 

   

     

  

. L, ,.§%§§_Lu éswr: “£33.

 

 

LIBRARY
Michigan St: 1
University

-—"-

 

 

This is to certify that the

dissertation entitled

The Role of Cognitive Modeling in Predicting Internet Use

presented by

Matthew S. Eastin

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Mass Media

 

' Major erfMU 0

Date é/30/0 /

MSU is an Afﬁrmative Action/Eq ual Opportunity Institution 0-12771

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE

Ni‘il/ ﬂ

 

6/01 c:/C|RCJDateDuo.p65-pl 15

 

THE ROLE OF COGNITIVE MODELING IN PREDICTING INTERNET USE
By

Matthew S. Eastin

A Dissertation

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY

Department of Telecommunication
2001

ABSTRACT
THE ROLE OF COGNITIVE MODELING IN PREDICTING INTERNET USE
By

Matthew S. Eastin

Using the framework provide by Social Cognitive Theory (Bandura, 1986), this
study. examined the relationships among behavior, cognitive, and environmental events
as mechanisms to Internet use. While prior research in this area focused primarily on
the cognitive and behavioral components, this study extends this to incorporate
environmental factors that inﬂuence the development of cognitive models of Internet
use. Here, both personal and non-personal experiences within an individual’s
environment are examined as antecedents of lntemet self-efficacy judgments. Results
assessing individual relationships indicated that lntemet self-efﬁcacy was signiﬁcantly
related to prior Internet experience, friend’s and parent’s lntemet use and lntemet
success, ﬁiends’ and parent’s lntemet encouragement, social group success, media
portrayals, social, personal and negative outcome expectancies, and lntemet anxiety.
Further, Internet use was signiﬁcantly related to social, informational, entertainment,
and negative outcome expectations, Internet self-efﬁcacy and lntemet anxiety. When
using path analysis to test the direct and indirect theoretical relationships among these
constructs, the data failed to ﬁt the model. However, a new model identifying only

substantial contributors was tested and found to be consistent with the data.

€0pyﬁght by
MATTHEW s. EASTIN
2001

ACKNOWLEDGEMENTS

There are several people who contributed in different and signiﬁcant ways to
the completion of my doctoral degree. I would ﬁrst and foremost like to thank my
wife Emily for her support. I love you and thank you for all that you have done to
help me earn this degree.

I would like to thank my father. I owe him my work ethic and integrity. He
will always be the measuring tool from which I will gage my success. I also would
like to thank my family. I know you didn’t always know exactly what I was
studying or why I found it to be important, but you always supported and trusted my
decision to continue with my education. Finally, I would like to thank my mother
and father in-law. They have supported me throughout this process. . .that really
meant a lot to me.

I would like to thank several faculty who have played key roles over the past
four years. Dr. Bradley Greenberg, as my advisor he has provided me with the
knowledge to develop and grow as a scholar. Dr. Robert LaRose, who I was
fortunate to have worked on several research projects with. He has taught me many
valuable research skills that I will continue to use throughout my career. Dr. Pamela
Whitten, an outstanding person and a great model for the tenure process.

While family and those who trained me played a crucial role in my success, I
can’t forget the friends of mine who supported me emotionally and showed me that
there is life outside of my office. I thank you so much for that. In particular, I want

to thank Dr. Sara Finney, Dr. Dana Mastro, Dr. Teny Daugherty, and Dr. Craig

Enders. I can’t imagine going through this experience without meeting and coming

to respect all of you.

To all those who I mentioned and forgot to mention. . .thank you so much?!

TABLE OF CONTENTS

INTRODUCTION ................................................................. 1
CHAPTER I
Understanding lntemet Usage .......................................... 3
Applying Social Cognitive Theory (SCT) ............................ 7
Internal Events That Shape Perception ....................... 9
Social Environmental Factors ................................ 14
CHAPTER 2
Method .................................................................... 23
Participants ...................................................... 23
Pretest ............................................................ 23
Procedure ........................................................ 24
Exogenous Variables .......................................... 25
Endogenous Variables ......................................... 32
CHAPTER 3
Results .................................................................... 39
Hypotheses ...................................................... 42
Path Model Analysis .......................................... 46
CHAPTER 4
Discussion ............................................................... 52
Future Research ................................................ 55
Implications ..................................................... 60
Limitations ...................................................... 61
APPENDICES
Appendix A
Pretest Questionnaire .......................................... 63
Appendix B
Actual Questionnaire .......................................... 78
REFERENCES ................................................................... 91

vi

LIST OF TABLES

TABLE 1
Prior Experience of lntemet Use .......................................... 39

TABLE 2
Amount and Type of lntemet Use ........................................ 40

TABLE 3
Pearson Correlation Coefﬁcients among the
Antecedents of lntemet Self-Efficacy .................................... 41

TABLE 4
Pearson Correlation Coefﬁcients among Outcome Expectations. . ..42

vii

LIST OF FIGURES

FIGURE 1
The triadic reciprocal causation model of human agency
within Social Cognitive Theory ........................................ 7

FIGURE 2
The incorporation of lntemet use as behavior and outcome
expectations and self-efﬁcacy judgments as conceptualized
in the triadic reciprocal causation model of human agency

within Social Cognitive Theory ........................................ 9
FIGURE 3

The relationship between self-efﬁcacy judgments and outcome

expectations as conceptualized within the cognitive domain ...... 13
FIGURE 4

The implementation of social environmental factors of thought
veriﬁcation as conceptualized in the triadic reciprocal causation
model of human agency within Social Cognitive Theory .......... 15

FIGURE 5
Research Model .......................................................... 22

FIGURE 6
Pearson Correlation Coefficients as Conceptualized with the

Research Model ......................................................... 45

FIGURE 7
A Social Cognitive Model of lntemet Use ........................... 47

FIGURE 8 . . .
A Revised Conceptual Model Examining the Socral Cognitive

Model of lntemet Use ................................................... 49
FIGURE 9 . .

A Revised Model Testing the Social Cognitive Model of

lntemet Use .............................................................. 50

FIGURE 10
Distinct Models of lntemet Use ....................................... 57

VIII t,

Introduction

Research investigating the lntemet has transitioned from basic questions
regarding who and how many are online to why people are going online and with what
effects. This shift has been led by research which suggests that being online increases an
individual’s level of social isolation (Nie & Embring, 2000), subsequently decreasing
psychological well being. In addition to psychological effects, researchers also have
begun to uncover a racial divide among lntemet users (Hoffman, Novak, and Schlosser,
2000; NTIA, 1999). Initially, access was thought to be the only barrier to minority
Internet use. However, recent ﬁndings also suggest that a lack of use by minorities could
also be due to psychological barriers (Scholﬁeld, Davidson, Stocks, & Fortman, in press).

Through the use of traditional models of media behavior such as Uses and
Gratiﬁcations, communication scholars have laid the foundation to understanding.
motivations for lntemet use (Charney & Greenberg, in press; Papacharissi & Rubin,
2000). In order to further this research, it is necessary to assess additional cognitive
mechanisms that lead to the behavior of lntemet use. To do this, Social Cognitive
Theory as posited by Bandura (1986) was tested. Social Cognitive Theory (SCT)
allowed for an investigation of the relationship among behavior, cognitive, and
environmental events as interacting determinants of each other. Here, it was possible to
investigate how internal reasoning as fOrmed through enactive and vicarious learning as
well as verbal persuasion affect behavior.

Therefore, through the framework of SCT, this study examines the

interrelationships among outcome expectancies, self-efficacy judgments, previous

lntemet experience, reference group Internet use, success and encouragement, social

group success, media portrayals, lntemet anxiety, and lntemet use.

CHAPTER I
Understandingﬁtemet Usage

Our understanding of Internet users has progressed over the past several years,
however, many important questions surrounding the decision to use the lntemet remain
unanswered. That said, this research evaluates how researchers have studied the Internet
by discussing who and why people are going online. Further, using social cognitive
theory, this study theoretically and empirically examines how internally formed
perceptual models inﬂuence how and why people use the lntemet.

Early research conducted to understand general computing technology use applied
the theory of reasoned action as posited by Fishbein and Ajazen (I975). Reasoned action
was the ﬁrst theory to gain acceptance in understanding why people adopt information
systems. This framework considers the beliefs that an individual has about a behavior
and the actions taken from those beliefs. Speciﬁcally, the theory of reasoned action
indicates that an individual would use an information system if they could associate
positive beneﬁts with use. While this approach has had success in predicting the use of
computing technologies (Davis, Bagozzi, & Warsaw, 1989), later studies indicated that
modifications to the model were needed. For example, by linking beliefs to emotion and
future consequences (Triandis, 1980; Thompson, Higgins & Howell, 1991) as well as
self-efﬁcacy judgments (Compeau & Higgins, 1995) researchers have increased the
ability to predict and understand the use of various information systems.

The latest information system to be subjected to investigations is the lntemet.
Currently, 51 percent of US households have at least one computer, and approximately

42 percent of them are connected to the lntemet (NTIA, 2000); producing an estimated

lO4-million adult US lntemet users (Pew Research Center, 2000). Of these, 30-million
are under the age of 18. Further, 73 percent of those between the ages of 12 and 17 are
active Internet users while 29 percent of those under the are of 12 have accessed the
lntemet at sometime. Overall, there are slightly more females (51%) on the Internet than
males (49%).

The number one use of the lntemet is email (Nie & Embring, 2000; Lubans, 1999;
NTIA, 2000) with a reported 80 - 90% of all users engaging in this activity. Other tOp
uses of the Internet include playing games, surﬁng, accessing databases and engaging in
interpersonal activities such as discussion groups and chat rooms (Lubans, 1999; Nie &
Embring, 2000). In addition to email use, information gathering also has been heralded
as a primary use of the lntemet (Lubans, 1999).

In an attempt to understand why people use the lntemet, many researchers have
applied traditional models of media behavior. Within the ﬁeld of mass communication,
Uses and Gratiﬁcations is the dominant paradigm for examining lntemet use (Charney &
Greenberg, in press; Kaye, I996; Korgaonkar & Wolin, 1999; Lin, 1999; Papacharissi &
Rubin, 2000). With its emphasis on active media use and its ability to explain both mass
and interpersonal communication, uses and gratiﬁcations is a useful paradigm from
which to begin assessing the Internet.

Kaye (1996) applied well-established gratiﬁcation factors for television viewing
to the lntemet. She found moderate correlations between the amount of weekly Internet
usage and entertainment, social interaction and escape gratiﬁcations. Korgaonkar and
Wolin (1999) found that factors of escapism, information control, interactive control

(relating to the ability to control the presentatiOn of information), socialization and

economic motivation successfully distinguished between low and high lntemet users.
Further, Lin (1999) also identiﬁed motivations for using online services such as
surveillance, companionship, identity, and entertainment as predictors of adoption.

In an attempt to give the lntemet its own uses and gratiﬁcations personality,
Chamey and Greenberg (in press) and Papacharissi and Rubin (2000) found three
common gratiﬁcation factors for the lntemet, keep informed, entertainment, and
communication. Other factors included, (I) develop a new identity, (2) improve mood,
(3) view aesthetically pleasing sights, (4) improve one’s career status, (5) improve social
standing (Charney & Greenberg, in press), (6) take up free time, and (7) make life easier
(Papacharissi & Rubin, 2000).

The development of research investigating who goes online and why people go
online has steadily advanced. As discussed, many psychologists and social scientists
have used the theoretical framework of uses and gratiﬁcations. The uses and gratiﬁcation
model starts with individual needs, suggesting that in order to use a medium, an
individual only has to deem it appropriate to fulﬁll its needs. However, the complex
nature of using the lntemet as well as information systems in general suggests that an
evaluation of self-regulatory mechanisms that inﬂuence lntemet use such as self-efﬁcacy
is essential.

Another route to understanding lntemet use is through the investigation of social
issues such as the digital divide. The digital divide or the difference that separates white,
high income lntemet users from minority, lower income non-users has been a primary
tOpic of discussion among both policy makers (NTIA, 2000, 1999) and social scientists

(Hoffman & Novak, 1998; Hoffman, Novak, & Schlosser, 2000). The digital divide has

been Operationalized primarily in terms of race and class discrimination that is reﬂected
in unequal access to computers and the lntemet. Currently, the divide is still compelling
with Aﬁican American and Hispanic homes recording 23 percent penetration; compared
to the 42 percent penetration nationally (NT IA, 2000). Further, only 19 percent of
African Americans and 16 percent of Hispanics report actually using the lntemet at
home.

While the importance of socio-economic and racial factors cannot be denied,
prospective lntemet users also face psychological barriers. Research has indicated that
new Internet users are less comfortable using the lntemet, are less satisﬁed with their
Internet skills and are more likely to encounter stress-inducing problem situations (GVU,
1999). So, while cost and lack of access are the most signiﬁcant barriers to initial use of
the lntemet, uncertainty about initial use and the perception that computers are too
complicated are also important compOnents (Katz & Aspen, I 996).

The most convincing evidence of this can be seen in the classroom (Scholﬁeld,
1997; Scholﬁeld, Davidson, Stocks, & Fortman, in press). As mentioned, access is
considered the primary barrier to minorities. However, when given access via the
classroom, usage patterns tend to mimic general society. Here, higher levels of
familiarity with computer applications (primarily due to home use) have caused
Caucasians to dominate lntemet use in technologically equipped classrooms. These
ﬁndings have led researchers to begin considering Internet use from a cogritive
perspective (Jackson, Barbatsis, Biocca, Fitzgerald, & VonEye, 2000).

Using the knowledge provided by the uses and gratiﬁcation research, this study

will expand current understanding of lntemet use by framing the decision to use within

Social Cognitive Theory (Bandura, 1986). By doing so, psychological factors that
inﬂuence internal perceptions and subsequent lntemet use will be assessed. This type of
investigation will provide a more encompassing picture of why people decide to use the

lntemet.

Applﬂng Social Cognitive Theory (SCT)

Through numerous empirical investigations, SCT has provided social scientists
with a model that enables the examination of individual behavior (Bandura, 1986).
Using a triadic reciprocal causation model, SCT examines the relationships among
behavior, cognitive, and environmental events as interacting determinants that inﬂuence

each other bi—directionally (see Figure 1).

Internal Events (Cognition)

/ That Shape Perception \

B h ' + Environment
e avror

 

Figure l. The triadic reciprocal causation model of human agency within Social
Cognitive Theory.

That is, (a) the environment in which an individual exists inﬂuences the internal events,
which then aﬁ'ects the environment; (b) behavior in a given situation is affected by
environmental elements, which are then affected by an individual’s behavior; and (c)

behavior is inﬂuenced by cognitive factors, which are inﬂuenced by the behavior. This

reciprocal relationship makes an individual both the product and producer of its
environment.

From this perspective, Bandura (1994) characterizes human nature as “a vast
potentiality that can be fashioned by direct and observational experience into a variety of
forms within biological limits” (p. 62). Therefore, humans contain the inherent
capability to examine a course of action, evaluate the consequences of that action against
previously held judgments, and then modify existing judgments accordingly. The triadic
model is mediated by these capabilities that transfonn sensory experiences into individual
cognitive models of actions.

According to Bandura (1986), SCT includes a complex causal structure that
establishes the development of competency and the regulation of action. Knowledge
structures created represent cognitive strategies of effective action which guide behavior.
These cognitive models afford individuals the ability to produce skills as wells as internal
standards needed to proﬁciently execute a behavior. This cognitive guidance toward
behavior is key in the developmental stages of a behavior. Understanding the importance
in cognitive development as an inﬂuence of behavior, this study only focuses on the
primary cognitive factors that inﬂuence individual behavior. That said, there are two
prominent expectations guiding behavior within SCT, outcome and self-efﬁcacy
judgmentsmandura, 1986). Both outcome expectations and self-efficacy are

conceptualized within the cognitive domain of SCT (see Figure 2).

Self-Efﬁcacy Judgments;
Outcome Expectations

/ \.

lntemet Use + > Environment

 

 

Figure 2: The incorporation of Internet use as behavior and outcome expectations and
self-efﬁcacy judgments as conceptualized in the triadic reciprocal causation
model of human agency within Social Cognitive Theory.

Internal Events That Shag Perception

{Outcome Expecfatibhs are deﬁned as the perceived likely consequences of
performing a behavior (Bandura, 1997). While rewards are typically thought of as
external (i.e., monetary rewards), SCT focuses on the idea that rewards are often internal.

Two distinct types of outcome expectations are recognized in SCT: social and personal.

Each form contains positive (i.e., incentives) and negative (i.e., disincentives)

expectations. Positive social outcomes include positive social reaction of others,

approval, social recognition, etc., while negative effects include items such as social
rejection and disinterest. Positive personal outcomes establish self-satisfaction, pride,
self-worth, etc., while negative outcomes produce self-dissatisfaction and self-

devaluation (Bandura, 1997).

Past research evaluating outcome expectations has indicated that perceived
usefulness of various information systems such as computers (Compeau, Higgins, &

Huff, 1999; Davis, 1989; Thompson, Higgins, & Howell, 1991) increases use of that

technology. Both performance and personally based expectations of computer use were
found to be signiﬁcantly related. In terms of the Internet, social outcomes would derive
from social encounters online, while personal outcomes would develop from online
personal achievements, such as obtaining information or being entertained. Guided by
Chamey and Greenberg (in press), LaRose, Mastro, and Eastin (in press) assessed the
relationship between outcome expectations and lntemet use. They found that expected
outcomes were signiﬁcantly related to lntemet use. Speciﬁcally, pleasing sensory, novel
sensory, and social outcome expectations were identiﬁed as being positively related to
Internet use, and negative outcomes as being negatively related to Internet use. Here,
negative outcomes were deﬁned as the likelihood of encountering negative effects
associated with lntemet use such as having your browser freeze-up and getting blocked
by password protected sites.

Given these results, the outcome expectations identiﬁed by Bandura ( l 986; 1997),
and the most common uses of the lntemet (Lubans, I999; Nie & Erbring, 2000), the

following relationships between outcome expectations and Internet use are hypothesized:

H1 The higher the social outcome expectations, the higher the lntemet use.

H2 The higher the personal outcome expectations in informational and entertainment
activities, the higher the Internet use. -

H3 The higher the negative outcome expectations, the lower the Internet use.

The second component of internal judgments is self-efficacy. As a central
component of SCT, self-efﬁcacy is deﬁned as the belief ”in one's capabilities to organize

and execute the courses of action required to produce given attainments" (Bandura, I997,

IO

p. 3). This cognitive construct has long been recognized as an important determinant of
individual behavior within SCT. Speciﬁcally, self-efﬁcacy is a form of self-evaluation
that inﬂuences decisions about what behaviors to undertake, the amount of effort and
persistence put forth when faced with obstacles, and ﬁnally, the mastery of the behavior.
Self-efﬁcacy is not a measure of skill; rather,it reﬂects what an individual believes they
can do with the skills they possess. Self—efﬁcacy can be distinguishable from self-
esteem, in that self-esteem is deﬁned as the general belief in one’s self-worth and general
beliefs about'oneself, while self-efﬁcacy is ’speciﬁc to a particular behavior domain (i.e.,
Internet use). Furthermore, aﬁer compiling and assessing several experimental studies on
self-efﬁcacy, Bandura (1982) found perceived efﬁcacy toward a given behavior to be a
better predictor of behavior than past performance.

From an information technology view, the relationship between computer self-
efﬁcacy and computer use has been studied oﬁen (Burkhart & Brass, 1990; Compeau &
Higgins, 1995; Oliver & Shapiro, 1993). Personal computers represent a complex and
somewhat troublesome technology, requiring considerable skill and extensive training to
operate successfully. Staying consistent with the concept of self-efﬁcacy, Compeau and
Higgins (1995) distinguished between component skills such as formatting disks and
booting up the computer and behaviors individuals can accomplish with these skills, such
as using software to analyze data.

Internet self-efﬁcacy focuses on what a person believes can be accomplished
online. It does not refer to a person’s skill at writing HTML, using a browser, or
transferring ﬁles. Instead, it assesses judgments of their ability to apply these types of

skills in'a more encompassing mode (e.g., ﬁnding information online). Using the lntemet

requires a set of skills more advanced than those required by traditional computer use,
which may be considered daunting to beginning users. These include establishing and
maintaining a stable lntemet connection, learning how to navigate on the lntemet and
how to search for relevant information. Thus, lntemet self-efficacy may be distinguished
from computer self—efﬁcacy as the beliefs that one can successfully perform a distinct set
of behaviors required to establish, maintain and effectively utilize the lntemet over and A
above basic personal computer skills.

Early research on Internet self-efficacy focused on the performance of speciﬁc
tasks such as entering World-Wide Web addresses, creating folders and bookmarks,
mailing pages, using the File Transfer Protocol (FTP) and telnet, constructing a hypertext
index, and moving bookmarks (N ahl, I996; Nahl, 1997). Recently, researchers have
evaluated the relationship between lntemet self-efﬁcacy and performance (N ahl, I996;
Nahl, 1997; Staples, Hulland, & Higgins, 1998) and prior use of the Internet (Eastin &
LaRose, 2000). Results were consistent with previous self-efﬁcacy literature, with self-
effrcacy perceptions positively related to task performance (N ahl, 1996; Nahl, 1997;
Staples et al., 1998) and prior use (Eastin & LaRose, 2000).

People who have little conﬁdence in their ability to use'the lntemet, or who are
uncomfortable using the Internet may be said to have weak self-efﬁcacy beliefs.
According to SCT, those with low lntemet self-efficacy should be less likely to use the

Internet than those with high degrees of self-efﬁcacy.

H4 The higher the Internet selﬁeﬂicacy judgments, the higher the Internet use.

12

In addition to the relationship between self-efﬁcacy and Internet use, self-efﬁcacy is
also considered a causal antecedent to outcome expectations (see Figure 3). As
previously stated, outcome expectations are deﬁned as the perceived likely consequences
of performing a behavior (Bandura, 1997). This relationship demonstrates the interplay
among variables within the cognitive domain of SCT. Oliver and Shapiro (l 993), found
that the stronger a person's self-efﬁcacy beliefs, the more likely they were to try to
achieve the desired outcome, thus, supporting the persistence characteristic inherent to
self-efficacy judgments. Furthermore, results from Compeau & Higgins (1995) analysis
of computer use indicated that computer self-efﬁcacy inﬂuenced expectations about the

future outcomes of computer use such as job performance and personal accomplishment.

Self-Efﬁcacy Judgments

l

Outcome Expectations

/ '\

lntemet Use 4 > Environment

 

 

Figure 3: The relationship between self-efﬁcacy judgments and outcome expectations .
as conceptualized within the cognitive domain.

In an assessment of online outcomes, Eastin and LaRose (2000) found expectations
such as informational to be signiﬁcantly related to lntemet self-efﬁcacy. In the present
context this means that Internet self-efﬁcacy should be positively related to the

expectation of positive outcomes from Internet use, such as obtaining useful information,

13

being entertained and social interactions. Conversely, negative outcome expectations are

expected to be lower when lntemet self-efﬁcacy perceptions are high.

H5 The higher the Internet self-efﬁcacy judgments, the higher the social outcome
expectations.

H6 The higher the Internet self-eﬂicacy judgments, the higher the personal outcome
expectations

H7 The higher the lntemet self-efﬁcacy judgments, the lower the negative outcome
expectations.

Social Environmental Factors

Individuals have the capacity of self-reﬂection. Through self-reﬂective means,
individuals establish, maintain, and reconstruct cognitive models such as self-efficacy
perceptions. People are able to understand causal relationships and increase their
knowledge through personal and non-personal experiences within their environment.
Hence, social factors within an individual’s environment inﬂuence how cognitive models
are developed. For example, when engaging in a behavior an individual develops a
positive or negative model of the experience. The individual then uses this model to
determine the extent of future participation. Social Cognitive Theory suggests that the
thought veriﬁcation process occurs within three modes, enactive, vicarious, and

persuasive (see Figure 4).

l4

Self-Efficacy Judgments

l

Outcome Expectancies

/ ‘ \

lntemet Use ‘ + Social Environmental Factors:
enactive, vicarious, persuasive

 

 

Figure 4: The implementation of social environmental factors of thought veriﬁcation
as conceptualized in the triadic reciprocal causation model of human agency
within Social Cognitive Theory.

Enactive inﬂuences are developed by an individual through actual experience
with a behavior. Positive interactions reinforce and increase previous ability models,
while negative interactions weaken previously constructed models. Past research has
indicated prior experience as an important antecedent of self-efﬁcacy (Lewis, 1985). For
example, math skills are needed in computer programming, and math skills and number
of math courses taken play an important role in individual judgments about their
programming ability (Bandura, 1997; Oliver & Shapiro 1993). Research on computer
self- efficacy supporting these ﬁndings indicates a signiﬁcant positive relationship
between computer self-efﬁcacy and computer usage (Burkhart & Brass, 1990; Oliver &
Shapiro, 1993).

The positive relationship between lntemet use and lntemet self-efﬁcacy has been
empirically validated (Eastin & LaRose, 2000; Staples, et al., 2000). Staples et al., (2000)
found that direct experience with remote work environments increased remote work self-
efﬁcacy, while Eastin and LaRose (2000) indicated that past Internet use (e.g., number of

months using the lntemet) positively inﬂuence Internet efﬁcacy judgments. Thus, the

IS

following relationship between lntemet experience and lntemet self—efﬁcacy is

hypothesized:
H8 The greater the prior Internet experience, the greater the Internet self-eﬂicary.

Vicarious learning allows an individual to shape cognitive models without physically
participating in a speciﬁc behavior. The behavior and subsequent effects they observe
can serve as a mechanism for which to compare previous interpretations. This modeling
mechanism can be broken into three different areas, use by others, success of others and
media portrayals. First, the inﬂuence of vicarious learning occurs when use observation
occurs within an individual’s reference group (Bandura, 1986). The behavior of an
individual’s reference group use is used to form and evaluate personal self-efﬁcacy
beliefs. Learning through observation has been established within SCT literature as a
, method individuals use to determine whether or not to participate in a given behavior
(Compeau & Higgins, 1995; Manz & Sims, 1986; Schunk, 1981). According to Bandura
, (l 997), the relationship between the inﬂuence of others’ use and adoption of the behavior
is mediated by self-efficacy. Given this, it is plausible to suggest that the amount of
exposure an individual has to others being online will inﬂuence their lntemet self-
efﬁcacy perceptions.

More speciﬁcally, research indicates that these effects are stronger when the
individuals being observed are in a person’s reference group (Bandura, 1986). An
individual’s reference group consists of people from whom the individual seeks guidance

(Bandura, 1986; Cross, 2000). Here, it is hypothesized:

l6

H9 The higher Internet use is within the reference group, the greater the level of
Internet self-efficacy.

The second source of modeling formation comes from the concept of visualizing
others’ attainments. Eﬁicacy believes are raised when an individual visualizes another
experience as being positive (Bandura, 1997). An individual uses perceptions of others
ability to successfully perform a behavior to persuade themselves that they too have the
ability (Bandura, 1982, Schunk, Hanson, & Cox, 1987, Staples, 1999). At the same time,
when an individual observes the failure of another, judgments in their own ability are
lowered. Again, it is important that the observed or visualized individual is seen as
similar, as the greater the similarity the greater the effect. In this case, while perceived
reference group success will have the greatest effect, it is also hypothesized that the
success of social group members thought to be similar, will also inﬂuence efﬁcacy
judgments.

H10 The higher the perceived success of Internet use within an individual ’s reference
group, the greater the level of Internet self-efficacy.
H11 The higher the perceived success of Internet use by social group members, the
' greater the level of Internet self-efficacy.

The third form of thought veriﬁcation is verbal persuasion. Verbal persuasion is
used to convince people that they possess the capabilities to achieve a sought behavior.
Within SCT, persuasion often takes on the form of encouragement by others (Bandura,
1986). Conceptually, this suggests that peOple rely on the beliefs of others to form
judgments of their abilities (i.e., self-efﬁcacy). Verbal persuasion is thought to be more

effective when the encouragement comes ﬁom someone in the individual’s reference

[7

group. Here, an individual’s reference group consists of people from whom the
individual seeks guidance.

Assessing computer use, Compeau and Higgins ( I 995) found that encouragement
from others did signiﬁcantly inﬂuence computer self-efﬁcacy. Internet research
investigating social support has yet to examine a person’s reference group in the physical
world. However, research does exist that suggests online relationships can develop and
create a support mechanism (Katz & Aspden, 1997). In some cases these online
relationships actually strengthen real world social ties (Wellman & Gulia, I999).
Activities identiﬁed include sending and receiving e-mail and participating in discussion
groups and chatrooms. In order to extend current understanding of support, the
relationship between lntemet self-efﬁcacy and real world social support as
Operationalized within SCT will be tested.

H12 The greater the encouragement to use the Internet by the reference group, the
greater the level of In ternet self-efﬁcacy.

Up to this point, modelng through vicarious learning has been deﬁned within the
behavior of others. However, throughout a given day, the amount of direct contact that
an individual has with others in the physical and social environment is limited by their
daily routines. This narrow avenue for experience increases the reliance on other
vicarious mechanisms of thought veriﬁcation. For example, vicarious learning occurs by
observing and attending to how a medium such as television portrays and defines a
phenomenon. As with other mechanisms of vicarious learning, this information is then
compared with currently held judgments. This type of inﬂuence allows people to surpass

the physical boundaries of their daily life by deveIOping cognitive models through others

l8

interpretations. Individuals can observe the “attitudes, styles of competency, and
attainments of members of different segments of their society...” (Bandura, 1997, p. 93).
Exposure to these types of models that demonstrate efﬁcient strategies as well as attitudes
toward a behavior, inﬂuences an individuals own belief in their abilities (Bandura, I982;
Schunk, 1987).

Given this and recent ﬁndings that indicate television as an inﬂuential component to
perceptions of computing technology (Cross, 2000), it is expected that perceptions of
media portrayal of the Internet as a positive phenomenon will increase lntemet use.

H13: Positive media portrayals of Internet use will increase Internet self-eﬂ‘icacy
judgments. .

Finally, SCT discusses the ability of efﬁcacy beliefs to regulate affective states
(Bandura, 1997). From one perspective, it is posited that people rely on psychological
and emotional states to form efficacy judgments. Here, anxiety is an antecedent of self-
efficacy. These states are considered most inﬂuential when people are faced with
physical activities, health functioning, and coping with stressful or taxing situations
(Bandura, 1997). However, another concept indicates that self-efﬁcacy affects
psychological states; here it is thought that people’s self-efﬁcacy beliefs affect the
amount of stress (i.e., anxiety) they encounter or perceive. Simply stated, those who have
high self-efﬁcacy judgments do not envision that they will encounter stressful situations
(Bandura, 1989; Compeau & Higgins, I995; Compeau et al., 1999). The level of stress a
person envisions subsequently inﬂuences whether or not they engage in a particular
behavior. In the case of lntemet use, an individual ’3 level of lntemet self-efﬁcacy is

thought to inﬂuence the amount of stress they perceive will be encountered while online.

I9

These levels of perceived stress will determine the level of lntemet use an individual will
engage in. The goal of this study is to understand lntemet use. Thus, while perceived
stress has been considered an antecedent of self-efﬁcacy, this study will evaluate stress
(i.e., anxiety) as a mediating variable between self-efﬁcacy judgments and level of
lntemet use.

That said, perceived stresses toward a behavior can cause arousal in the form of
anxiety. Anxiety is deﬁned “as a state of anticipatory apprehension over possible
deleterious happening” (Bandura, 1997, p. 137). Initially, the majority of support for this
relationship was found between low mathematical efﬁcacy and high math anxiety
(Bandura, I997). Recent studies assessing cognitive models and new technologies have
found that people with high levels of computer (Compeau & Higgins, 1995) and remote
wOrking (i.e., telecommuting) (Staples, et al., 1998) self-efﬁcacy have low levels of
computer anxiety.

Feelings of anxiety are also found to predict behavior. For example, Webster,
Heian, and Michelman (1990), Compeau and Higgins (1995), Compeau et al., (1999)
found a strong relationship between level of computer anxiety and computer use. While
the relationship between computer anxiety and computer use has received a considerable
amount of attention, separating out anxious feeling toward the lntemet has not.
Therefore, given the relationships between self-efficacy, anxiety, and lntemet use the
following are hypothesized.
H14 The higher the Internet self-efficacy judgments, the lower the anxieties toward

using the Internet.

H15 The lower the anra'eties toward using the Internet, the higher the Internet use.

20

Beyond the investigation of bivariate relationships, this research also considers how
these constructs inﬂuence each other both directly and indirectly. To accomplish this, a
social cognitive model demonstrating the deveIOpment of lntemet use was constructed.
The research model guiding this study was developed from the conceptual framework of
social cognitive theory as well as previous research on Internet use (Bandura, 1986;
Bandura, 1997; Compeau et al., 1999; Eastin & LaRose, 2000) (see Figure l, 2, 3, and 4).
The model assesses lntemet use and self-efﬁcacy by identifying inﬂuential factors (i.e.,
antecedents) such as prior experience, reference group Internet use, ability and support,
social group perceptions, media portrayals social outcome expectations and anxiety. By
assessing the direct and indirect relationships among previous lntemet experience,
reference group lntemet use, ability and verbal support, social group success, media
portrayals, Internet anxiety, lntemet self-efﬁcacy judgments, outcome expectancies and
lntemet use a better understanding of cognitive mapping will be developed (see Figure
5).

Speciﬁcally, lntemet experience, parent’s and ﬁiend’s Internet use, success, and
encouragement, social group success, and media portrayals are the eight exogenous
predictors of Internet self-efficacy. These efﬁcacy judgments are then used to predict
social, entertainment, informational, and negative outcome expectations, lntemet anxiety,
and lntemet use. Social, entertainment, informational, and negative outcome expectations

and lntemet anxiety also are used to predict Internet use.

21

ﬁgure 5. ResearcLMMel

 

Prior lntemet
Experience

 

Friend’s lntemet
Use

 

Parent’s Internet
Use

 

Friends’
Encouragement of
lntemet Use

 

Parent’s
Encouragement of
Internet Use

 

Friend’s Success at
lntemet Use

 

Parent’s Success at
lntemet Use

 

Success, of Social
Group

 

 

Media Portrayal of
Internet Use

\\\W//

 

 

 

 

lntemet Self-

Efﬁcacy +

 

 

 

+

 

 

22

 

+

j

 

 

Social

Outcome _,

 

Expectations

 

 

 

lntemet
Anxiety

 

 

 

lntemet
Use

 

 

.ﬁ Entertainment

Outcome
Expectations

 

 

Information

’ Outcome

 

Expectations

 

 

 

 

Negative
Outcome
Expectations

 

 

 

 

CHAPTER 2
Methods
In order to examine the hypothesized relationships among environmental
inﬂuences, cognitive models and lntemet use, the present study was conducted. Pearson
Correlations were used to test each of the hypothesized relationships, while path analysis
was used to test the proposed theoretical model. Corrections for attenuation were made

when testing the path model.

Participants

A total of 260 high school students from Midwest and Southwest high schools
participated in the study. Sixty-three of the Midwest students took part in the pretest
while the remaining 197 students from the Southwest school were used in actual
analyses. Of those used in the ﬁnal analyses, 106 (54%) were females and 91 (46%) were
male, with a mean age of 17 years old. Respondents’ race was distributed as follows: 57
(29%) were Hispanic/ Latino, 101 (51%) Caucasian, 25 (13%) African American, 5 (3%)

Asian, 4 (2%) Native Americans, and 4 (2%) were identiﬁed as ‘other.’

m

Prior to the actual data collection, a pretest was conducted with students from a
Midwest school (N = 63). Speciﬁcally, an empirical investigation was performed on
each of the constructs under consideration. Other logistical survey considerations such as
survey length and question wording were also assessed. All constructs were tested using
conﬁrmatory factor analysis (CFA) (Hunter & Gerbing, 1982). CFA tests a speciﬁed

factor structure based on theory-driven indicators. The pretest instrument can be found in

23

Appendix A. The conﬁrmatory factor process also was conducted on actual study
participants; including tests of internal consistency and parallelism where needed.
Changes made from the pretest include (I) all reversed items were changed to
more clearly identify negatively phrased questions. For example, items changed from
‘unsuccessful’ or ‘unfavorable’ to ‘NOT successﬁrl’ and ‘NOT favorable.’ (2) At the
beginning of each section, participants read a passage that clearly stated that some of the
items were negatively phrased. (3) All usage items were changed from a seven-point
scale to open-ended (Schwarz, 1999). (4) Items were changed to more clearly describe
the intended meaning where deemed necessary. And, (5) items breaking weekend
Internet use into Saturday and Sunday were collapsed into one question asking about total

weekend lntemet use.

Procedure

All questionnaires were administered at the beginning of each class period. A
total of 12 junior and senior classes were surveyed. Classes contained between 20 and 30
students. Data collection was conducted over a two-day period. Participants spent
between 20 and 30 minutes completing the questionnaire. Upon completion, the
participants were debriefed on the nature of the study.

The ﬁnal questionnaire can be found in Appendix B. With the exception of
parent’s success (N = 152), the number of subjects used in each construct ranges ﬁom

191 to I97.

24

Exogenous Variables

Prior lntemet erggerience. lntemet experience was measured by determining the number
of years each participant had used the lntemet. Past research has measured prior lntemet
experience as the amount of time an individual has been using the lntemet (Eastin &
LaRose, 2000). However, this overall measure does not provide all the necessary
information to adequately understand what peOple have been doing online in the past. In
addition, categorizing activities can serve as a cueing mechanism increasing response
accuracy.

Therefore, this study assessed how long each participant has been accessing the
Internet as well as the activities in which they engage (e. g., playing games, searching for
information, listening to music, etc.). Speciﬁcally, what and how long a person has been
participating in each activity was measured. Five Open-ended items were used to assess
the number of years a person has used the lntemet to (I) gather information, (2) play
games, (3) listen to music, (4) meet people, and (5) email. These items were aggregated

to produce a prior Internet experience (M = 10.33, SD = 6- 13)-

How many years have you used the Internet for gathering information? year(s)

How many years have you been playing video games on the Internet? __ycar(s)

How many years have you been listening to music on the Internet? year(s)
How many years have you been meeting people (e.g., chat rooms

and discussion groups) on the lntemet? ___)’33T(S)
How many years have you been emailing people? _)’€af(8)

25

Teenage Reference Groum: An individual’s reference group consists of people from
whom the individual seeks guidance. Research suggests that a teenagers reference group
consist of a small group of ﬁiends (i.e., peers) and parents (Cross, 2000). Therefore,
when measuring the amount of Internet use and encouragement from an individual’s
reference group, items focused on a close ﬁiend and a parent. A main concept driving
this study is cognitive modeling through perception, thus, when measuring Internet use,
ability and encouragement by an individual’s reference group, this study assessed
individual perception. Here, it is irrelevant if a parent uses the Internet 20 hours per week
if their child perceives them as spending 10 hours online weekly.

That said, reference group support was deﬁned by six constructs, (I) a friend’s
use; (2) a friend’s ability; (3) ﬁiends’ encouragement; (4) a parent’s use; (5) a parent’s
ability, and (6) parents’ encouragement.

Friend use. A ﬁiend’s use was measured with three open-ended items assessing average
weekday lntemet use at home and school, and home use on the weekend (M = 8. 75
hours, SD = 6.56 hours). Items included:

Think of a close friend, on an average weekday (Monday through Friday), how many
hours does that friend spend on the lntemet at home? _ Hour(s)

Considering that same friend, on an average weekday, how many hours does that friend
spend on the Internet at school? Hour(s)

Considering that same ﬁiend,on an average weekend, how many hours does that ﬁiend
spend on the lntemet? . _ ._..__HOUI'(S)

26

 

 

Friend Success. A friend’s success was measured with three seven point Likert-type
items (a = .83) assessing the perceived success of a close ﬁiend when using the Internet

for general activities (M = 16.09, §_D_ = 4.49). Items included:

Considering that same friend, how successful is that ﬁiend at using the lntemet to
complete school assignments?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

Considering that same ﬁiend, how successful is that ﬁiend at using the lntemet for
entertainment purposes (for example, Internet use that is not schoolwork related)?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

Considering that same ﬁiend, how easy is it for that ﬁiend to use the lntemet for
schoolwork?
1 2 3 4 5 6 7

Not at All Completely
Easy Easy

Friend Encouragement. Friends’ encouragement was measured with three seven-point
Likert-type items (or = .85) ranging ﬁom Strongly Disagree (score = l) to Strongly Agree
(score = 7). This construct measured the amount of encouragement received from ﬁiends

(M = 13.37, SD = 5.37). Items used were as follows:

My ﬁiends encourage me to use the lntemet.
l 2 3 4 5 6 7

Strongly Strongly
Disagree Agree

27

 

My friends discuss how being able to use the Internet will help me in the future.
1 2 3 4 5 6 7
Strongly Strongly
Disagree Agree
My friends tell me that there are a lot of fun things to do on the lntemet.
l 2 3 4 S 6 7
Strongly Strongly
Disagree Agree
Parent use. A parent’s use was measured with three Open-ended items assessing average
weekday lntemet use at home and work, and home use on the weekend (M = 6.42, SD =

6.84). Reports of parents never using the Internet were coded as zero and leﬁ in analyses.

Items included:

Think of your parent who uses the lntemet the most, on an average weekday, how many
hours does that parent spend on the lntemet at home?

Hour(s)

Considering that same parent, on an average weekday, how many hours does that parent
spend on the lntemet at work?

Hour(s)

Considering that same parent, on an average weekend, how many hours does that parent
spend on the lntemet? ‘

Hour(s)

Parent Success. A parent’s ability was measured with four seven point Likert-type items
(or = .91) assessing the perceived successfulness of a parent when using the lntemet for
general activities (M = 19.96, SD = 6.82). These items where not completed by

respondents who reported their parents had never use the lntemet. Items used included:

28

Considering that same parent, how successful is that parent at using the lntemet for work
related (for example, information gathering) projects?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

Considering that same parent, how successful is that parent at using the lntemet for
entertainment (for example, lntemet use that is not work related)?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

Considering that same parent, how easy is it for that parent to use the Internet for work
related projects?

1 2 3 4 5 6 7
Not at All Completely
Easy Easy

Considering that same parent, how easy is it for that parent to use the lntemet for
entertainment (for example, lntemet use that is not work related)?

1 2 3 4 5 6 7

Not at All Completely
Easy Easy

Parent Encouragement. A parent’s encouragement was measured with three seven point
Likert-type items (or = .85) ranging ﬁom Strongly Disagree (score = l) to Strongly Agree
(score = 7). This construct measured the amount of encouragement received from a

parent (M = 12.29, S_D = 5.39). Items used are as follows:

My mom or dad tells me that the lntemet is a great source of information.
1 2 3 4 5 6 7

Strongly Strongly
Disagree Agree

29

My mom or dad discusses how being able to use the lntemet will help me in the future.

1 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

My mom or dad tells me that there are a lot of ﬁrn things to do on the lntemet.
1 2 3 4 5 6 7 ‘

Strongly Strongly
Disagree Agree

As stated, while reference group support is thought to play a key role in cognitive
development, there are other more general social inﬂuences that also need to be
considered (Compeau & Higgins, 1995, Bandura, 1986). These other support
mechanisms are identiﬁed as perceived success of social group at using the Internet and

media portrayals of the Internet.

Social Group Success. Perceptions of social group success was measured with four
Likert-type items (or = .91) assessing the success of kids ‘the same age’ have when using

the lntemet (M = 22.60, S_D = 5.09). Items included:

In general, how successful are kids your age at using the lntemet to help them with their
schoolwork?

l 2 3 4 5 6 7
Not at All Completely
Successful Successful

In general, how successful are kids your age at using the lntemet for entertainment
pmroses? . '

l 2 3 4 5 6 7
Not at All Completely
Successful Successful

30

In general, how easy is it for kids your age to use the lntemet for schoolwork related
projects?

1 2 3 4 5 6 7
Not at All Completely
Easy Easy

In general, how easy is it for kids your age to use the Internet for entertainment purposes
(for example, lntemet use that is not work related)? '

l 2 3 4 5 6 7
Not at All Completely
Easy Easy

Media Portrayal. Media portrayal was measured with four Likert-type items ranging from
Not at All (score I) to A Lot (score 7) (or = .83), which assessed how the media portray

lntemet use (M = 17.61, _S_D = 5.65). The following items were included:

In general, how often do you see mph on television programs and commercials (not
including news programs) use the lntemet to help them?

1 2 3 4 5 6 7
Not at All . A lot

In general, how often do you see kids your age on television and commercials (not
including news programs) use the lntemet to help them?

1 2 3 4 5 6 7
Not at All A lot

In general, how often do news programs talk about the lntemet in a favorable way?

1 2 3 4 5 6 7
Not at All A lot

31

In general, how often do you see people in movies use the lntemet to help them?

1 2 3 4 5 6 7
Not at All A lot

Endogenous Variables

A total of four outcome expectancies were measured. These included social,
personal, which contained both information and entertaimnent expectancies, and negative
outcome expectancies. Items created for each of the outcome expectancies were
borrowed from the research of Chamey and Greenberg (in press) and Eastin and LaRose

(2000).

Social Outcome Expectancies. Social expectancies were constructed using four seven
point Likert items ranging from Strongly Disagree (score I) to Strongly Agree (score 7)
(a = .73). This construct measured the perceived likelihood of developing relationships

over the lntemet (M = 16.74, S_D = 6.22). Items used included:

I am likely to establish a romantic relationship on the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am likely to get in touch with people I know on the lntemet.

, l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am NOT likely to meet new ﬁiends on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

32

I am NOT likely to visit chat areas to talk with other people.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

Both information and entertainment outcome expectancies were constructed using

ﬁve Likert items ranging ﬁom Strongly Disagree (score I) to Strongly Agree (score 7).

Information Outcome Expectancies. The information outcome expectancy construct

included three items which assessed the likelihood of obtaining information on the

lntemet (M = 16.28, S_D = 4.59) (or = .72). Items included:

I am likely to get immediate information about big news events on the lntemet.

l 2 . 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am likely to ﬁnd information to complete a class assignment on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

1 am likely to get information about products such as cars or clothes on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

Entertainment Outcome Expectancies. The entertainment outcome expectancy construct

was measured with two items assessed the likelihood of being entertained while on the

lntemet (M = 10.74, _S__l_)_ = 3.42) (or = .79). Items included:

33

I am likely to ﬁnd a way to pass time on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am likely to be entertained on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

Negative Outcome Expectancies. Three items ranging from Strongly Disagree (score I)
to Strongly Agree (score 7) measured negative outcome expectancies (a = .73). This
construct assessed the likelihood of encountering negative outcomes associated with
lntemet use (M = 11.28, S_D = 4.67). Items included:

I am likely to ﬁnd that search engines do NOT have enough detailed to quickly ﬁnd what
I am looking for.

1 2 3 4 5 v 6 7
Strongly Strongly
Agree Disagree

I am likely to have problems opening large documents found online.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am likely to have long download times on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

34

lntemet Self-EfﬁcacL Using items created by Eastin & LaRose (2000), Internet self-
efficacy was measured with seven Likert items ranging from Strongly Disagree (score I)
to Strongly Agree (score 7) (a = .90). The efﬁcacy construct is thought to measure a

person’s belief in their ability to organize and execute the courses of action required to

produce given behavior (M = 30.20, SQ = 11.15). Items included:

I am conﬁdent understanding terms/words relating to lntemet hardware.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am conﬁdent understanding terms/words relating to lntemet software.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

1 am conﬁdent describing functions of Internet hardware.

1 f 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am conﬁdent trouble shooting lntemet problems.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

1 am conﬁdent explaining why a task will not run on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

35

I am conﬁdent using the lntemet to gather data.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I am conﬁdent learning advanced skills in a speciﬁc lntemet program.

I 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

lntemet Anxiety. AdOpting items from the Compeau et al., (1999) computer anxiety
scale, an lntemet anxiety construct was created using four Likert items ranging ﬁom
Strongly Disagree (score 1) to Strongly Agree (score 7) (a = .90). This construct assessed

the level of anxiety a person feels toward the lntemet (M = 11.07, SD = 6.86). Items

were:

The Internet is somewhat intimidating to me.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I’ hesitate to use the Internet for fear of making a mistake I cannot correct.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

I’m afraid that I might destroy/ lose information on my computer by accessing corrupted
ﬁles while on the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

36

a;

 

Using the lntemet makes me nervous because I’m not sure how much personal
information (for example, age, gender, sites I visit, etc.) is being collected electronically.

1 2 3 4 S 6 7
Strongly Strongly
Agree Disagree

lntemet Use. Internet use was measured with ten Open-ended items. Five items assessed
amount of Internet use on an average weekday, while the other ﬁve items assessed use on
an average weekend. Each unit of measure assessed use in the following areas: (I) gather
information, (2) play games, (3) listen to music, (4) meet people, and (5) email. All ten

items were then aggregated to produce Internet use (M = 9.19, S_D = 7.05).

On an average weekday (Monday through Friday), how many hours do you use the
lntemet to gather information? hour(s)

On an average weekday, how many hours do you use the lntemet to play video games?
hour(s)

On an average weekday, how many hours do you use the Internet to listen to music?

hour(s)

On an average weekday, how many hours do you use the lntemet to meet people (e. g.,
chat rooms and discussion groups)? hour(s)

On an average weekday, how many hours do you use the lntemet to email?

hour(s)

On an average weekend, how many hours do you use the lntemet to gather information?

hour(s)

On an average weekend, how many hours do you use the Internet to play video games?

hour(s)

On an average Weekend, how many hours do you use the Internet to listen to music?

hour(s)

37

On an average weekend, how many hours do you use the lntemet to meet people (e.g.,
chat rooms and discussion groups)? hour(s)

On an average weekend, how many hours do you use the lntemet to email?
hour(s)

38

CHAPTER 3
Results

Overall, ninety-three percent of the participants reported having a computer at
home, and 86 percent reported having home Internet access. Seventy percent reported
their lntemet use occurred either ‘mostly in the house (52%) or ‘only in the house (18%).
Other Internet use occurred ‘only outside the house’ (16%), ‘about equally in the house
and outside the house’ (9%), and ‘mostly outside the house’ (4%).

Only ﬁve percent of the participants reported no prior experience with
information seeking via the lntemet. Thirteen percent reported never using email,
followed by chat room use (27%), listening to music (32%) and online video game
playing (49%). Prior lntemet experience was led by information seeking with an average
of 2.98 years; followed by prior email use, chat room use, listening to music, and ﬁnally

playing games (see Table 1).

Table 1

Prior Experience of lntemet Use (N = 197)

 

 

ML Mew) SD
Information Seeking 2.98 1.54
Email 2.50 1 .65
Chat room 1.95 1 .71
Music 1 .25 1 .23
Game Playing 1.24 1.66

 

Information seeking was the most common type of current use with 97 percent of

the participants reporting spending at least some looking for information on the Internet.

39

Email use (80%) is the second most common type of use, followed by music (61%), chat
room (58%), and video game playing (47%). Amount of lntemet use was led by
information seeking, followed by email, music, chat room, and online video games (see

Table 2).

Table 2

Amount and Type of lntemet Use (N = 197)

 

 

We MW SD
Information Seeking 2.50 2.26
Email 1.84 1.70 .
Music 1.82 2.44
Chat room 1.60 2.21
Game Playing 1.35 2.12

Total Use 9. 10 6.78

 

The initial concept was that reference groups would vicariously inﬂuence efﬁcacy
judgments. To better understand how the different components within an individual’s
reference group are related, an analysis of each bivariate relationship was assessed. The
' data indicated that the largest correlation coefﬁcients are between, (1) friends’ and
parent’s encouragement (r = .63), (2) ﬁiend’s success and social group success (r = .67),
and (3) friend’s success and ﬁiends’ encouragement (r =' .61) (see Table 3). These strong
relationships are not surprising as it is expected that a friend who is successful at using
the Internet also is going to be more encouraging. Likewise, when assessing social group
success, it is expected that an individual will include his/her ﬁiends as part of their social

group, thus, increasing the relationship between the two constructs. The only non-

40

signiﬁcant relationship exists between ﬁiend’s use and parent’s encouragement (r = .08).
The remaining correlation coefﬁcients are low to moderate in size. All other relationships

among the antecedents of Internet self—efﬁcacy can be found in Table 3.

Table 3
Pearson Correlation Coefﬁcients amongthe Antecedents of lntemet Self-efﬁcacy

FU FS FE PU PS PE SGS MP IE

Friend

Use (FU)

Friend

Success(FS) .35“

Friend

Encourage .28" .61**

(FE)

- Parent

Use(PU) .30** .25" .14

Parent

Success .19* .51** .43** .50**

(PS)

Parent

Encourage .08 .35" .63" .26" .53**

(PE)

Social

Group .35M .67" .54** .18“ .47" .35"
Success

(SGS)

Media

Portrayal .28" .47** .55“ .24** .47” .49** .50**
(MP)

lntemet

Experience .26" .35” .31" .32" .31" .21** .37" .29”
(IE)

" p<.05

**p<.01

Four outcome expectations were assessed, social, entertainment, informational,
and negative. Each was designed to measure different expectations of using the lntemet.

Data indicated that all variables are signiﬁcantly correlated. A substantially large

41

correlation coefﬁcient was found between information and entertainment outcome

expectations (r = .70) (see Table 4). This suggests that those who have high information

expectations also have high entertainment expectations. This also could indicate that

information uses of the lntemet are considered entertaining, as deﬁned in this study. That

said, these constructs empirically demonstrate multidimensionality and have been

conceptually distinct constructs in pervious research (Charney & Greenberg, in press;

LaRose, Mastro, & Eastin, in press) thus, they remain as two different constructs in

ﬁrrther analyses.

Table 4

Pearson Correlation Coefﬁcients among Outcome Egmctations

SE 1E

Social
Expectations
(SE)
Information
Expectations .50"
(IE)
Entertaimnent
Expectations .54" .70**
(EE)
Negative
Expectations -.30** -.32**
(NE)
* p < .05
** p < .01

EE NE

434**

Zero order correlations were used to initially test each of the hypothesized

relationships. Path analysis was used to test the combined effect of their direct and

indirect theoretical relationships.

42

Hypotheses

Hypothesis 1 posited that the greater the social outcomes expectancies from
lntemet use, the greater the actual Internet use; this relationship was supported by the
data (r = .42, p < .001).

Also supported were the relationships between entertaimnent expectations and
Internet use (r = .33, p < .001) and information expectations and Internet use (r = .34, p <
.001) (see Hypothesis 2).

Hypothesis 3, which indicated that an increase in negative outcome expectancies
would decrease lntemet use was supported (r = - .25, p < .001). Combined, these ﬁndings
support the connection between cognitive models of outcome expectations and total
lntemet use.

Another cognitive model, whichihypothesized a positive relationship to lntemet
use, was lntemet self-efﬁcacy (Hypothesis 4). Here, lntemet self-efﬁcacy was found to
be positively related to lntemet use (r = .43, p < .001).

Hypotheses 5, 6, and 7 posited that lntemet self-efﬁcacy would be positively
related to social, entertainment, and information outcome expectations, and negatively
related to negative outcome expectations. Data indicated that these relationships were all
signiﬁcantly related to lntemet self-efﬁcacy. lntemet self-efﬁcacy was positively related
to social outcome expectancies (r = .48, p < .001), entertainment outcome expectancies (r
= .58, p < .001), and information outcome expectancies (r = .56, p < .001), and negatively
related to negative outcome expectations (r = - .48, p < .001).

Supporting Hypothesis 8, prior Internet experience also indicated a signiﬁcant

positive relationship with lntemet self-efﬁcacy judgments (r = .40, p < 001)-

43

Tuming to the environmental inﬂuences of Internet self-efﬁcacy, results indicate '
a signiﬁcant relationship between ﬁiend (r = .31 , p < .001) and parent (r = .21, p < .001)
(i.e., reference group) lntemet use and self-efficacy judgments, supporting Hypothesis 9.
The success of an individual’s reference and social group at using the lntemet was
posited to have a positive relationship with Internet self-efﬁcacy. Here, ﬁiend (r = .46, p
< .001), parent (r = .43), and social group (r = .49, p < .001) success at using the lntemet
were signiﬁcantly related to efﬁcacy judgments; thus indicating support for Hypotheses
10 and 11.
Data also support the relationship between efﬁcacy judgments and encouragement
from ﬁiends’ (r = .49, p < .001) and a parent (r = .31, p < .001) (see Hypothesis 12).
Hypothesis 13, which assessed the relationship between vicarious learning
through television portrayals was found to increase efﬁcacy judgments (r = .38, P< .001).
Finally, as Hypotheses 14 and 15 state, signiﬁcant relationships between lntemet
anxiety and efﬁcacy judgments (r = - .57, p < .001) and lntemet anxiety and lntemet use
(I = - .22, p < .001) were found to be signiﬁcant.
These correlation coefﬁcients can be viewed within the context of the

theoretically proposed model (see Figure 6).

44

we 6. Pearson Correlation Coefﬁcients as Conceptualized within the Research
Model A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Prior Internet r = .40
Experience
. lntemet
Fnend’s lntemet ,
Use = - .57 Anxrety
lntemet Self-
Parent’s lntemet Efﬁcacy r _ 43 r = - .22
I Use' — °
Friends’ r = .58 r = .48
Encouragement of /v r = .49 = 42
lntemet Use Social r '
Outcome —’ lntemet
Parent’s r = 31 Expectations SC
Encouragement of f ' r = .33
lntemet Use
_> Entertainment /
Friend’s Success at /4 r = .46 Outcome
lntemet Use Expectations
r = .34
r = .56
Parent’s Success at /4 r = .43
lntemet Use _, Information
r = .49 Outcome
Success of Social /‘ Expectations 25
Group r= - .48 _ - '
. r = .38 .
Media Portrayal of f _, Negatrve
lntemet Use Outcome
Expectations

 

 

 

 

45

Path Model Analysis

Using path modeling analysis, the second analysis tests both the direct and
indirect relationships among each of these variables as posited by Social Cognitive
Theory. Parent’s ability was dropped ﬁom the model because 23 percent of the
respondents indicated that their parents had never used the Internet; thus, ability for those
parents could not be assessed. Results suggest that the data did not ﬁt the model, [(46)
= 105.89, p < .001 (see Figure 7). Only three of the antecedents of self-efﬁcacy
demonstrated substantial coefﬁcients; prior Internet experience (,6 = .19), ﬁiends’
encouragement of Internet use ([3 = .24), and social group success ([3 = .20). All other
antecedents of Internet self-efﬁcacy, ﬁiend’s use (3 = .08), parent’s use (6 = .03), parents
encouragement ()6 = - .01), friend’s success ([3 = .06), and media portrayal (,8 = .04)
indicated only negligible contributions. lntemet self-efﬁcacy however displayed strong
relationships with social outcomes expectancies (,6 = .48), entertainment outcomes
expectancies (,6 = .58), information outcomes eXpectancies (,6 = .56), negative outcomes
expectancies (,6 = - .48), lntemet anxiety ([2 = - .57), and lntemet use (,6 = .30).

Other predictors of Internet use indicated mixed results. While social outcome
expectations indicated a substantial relationship (,6 = .27), entertainment (,8 = .00),
information (,8 = .06) and negative (,6 = - .04) outcome expectations did not. Finally,

lntemet anxiety indicated a small positive relationship (6 = .09) with Internet use. Note

that this relationship was hypothesized to be negative.

46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fi re 7: A Social Co itive Model of Internet Use
Prior lntemet B = J9"
Experience \
B _ 57 lntemet
Friend lntemet 3 = ~08 — -' Anxrety
Use ‘5
Internet Self- B = .09
Parents Internet B = 03 Efﬁcacy = 30
Use I
- .48
Friends B= 24* ”f "
Encouragement
of lntemet Use _ Social B = .27
Parents B = - 01 Outcome ’ lntemet
' Use
Encouragement / Expectations
of lntemet Use
B = .06 7’ Entertainment 3: .00
Friends Success / Expectations
at lntemet Use
— *
,6 .20 B = .56
Success of . B: 06
Similar Others , - 04 —> Information
, '6 .. ° Expectations
Media Portrayal
of lntemet Use 5 = - .48 B =
H Negative
Outcome
Expectations

 

 

47

There are several reasons why this model failed. First, variables such as friend’s
and parent’s lntemet use, parent’s encouragement, friend’s success, media portrayals, and
entertainment, information, and negative outcome expectancies did not produce
meaningful relationships. These weak relationships are due to high levels of multi-
collinearity among the predictors of lntemet self-efﬁcacy and lntemet use (see Tables 3
and 4). Second, lntemet anxiety does not mediate the relationship between Internet self-
efﬁcacy and Internet usage. This is partially because Internet anxiety has a stronger
relationship with Internet self-efﬁcacy than it does with amount of lntemet use.
Therefore, anxiety is not a mediator of lntemet use as suggested by some research
investigating computer use (Compeau etal., 1999; Compeau and Higgins, 1995), but
rather, it is an antecedent to lntemet self-efﬁcacy as early research on self-efﬁcacy would
suggest (Bandura, 1997). Understanding these errors in the model, a post hoc model
adjusting for each problem‘was tested (see Figure 8).

The new model, as seen in Figure 8, has prior lntemet experience, ﬁiends’
encouragement, social group success, and lntemet anxiety as direct antecedents to
Internet self-efﬁcacy. lntemet self-efﬁcacy then has both a direct and indirect relationship
with Internet use through social outcome expectancy. Finally, social outcome expectancy

is modeled as having an inﬂuence on lntemet use.

48

Figure 8: A Revised Conceptual Model Examining the Social Cogﬂm Model of
Internet Use '

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Prior lntemet +
Experience \
Friends’ +
Encouragement —’ lntemet Self- +
Efﬁcacy \
4.
Success of Social / + Internet
Group + Use
Social /
lntemet Anxiety Outcome
Expectations

 

 

 

 

 

49

Results suggest that the data are consistent with the model, f(8) = 13.81, p > .05

(see Figure 9). With the exception of prior lntemet experience, the antecedents of

lntemet self-efﬁcacy produced substantial coefﬁcients. Friends’ encouragement (B =

.24), social group success (B = .16), and lntemet anxiety (B = - .38) all individually

contributed substantially to lntemet self-efﬁcacy. lntemet self-efﬁcacy demonstrated

strong relationships to both social outcome expectancy (B = .48) and lntemet use (B =

.30). Social outcome expectancy also obtained a substantial relationship with lntemet use

(a = .28).

Figure 9: A Revised Model Testhg the Social Cogitive Model of Internet Use

 

 

 

 

 

 

 

Prior lntemet B = .12
Experience \
Friends’ B = .24
Encouragement ’

 

 

 

 

Success of Social
Group

 

 

 

 

lntemet Anxiety

 

 

 

 

 

 

 

 

lntemet Self- 5 =
Efﬁcacy \‘
l B = .48
B = .28
Social /
Outcome
Expectations

 

 

50

.30

 

lntemet
Use

 

 

 

In Figure 9 prior Internet experience does not obtain a substantial coefﬁcient (B =
.12). However, because of strong empirical evidence supporting this relationship, leaving
it in the model is beneﬁcial for future research. While this model does stay within the
theoretical framework presented, it is data driven and thus should be retested in order to
validate these ﬁndings. In addition, large error terms also were found between ﬁiends’
encouragement and social outcome expectations (.19), and social group success and
social outcome expectations (.19); indicating that the observed correlation coefﬁcients
were larger than the reproduced. This could suggest that while self-efﬁcacy is a mediator

of lntemet use, it might not be the sole mediator in this type of model.

51

CHAPTER 4
Discussion

It should be noted that the simple bivariate relationships hypothesized are
counterintuitive to the path model tested. For example, the model contains no simple
bivariate relationships and thus, hypothesizing relationships while not controlling for all
other predictors is counterintuitive. Nonetheless, when evaluating the hypothesized
relationships, support was found for all 15 bivariate relationships tested. Internet use was
signiﬁcantly related to social, informational, entertaimnent, and negative outcome
expectations, lntemet self-efﬁcacy and lntemet anxiety. lntemet self-efﬁcacy was
signiﬁcantly related to prior lntemet experience, friend’s and parent’s lntemet use,
ﬁiend’s and parent’s Internet success, friends’ and parent’s encouragement, social group
SUCCCSS, media portrayals, social, personal and negative outcome expectancies, and
lntemet anxiety. That said, it is concluded that SCT does provide a good framework
ﬁ'om which to understand the environmental, cognitive and behavioral mechanisms of
lntemet use when assessing adolescents.

However, the model evaluating the direct and indirect relationships among each
of the variables (see Figure 5) was not supported. In testing the model, it was evident
which variables played a substantial role in both efﬁcacy judgments and Internet use. As
shown in the post hoc model, prior experience, friends’ encouragement, social group
success, and lntemet anxiety all uniquely contribute to lntemet self-efﬁcacy judgments.
lntemet self-efﬁcacy was then related to social outcome expectancies and lntemet usage.

Finally, social outcome expectancies indicated a substantial relationship with lntemet

52

use. Thus, while the original model failed, the new model indicates that key components
within SCT do play a major role in adoption by adolescents.

Why did the original model fail? To begin, the variables ﬁiend’s and parent’s
Internet use, parent’s encouragement, ﬁiend’s success, media portrayals, and
entertainment, information, and negative outcome expectancies did not produce
meaningful relationships. These weak relationships are partially due to moderate to high
levels of multi-collinearity among the predictors of lntemet se1 f-efﬁcacy and Internet use
(see Tables 3 and 4). Also, it could be argued that while the constructs measured in the
present study separated information and entertainment, the two might not be distinct at
any given time for the user. For example, an individual could be at a music site looking
up information about his/her favorite group; the user could perceive this behavior as
either entertaining or informative. Here, it would be advantageous to question users
regarding their motivation (e.g., information or entertainment purposes) while in the act
of searching, and then apply the appropriate e'Xpectancy scale. Finally, lntemet anxiety
did not mediate the relationship between lntemet self-efﬁcacy and lntemet usage as
posited. In part, this is because Internet anxiety had a stronger relationship with lntemet
self-efﬁcacy than amount of lntemet use.

From an item analysis, it could be argued that asking participants to think of “a
friend” or “a parent” was too limiting. The diverse applications available on the lntemet
allow users to customize their experience (Cody, 1999; Gervey, 2000). For example, a
friend may only use the Internet for gaming, leaving the remaining items that question
school uses irrelevant. This same ﬁiend may be high on entertainment success and low

on school success; providing misleading results for the relationship between lntemet self-

53

efﬁcacy and friend’s success. This can also be argued for parents. A parent may use the
lntemet all day at work, making them the “high” lntemet user, however, this same parent
may never use the lntemet at home, making it impossible to accurately perceive their
success. Perhaps this is why social group success worked so well; participants were able
to evaluate more generally and thus, incorporate different types of users. Given this, the
inclusion of more than one friend and both parents should provide a more accurate and
complete picture.

Further, while parent’s encouragement did not obtain a substantial relationship,
friends’ encouragement did. Friends’ encouragement allowed respondents to consider
more than one ﬁiend, while parental encouragement was limited to a single parent. By
allowing more than one friend to be considered, respondents were able to consider a
broader range of encouragement. That said, giving respondents the opportunity to
consider both parents could increase the relationship. The media portrayal questions were
also broad in scope, allowing for many different media sources to be considered.
However, with media portrayals a more speciﬁc or relevant example of television could
have been given. Instead of listing news portrayals on the lntemet, MTV portrayals
would prompt a more relevant frame of reference for teens (Cross, 2000). Of course all
of these rationales are only speculative and should be empirically tested in future
research.

lntemet self-efﬁcacy was found to inﬂuence lntemet use and social outcome
expectations. These ﬁndings reinforce previous research identifying efﬁcacy judgments
as a major component in the adoption and utilization of information systems such as the

lntemet. Current research indicates that a primary use of the lntemet is social (e.g., email)

54

(Nie & Embring, 2000; Lubans, 1999; NTIA, 2000), thus, it is not surprising that social
outcome expectations are a driving and dominant antecedent to overall lntemet use.
However, because this sample only consisted of adolescents, the model may differ when
considering an adult sample. For example, obtaining information through social and
general use is a primary expectation among adult users (Cody, 1999); therefore,

information outcome expectancies may drive a model using an adult sample.

Future research

The ﬁrst step for future research should be to validate the post hoc model
supported in this study. Since research in this area will most likely continue to evaluate
online users as a mass audience, it is important to have a valid model for research only
interested in assessing amount of general use. Constructs such as lntemet self-efﬁcacy
provide online research with a perceived ability control for novice users when
experimentally testing constructs such as navigation ability; thus, the validation of a self-
efﬁcacy construct is important.

However, according to social cognitive theory, self-efﬁcacy is most effective
when evaluated within a given domain or behavior. This study looked at self-efﬁcacy
within the domain of Internet use. A more speciﬁc breakdown of lntemet self-efﬁcacy as
well as lntemet use would provide a greater understanding of ad0ption (Bandura, 1997).
Speciﬁcally, by creating distinct models of lntemet use, new cognitive models can be
developed, thus creating opportunities for variables such as entertainment and
information outcome expectations to play a more substantial role in predicting Internet

use.

55

This type of model would (a) evaluate different types of prior lntemet experience,
(b) assess task speciﬁc self-efﬁcacy judgments such as social, informational, and
entertainment, (c) assess speciﬁc types of outcome expectancies such as social,

entertainment, and informational; and ﬁnally, (d) identify speciﬁc types of actual use

(Figure 10).

56

figure 10. Distinct Models of lntemet Use

 

Information
Experience

 

 

Social
Experience

 

 

Entertainment
Experience

 

 

 

Social Group
Success

 

if.

 

Friend
Encourage

II

 

 

lntemet
Anxiety

 

 

 

 

 

Information
Seeking Efﬁcacy

l

}_,

 

Information
Outcome
Expectation

 

 

 

 

lntemet Social
Efﬁcacy

 

 

l

 

 

Social Outcome
Expectations

 

 

 

 

Entertainment
Efﬁcacy

 

 

I

 

 

Entertainment
Outcome
Expectations

J__

 

 

57

 

 

Information
Use

 

 

 

 

Social Use

 

 

 

Entertainment
Use

 

 

 

A simple evaluation of the relationships within the current data set demonstrates
the plausibility of the model. Here, strong relationships are observed between social use
(i.e., email and chatroom use) and social outcome expectations (r = .48, p < .001),
informational use and informational outcome expectations (r = .37, p < .001), and
entertainment use (i.e., online game playing and music) and entertainment outcome
expectations (r = .28, p < .001). With the exception of entertainment, all of these outcome
expectations have a stronger relationship with their corresponding use measure than with
general Internet use.

The lntemet has long been considered a non-restricted medium, which allows
users to customize their online experiences. To advance research in this area, it will be
essential to begin micro-analyzing the user and their unique experiences. Figure 10
represents a step toward this type of Understanding. Research designed to test this model
should experimentally track users under different conditions. For example, by
implementing interventions designed to increase efﬁcacy judgments, researchers will be
able to increase certain types of use that are seen as more beneﬁcial (e.g., informational
use) to the user. Through this process, more robust and effective measures can be
deveIOped for each of the models.

Increasing the amount that people use the lntemet is only the ﬁrst step to a more
effective digital society. With a growing nurnber of antisocial activities available on the
web, it is important to understand howuse differs. A practical starting point for this type
of analysis is to assess difference in types of use across gender and race. The present data

set indicated no difference between gender and actual use (r = .02, p > .05). However,

58

women were found to be signiﬁcantly higher social users (r = .16, p < .05) and
signiﬁcantly lower in their online game playing (r = - .29, p < .05). Finally, there was
virtually no difference between gender and information (r = .04, p > .05) or music
Internet use (r = -.06, p > .05).

According to current understanding of Internet use and subsequent social isolation
(Nie & Embring, 2000), these results indicate that males are more likely to suffer possible
negative psychological effects such as those resulting ﬁ'om social isolation than females
due to their increased levels of online gaming. These ﬁndings also suggest that females
are more likely to be experiencing the social beneﬁts of the lntemet because of their
increased levels of social use. While these types of differences also should be assessed by
race, sample limitations in this study prevented these differences from being explored.
This would. allow digital divide researchers to begin questioning how certain minority
groups are using the lntemet and with what effects.

Future research should also look to past models of computer use to further
understanding of the lntemet. For example, rather than limiting outcome expectations to
the immediate future, future research should consider the long-term outlooks of use. A
study conducted by Thompson et al. (1989) found a signiﬁcant relationship between
long~term consequences such as increased future job opportunities and personal computer
use. From an lntemet use perspective, researchers should consider the relationship
between the different types of lntemet use and their ﬁrture consequences. This type of
analysis would afford policy makers with methods from which to promote the beneﬁts of

various types of lntemet use.

59

__ “—‘

 

Implications

By incorporating environmental factors, this study does build on previous
research using SCT to understand lntemet adoption. This knowledge presents new
opportunities for old problems still facing the lntemet. For example, the most common
strategy implemented to understand and decrease the digital divide has been to provide
access. By focusing on access, questions surrounding inﬂuences of adoption and level of
adoption have been primarily ignored. The goal of this research was to increase
understanding of lntemet use by increasing understanding of relevant psychological
variables. From this investigation, new approaches can be explored to (a) help those who
have not adopted, (b) increase the use of current users, and ﬁnally, (c) understand how to
increase different types of use. For example, in addition to providing access,
demonstrating the success of others within an individual’s social group will provide an
added source for efﬁcacy development. This type of reinforcement of use could increase
encouragement among individuals; subsequently, increasing efﬁcacy judgments and use.

Understanding this, policy makers can implement new methods to disseminate
information about the Internet to novice and non-users. Here, in addition to providing
lntemet access, resources afforded to non-users should include support mechanisms.
From a research perspective, creating lntemet labs in setting with a large numbers of low
users could provide individuals with access and the opportunity to vicariously observe
others’ successﬁrl Internet use. These labs could also be a source of encouragement from
fellow lab users. Longitudinally tracking changes in use and support will provide a more

complete understanding of user cognitive development over time.

Limitations
Several limitations of the study need to be considered in future tests of the model

and/ or new approaches to understanding lntemet use. First, all usage statistics are based
on respondent recall. Some researchers have suggested that a better assessment can be
obtained by electronically monitoring individual use (Kraut et a1. 1999; Jackson, et al.
2000). However, the. more advanced an lntemet user becomes the more likely he/she is
to encounter larger more cumbersome applications (e.g., audio and video ﬁles), causing
usage estimates to be negatively skewed. Speciﬁcally, long download times and
unlimited access (e.g., cable modems) present a problem for eleCtronic monitoring.
Second, while the sample was relevant for this investigation of lntemet use, respondents
were from a single high school; thus, the generalizability of the results toward other users
of the same age group remains unanswered. Third, the sample was limited to one age
group, preventing the ﬁndings from being generalized to a more general group of lntemet
users.

Forth, items should be added to all outcome expectancy measures; this would help
to increase the reliability of each construct. For example, the entertainment outcome
expectation construct was created with two items, forcing assumptions to be made with
testing both internal consistency and parallelism. Another limitation deals with the
reciprocal causation model speciﬁed in SCT. Without longitudinal data it is hard to
distinguish cause and effect ordering (Pedhazur, 1982). The reciprocal causation is an
important consideration with respect to the lntemet, since the Internet itself is Continually

changing and no two destinations on the lntemet are exactly alike. lntemet users

61

therefore continually modify their lntemet self-efﬁcacy beliefs with each online

experience.

Conclusion

In conclusion, this research has provided a number of contributions, the most
signiﬁcant being the inclusion of the environmental factors of lntemet use. Researchers
can use this study to hone and experimentally test inﬂuences of self-efﬁcacy judgments

and lntemet use. Further, this study has identiﬁed the prominent constructs for a general

 

model and possible individual models of lntemet use. Based on these ﬁndings, research
investigating distinct models of lntemet use as well as the psychological consequences of

lntemet use can use social cognitive theory as a foundation for investigation.

62

Appendix A

1. How long have you used the lntemet to gather information?

Never ‘6 l l 'A 2 2 'A 3 3 '6 4 or
year year years years years years years more
years

2. How long have you been playing video games on the Internet?

Never '15 l l 'A 2 2 'A 3 3 ‘A 4 or
year year years years. years years years more
years

3. How long have you been listening to music on the lntemet?

Never 99 l l 'A 2 2 ‘A 3 3 'A 4 or
year year years years years years years more
years

4. How long have you been meeting people (e.g., chat rooms and discussion groups) on
the lntemet?

Never 'A l 1 'A 2 2 ‘A 3 3 'A 4 or
year year years years years years years more
years

5. How long'have you been emailing people?

Never ‘14 l l 'A 2 2 'A 3 3 'A 4 or
year year years years years years years more
years

6. How long have you been participating in e-commerce (e.g., shopping online) on the '
lntemet? ‘

Never ‘A l l 'A 2 2 'A 3 3 ‘A 4 or
year year years years years years years more
years

We would now like you to consider how others use of the lntemet.

7. Think of a close friend, on an averageweekday (Monday
through Friday) how much time does that friend spend on the
lntemet at home? ’

None ‘A l l 'A 2 2 'A 3 3 ‘6 4 or
hour hour hours hours hours hours hours more
hours

63

8. Considering that same ﬁiend, on an average weekday how much time does that
ﬁiend spend on the lntemet at school?

None ‘A l l 95 2 .2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

9. Considering that same friend, on an average Saturday how much time does that
ﬁiend spend on the lntemet?

None 95 l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

10. Considering that same ﬁiend, on an average Sunday how much time does that friend
spend on the Internet?

None 'A l l 'A 2 2 'A 3 3 ‘A 4 or
hour hour hours hours hours hours hours more
hours

1 1. Considering that same ﬁiend, how successful is that friend at using the Internet to
complete school assignments?

l ’ 2 3 4 5 6 7
Not at All Completely
Successful Successful

12. Considering that same ﬁiend, how successful is that ﬁiend at using the Internet for
entertainment! social purposes?

234567

1.
Not at All Completely
Successful Successful

13. Considering that same friend, how satisﬁed is that friend with how easy they are able
to use the Internet for schoolwork?

l 2 3 4 5 6 7
Not at All Completely
Satisfied Satisfied

14. Considering that same friend, how satisﬁed is that ﬁiend with how easy they are able

to uSe the lntemet for entertainment/ social purposes?

64

l 2 3 4 5 6 7
Not at All Completely
Satisfied Satisﬁed

15. Does at least one of your parents use the lntemet at work or at home (please circle
your response)?
Yes No (if no, skip to question #24)

16. Think of your parent who uses the Internet the most, on an average weekday, how
much time does that parent spend on the lntemet at home?

None 96 l l 'A 2 2 'A ' 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

17. Considering that same parent, on an average weekday, how much time does that
parent spend on the Internet at work?

None ‘A 1 l 'A 2 2 'A 3 3 '6 4 or
hour hour hours hours hours hours hours more
hours

18. Considering that same parent, on an average Saturday, how much time does that
parent spend on the lntemet?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

19. Considering that same parent, on an average Sunday, how much time does that
parent spend on the lntemet?

None 'A l l 'A 2 2 'A 3 3 14 - 4 or
hour hour hours hours hours hours hours more
hours

20. Considering that same parent, how successful is that parent at using the lntemet at
work related (for example information gathering) pr0jects.

l 2 3 4 5 6 7

Not at All Completely
Successful Successful

65

21. Considering that same parent, how successful is that parent at using the lntemet at for
entertainment/ social purposes?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

22. Considering that same parent, how satisﬁed is that parent with the case at which they
are able to use the Internet for work related projects?

1 2 3 4 5 ' 6 7
Not at All Completely
Satisfied Satisfied

23. Considering that same parent, how satisﬁed is that parent with how easy they are able
to use the lntemet for entertainment! social purposes?

1 2 3 4 5 6 7
Not at All Completely
Satisfied Satisfied

24, In general, how successful are kids your age at using the lntemet to assist with their
schoolwork?

l 2 3 4 5 6 7
Not at All Completely
Successful Successful

25. In general, how successful are kids your age at using the lntemet for entertainment/
social purposes?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

26. In general, how often do you see peeple on television (including programs and
commercials) use the Internet to help them?

1 2 3 4 5 6 7
Not at All A lot

27. In general, how often do you see kids your age on television (including programs and
commercial) use the lntemet to help them?

1 2 3 4 5 6 7
Not at All A '0‘

66

28. In general, how often do news programs show how the Internet is being used in an
unfavorable way?

1 2 3 4 5 6 7
Not at All A lot

29. In general, how often do news programs talk about the lntemet in a favorable way?

1 2 3 4 5 6 7
Not at All A lot

30. In general, how often do people on television (i.e., other than news programs) discuss
the lntemet in an unfavorable way?

1 2 3 4 5 6 7
Not at All A lot

31. In general, how often do you see people in movies use the Internet to help them?
1 2 3 4 5 6 7
Not at All A lot
Now, we would like you to think about discussions you have had with various people

about the Internet.

32. My ﬁiends encourage me to use the lntemet.

l - 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

33. My ﬁiends tell me that the lntemet is a terrible source for information.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree.

34. My ﬁiends talk with me about using the lntemet.

1 2 3 ' 4 5 6 7
Strongly Strongly
Agree Disagree

67

35. My friends discuss how being able to use the lntemet will help me in the future.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

36. My ﬁiends tell me that there are a lot of fun things to do on the lntemet.

l 2 3 4 S 6 7
Strongly Strongly
Agree Disagree

37. My mom or dad does not encourage me to use the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

38. My mom or dad tells me that the lntemet is a great source for information.

1 2 3 4 5 6 7
V Strongly Strongly
Agree Disagree

39. My mom or dad does not talk to me about using the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

40. My mom or dad discusses how being able to use the lntemet will help me in the
future.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

41. My mom or dad tells me that there are a lot of fun things to do on the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

68

42. In general, the people who are the most important to me encourage me to use the
lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

43. In general, most people I know think that lntemet use is not important.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

44. In general, most people in society do not think that using the lntemet is important.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

The following questions will ask you about the types of experiences and
information you encounter while online. Given your past lntemet experiences,
please identify the likelihood of them happening to you when you use the
Internet.

45. I am likely to establish a romantic relationship on the lntemet.

l 2 3 4 S 6 7
Strongly ' Strongly
Agree Disagree

46. I am likely to get immediate information about big news events on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

47. I am unlikely to meet new ﬁiends on the lntemet.

l 2 3 4 5 6 7
Strongly 3?”le
Agree . Disagree

69

48. I am likely to ﬁnd a way to pass time on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

49. I am likely to ﬁnd information to complete a class assignment on the lntemet.

l 23 4 5 6'7

Strongly Strongly
Agree Disagree
50. I am unlikely to have problems opening a large video ﬁle found online.
1 2 3 4 5 6 7
Strongly . Strongly
Agree Disagree

51. I am likely to get in touch with people I know on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

52. I am likely to receive email I do not want.

1 2 3 4 5 6 7
Strongly Strongly
Agree . Disagree

53. I am unlikely to have fun on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

54. I am likely to meet someone in person whom I met on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

55. I am likely to have long download times on the lntemet.

1 2 3 4 5 ' 6 7
Strongly Strongly
Agree Disagree

70

 

56. I am unlikely to visit chat areas to talk with other people.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

57. I am likely to ﬁnd current information like time, weather, stock prices, and sports
scores on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

58. I am likely to have my computer freeze while I am searching the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

59. I am unlikely to get information about products such as cars or clothes on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

60. I am likely to ﬁnd information that is current on the lntemet.

1 2 3 4 S 6 7
Strongly Strongly
Agree Disagree

61. I am likely to feel entertained on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

62. I am likely to be bored on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

71

”-11!— w

 

63. I am likely to receive too much information when I am looking for something.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

64. I am likely to ﬁnd that search engines don’t have enough detail to quicklyﬁnd what I
am looking for.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

65. I am likely to have problems opening a large text document I find online.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

66. I am unlikely to have my personal information such as age, spending habits, sites I
visit, etc. taken from my computer when I am surﬁng the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

67. I am likely to have problems opening a large audio ﬁle found online.

1 2 3 4 5 6 ' 7
Strongly Strongly
Agree Disagree

68. While using the lntemet I am likely to catch a computer virus.

1 2 3 4 5 6 7
Strongly 5"?”le
Agree Disagree

72

 

How do you feel about using the Internet? Internet use includes sending or receiving
electronic mail, visiting chat rooms, participating in discussion groups and visiting
sites on the World Wide Web. We would like you to answer each of the following
questions even if you are not an lntemet user.

69. I am conﬁdent understanding terms/words relating to lntemet hardware.

1 2 3 4 S 6 7
Strongly Strongly
Agree Disagree

70. I am conﬁdent using the lntemet to gather data.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

71. I am not conﬁdent understanding terms/words relating to lntemet software.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

72. I am conﬁdent describing functions of lntemet hardware.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

73. I am conﬁdent trouble shooting lntemet related problems.

1 2 3 4 5 6 _ 7
Strongly Strongly
Agree Disagree

74. I am conﬁdent explaining why a task will not run on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

73

75. 1 am conﬁdent learning advanced skills in a speciﬁc lntemet program.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

76. I am conﬁdent turning to an on-line discussion group when help is needed. '

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

77. I am not apprehensive about using the lntemet.

 

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

78. The Internet is somewhat intimidating to me.

1 2 3 4 5 6 7 _
Strongly Strongly
Agree Disagree

79. I hesitate to use the lntemet for fear of making a mistake I cannot correct.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

80. I’m afraid that I might destroy/ lose information on my computer by accessing
corrupted ﬁles while on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

81. Using the lntemet makes me nervous because I’m not sure how much personal
information such as age, gender, sites I visit, etc. rs being collected on me

electronically.
1 2 3 4 5 6 7
Agree Disagree

74

F’mally, we are interested in how often you currently use the lntemet for various
types of activities. Please read each questions carefully.

82. On an average weekday (Monday through Friday), how many hours do you use the lntemet
to gather information?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

83. On an average weekday, how many hours do you use the Internet to play video games?

None 'A l l 'A 2 2 'A 3 3 '6 4 or
hour hour hours hours hours hours hours more
hours

84. On an average weekday, how many hours do you use the lntemet to listen to music?

 

None 'A l l 'A 2 2 ‘A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

85. On an average weekday, how many hours do you use the lntemet to meet people (e.g., chat
_ rooms and discussion goups)?

None 'A 1 l ‘A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

86. On an average weekday, how many hours do you use the Internet to email?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

87. On an average Saturday, how many hours do you use the lntemet to gather information?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours :rore
ours

88. On an average Saturday, how many hours do you use the Internet to play video games?

None ‘A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours lmore
ours

89. On an average Saturday, how many hours do you use the Internet to listen to music?

'A 4 or
None 'A l l ’A 2 2 'A 3 3
hour hour hours hours hours hours hours 21:5;

75

90. On an average Saturday, how many hours do you use the Internet to meet people (e.g., chat
rooms and discussion groups)?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

91. On an average Saturday, how many hours do you use the lntemet to email?

None 'A l 1 'A 2 2 K 3 3 '6 4 or
hour hour hours hours hours hours hours more
hours

92. On an average Sunday, how many hours do you use the lntemet to gather information?

None 'A l l 'A 2 2 ‘6 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

93. On an average Sunday, how many hours do you use the lntemet to play video games?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

94. On an average Sunday, how many hours do you use the lntemet to listen to music?

None 'A l l 'A 2 2 ‘A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

95. On an average Sunday, how many hours do you use the lntemet to meet peOple (e.g., chat
rooms and discussion groups)?

None 'A l l 'A 2 2 'A 3 3 'A 4 or
hour hour hours hours hours hours hours more
hours

96. On an average Sunday, how many hours do you use the lntemet to email?

None '6 l 1 ‘6 2 2 ‘A 3 3 ‘A 4 or
hour hour hours hours hours hours hours lmore
ours

97. Do you have a computer at home?

yes 110

98. Do you have Internet access at home?

yes __________nO

 

76

 

99. In general, where do you most often use the lntemet?

Only outside my home (e. g., at school)

Mostly outside my house

About equally in my home and outside my home
Mostly in my home

Only in my home

{DP-99"!”

Now, just a few personal questions.
100. Sex

1) Female
2) Male

101. What is your age

 

102. What ethnicity are you?

1) Aﬁican American

2) Caucasian

3) Asian (including Chinese, Korean, Japanese and Southeast Asians)
4) Paciﬁc Islander

5) Native American or Alaskan native

6) Spanish, Hispanic, or Latino

7) other?

 

77

Appendix B

First we would like you to think about how long you have been using the lntemet.

1.How many years have you used'the lntemet for gathering information? year(s)
2. How many years have you been playing video games on the lntemet? year(s)
3. How many years have you been listening to music on the lntemet? year(s)

4. How many years have you been meeting people (e.g., chat rooms
and discussion groups) on the lntemet? year(s)

5. How many years have you been emailing people? year(s)
We would now like you to consider how others that you know use the Internet.

6. Think of a close ﬁiend, on an average weekday (Monday through Friday), how many
hours does that ﬁiend spend on the lntemet at home?

Hour(s)

7. Think of a close friend, on an average weekday, how many hours does that ﬁiend
spend on the lntemet at school? ‘

Hour(s)

8. Think of a close ﬁiend, on an average weekend, how many hours does that ﬁiend
spend on the lntemet?

Hour(s)

9. Considering that same friend, how successful is that ﬁiend at using the lntemet to
complete school assignments?

1 2 3 4 5 6 7
Not at All Completely

Successful Successful

10. Considering that same ﬁiend, how successful is that ﬁiend at using the lntemet for
entertainment purposes (for example, lntemet use that is not schoolwork related)?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

78

1 1. Considering that same friend, how easy is it for that friend to use the lntemet for
schoolwork?
l 2 3 4 5 6 7 '
Not at All Completely
Easy Easy

12. Considering that same friend, how easy is it for that ﬁiend to use the lntemet for
entertainment purposes (for example, Internet use that is not schoolwork related)?

1 2 3 4 5 6 7
Not at All Completely
Easy Easy

13. Does at least one of you parents use the lntemet at work or at home (please circle you
response)?

Yes (if yes, continue with question 14) No (if no, skip to question #21)

14. Think of your parent who uses the Internet the most, on an average weekday, how
many hours does that parent spend on the lntemet at home?

Hour(s)

15. Considering that same parent, on an average weekday, how many hours does that
parent spend on the lntemet at work?

Hour(s)

16. Considering that same parent, on an average weekend, how many hours does that
parent spend on the lntemet?

Hour(s)

17. Considering that same parent, how successful is that parent at using the lntemet for
work related (for example, information gathering) projects?

1 2 3 4 5 6 7
Not at All Completely

Successful Successful

79

 

18. Considering that same parent, how successful is that parent at using the lntemet for
entertainment (for example, lntemet use that is not work related)?

1 2 3 4 5 6 7
Not at All Completely
Successful Successful

19. Considering that same parent, how easy is it for that parent to use the Internet for
work related projects?

1 2 3 4 5 6 7
Not at All Completely
Easy . Easy

20. Considering that same parent, how easy is it for that parent to use the lntemet for
entertainment (for example, lntemet use that is not work related)?

1 2 3 4 5 6 7
Not at All Completely
Easy , Easy

Now just a few general questions about how other are using the
Internet.

21. In general, how successful are kids your age at using the lntemet to help them with
their schoolwork?

l 2 '3 4 5 6 7
Not at All Completely
Successful Successful

22. In general, how successful are kids your age at using the lntemet for entertainment
purposes?

1 2 3 4 5 6 7

Not at All Completely
Successful Successful

23. In general, how easy is it for kids your age to use the lntemet for schoolwork related
projects?

1 2 3 4 5 6 7
Not at All ComPMely

Easy Easy

80

24. In general, how easy is it for kids your age to use the lntemet for entertainment
purposes (for example, lntemet use that is not work related)?

1 2 3 4 5 6 7
Not at All Completely
Easy Easy

Now, just a few questions about other Internet related issues. Please ‘
note that some items are negatively phrased!

25. In general, how often do you see people on television programs and commercials (not
including news programs) discuss the lntemet in an unfavorable way?

1 2 3 4 5 6 7
Not at All A lot

26. In general, how often do you see people on television programs and commercials (not
including news programs) use the lntemet to help them?

1 2 3 4 5 ' 6 7
NotatAll Alot

27. In general, how often do you see kids your age on television and commercials (not
including news programs) use the lntemet to help them?

1 2 3 4 5 6 7
Not at All A lot

28. In general, how often do news programs show how the Internet is being used in an
unfavorable way?

1 2 3 4 5 6 7
Not at All A '0‘

29. In general, how often do news programs talk about the lntemet in a favorable way?

1 2 3 4 5 6 7
Not atAll Mat

81

30. In general, how often do you see pe0ple in movies use the lntemet to help them?

1 2 3 4 5 6 7
Not at All A lot

Now we would like you to think about discussions you have had with
various people about the Internet. Please not that some of the items are
negatively phrased!

31. My ﬁiends encourage me to use the Internet. 1

l 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

 

32. My ﬁiends tell me that the lntemet is a terrible source for information.

1 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

33. My ﬁiends discuss how being able to use the lntemet will help me in the future.

1 2 3 4 S 6 7
Strongly Strongly
Disagree Agree

34. My ﬁiends talk with me about using the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

35. My friends tell me that there are a lot of ﬁrn things to do on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly

Disagree Agree
36. My mom or dad tells me that the Internet is a great source of information.
1 2 3 4 5 6 7

Strongly Strongly
Disagree Agree

82

37. My mom or dad does NOT encourage me to use the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

38. My mom or dad tells me that the lntemet is a great source for information.

1 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

39. My mom or dad does NOT encourage me to use the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

40. My mom or dad discusses how being able to use the Internet will help me in the

future.
1 2 3 4 5 6 7

Strongly Strongly
Disagree Agree

41. My mom or dad tells me that there are a lot of fun things to do on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Disagree Agree

The following questions will ask you about the types of experiences and
information you encounter while online. Given your Internet .
experiences, please identify the likelihood of each of them happenmg to

you when you use the Internet. Please not that some of the items are
negatively phrased!

42. I am likely to establish a romantic relationship on the lntemet.
1 2 3 4 5 6 7

Strongly Strongly
Agree Disagree

83

43. 1 am likely to get immediate information about big news events on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

44. I am NOT likely to meet new ﬁiends on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

45. I am likely to ﬁnd a way to pass time on the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

46. I am likely to ﬁnd information to complete a class assignment on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

47. I am likely to get in touch with people I know on the lntemet.

1 2 3 4 5 6 7
Strongly ’ Strongly
Agree Disagree

48. I am likely to receive email I do not want.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

49. I am NOT likely to have fun on the lntemet.

l 2 3 4 5 6 . 7
Strongly Strongly
Agree Disagree

84

50. I am likely to meet someone in person whom I ﬁrst met on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree ‘ Disagree

51. I am likely to have long download times on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

52. I am NOT likely to visit chat areas to talk with other people.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

53. I am likely to ﬁnd current information like time, weather, stock prices, and sports
scores on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

54. I am likely to have my computer freeze up while I am on the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

55. I am likely to get information about products such as cars or clothes on the lntemet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

56. 1 am likely to be entertained on the Internet.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

85

57. I am likely to be bored on the lntemet.

l 2 3 4 5 6 7
8&0“le Strongly
Agree ‘ Disagree

58. I am likely to receive too much information when I am looking for something on the
lntemet.

1 2 3 4 5 6 7
Strongly ‘ Strongly
Agree Disagree

 

59. I am likely to ﬁnd that search engines do NOT have enough detailed to quickly ﬁnd
what I am looking for.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

60. I am likely to have problems opening large text documents found online.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

61. I am likely to have problems opening an audio ﬁle (for example, a music ﬁle) found
online.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

62. While using the lntemet I am likely to catch a computer virus.

1, 2 3 4 5 6 7
Strongly Strongly
Agree ' Disagree

86

Internet use includes sending and receiving email, visiting chat rooms,
participating in discussion groups and visiting sites on the World Wide
Web. We would now like you to answer each of the following questions
even if you are not an Internet user. Please note that some of the items
are negatively phrased!

63. I am conﬁdent understanding terms/words relating to Internet hardware.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

64. I am conﬁdent understanding terms/words relating to lntemet software.

1 2 . 3 4 5 6 ' 7
Strongly Strongly
Agree . Disagree

65. I am conﬁdent describing functions of lntemet hardware.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

66. I am conﬁdent trouble shooting lntemet problems.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

67. I am conﬁdent explaining why a task will not run on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

68. I am conﬁdent using the lntemet to gather data.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

87

69. I am conﬁdent learning advanced skills in a speciﬁc lntemet program.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

70. I am conﬁdent turning to an online discussion group when help is needed.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

71. I am not apprehensive about using the lntemet.

1 2 3 4 5 6 7
Strongly ‘ Strongly
Agree Disagree

72. The lntemet is somewhat intimidating to me.

1 2 3 4 5 6 7
‘ Strongly 5"“le
Agree Disagree

73. I hesitate to use the Internet for fear of making a mistake I cannot correct.

1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

74. I’m afraid that I might destroy/ lose information on my computer by accessing
corrupted ﬁles while on the lntemet.

l 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

75. Using the lntemet makes me nervous because I’m not sure how much personal

information (for example, age, gender, sites I visit, etc.) is being collected electronically.
'1 2 3 4 5 6 7
Strongly Strongly
Agree Disagree

88

76. On an average weekday (Monday through Friday), how many hours do you use the
lntemet to gather information?

hour(s)

77. On an average weekday, how many hours do you use the lntemet to play video
games?

hour(s)
78. On an average weekday, how many hours do you use the lntemet to listen to music?
hour(s)

79. On an average weekday, how many hours do you use the Internet to meet people
(e.g., chat rooms and discussion groups)?

hour(s)
80. On an average weekday, how many hours do you use the lntemet to email?
hour(s)

81. On an average weekend, how many hours do you use the lntemet to gather
information?

hour(s)

82. On an average weekend, how many hours do you use the lntemet to play video
games?

hour(s)
83. On an average weekend, how many hours do you use the lntemet to listen to music?
hour(s)

84. On an average weekend, how many hours do you use the lntemet to meet people
(e.g., chat rooms and discussion groups)?

hour(s)
85. On an average weekend, how many hours do you use the lntemet to email?

hour(s)

89

 

86. Do you have a computer at home?

yes no

87. Do you have lntemet access at home?

yes

 

no
88. In general, where do you most often use the lntemet?

a. Only outside my house (e.g., at school)
b. Mostly outside my house

c. About equally in my home and outside my home
(1. Mostly in my home
e. Only in my home

89. Sex

a. Female

b. Male

90. What is your age

 

91. What ethnicity are you?
1) African American
. 2) Caucasian

3) Asian (including Chinese, Korean, Japanese and Southeast Asians)
4) Paciﬁc Islander

5) Native American or Alaskan native

6) Spanish, Hispanic, or Latino
7) other?

 

 

References

Bandura, A. (1982). Self- efﬁcacy mechanisms in human agency. Ameri___c_an
schologist, 37, 122-147.

Bandura, A. (1986). Social foundations of thought and pgtion: a social cognitive
theogg. Englewood Cliffs, N.J.: Prentice-Hall.

Bandura, A. (1989). Human Agency in Social Cognitive Theory. American
Psychologist, 44, 1 175-1 184.

Bandura, A. (1994). Social Cognitive Theory of Mass Communication. In J.
Bryant and D. Zillman (Eds) Media Effects. Hillsdale, New Jersey: Lawrence Erlbaum
Associates.

Bandura, A. (1997). Self-efﬁcacy: the exercise of control. New York: W.H.
Freeman.

Burkhart, M. & Brass, D. (1990). Changing Patterns of Patterns Changing: The
Effects if a change in Technology on Social Network Structure and Power.
Administrative Science Quarterly, 35, 104-127.

Charney & Greenberg (in press). Uses and Gratiﬁcation of the Internet.
Communication, Technology and Science. In C. Lin and D. Atkin (Eds) Communication,

 

Technology and Society: New Media Adoption and Uses. Hampton Press.

Cody, M. (1999). Silver Surfers: Training and Evaluating lntemet Use among
Older Learners. Communication Education 48(4), 269-286.

 

Compeau, D. & Higgins, C. (1995). Computer self~efﬁcacyz Development of a
measure and initial test. MIS Quarteply 19, 189-211.

 

Compeau, D. & Higgins, C. Huff, S. (1999). Social Cognitive Theory and
individual reactions to computing technology: a longitudinal study. MIS Quarterly, 23,
145-1 58.

 

Cross, T. (2000). Gifted Students’ Social and Emotional Development in the 218’
Century. Gifted Child Today. 23(2), 14,15,52.

Davis, FD. (1989). Perceived Usefulness, Perceived Ease of Use, and User
Acceptance of Information Technology. MIS (marterlyLl 3(3), 319-340.

 

91

Davis, F.D., Bagozzi, R.P. & Warsaw, PR. (1989). User Acceptance of Computer
Technology: A Comparison of Two Theoretical Models. Man_agement Science, 35(8),
983-1003.

Eastin, M.S. & LaRose, R. (2000). Internet Self-Efﬁcacy and the Psychology of
the Digital Divide. Journal of Computer Mediatﬁed Communication, 6(1), [Online]
Available http://www.Jasusc.org[Lcmc/vol6/issue1/eastin.htrpl,, February 26, 2001

Fishbein, M. Ajazen, I. (1975). Belief, Attitude, Intentions and Behavior: An
Introductory to Theory and Research, Addison-Wesley, Boston, MA.

Gervey, B. (2000). The age factor: How lntemet use varies from teens to seniors.
Advertising Age 71(16), 22

GVU (Graphic, Visualization and Usability Center). (1999). GVU's tenth annual
WWW user’s survey. Geoggia Institute of Technology. [On-line], Available
http://www.cc.gatech.edu/gvu/user surveys/survey-l998-lO/graphs/usel, October 20,
2000 ‘

Hoffman, D.L. & Novak, T.P. (1998). Bridging the Racial Divide on the Internet,
Science, April 17.

Hoffman, D.L., Novak, T.P. & Schlosser, AB. (2000). The Evolution of the
Digital Divide: How Gaps in Internet Access May Impact Electronic Commerce. Journal
of Computer Mediated Communicatiom 5(3), 1-60.

Hunter, J.E. & Gerbing, D.W. (1982). Machiavellian beliefs and personality:
Construct invalidity of the machavellianism dimension. Journal of Personality and Social
Psychology, 43, 1293 — 1305.

Jackson, L, Barbatsis, G., Biocca, F., Fitzgerald, H., & VonEye, A. (2000).
HomeNetToo: Motivational, affective, and cognitive factors ﬁend Internet use: Explaining
the digital divide And the Internet paradox. Accepted grant proposal to the National

Science Foundation, Information Technology Research, Computation and Social
Systems.

Katz, 1. E & Aspden, P. (1997). A nation of strangers? Communicationgf the
ACM 40(12), 81-86.

 

Kaye, B. (1996). Uses and gratifycgtions of the World Wide Web. Presented to the .
Communication Technology and Policy Division of. the Association for Education in
Journalism and Mass Communication.

Korgaonkar, P. & Wolin, L. (1999). A multivariate analysis of Web usage.
Journal of Advertising Research, 39, 53-68.

92

Kraut, R, Lundmark, V, Patterson, M, Kiesler, S., Mukophadhyay, T. & Scherlis,
W. (1998). lntemet paradox: A social technology that reduces social involvement and
psychological well-being? American Psychologist, 53, 1017-1031.

LaRose, R. L., Gregg, J. L. & Eastin, MS. (1999). Audiographic Telecourses for
the Web: An Experiment. Journal of Computer Mediated Communication, 4(2),

[Online] Available http://jcmchtn'i.ac.il/vol4/issue2/, February 26, 2001

LaRose, R. Mastro, D., & Eastin, MS. (in press). Understanding lntemet Usage.
Social Science and Computer Review.

Lewis, C. (1985). An exploratory study of the inﬂuences of self-eﬁicay on the
personal computer learner. Dissertation Abstracts International, 46, As reported in Oliver,
T.A. & Shapiro, F. (1993). Self-Efficacy and Computers. Journal of Comppter-Based
Interactions. 20, 81-85.

Lin, C. (1999). Online-service adoption likelihood. Journal of Advertising
Research, 39, 79-89.
Lubans, J. (1999). When students hit the surf. School of Library Journal, 45(9),
144-147.

Manz, C.C. & Sims, HP. (1986). Beyond Imitation: Complex Behavioral and
Affective Linkages Resulting from Exposure to Leadership Training Models. Journal of
Applied Psychology. 71(4), 571-578.

Nahl, D. (1996). Affective monitoring of Internet learners: Perceived self-efﬁcacy
and success. Journal of American Society for Information Sciences, 33, 100-109.

Nahl, D. & Meer, M. P. (1997). User-centered assessment of two Web browsers:
Errors, perceived self-efficacy, and success. Journal of American Society for Information
Sciences, 34, 89-97.

National Telecommunication and Information Administration, NTIA (2000).
Falling Through the Net: Toward Digital Inclusion. US Department of Commerce,

Washington, DC. [Online] Available httpyzllwwwntiadocgov, October 21, 2000

National Telecommunication and Information Administration, NTIA (1999).

Falling Through the Net: Deﬁning the digital divide. US Department of Commerce,
Washington, DC. [Online] Available http://wwwntiadocgov

Nie, N. n. & Embring, L. (2000). lntemet and Socieg: A Preliminary Rgpprt.
Stanford Institute for the Quantitative Study of Science. [Online] Available

http://www.stanford.edu/grorm/sicris/PrengeIease/Prelimipm Reportpdf, September
27, 2000

93

Nielsen NetRatings, (1999). Nielsen NteRatings. [Online] Available
http://wwwnielsennetratingscom, December 15, 1999

Oliver, T.A. & Shapiro, F. (1993). Self-Efﬁcacy and Computers. Journal of
Computer-Based Interaptionsio, 81-85.

Papacharissi, Z. & Rubin, A. (2000). Predictors of lntemet use. Journal of
Broadcasting & Electronic Media, 44(2), 175-196.

Pedhazur, E. (1994). Multiple regression in behavioral research. Fort Worth:
Harcourt Brace College Publishers.

Pew Research Foundation (2000). 1ntemet& American Life. [Online] Available
http://wwwpewinternetorg, December 10, 2000

Schoﬁeld, J.W. (1997). Psychology. Computers and classroom social processes —
A review of the literature. Social Science Compﬁuter RevierS, 27-39.

Schoﬁeld, J.W., Davidson, A., Stocks, J.E., & Futoran, G. (in press). The lntemet
in schools: A case study of educator demand and its precursors. In S. Kiesler (Ed)
Research milestones on the information highway: Social scientists look at electron
communication. Hillsdale, NJ: Erlbaum.

Schwarz, N. (1999). Self-Reports: How the Questions Shape the Answers.
American Psychologist, 54(2), 93-105.

Schnuk, DH. (1981). Modeling and Attributional Effects on Children’s
Achievment: A Self-Efﬁcacy Analysis. Journal of Educational Psycholgy, 73, 93-105.

Staples, D.S., Hulland, J .S., & Higgins, CA. (1998). A Self-Efﬁcacy Theory
Explanation for the Management of Remote: Workers in Virtual Organizations. Journal of

Computer Mediated Communication,3(4), 1-38.

Thompson, R.L., Higgins, C.A., & Howell, J. (1991). Personal Computing:
Towards a Conceptual Model of Utilization. MIS Quarterly, 14, 125-143.

Triandis, HQ (1979). Values Attitudes and Interpersonal Behavior. Nebraska
Symposium on Motivation: Beliefs, Attitude and Values. University of Nebraska Press,

Lincoln, NE, p 195-259.

US Census Bureau (2000). National Estimates Annual Population Estimates In!
Sex, Race and Hispanic Origin, Selected Years from 1990 to 2000. [Online] Available
hmﬂ/www.censusggpv/population/www/estimates/nation3.htrnl December 10, 2000

 

94

Wellman, B. & Gulia, M. (1999). Virtual communities as communities: Net
surfers don't ride alone. In. Smith, MA & Kollock, P. Communities in CyberSpace. NY:
Routledge.

Webster, J ., Heian, J .B., & Michelman, J .E. (1990). Mitigating the effects of

Computer Anxiety Through Training. In J. Degross, M. Alvia, & H. Oppelland (Eds)
Proceedings of the Eleventh International Conference on Information Systerg.

95 y

 

 

     

AN STATE LIB

IT | Ill ITI Tll‘ll Tull

31293 02177 7820