter, in looking into

the camera, looks into the eyes of the viewer, speaks directly

and personally to each individual viewer, establishing with

him a rapport which is falsely based on the myth of a real,

interactive presence" (p. 155).

Operationalization of PSI

A number of different scales have been developed over the

years to measure viewers' parasocial interaction with


'Newton and Buck (1985) look at television from a symbolic
interactionist perspective as a type of significant other,
examining the associations between self-perceptions and TV.






42

television personalities (Auter & Palmgreen, 1992; Houlberg,

1984; Nordlund, 1978; Levy, 1979; Rosengren & Windahl, 1972;

Rosengren, Windahl, Hakansson, & Johnsson-Smaragdi, 1976;

Rubin & Perse, 1987a; Rubin, Perse, & Powell, 1985). As Auter

and Palmgreen (1992) argued:

most PSI scales have not been able to provide any
solid evidence of the development of parasocial
interaction over repeated exposures to program
persona. This may be due to the fact that these
scales are not specifically designed to measure the
development of parasocial interaction between
viewer's predisposition to parasocially interact
with their favorite persona. As such, they seem to
tap a personality trait, rather than a
developmental process (p. 2).

Rosengren and Windahl (1972) first measured parasocial

interaction in their four-cell "degree of involvement"

typology. In developing a theoretical model of parasocial

interaction, Rosengren and Windahl (1972) provide a typology

of social relations between the audience and television

personality: 1) detachment (neither interaction nor

identification with the TV personality), 2) solitary

identification (no interaction but identification with the TV

personality), 3) parasocial interaction (imaginary interaction

but not identification with the TV personality), and 4)

capture (both interaction and identification with the TV

personality). The typology represents levels of

identification ("the act of imagining oneself to be in the

place of another person") and interaction ("mutual stimulation

and response") that the viewers have with the television







43

personalities (p.172). The authors pointed out that they were

not satisfied with the measure.

A much stronger measure was developed by Rosengren,

Windahl, Hakansson, and Johnsson-Smaragdi (1976) who found

some validity (convergent and discriminant) for three scales

(PSI, capture, and long-term identification ) by means of the

multitrait-multimethod technique. The scales of PSI (e.g., I

often think that the people I see in this program almost

become old friends), capture (e.g., Sometimes when I'm

watching this program, I believe that I'm really one of the

people in the story), and long-term identification (e.g.,

Often when I've seen this program, I think about it for a long

time afterwards) were significantly related to each other.

The authors found that the PSI scale had the highest validity,

followed by the long-term identification scale and the capture

scale. There was only a weak correlation between PSI and

television exposure. Their operationalization of parasocial

interaction was problematic. Recently, Auter and Palmagreen

(1992) summed up the problems:

It is possible that the 3-item PSI measure was not
sensitive enough to capture the relationship
between amount of TV viewing and PSI in the cross-
sectional survey setting. Rosengren et al. also
found that their new PSI scale did not correlate
with their 1972 measure. While it seems that
Rosengren et al.'s (1976) series of 'degree of
involvement' scales could be useful uses and
gratifications measures, they have not been used in
further published studies (pp. 10-11).

Nordlund (1978) examined relationships between four key

variables and parasocial interaction: "(a) those






44

social/psychological structures that affect need fulfillment

possibilities, relating these to (b) mass media exposure, and

(c) media relations including especially media interaction and

ending with (d) some consequences of the latter" (p.154). He

provides his own media interaction (parasocial interaction)

model: 1) If social and personality structure prevents

fulfillment of the social interaction need, one will use more

media with a high degree of parasocial interaction with media

personalities (a & b); 2) there will be a positive

relationship between media exposure and PSI (c) ; 3) If one

uses media for parasocial interaction, one will be more

dependent on the mass media (d). The social structural

variables in the study were "social interaction" ("a measure

of the number of persons that comprised a respondent's friends

and acquaintances"), and "leisure activity style" ("a measure

of the variety and extent of the respondent's customary

leisure pursuits" (p.158). The psychological structural

variable was "level of neuroticism" as measured by one

dimension of Eysenck and Eysenck's (1969) EPI. The dimensions

underlying the media interaction [PSI] items included "showing

interest in persons in the content, talking with others about

persons in the content, relaxing, withdrawing from reality,

'participating' in what is happening in the media, and

'knowing' the characters in the media content" (p. 159). The

two "consequence" variables included the mass media dependency

and the extent to which different media are used when lonely.






45

Levy (1979) developed a scale of television news PSI

after focus group interviews with regular TV news viewers.

From a propositional inventory of 42 gratifications

statements, Levy administered seven parasocial interaction

items to 240 adults in the Albany County, New York region.

But, Levy's PSI measure has not been subjected to extensive

tests of its validity, and its reliability (a=.68) was low.

Based on focus group interviews with 32 adults, Houlberg

(1984) developed an 18 item Likert-type scale, consisting of

6 PSI items, 6 perceived newscaster professional

qualifications, and 6 newscaster physical attributes.

Houlberg factor analyzed the 18 items and found the PSI items

as the strongest factor, accounting for 26.7 percent of the

total variance.

Rubin, Perse, and Powell (1985) developed the most

reliable parasocial interaction scale (see Appendix B). Their

original study looked at PSI and local television news

viewing. They administered their PSI scale to 329 local

television news audience members. The reliability coefficient

for the scale was .93. A revised 10-item version of PSI scale

was later introduced by Rubin and Perse (1987a) The revised

PSI scale had a .88 Cronbach alpha and correlated highly with

the 1985 PSI scale (r=.96, p<.001). The two measures have

been used to study both entertainment and news programs.

Recently, Auter and Palmgreen (1992) suggested that PSI

research should focus on "particular characters within a






46

specific program, rather than on a favorite character within

an entire content genre or on a predisposition to interact

with an entire range of characters across different media" (p.

15). They also argued that parasocial interaction is a

construct consisting of at least four dimensions rather than

unidimension suggested by Rubin et al. (1985): 1)

Identification with Favorite Character, 2) Interest in

Favorite Character, 3) Group Identification/ Interaction, and

4) Favorite Character's Problem Solving Ability. Auter and

Palmgreen (1992) administered their Audience-Persona

Interaction (API) scale to 417 undergraduate students after

showing an episode of Murphy Brown. Factor analysis provided

the creation of a 22-item, four-factor API scale with

acceptable levels of reliability.



Eysenck's Personality Dimension

There have been several conceptual reviews on personality

and interpersonal communication (Anderson, 1987; Daly, 1987;

Steinfatt, 1987), but no effort has been made to increase

understanding of the relationship between personality and mass

communication behavior.

Eysenck and Eysenck (1985) defined the three basic

personality dimensions as: 1) Intraversion-Extraversion (E),

2) Neuroticism or Emotional Stability (N), and 3) Psychoticism

(e.g., tough-minded, antisocial tendencies vs. socialized

humaneness). The Psychoticism dimension was not included in






47

the Eysenck Personality Inventory (EPI), and a questionnaire

version of the dimension was not created until later as it was

included in the Eysenck Personality Questionnaire (EPQ)

(Eysenck & Eysenck, 1975). Eysenck's hierarchical model

considers Extraversion (E), Neuroticism (N), and Psychoticism

(P) as broad "supertraits" consisting of narrower traits. The

personality traits subsumed under E, N, and P are as follows

(Eysenck & Eysenck, 1985):

E: sociable, lively, active, assertive, sensation
seeking, carefree, dominant, surgent, venturesome

N: anxious, depressed, guilt feelings, low self-
esteem, tense, irrational, shy, moody, emotional

P: aggressive, cold, egocentric, impersonal,
impulsive, antisocial, unemphatic, creative, tough-
minded.

Note that sensation seeking and venturesomeness are

listed under Extraversion and impulsivity is under

Psychoticism. Eysenck and Eysenck (1985) found that

Venturesomeness and Impulsiveness correlated positively with

Psychoticism and Extraversion, but while Venturesomeness

correlated negatively with Neuroticism, Impulsiveness

correlated positively with Neuroticism. They originally

perceived Impulsiveness as a part of Extraversion with

sociability and liveliness (Eysenck & Eysenck, 1963), but

later they found two components of Impulsivity--

Venturesomeness corresponding to Extraversion and

Impulsiveness to Psychoticism (Eysenck & Eysenck, 1978, 1980).

Corulla (1988), in his factor analysis and multiple regression






48

analysis, found that Psychoticism, Extraversion, and

Neuroticism dimensions represent orthogonall" dimensions of

personality.

The Eysenck Personality Inventory (EPI) contained 57

items, the Eysenck Personality Questionnaire (EPQ) contained

90 items, and the Revised Eysenck Personality Questionnaire

contained 100 items. Eysenck and Barrett (1985) devised a

short form of the EPQR-S(the Revised Eysenck Personality

Questionnaire). It included the four indices of Extraversion

(alpha=.86), Neuroticism (alpha=.82), Psychoticism (alpha=.62)

and the Lie scale (alpha=.75), each consisting of 12 items.

Recently, Francis, Brown, and Philipchalk (1992) developed an

abbreviated form of the EPQR-A, each dimension measured by 6

items. They argue that there are some practical disadvantages

in long-item tests for measuring a personality trait. All the

items in the EPQ (Eysenck & Eysenck, 1975) and the EPQR-S

(Eysenck, Eysenck, & Barret, 1985) were administered to 685

undergraduate students from the U.S.A., England, Canada, and

Australia. Based on the item rest of test correlations for

the four scales of the EPQR-S, Francis et al. (1992) selected

the 6 significant items from each scale. The EPQR-A Scale's

alpha coefficients were as follows: Extraversion (.74 .84);

Neuroticism (.70 .77); Psychoticism (.33 .52); Lie Scale

(.59 .65). Correlations between the EPQR-S and the EPQR-A

of the Extraversion (r=.93 .95), Neuroticism (r=.92 .94),

Psychoticism (r=.80 .87), and the Lie Scale (r=.90 .92).





49

The pattern of inter-correlations between the short form of

EPQR and EPQR-A lends some support to the abbreviated form of

EPQR.



The Person or the Situation

Social psychologists assume that situational factors are

the main determinants of social behavior and attempt to

identify the cause by controlled experiment. On the other

hand, personality psychologists assume that a person brings

enduring dispositions into the situation and employ

correlation analysis to understand individual differences. At

the center of the controversy, Mischel (1968) criticized

personality research yielding correlations with behavior as

low as .30, doubting their particular consistency,

reliability, and validity along with self-report techniques in

general.

Responding to Mischel's criticisms, Daly (1987) argues

that "when characteristics of situations and people are

considered together, correlations increase" (Daly, 1987, pp.

19-20). Likewise, Buss (1989) found further evidence that the

share of variance from personality traits may be increased

through aggregation of responses, situations, and time. Buss

conceptualized personality traits as classes of single

responses involving individual differences. He suggests the

"act frequency" approach to identify clusters. While

personality traits may be variable in breadth, so situational





50

contexts may affect manifestation of traits. Because

situations elicit behavioral changes, Buss (1989) argued that

imposing cross-situational consistency on traits is of no use.

Narrowly defined personality traits predict behavior better.

On the other hand, broader classes of traits allow greater

frequencies of behavior. Buss concluded that methods employed

to study personality traits should vary considerably and

points to active and passive models of behavior used in recent

research in genetics, evolution, and learning to study

determinants of personality.

Blass (1984) argue that social and personality psychology

are drawing closer to each other, and he provides some

evidence of convergence between the two fields. First, more

social psychologists are giving more attention to personality

variables. Second, there seems to be a growing consensus from

both fields about "the desirability of an interactional

perspective" toward social behavior research. Finally, there

seems to be a commonality of the key role played by a similar

disposition in both fields--attitude and trait. Blass notes

a new flexibility among social psychologists about

incorporating personality variables into their research

strategies. And personality psychologists increasingly

recognize situational influences.






51

Mass Comm. from Individual Difference Perspective

Recently, mass communication researchers re-examined the

mass communication processes from an individual difference

perspective (Donohew, et al., 1989; Ferguson, Cho, Darlington,

& Valenti, 1990; Ferguson & Valenti, 1990a, 1990b; Finn, 1990;

Finn & Gorr, 1988; Gunter & Furnham, 1983, 1984, 1986; Gunter,

Furnham & Jarrett, 1984; Litle & Zuckerman, 1986; Perloff,

Quarles & Drutz, 1983; Rubin, Perse, & Powell, 1985). Several

of these research programs are discussed below.

Extraversion, Neuroticism, Psychoticism, and TV

In the 1980s, Gunter and his associates examined the

relationships between personality factors and mass

communication processes (Gunter & Furnham, 1983, 1984, 1986;

Gunter, Furnham, & Jarrett, 1984; Wober & Gunter, 1982).

Their research on personality types associated with differing

perceptions of portrayal of violence on TV used Gerbner's

normative definition of violence--"the overt expression of

physical force against self or other, impelling action against

one's will on pain of being hurt or killed, or actually

hurting or killing" (Gerbner, 1972, p. 31). Based on

Gerbner's cultivation studies suggesting that the viewing of

TV violence can result in increased levels of personal

aggressiveness of individuals and misconceptions about social

reality such as likelihood of falling victim to crimes, Gunter

and Furnham (1983) examined the links between personality

scores (EPQ) and their perceptions of TV violence portrayals.






52

Individuals high in Neuroticism, Extraversion and

Psychoticism dimensions were more likely to state the belief

that violent TV scenes tend to disturb viewers in general.

Those high in Neuroticism and Psychoticism were more disturbed

by some violent scenes than those low in these constructs.

Neuroticism was "most powerfully and consistently" correlated

with viewers' perceptual evaluations of TV violence,

indicating that "violence in more authentic or realistic

contexts can be especially upsetting to more anxious or

sensitive individuals" (Gunter & Furnham, 1983, p.320).

Gunter and Furnham (1984) also explored the mediating

effects that the personality factor "aggressiveness" has on

audience perceptions of violent TV portrayals. Their

canonical analysis revealed a significant personal

aggressiveness difference in viewers' perception of TV

violence. Viewers with high propensity to aggression tended

to rate American crime drama as "less realistic, less

frightening, less personally disturbing and less likely to

disturb others, and more suitable for children" (Gunter &

Furnham, 1984, p. 159). They also found that viewers who

watched assault on TV perceived stabbings and fist-fights as

less violent, frightening and disturbing, and shootings and

explosions as more serious.

Gunter and Furnham (1986) looked at the independent

effects of sex, TV "modality" (audio-visual, audio-only and

print) and personality extraversionn, neuroticism and





53

psychoticism) on recall of violent and non-violent TV news.

They found that news recall was significantly better among

extroverts (beta=.26, p<.001) and those with low Neuroticism

scores (beta=-.22, p<.01). There were significant sex

differences (beta=-.31, p<.001) in recall of violent news, but

no difference in non-violent news recall. Extraversion and

Neuroticism were significant predictors of both violent news

recall and non-violent news recall. Under the audio-visual

presentation condition, Extraversion (beta=.36, p<.01),

Neuroticism (beta=-.28, p<.05) and Psychoticism (beta=.40,

p<.01) all were significantly related to violent news recall,

but not to non-violent news recall. Under the audio-only

condition, only Neuroticism (beta=-.42, p<.01) was

significantly associated with violent news recall.

Eysenck (1967) hypothesized that extroverts are

functionally less aroused than introverts under mild sensory

deprivation, so extroverts need to seek external sources of

stimulation and make more efforts to get stimulation.

Based on Eysenck's (1967) hypothesis, Gunter, Furnham and

Jarrett (1984) compared introverts and extroverts on the

subjects' abilities to recall TV news content at three

different times during the day. Gunter, Furnham and Jarrett

(1984) found significant differences in recall between morning

(9:30 AM) and early afternoon (1:30 PM) and between morning

(9:30 AM) and late afternoon (4:30 PM) (p<.01). Two-way

(Extraversion x Time of Day) ANOVA yielded significant main





54

effects for free recall (F(2,84)=6.86, 2<.O1) and for

depth-cued recall: people/places (F(2,84)=16.20, p<.01),

suggesting that introverts were better in delayed retention

than extroverts during the late afternoon.

Finally, exploring limitations of Gerbner's cultivation

analysis, Wober and Gunter (1982) offered an alternative

hypothesis that both the amount of television viewing and

social anxiety may be a function of a personality factor.

They reported that locus of control (Rotter, 1965, 1967) was

significantly related to measures of social anxiety and

mistrust. Particularly, Wober & Gunter (1982) demonstrated

independent relationships of Rotter-type social anxiety items

and Gerbner-type social anxiety items with television-watching

behavior.

Sensation Seeking and Media Use

Donohew et al. addressed "how to reach out in an

effective manner via televised public service announcements

(PSAs) to particular 'at risk' audiences to motivate

participation in other phases of drug abuse prevention

program" (1989, p. 1). They found no significant main effects

of sensation seeking, message sensation value and drug use on

the behavioral intention index, but a strong main effect for

motivational introduction (p=.011). Their findings suggest

that it is important to pay attention to the introductory

audio portion of TV anti-drug PSAs. The behavioral intention

of low sensation seekers was more influenced by the low







55

sensation message (p=.056), while high sensation seekers were

not affected by the high sensation message. High sensation

seeking drug users, who were regarded as the most impervious

to media effect, seem to be "vulnerable to appeals based on

exciting alternatives to drug use and to appeals which promise

peer-resistance skills" (Donohew et al., 1989, p. 26).

Litle and Zuckerman (1986) found a significant

relationship between a Musical Preference Scale (MPS) and a

Sensation Seeking Scale (SSS) form 5. Their findings

suggested that certain kinds of music provide another way of

satisfying sensation seekers' needs for arousal and new

experiences.

Glasgow, Cartier, and Wilson (1985) reported a

significant positive relationship between Wilson-Patterson's

conservatism and music preferences but a relationship between

Zuckerman's sensation seeking (SS) and music preferences was

not significant, though SS was negatively related to

conservatism (r=-.48, p<.01).

Personality, Drugs, and TV Use

Based on the assumption that motivations for television

watching and drug use represent "overlapping domains," Finn

(1990) found significant relationships between TV watching and

three measures consistently identified as predictors of drug

use. He found: (1) negative relationships between television

use and sensation seeking (r=-.ll to -.18); (2) positive

relationships between hostility and television use (r=.ll to







56

.26); and (3) positive relationships between sensation seeking

and drug use (r=.43 for marijuana use, and r=.51 for alcohol

consumption). Finn's findings indicate that television

watching decreases as drug use and stimulating activities

increase.

Finn and Gorr (1988) also examined the relationship

between individual differences and motivations for television

watching. They found shyness was positively correlated with

social compensation motives (r=.30, p<.001), but loneliness

was not significantly related to the motives (r=.09, p>.05).

Self-esteem was negatively related to the social compensation

motives (r=-.24, p<.001), while Appraisal (r=.20, p<.001) and

Belonging (r=.ll, p<.05) correlated positively with the mood

management motives such as relaxation, entertainment, arousal,

and surveillance.

Risk-Taking, Information Seeking, and Media Gratifications

Recently, Ferguson and her colleagues have examined the

relationships between the "Big Five" personality traits and

media use motives in interaction with information processing.4

Ferguson et al. (1990) reported a significant interaction

effect of "openness to experience" with use of the media for

surveillance on persuasive messages about drug use. The study

found that closed-minded people who saw the message in

Executive typeface were significantly more concerned about


4The "Big Five" personality dimension is as follows: 1)
Extraversion/Intraversion; 2) Neuroticism; 3) Conscientiousness; 4)
Agreeableness; 5) Openness to Experience.







57

drug use than open-minded people who saw the message in the

same typeface.

Based on a series of risk-taking predispositions and

information processing studies, Ferguson and Valenti (1988,

1990a, 1990b) identified three risk factors with strong

effects in interaction with message variation: 1) rebellious

risk taking (similar to disinhibition by Zuckerman); 2)

impulsive risk taking (similar to Eysenck's impulsivity in the

narrow sense [IMPn]); and 3) adventurous risk taking (similar

to Eysenck's impulsiveness). In general, they found that

people high in rebellious risk taking and low in impulsive

risk taking were more concerned about all types of risk taking

including the greenhouse effect, ozone destruction, food

irradiation, etc. They found a significant interaction effect

between rebellious and impulsive risk taking on information

seeking about health and environmental risks: those high in

rebellious risk taking and low in impulsive risk taking were

more likely to read information or view TV programs about

health issues such as syphilis and AIDS.

Personality & Communication-Related Behaviors

Anderson (1987) points out that many personality studies

have been conducted with little awareness of how the research

fits with overall communication research processes. McCroskey

and Daly (1987), Personality and Interpersonal Communication,

provide a thorough review of personality trait research in the

field of interpersonal communication.







58

Addressing his own question about why communication

scholars ought to be interested in personality in the first

place, Daly (1987) notes that personality variables play a

number of roles in communication research: (1) they are

"useful covariates" in communication studies--"By

statistically covarying certain dispositions, better evidence

of the role of situational characteristics or experimental

manipulations can be demonstrated" (partialling out); (2) they

are predetermined factors--Incorporating a personality

variable allows the researcher to look at not only the

independent impact of the trait on the criterion variable, but

also the interaction effects of trait and situation on the

dependent variable; (3) they are used in the validation of a

personality construct--Correlation between a particular

disposition and other personality variables presumed to be

relevant is used to validate traits; and (4) they are used in

the relationships among "interactants as a function of their

traits" (pp. 29-31).

Steinfatt (1987) reviewed six personality variables

relevant to communication behavior: authoritarianism,

dogmatism, rigidity, intolerance of ambiguity,

Machiavellianism and locus of control. Steinfatt (1987)

suggested each of these six variables has some relationship

with communication behaviors: (1) authoritarianism--"its

prediction of sexual content differences in media

preferences"; (2) dogmatism--"the apparent inability of high







59

dogmatics to discriminate between the source and content of a

message"; (3) rigidity--"an analytic block in cognitive

functioning; (4) intolerance of ambiguity--"openness to

information"; (5) Machiavellianism--"ease of engaging in

counter-attitudinal advocacy"; and (6) locus of

control--"responses to internal versus external persuasive

attempts" (p. 108).

Harvey and Hays (1972) reported that closed-minded

individuals treated message content as dominated by the source

(peripheral route in Petty & Cacioppos's ELM), whereas

open-minded persons better discriminated between judgments

about a source and judgments about the source's messages

(central route in ELM). Rosenman (1967) found that

open-minded people rated a nightmare comedy movie (Dr.

Strangelove) as more to their liking than did closed-minded

(dogmatic) people.

Miller and Bacon (1971) found evidence that closed-minded

people took longer to recognize the humorous element in a

picture from the 1966 Harvard Lampoon Parody of Playboy than

did open-minded people. Kelley (1985) demonstrated that those

high in authoritarianism had significantly more negative

feelings about pornography than did those low in the

construct. Greendlinger (1985) found that individuals high in

authoritarianism responded more negatively to films depicting

homosexuals than those low in the construct. In general,

correlations indicate that dogmatism and authoritarianism







60

share about 25% to 40% of their variance, suggesting a strong

relationship between different constructs.



Local Television News

Everyday, tens of millions of people watch the news on

television. Television ratings and market research have

traced out demographic characteristics of the TV news

audience. Advertisers pay money ranging from $10,600 to

$61,100 to enjoy the attention of this audience for a few

seconds. More than 40% of a network-affiliated TV station's

annual revenues come from the news operations (Boemer,

1987).

While television news serves the profit motive, it also

helps to preserve the democracy by informing the citizenry.

An understanding of the TV news viewing audience is

important because it determines TV news programming

decisions (Babrow, 1989).

Wilbur Schramm, in his 1949 article, "The Nature of

News," viewed news consumption as guided by either "reality"

motives or "pleasure" motives. As Schramm suggested,

"reality" motives have delayed rewards for audience, while

"pleasure" motives have immediate rewards. The uses and

gratifications approach to political news was brought about

by Blumler and McQuail's (1969) study of the 1964 British

general election. Influenced by Blumler and McQuail, many

uses and gratifications researchers in the 1970s (Becker,






61

1976, 1979; McLeod & Becker, 1974; McLeod, Durall, Ziemke, &

Bybee, 1979) have focused on news gratifications of a

political nature.

In the 1980s, most uses and gratifications research has

examined gratifications from television news, although some

studies focused on the gratifications from newspapers

(Becker, 1979; Weaver & Buddenbaum, 1980). Research on the

gratifications associated with television news examined

various types of television news programming, ranging from

regularly scheduled newscasts (Gantz, 1978; Levy, 1983;

Palmgreen, Wenner, & Rayburn, 1980, 1981; Palmgreen &

Rayburn, 1982; Perse, 1990a; Rayburn & Palmgreen, 1984) to

news interview programs (Levy, 1978), and news magazine

programs such as 60 Minutes (Rubin, 1981; Wenner, 1982).

Specifically, research on TV news gratifications has

focused on the newscast regularly scheduled, that is, one of

the three nightly network newscasts or network-affiliated

local news casts. Levy (1978) found that gratifications

associated with watching television newscasts would

influence continuing rates of exposure, attitude toward TV

news, and public affairs knowledge. Recently, Perse (1990a)

found that a "utilitarian" viewing motivation was

significantly related to higher "cognitive involvement,"

while "diversionary" viewing motivation was correlated with

"feeling happy" when watching local television news.






62

Rubin, Perse, and Powell (1985) suggest that local

television news offers "salient media content" for examining

the parasocial interaction process. Local television news

is broadcast three times during a typical weekday. Local

newscasts provide "carefully created personae trained to

exhibit qualities necessary for the development of

parasocial interaction, encourage the bond of familiarity

and dependence between anchor and viewer" (Rubin et al.,

1985). In local television news watching, Rubin et al.

(1985) argued that "parasocial interaction is part of the

active, more goal-directed pattern of instrumental

television use" (p. 175). Rubin and his associates

developed a reliable parasocial interaction scale to measure

feelings of audience relationship with local television news

personalities. They noted that "sizable correlations for a

sample of this scope between the parasocial interaction

scale and attitudes of affinity and perceived realism about

the local television news support the construct validity of

the scale" (p. 176).

Most local news formats feature interpersonal

interaction of news team members, as newscasters address

viewers directly and induce a sense of camaraderie with the

news team members. As Rubin et al. (1985, p. 161) pointed

out, "local news promotions typically emphasize the

'natural, down-to-earth' news personalities and their local

community involvement. Newscasters often become celebrities






63

in a broadcast market, appearing on local interview

programs, making personal appearances, and they are

discussed in the local print media." Recently, Lin (1992)

demonstrated some evidence that heavy local news viewers are

more likely to watch local television news for parasocial

interaction. In particular, local viewers wanted to

personally identify with weather anchors with "good

performance."

Finally, Wright (1992) pointed out:

the emergence of new news gathering and reporting
technologies and the increased importance of local
news segments in overall station revenues and
image in the community have produced qualitative
and quantitative changes in local news. Cable
television has extended coverage areas well beyond
the city license. Satellite feeds have increased
the number of regional and national-international
stories. Statewide or regional network
affiliations (such as The Florida News Network)
have also impacted local story availabilities.
Image-oriented, conversational styles of
presentation have been complemented by reduced
emphasis on issue orientation and
editorialization. And many stations have doubled
or tripled the amount of air time devoted to local
news. For example, WJXT, a CBS affiliate in
Jacksonville, Florida, has increased its early
evening news show from 30 to 90 minutes. Soft
news stories are increasingly emphasized as
stations expand news segments (p. 1).

More recently, many CBS affiliates including Detroit and

Atlanta are canceling CBS morning news program for their own

local morning news shows.












CHAPTER 3
RESEARCH DESIGN


Proposed Model of PSI



Despite the spate of recent research, findings are

inconsistent about why people watch television news for

parasocial interaction. This lack of significant findings

may exist because researchers have not included in their

parasocial interaction (PSI) model both sociological and

psychological variables. Individual differences such as

need for social interaction as well as situational factors

such as loneliness may help explain variability in

parasocial interaction.

This study has two main objectives. First, in contrast

to earlier research that focused on sociological antecedents

of parasocial interaction, it is assumed that individual

difference variables such as Extraversion, Neuroticism, and

Need for Social Interaction as well as the situational

variable Loneliness could explain parasocial interaction.

The second goal is to increase understanding of parasocial

interaction and the concept's relevance for local television

news watching. In particular, this study attempted to test







65

hypotheses about social and psychological origins of

parasocial interaction.

What factors or variables contribute to parasocial

interaction? What are the social and psychological origins

of parasocial interaction? In particular, what role does

personality play in parasocial interaction with local

television newscasters? Does the personality dimension

explain the development of parasocial interaction better

than situational factors, or vice versa?

The purpose of this study is to empirically test the

proposed model of parasocial interaction in the domain of

local television news, a model in which parasocial

interaction may be influenced by several determinants: 1)

personality (Extraversion, Neuroticism, and Need for Social

Interaction), and 2) situational factors (Loneliness).

To explicate the roles of individual difference and

situation in media uses and effects, this study focused on

parasocial interaction with local TV news personalities.

Local television news was targeted for several reasons.

First, local television news is a trusted and relied upon

source of information. Second, local news is widely viewed

by people. Third, local television news is a mixture of

information and entertainment. Recently local television

newscast has provided human and local interest stories as

well as hard news stories. Finally, parasocial interaction









0~


(0

co
L
o 2-
0

z.E



E
O
'o
L.)



C
0
z




0



L.
jr

cm


a)
~oU)


'~L.
.Eo>


U)



>)
- c


"ilio


p


c .u
*= Q
(0 (D
.:p m
+CO
-',







67

has been empirically demonstrated in local television news

watching as well as in general news watching.

My proposed parasocial interaction model is a moderator

model (see Figure 3-1). This process approach attempts to

identify moderators of media exposure/reliance-PSI

relationship.

Baron and Kenny (1986) distinguish between a moderator

and a mediator model.

Moderator Model

Predictor----------------- >

Moderator----------------- > Dependent Variable

Predictor x Moderator----->

Within a correlational analysis, a moderator is a third

variable affecting the zero-order correlation between two

other variables. In ANOVA terms, a moderator effect can be

regarded as an interaction between an independent variable

and a factor specifying the appropriate conditions for its

operation. The moderator hypothesis is supported if the

interaction is significant. Unlike the mediator-predictor

relation (where the predictor is causally antecedent to the

mediator), moderators and predictors are at the same level

regarding their role as causal variables antecedent to

certain dependent variables. That is, moderator variables

always function as independent variables, whereas mediators

shift roles from effects to causes, depending on the focus

of the analysis. Moderation implies that the causal






68

relation between two variables changes as a function of the

moderator variable. The statistical analysis must test the

differential effect of the independent variable on the

dependent variable as a function of the moderator.

Mediator Model

Independent var.--->Mediator(M)--->Dependent var.

In the mediator model, a mediator intervenes between

the stimulus (independent variable) and response (dependent

variable). A mediator should meet the following condition:

when paths between X and M, and M and Y are controlled, a

previously significant relationship between X and Y is no

longer significant. To test for mediation effect, one

should estimate the following regression equations: 1)

regressing M on X; 2) regressing Y on X; 3) regressing Y on

both X and M. Separate coefficients for each equation

should be tested. Perfect mediation holds if the X has no

effect when the mediator is controlled.

Whereas moderator variables specify when some effects

will hold, mediators speak to how and why such effects

occur. Moderators are introduced when there is weak or

inconsistent relation between X (media use) and Y

(parasocial interaction) (e.g., a relation holds in one

setting but not in another, or for one subpopulation but not

for another). Mediation is best done in the case of a

strong relation between X and Y.






69

Hypotheses

This study hypothesized that there are two

possibilities for satisfying the need for social

interaction: individual and situational. They result in

four possible combinations of degree of dependence on mass

media as functional alternatives for the social interaction

need: high individual-high situation, high individual-low

situation, low individual-high situation, and low

individual-low situation. When a person does not have

satisfactory individual and situational possibilities to

satisfy the need for social interaction, then as a

substitute for the need for social interaction one may

attend to mass media such as television. In other words, a

person with low individual and situational potentials for

social interaction is more likely to watch television for

parasocial interaction than one with high individual and

situational possibilities.

The first two hypotheses are tests of relationships

between local television news watching and parasocial

interaction. Levy (1979) found a positive relationship

between parasocial interaction and television news watching.

Recently, Auter and Palmgreen (1992) found a significant

association between parasocial interaction and overall

exposure to television. It is expected that parasocial

interaction will be positively related to local television

news exposure and reliance.









HI: Local television news exposure will be positively
related to parasocial interaction.

H2: Local television news reliance will be positively
related to parasocial interaction.

Nordlund (1978) found a positive correlation between

neuroticism and parasocial interaction. There exist some

evidence that extraversion and neuroticism are associated

with parasocial interaction (Rosengren & Windahl, 1977). As

argued earlier, need for social interaction should also be

positively related to parasocial interaction. The next

three hypotheses (H3, H4, and H5) are tests of whether or

not individual difference variables explain parasocial

interaction.


H3: A person high in need for social interaction will be
more likely than a person low in need for social interaction
to watch local television news for parasocial interaction.

H4: An intraverted person will be more likely than an
extraverted person to watch local television news for
parasocial interaction.

H5: A person high in neuroticism will be more likely than a
person low in neuroticism to watch local television news for
parasocial interaction.

Rubin, Perse, and Powell's (1985) parasocial

interaction model predicted that an individual with a

greater degree of loneliness will more likely to watch local

television news for parasocial interaction as functional

alternatives for social interaction. Hypothesis six is a

test of whether or not the situational variable explains

parasocial interaction.






71

H6: A person experiencing a high degree of loneliness will
be more likely than a person experiencing a low degree of
loneliness to watch local television news for parasocial
interaction.


The next eight hypotheses (H7a Hl0b) test the

following research questions: What roles do individual

variables (NSI, extraversion, and neuroticism) and a

situational variable (loneliness) play as moderators in the

relationship between parasocial interaction and local

television news exposure and reliance. In other words,

these hypotheses incorporate individual and situational

variables with local television news exposure and reliance

to more fully investigate parasocial interaction process.


H7a: Need for social interaction will moderate the
relationship between local television news exposure and
parasocial interaction.

H7b: Need for social interaction will moderate the
relationship between local television news reliance and
parasocial interaction.

H8a: Extraversion will moderate the relationship between
local television news exposure and parasocial interaction.

H8b: Extraversion will moderate the relationship between
local television news reliance and parasocial interaction.

H9a: Neuroticism will moderate the relationship between
local television news exposure and parasocial interaction.

H9b: Neuroticism will moderate the relationship between
local television news reliance and parasocial interaction.

Hl0a: Loneliness will moderate the relationship between
local television news exposure and parasocial interaction.

Hl0b: Loneliness will moderate the relationship between
local television news reliance and parasocial interaction.






72

Procedure

Survey research is the predominant uses-and-

gratifications methodology. Data are gathered via self-

report questionnaires to analyze psychological variables and

media gratifications.

Student data were collected by administering 302

questionnaires to 280 undergraduate students in six

communication courses and 22 graduate students in a mass

communication course at the University of Florida.

Non-student data for this study were gathered by

trained graduate students during the last weekend of spring

term (May 1, 1993). A data-gathering location was needed

that would permit many interviews in one location. With the

approval of the management of a large shopping mall in the

area, paid interviewers screened for non-student

respondents. As an incentive, potential mall respondents

received a two-dollar bill in return for a completed

questionnaire. Five hundred eighty-seven (303 students and

284 non-students) usable responses were obtained in this

study.



Measures

Local television news exposure and reliance are the

independent variables, and the parasocial interaction is the

dependent variable. The individual difference variables,

Extraversion/Intraversion, Neuroticism, and Need for Social






73

Interaction, and the situational variable Loneliness are

moderators.

Local TV News Exposure

An overall local TV news exposure index is obtained by

summing and averaging standardized z-scores of the following

three measures: 1) days of local TV news watching (TVDAY),

2) hours and minutes of local TV news watching (TVHR), and

3) attention paid to local TV news watching (TVATN).' TVDAY

was measured by asking respondents how many times they

watched local television news during a typical week on their

most regularly watched channel. TVHR was measured with the

following question: "During a typical day, how many hours

and minutes do you spend watching local television news?"

For the TVATN measure, respondents were asked: "How much

attention do you pay to local television news stories when

you watch local television news?" Responses for the TVATN

measure were recorded on a four point scale, with 4 being a

lot of attention and 1 being no attention at all. Cronbach's

standardized alpha is .73 for the local TV news index,

suggesting moderately high internal consistency (see Table

3-1 for reliability coefficients for measures).






'Standardized z scores were obtained by formula z=score -
mean / standard deviation: TVDAY(z) = TVDAY 4 / 2.03;
TVHR(z) = TVHR 55.31 / 45.11; TVATN(z) = TVATN 3.25 / .73.
TVDAY(z), TVHR(z), and TVATN(z) were summed and divided by 3
to form a single television news exposure index.









Table 3-1
Reliability Coefficients for Key Measures



Scale Cronbach's Standardized Alpha


Television News Exposure .73

Television News Reliance .57

Loneliness .95

Extraversion .85

Neuroticism .80

Need for Social Interaction .69

Parasocial Interaction .88











Local TV News Reliance

Local TV news reliance was measured with the following

two items: 1) "On a scale of 1-7 with 7 being high reliance,

and 1 being low reliance, how much would you say you rely on

television as source of news and information about local

issues or problems?"; 2) "On a scale of 1-7 with 7 meaning

you would miss it very much and 1 meaning you would not miss

it at all, how much would you miss watching local television

news if you were not able to watch it?" Overall local TV

news reliance was calculated by summing and averaging the

two items. Cronbach's standardized alpha is .57 for local

TV news reliance, suggesting low to moderate internal

consistency.

Parasocial Interaction

This study selected Rubin's revised PSI scale for two

reasons. First, the scale is the most widely used. Second,

previous research offers consistent evidence of validity and

reliability. In a series of studies, Rubin and Perse (1987)

developed a 10-item parasocial interaction scale choosing

from their original 20 PSI items. The revised 10-item PSI

scale had a .88 Cronbach alpha and correlated highly with

the original 20-item PSI scale (r=.96, p<.001). The two

measures have been used to study both entertainment and news

programs.







76

In this study, parasocial interaction is measured with

Rubin's revised 10-item version of the PSI scale, which

expresses relationships with local television newscasters.

Responses were recorded on a seven-point scale ranging from

"very strongly disagree" (1) to "very strongly agree" (7)

(e.g., "My favorite newscasters make me feel comfortable, as

if I am with friends; "If my favorite newscaster appeared on

another television program, I would watch that program,"

etc.). Appendix B presents the wording of the 10 PSI items.

An overall measure of parasocial interaction was calculated

by summing and averaging the 10 items. Cronbach's

standardized alpha is .88 for PSI, indicating high internal

consistency.

Loneliness

Loneliness is measured by the revised UCLA Loneliness

Scale, which consists of 20 statements, 9 expressing

satisfaction (e.g., I feel I am in tune with the people

around me; I feel I can find companionship when I want it)

and 11 dissatisfaction with social relationships (e.g., I

feel isolated from others; I feel there is no one I can turn

to) (see Appendix B for all 20 loneliness statements).

Responses were recorded on a seven-point scale ranging from

"very strongly disagree" (1) to "very strongly agree" (7).

An overall measure of loneliness was computed by summing and

averaging 20 items. Cronbach's standardized alpha for






77

Loneliness scale in this study is .95, indicating very high

internal consistency.

Extraversion

In order to examine the hypothesized relationship

between the three individual difference variables

(Extraversion, Neuroticism, and NSI) and parasocial

interaction, 24 items (including six items for lie scale)

from the individual difference literature (Eysenck &

Barrett,1985; Eysenck & Eysenck, 1975; Francis, Brown, &

Philipchalk 1992) were measured on a seven-point scale and

factor analyzed (see Appendix B).2 The principal axis

factors were rotated by the varimax method (see Table 3-2).3

The Extraversion factor revealed an eigen value of 6.2 and

contained the following items: 1) I am a talkative person;

2) I am rather lively; 3) I can easily get some life into a

rather dull party; 4) I tend to keep in the background on

social occasions; 5) I am mostly quiet when I am with other

people; 6) Other people think of me as being very lively.

And Cronbach's alpha reliability coefficient is .85,

suggesting high internal consistency.




2Six-item Lie scale was used for detecting personality response
consistency from respondents in this study (see Appendix B for Lie
scale statements).

3A varimax orthogonall) rotation was used for 24 individual
difference statements. While the underlying assumptions of a
varimax method differ from those of an oblique method, the
resulting four factors are very similar in structure to those of an
oblique solution.











Table 3-2
Factor Analysis of Individual Difference Items


Factor Loadings Mean SD
1 2 3 4

Extroversion
1.Being lively .75 5.0 1.3
2.Being quiet .72 4.9 1.5
3.Talkative .70 5.0 1.5
4.Rather lively .68 5.2 1.2
5.Some life .64 4.6 1.4
6.Background .63 4.7 1.5
7.Seeking out -.60 3.9 1.3
8.Soc. unskilled -.49 2.3 1.3
9.Secure -.47 .45 2.9 1.4

Neuroticism
10.Nervous .72 3.2 1.6
11.Suffering 'nerves' .69 3.1 1.6
12.Worrier .67 4.2 1.7
13.Up and down mood .50 4.2 1.6
14.Feeling fed-up .46 4.2 1.6
15.Feeling lonely .43 .40 2.9 1.5

Lie Scale
16.Taking advantage .77 4.3 1.7
17.Cheating .69 4.0 1.7
18.Stealing .60 4.1 1.7
19.Being greedy .53 4.2 1.5
20.Blaming someone .52 4.7 1.6

NSI
21.Being satisfied -.36 .53 3.0 1.5
22.Social competence .46 3.2 1.5
23.Practicing what I preach -.34 4.3 1.3
24.Seeing enough of friends .29 4.0 1.6

Percent of Variance 23.4 12.2 4.9 3.0
Eigenvalues 6.2 3.5 1.7 1.3
Cronbach's alpha .85 .80 .76 .69

Note: Variamax-rotated principal axis factoring for N = 587; a 7-
point scale ranging from "Very Strongly Disagree" (1) to "Very
Strongly Agree" (7); See Appendix B for complete wording for 24
Individual Difference items; Items 7, 8, and 9 are included in NSI
index and item 23 is included in Lie Scale index; For the sake of
parsimony, factor loadings less than .30 are not reported.






79

Table 3-2 shows that the first factor consists of six

extraversion-related items and three items from need for

social interaction dimension (I enjoy being around other

people, and seek out social encounters frequently; I would

describe myself as socially unskilled; I feel secure in

social situations). The ninth item (I feel secure in social

situation) also has a loading (.45) on the fourth factor

"need for social interaction." For extraversion, the first

six items are included in the analysis because the items 7,

8, and 9 did not increase Cronbach's alpha.4

Neuroticism

Table 3-2 shows that the second factor "neuroticism" is

a composite of 6 items. Thus, an overall measure of

neuroticism is calculated by summing and averaging the 6

items (see Table 3-2). The Neuroticism factor revealed an

eigen value of 3.5 and reliability coefficient of .80,

indicating high internal consistency. The "I often feel

lonely" factor also has a factor loading (.40) on the fourth

factor Need for Social Interaction.





'Also, the three items ("I enjoy being around other people, and
seek out social encounters frequently"; "I would describe myself as
socially unskilled"; "I feel secure in social situations") were
conceptually unnecessary -- perhaps replication of other
Extraversion items.

iExact wordings for six Neuroticism items are as follows: 10)
"I would call myself a nervous person"; 11) "I suffer from
'nerves'"; 12) "I am a worrier"; 13) "My mood often goes up and
down"; 14) "I often feel 'fed-up'"; 15) "I often feel lonely."








Need for Social Interaction

Need for social interaction is defined as an

individual's tendency to engage in social activity. This

characteristic will be predictive of the manner in which

people search for functional alternatives for social

interaction. In this study, a Need for Social Interaction

scale (NSIS) is developed to explain parasocial interaction

process.

Six NSIS items were included in varimax-rotated

principal axis factor analysis (see Table 3-2). But, only

three items (I am satisfied with my social life today; I

have no doubts about my social competence; I see enough of

my friends and relatives) have factor loadings of .29 or

higher. For need for social interaction, six items (7, 8,

9, 21, 22, & 24) are included in analysis because the 7th,

8th, and 9th items did not decrease Cronbach's alpha.6 Need

for social interaction has an eigen value of 1.3 and .69

Cronbach's alpha. The strong correlation (r = .37, p<.01)

between the NSIS and extraversion scale also indicates

concurrent validity.





6Exact wordings for the six NSI statements are as follows: 1)
"I enjoy being around other people, and seek out social encounters
frequently"; 2) "I would describe myself as socially unskilled"; 3)
"I feel secure in social situations"; 4) "I am satisfied with my
social life today"; 5) "I have no doubts about my social
competence"; 6) "I see enough of my friends and relatives."
Conceptually the 1st, 2nd, and 3rd items seem more likely to fit in
NSI than Extraversion.












CHAPTER 4
RESULTS


The average age of respondents was 28.8 years old. The

oldest was 79 and the youngest was 17. More than half of

respondents were women (58.8%). Of the 587 respondents 302

(280 (47%) undergraduate and 22 (4%) graduate students) were

students, and 285 (49%) were non-students.' About 90 percent

of the sample had some college education. A little more than

one third of those 285 non-student respondents had

professional/ managerial/teacher occupations, about 40 percent

were clerical/blue color workers, and about 16 percent were

housewives, retired, and others.

Table 4-1 shows sample sizes, means, standard deviations,

and the minimum and maximum values for the independent and

dependent, and moderator variables. The frequency

distribution for all the variables but TV attention

approximated a normal curve. The means for TV attention (M =

3.25, SD = .73) were skewed more toward the positive than the

negative value.

Before testing the hypothesized relationship between

local television news watching and parasocial interaction it


'In this study, students refer to college undergraduates at the
University of Florida, while non-students refer to persons over 18
years old interviewed at the mall.












Table 4-1
Means, Standard Deviations and Ranges


of Variables


Variable N* Mean Std. Dev. Min. Max.


TVDAY 586 4.00 2.03 0.00 7.00
TVHR 587 55.31 45.11 0.00 240.00
TVATN 584 3.25 .73 1.00 4.00
TVREL 587 4.25 1.96 1.00 7.00
LONELY 552 3.83 .40 2.65 6.25
EXTRA 576 4.37 .57 1.83 7.00
NEURO 576 3.64 1.13 1.00 6.67
NSI 576 4.36 .73 2.00 6.50
LIE 568 3.81 .96 1.00 6.50
PSI 577 4.34 .95 1.00 7.00

*For some variables, N is smaller than 587 because of missing
values.

Note: TVDAY = Days of local TV news watching
TVHR = Minutes of local TV news watching
TVATN = Attention paid to local TV news watching
TVREL = Local TV news reliance
LONELY = Loneliness
EXTRA = Extroversion
NEURO = Neuroticism
NSI = Need for Social Interaction
LIE = Lie scale
PSI = Parasocial Interaction














Table 4-2
Zero-Order Correlations for TV News Variables


Variables TVDAY TVHR TVATN TVREL


TVHR .57** --


TVATN .52** .35** --


TVREL .55** .43** .50**

Note: TV Exposure & TV Reliance Variables

TVDAY = Days of local TV news watching
TVHR = Minutes of local TV news watching
TVATN = Attention paid to local TV news watching
TVREL = Local TV news reliance

*p<.05
**p<.01






84

was necessary to examine the correlations between individual

TV exposure variables (TVDAY, TVHR, and TVATN) and TV reliance

for multicollinearity problems.2 Table 4-2 indicates that the

zero-order correlations among each of the TV news exposure

measures with TV news reliance ranged between .35 and .57 and

all coefficients were significant at p<.01. Thus, a local TV

news exposure index was created by summing and averaging

standardized z-scores of TVDAY, TVHR, and TVATN.

An analysis of between-group comparisons (students vs.

adults) may provide some insight into data analysis. Table 4-

3 shows means and standard deviations of key variables for

each group. Table 4-4 summarizes t-tests between adults and

students for the main variables. When comparing adults to

students, the t-test revealed higher means for adults (M =

.37) than students in local television news exposure (M = -

.33, t(551.06) = -11.61, p<.000). A similar result occurred

for local TV news reliance in that adults (M = 4.98) reported

a significantly higher TV news reliance than students (M =

4.39, t(562) = -4.83, R<.000). For Neuroticism, students (M

= 3.80) had significantly higher means than adults (M = 3.49,

t(554) = 3.23, R<.001). Table 4-3 also shows that adults (M





2TVDAY was measured by asking respondents how many days they
watch the local television news during a typical week; TVHR was
measured by the following question "During a typical day, how many
hours and minutes do you spend watching local television news?";
TVATN asked, "How much attention do you pay to local television
news stories when you watch the local television news?"







85

= 4.51) had significantly higher parasocial interaction means

than students (M = 4.22, t(545.21) = -3.81, p<.000).

An analysis of comparisons between adults and students

revealed significant mean differences in TV Exposure, TV

Reliance, Neuroticism, and parasocial interaction. But, no

significant between-group mean differences emerged for

extraversion, need for social interaction, and loneliness (see

Table 4-3 & 4-4). Since there are four significant between-

group differences out of seven key variables, separate

statistical tests will be conducted for each group (adults and

students).



Relationships between Local TV News Watching & PSI

The first two hypotheses are tests of associations

between local television news watching and parasocial

interaction. It was hypothesized that parasocial interaction

will be positively related to local television news exposure

and reliance.


HI: Local television news exposure will be positively related
to parasocial interaction.

H2: Local television news reliance will be positively related
to parasocial interaction.

As Table 4-5 shows, parasocial interaction is

significantly correlated with both local TV news exposure (r

= .34, p<.01) and reliance (r = .34, p<.01). Hence,

Hypotheses 1 and 2 are supported. When only adult respondents

(N = 285) are included in the analysis, the correlation













Comparisons of


Table 4-3
Means betw. Students and Adults for Key
(S = Students; A = Adults)


Variables


Variables N Mean SD

S A S A S A

TV Exp. 281 282 -.33 .37 .67 .77
TV Rel. 282 282 4.39 4.98 1.41 1.52
Lonely 276 256 3.83 3.83 .30 .48
Extra. 281 275 4.36 4.38 .49 .64
Neuro. 279 277 3.80 3.49 1.09 1.16
NSI 281 275 4.36 4.37 .70 .76
PSI 276 281 4.22 4.51 .85 .99
Lie 281 267 3.99 3.66 .90 1.00


Table 4-4
T-Tests betw. Adults & Students for Key Variables
(For Adults, N=285; For Students, N=281)


Variables t-value DF Sig. Level (p-value)

TV Exposure -11.61 551.06 .000

TV Reliance -4.83 562.00 .000

Loneliness -.24 426.00 n.s.

Extraversion -.41 511.58 n.s.

Neuroticism 3.23 554.00 .001

NSI -.16 554.00 n.s.

PSI -3.81 545.21 .000

Lie Scale 4.09 546.00 .000













Table 4-5
Zero-Order Correlations for Key Variables
(N = 587)

Var. TE TR LO EX NE NSI Age


TR .61** ---



LO .06 .05 ---

EX .09* .09 .08 ---

NE -.09* -.03 .36** .05 ---

NSI .08 .08 -.10* .37** -.31** ---



AGE .36** .19 -.07 -.11** -.20** .04 ---



PSI .34** .34** .15** .05 -.03 .10* .20**

Note: TV Exposure & TV Reliance Variables
TE = Local TV News Exposure = Mean Standardized Scores of
3 TV News Watching Measures
TR = Local TV news reliance

Individual Difference Variables
LO = Loneliness
EX = Extroversion
NE = Neuroticism
NSI = Need for Social Interaction

Dependent Variable
PSI = Parasocial Interaction

*p<.05
**p<.01














Table 4-6
Zero-Order Correlations for Key Variables
(N = 281; Adults)

Var. TE TR LO EX NE NSI Age


TR .58** ---

---------------------------------------------------------------------

LO .13* .06 ---

EX .12* .07 .10 ---

NE -.05 -.05 .34** .09 ---

NSI .12 .10 .02 .38** -.28** --

---------------------------------------------------------------------

AGE .20** .15* -.08 -.17** -.21** .06 ---

---------------------------------------------------------------------

PSI .44** .41** .17** .02 -.03 .13* .19**

Note: TV Exposure & TV Reliance Variables
TE = Local TV News Exposure = Mean Standardized Scores of
3 TV News Watching Measures
TR = Local TV news reliance

Individual Difference Variables
LO = Loneliness
EX = Extroversion
NE = Neuroticism
NSI = Need for Social Interaction

Dependent Variable
PSI = Parasocial Interaction

*p<.05
**p<.01













Table 4-7
Zero-Order Correlations for Key Variables
(N = 280; Students)

Var. TE TR LO EX NE NSI Age


TR .59** ---



LO -.01 .03 ---

EX .09 .16** .07 --

NE -.01 .05 .41** .04 --

NSI .05 .07 -.26** .36** -.34** --



AGE .06 .00 -.21** -.10 -.13** .09 --



PSI .15* .18** .08 .15 -.01 .09 -.07

Note: TV Exposure & TV Reliance Variables
TE = Local TV News Exposure = Mean Standardized Scores of
3 TV News Watching Measures
TR = Local TV news reliance

Individual Difference Variables
LO = Loneliness
EX = Extroversion
NE = Neuroticism
NSI = Need for Social Interaction

Dependent Variable
PSI = Parasocial Interaction

*p<.05
**p<.01