Testing a moderator-type research model on the use of high speed Internet.
Fillion, Gerard ; Ekionea, Jean-Pierre Booto
INTRODUCTION AND BACKGROUND FOR THE STUDY
The vast technological possibilities of the Internet are at the
basis of the fast progress of the information society (Al-Omoush &
Shaqrah, 2010). It has become one of the most important means of new
forms of cooperation and competition in the various subsystems of
society (Al-Omoush & Shaqrah, 2010). Anderson (2008) argues that
Internet has a great influence on people's connections to friends,
families, and their communities, on the social system of formal and
informal support, and on the working of groups and teams. It is also the
valuable instrument of scientific, social, marketing researches, and
business development (Al-Omoush & Shaqrah, 2010). In addition, the
Internet, as an information and entertainment technology, affects
education, government, publishing, retail industry, banking, broadcast
services, health care delivery, and so on (Al-Omoush & Shaqrah,
2010). So, the scope of the Internet is now worldwide and in all sectors
of the society, and then forces to deliver this essential resource to
people in households. In order to provide the reader with a good overall
view of the actual Internet population, we have regrouped in Table 1 the
different generations (as conventionalized by Strauss and Howe's
book: Generations: The History of America's Future, 1584 to 2069)
and the percentages of adult both total and online.
As shown in Table 1, "Contrary to the image of Generation Y as
the 'Net Generation,' Internet users in their twenties do not
dominate every aspect of online life. Generation X is the most likely
group to bank, shop, and look for health information online. Boomers are
just as likely as Generation Y to make travel reservations online. And
even Silent Generation Internet users are competitive when it comes to
e-mail (although teens might point out that this is proof that e-mail is
for old people). The Web continues to be populated largely by younger
generations, as more than half of the adult Internet population is
between 18 and 44 years old. But larger percentages of older generations
are online now than in the past and they are doing more activities
online, according to the Pew Research Center's Internet &
American Life Project surveys taken from 2006-2008." (Jones &
Fox, 2009, p. 1)
As the Internet population is continually growing since its
infancy, the need for always faster and performing telecommunications
networks allowing people to communicate and to perform their daily
activities at a satisfying pace has become an everyday concern for all
the countries, hence the apparition of high speed Internet in the middle
of the 1990s. For example, for more than a decade, the Government of
Canada has been developing strategies to enable Canadians to become
participants in the information society (Government of Canada, 1999;
Government Online Advisory Panel, 2003; quoted in Middleton &
Ellison, 2006, p. 1). As part of this strategy, it was recommended that
broadband Internet access be made available to all Canadian households
(National Broadband Task Force, 2001; quoted in Middleton & Ellison,
2006, pp. 1-2). There are still many not served and underserved areas in
the country (CRACIN, 2005), and the Telecommunications Policy Review
Panel has urged the federal government to "reaffirm its commitment
to maintaining Canada's global broadband leadership and to ensuring
that broadband access is available everywhere in the country"
(Telecommunications Policy Review Panel, 2006, p. 8-5; quoted in
Middleton & Ellison, 2006, p. 2). Similarly, as a public issue,
broadband has taken on a higher profile in recent months because of
President Obama's decision to include funding for broadband in the
American Recovery and Reinvestment Act (ARRA); also ARRA included $7.2
billion for broadband with the goal of accelerating the deployment of
broadband Internet access in the United States (Horrigan, 2009). These
are just two examples here of how much all of the countries are putting
a strong emphasis on strategies capable of promoting the acceleration of
the deployment of high speed Internet everywhere in the world. As a
result, the number of households with a broadband Internet connection
continues to grow with one in five households worldwide expected to have
a fixed broadband connection by the end of 2009, according to Gartner
Research (Digital Home, 2009).
In 2009, Canada ranked fifth in broadband penetration. In fact,
leading the high speed Internet race at the beginning of 2009 was South
Korea at 86%, followed by the Netherlands (80%), Denmark (75%), and Hong
Kong (72%), followed by Canada and Switzerland (69%). Rounding out the
top ten were Norway (67%), New Zealand (65%), France, Singapore and the
UK (62%). The United States ranked 14th at 60%. Overall, in 2009,
approximately 21 countries had high speed Internet connections in at
least 50 percent of homes. (Digital Home, 2009)
But, what is high speed Internet?
According to Gill (2010), ever wonder what a company means when it
says its Internet service is "high speed"? Then check out
Table 2 that documents the plethora of technologies that the Federal
Communications Commission (FCC) counts as "broadband"--be
warned, speeds can vary by as much as 2,000 percent (for example,
regarding Canada, Dunn (2010) argues that high speed Internet access is
expensive and slow)! In short, "broadband" is defined by the
FCC as anything other than "dial up"--and "high
speed" has no commonly-agreed-to definition (Gill, 2010). So, in
this paper, when we talk about "high speed" Internet or
"broadband", we then consider the two as the same
telecommunications technology which is faster than "dial up".
What are the anticipated benefits of high speed Internet?
Here we provide the reader with only a brief overview of the
anticipated benefits of high speed Internet given that the studies
presented in Table 3 are further putting in evidence some benefits.
Thus, we describe in the next paragraphs three relevant examples of
benefits that high speed Internet can bring: one at the environmental
level, one at the social level, and one at the economic level.
First, a study realized by Fuhr and Pociask released in October
2007 by the American Consumer Institute praises the benefic impact that
a widely spread use of broadband could have on the environment by
cutting greenhouse gas emissions. The study focuses on the different
behaviors that the use of high speed Internet allows, such as buying
online, telecommuting, e-materializing, teleconferencing,
videoconferencing, as well as distance learning, and converts their
benefits into saving of greenhouse gas emissions by mainly cutting on
energy. According to Fuhr and Pociask (2007; quoted in Labriet-Gross,
2007), using high speed Internet mainly influences the amount of travel,
space and material needed when you buy, work or lean based on rather
simple findings. "Indeed, instead of going to five or 6 stores to
find who has a product or who has the best price, you can just search on
the Internet and buy online, so it cuts back on the pollution linked to
the commute", explain these authors. As for the supply chain, it
decreases the inventory, which means less storage facilities, so less
need for heat, air conditioning, and lighting. For instance, Dell has
increased its sales by 36 times, while its facilities space has been
reduced by 4. Overall, e-commerce already cuts 37.5 million tons of CO2
emissions (and only in business to business (B2B) and business to
consumers (B2C) transactions), and could save 206.3 million tons in
2017, argue Fuhr and Pociask (2007; quoted in Labriet-Gross, 2007).
Varon (2010) relates the experience of Case Western Reserve
University who launched the University Circle Innovation Zone, a large
project deploying gigabit fiber optic connections to residents of
Cleveland's poorest neighborhoods. Beginning with 104 homes, local
institutions, including hospitals, schools, electric utilities, and
public safety agencies, will use the network to deliver cutting edge
services to residents. Over the next 18 months, Case Western researchers
will study changes in residents' health and other indicators of
their standard of living. One project will use videoconferencing
technology provided by LifeSize to enable residents with chronic
conditions such as diabetes to consult with healthcare providers over
high definition video. Patients will also be given devices that
automatically monitor their health and transmit data to medical
professionals. Broadband access could enable residents to take better
care of them when they cannot visit their doctors. Other projects would
provide science and math ... materials to students, deliver video feeds
to police, and collect data to help residents manage energy usage. Case
Western's technology and service provider partners are financing
this social initiative.
And, a study conducted by Orazem (2005) measuring the impact of
high speed Internet access on local economic growth suggests that this
one increases growth in earnings per worker, aggregate earnings and the
number of firms, but it lowers the rate of growth of employment. All of
these are consistent with the presumption that high speed Internet
access can lower firm costs, improve information flows with suppliers
and consumers, and, at the same time, lower the need for employees
specializing in sales or procurement. And all of these effects were
larger in less density populated areas, then suggesting that rural areas
do benefit disproportionately from high speed Internet access. A more
recent study made by Majumdar (2008) is consistent with Orazem's
(2005) findings concerning earnings per worker and employment, that is,
he found that broadband diffusion within and between the firms over time
has a positive and significant impact on wage levels, but its impact on
employment is negative.
On the other hand, what are the main barriers to high speed
Internet adoption?
According to the Pew Internet & American Life Project (see
Horrigan, 2009), the factors positively correlated with home broadband
adoption (in order of importance) are: income (household income greater
than $75,000 annually); having college degree or more; parent of minor
child in household); married or living with partner; and full-time
employee. As for the factors negatively correlated with home broadband
adoption (in order of importance), they are: having less than a high
school degree; senior individual (aged of 65 or over); living in rural
America; having a high school degree; and being African American
(non-Hispanic).
Telecommunications industry is continually in a shift of change,
alimented by technological innovation and consumers' demand for
always better and faster communication tools. High speed Internet is now
an integral part of everyday life of more than a billion people. And, as
the tendency is showing up, its use will be still increasing in the
future. Thus, this technology has and will continue to have major social
and economic impacts. Individual adoption of technology has been studied
extensively in the workplace, but far less attention has been paid to
adoption of technology in household (Brown & Venkatesh, 2005). So,
few studies have been conducted until now to verify satisfaction of
household people using high speed Internet. It is therefore crucial to
more deeply examine the determining factors in satisfaction of using
high speed Internet by people in household. This is the aim of the
present study. The related literature on the actual research area of
high speed Internet is summarized in Table 3.
As we can see in the summary of literature related to high speed
Internet presented in Table 3, very few studies until now examined the
determining factors in satisfaction of using high speed Internet by
people in household. Thus, the present study brings an important
contribution to fill this gap given it allows a better understanding of
the impacts of high speed Internet usage in people's everyday life.
It focuses on the following research question: What are the determining
factors in satisfaction of using high speed Internet by people in
household?
The paper builds on a framework suggested by Fillion (2004) in the
conduct of hypothetico-deductive scientific research in organizational
sciences, and it is structured as follows: first, the theoretical
development of the study is presented; second, the methodology followed
to conduct the study is described; finally, the results of the study are
reported and discussed.
THEORETICAL DEVELOPMENT
This study is based on the theoretical foundations developed by
Venkatesh and Brown (2001) to investigate the factors driving personal
computer (PC) adoption in American homes as well as those developed by
Brown and Venkatesh (2005) in order to verify the determining factors in
intention to adopt a PC in household by American people. In fact, Brown
and Venkatesh (2005) performed the first quantitative test of the
recently developed model of adoption of technology in households (MATH)
and they proposed and tested a theoretical extension of MATH integrating
some demographic characteristics varying across different life cycle
stages as moderating variables. And Brown et al. (2006) tested the same
integrated model in the context of PC use. As pointed out by Brown et
al. (2006), even though the technology of interest in MATH is PC, the
model is expected to generalize to other information technology (IT)
products and systems in the household context. Also, with the exception
of behavioral intention (we included user satisfaction instead of
behavioral intention given people investigated in this study already
have high speed Internet access), all the variables proposed and tested
by Brown and Venkatesh (2005) are used in this study. And we added a new
variable, mobility, in order to verify whether or not it is a factor of
satisfaction of household people using high speed Internet. The
resulting theoretical research model is depicted in Figure 1.
Figure 1 shows that Brown and Venkatesh (2005) integrated MATH and
Household Life Cycle in the following way. MATH presents five
attitudinal beliefs grouped into three sets of outcomes: utilitarian,
hedonic, and social. Utilitarian beliefs are most consistent with those
found in the workplace and can be divided into beliefs related to
personal use, children, and work (we added beliefs related to mobility).
The extension of MATH suggested and tested by Brown and Venkatesh (2005)
presents three normative beliefs: influence of friends and family,
secondary sources, and workplace referents. As for control beliefs, they
are represented in MATH by five factors: fear of technological advances,
declining cost, cost, perceived ease of use, and self-efficacy. And,
according to Brown and Venkatesh (2005), integrating MATH with a life
cycle view, including income, age, child's age, and marital status,
allows to provide a richer explanation of household PC adoption
(household high speed Internet usage in this study) than those provided
by MATH alone. Finally, as shown in Figure 1, the dependant variable of
the theoretical research model developed is related to user satisfaction
(satisfaction in the use of high speed Internet by people in household).
All of the variables integrated in the theoretical research model
depicted in Figure 1 are defined in Table 4.
We can see in Table 4 that the definitions of MATH variables
integrated in the theoretical research model proposed in Figure 1 are,
in the whole, adapted from the theoretical foundations developed by
Venkatesh and Brown (2001) to investigate the factors driving PC
adoption in American homes. As for the definitions of the variables
related to the household life cycle, they were taken from Danko and
Schaninger (1990) as well as Wagner and Hanna (1983), respectively. And
the definition of the new independent variable that we added to the
model is from our own. In fact, we defined this variable in accordance
with which we wanted to measure regarding mobility before developing and
validating items that measure the variable on the basis of the
definition formulated.
[FIGURE 1 OMITTED]
In the reminder of the section, we develop eight research
hypotheses (H1-H8) related to the model suggested in Figure 1.
H1: Marital status and age will moderate the relationship between
applications for personal use and satisfaction of using high speed
Internet at home.
H2: Child's age will moderate the relationship between utility
for children and satisfaction of using high speed Internet at home.
H3: Age will moderate the relationship between utility for
work-related use and satisfaction of using high speed Internet at home.
H4: Age will moderate the relationship between applications for fun
and satisfaction of using high speed Internet at home.
H5: Age will moderate the relationship between status gains and
satisfaction of using high speed Internet at home.
H6: Age, marital status, and income will moderate the relationship
between the normative beliefs ((a) friends and family influences; (b)
secondary sources' influences; and (c) workplace referents'
influences) and satisfaction of using high speed Internet at home.
H7: Age and income will moderate the relationship between the
external control beliefs ((a) fear of technological advances; (b)
declining cost; and (c) cost) and satisfaction of using high speed
Internet at home.
H8: Age will moderate the relationship between the internal control
beliefs ((a) perceived ease of use; and (b) self-efficacy) and
satisfaction of using high speed Internet at home.
In the next section of the paper, the methodology followed to
conduct the study is described.
METHODOLOGY
The study was designed to gather information concerning high speed
Internet satisfaction in Atlantic Canadian households. Indeed, the focus
of the study is on individuals who have high speed Internet access at
home. We conducted a telephone survey research among individuals of a
large area in Atlantic Canada. In this section, we describe the
instrument development and validation, the sample and data collection,
as well as the data analysis process.
Instrument Development and Validation
To conduct the study, we used the survey instrument developed and
empirically validated by Brown and Venkatesh (2005) to which we added
two new scales, the first one measuring another dimension in
satisfaction of using high speed Internet by people in household, that
is, mobility, and the last one measuring user satisfaction as such. The
survey instrument was then translated in French (a large part of the
population in Atlantic Canada is speaking French) and both the French
and English versions were evaluated by peers. This review assessed face
and content validity (see Straub, 1989). As a result, few changes were
made to reword items and, in some cases, to drop items that were
possibly ambiguous, consistent with Moore and Benbasat's (1991) as
well as DeVellis's (2003) recommendations for scale development.
Subsequent to this, we distributed the survey instrument to a group of
MBA students for evaluation. Once again, minor wording changes were
made. Finally, we performed some adjustments to the format and
appearance of the instrument, as suggested by both peers and MBA
students, though these minor changes had not a great importance here
given the survey was administered using the telephone. As the instrument
was already validated by Brown and Venkatesh (2005) and showed to be of
a great reliability, that we used the scale developed by Hobbs and
Osburn (1989) and validated in their study as well as in several other
studies to measure user satisfaction, and that we added only few items
to measure the new variable mobility, then we have not performed a
pilot-test with a small sample. The evaluations by both peers and MBA
students were giving us some confidence that we could proceed with a
large-scale data collection.
Sample and Data Collection
First, in this study, we chose to survey people in household over
18 years from a large area in Atlantic Canada, which have high speed
Internet access. To do that, undergraduate and graduate students
studying at our faculty were hired to collect data using the telephone.
A telephone was then installed in an office of the faculty, and
students, one at a time over a 3- to 4-hour period, were asking people
over the telephone to answer our survey. And, to get the more
diversified sample as possible (e.g., students, retired people, people
not working, people working at home, and people working in enterprises),
data were collected from 9 a.m. to 9 p.m. Monday through Friday over a
5-week period. Using the telephone directory of the large area in
Atlantic Canada chosen for the study, students were randomly selecting
people and asking them over the telephone to answer our survey. The
sample in the present study is therefore a randomized sample, which is
largely valued in the scientific world given the high level of
generalization of the results got from such a sample. Once an individual
had the necessary characteristics to answer the survey and was accepting
to answer it, the student was there to guide him/her to rate each item
of the survey on a seven points Likert-type scale (1: strongly disagree
... 7: strongly agree). In addition, the respondent was asked to answer
some demographic questions. Finally, to further increase the response
rate of the study, each respondent completing the survey had the
possibility to win one of the 30 Tim Hortons $10 gift certificates which
were drawn at the end of the data collection. To that end, the phone
number of each respondent was put in a box for the drawing. Following
this data collection process, 322 people in household answered our
survey over a 5-week period.
Data Analysis Process
The data analysis of the study was performed using a structural
equation modeling software, that is, Partial Least Squares (PLS-Graph
3.0). Using PLS, data have no need to follow a normal distribution and
it can easily deal with small samples. In addition, PLS is appropriate
when the objective is a causal predictive test instead of the test of a
whole theory (Barclay et al., 1995; Chin, 1998) as it is the case in
this study. To ensure the stability of the model developed to test the
research hypotheses, we used the PLS bootstrap resampling procedure (the
interested reader is referred to a more detailed exposition of
bootstrapping (see Chin, 1998; Efron & Tibshirani, 1993)) with an
iteration of 100 sub-sample extracted from the initial sample (322
Atlantic Canadian people). Some analyses were also performed using the
Statistical Package for the Social Sciences software (SPSS 13.5). The
results follow.
RESULTS
In this section of the paper, the results of the study are
reported. First, we begin to present some characteristics of the
participants. Then we validate the PLS model developed to test the
research hypotheses. Finally, we describe the results got from PLS
analyses to test the research hypotheses.
Participants
The participants in this study were relatively aged, with a mean of
40 years and a standard deviation of 13.7 years. These statistics on the
age of the participants are, in fact, consistent with the growing old
population phenomenon. Near from two third of the participants were
female (62.7%). Near from 80% of the participants were married (52.1%)
or single (26.8%). The gross yearly income of the respondents in the
study was in the range of $0 to $60,000. Indeed, 80.9% of the
respondents were winning between $0 and $60,000, and, from this
percentage, 47.3% were winning between $30,000 and $60,000. And 5.3% of
the respondents were winning $100,000 or over. Concerning the level of
education, 20.3% of the participants in the study got a high-school
diploma, 28.3% got a college degree, 37.3% completed a baccalaureate,
and 10.3% got a master. Only 2.9% of the participants completed a
doctorate, which is relatively consistent with the whole population in
general. Finally, the respondents in our study were mainly full-time
employees (46.8%), part-time employees (14.6%), retired people (13.3%),
self employed (8.9%), unemployed (7.6%), and students (5.4%).
Validation of the PLS Model to Test Hypotheses
First, to ensure the reliability of a construct or a variable using
PLS, one must verify the three following properties: individual item
reliability, internal consistency, and discriminant validity (for more
details, see Yoo and Alavi, 2001).
To verify individual item reliability, a confirmatory factor
analysis (CFA) was performed on independent and dependent variables of
the theoretical research model. A single iteration of the CFA was
necessary given all loadings of the variables were superior to 0.50 and
then none item was withdrawn nor transferred in another variable in
which the loading would have been higher. Indeed, in the whole, items
had high loadings, which suppose a high level of internal consistency of
their corresponding variables. In addition, loadings of each variable
were superior to cross-loadings with other variables of the model. Hence
the first criterion of discriminant validity was satisfied.
And to get composite reliability indexes and average variance
extracted (AVE) in order to satisfy the second criterion of discriminant
validity and to verify internal consistency of the variables, we used
PLS bootstrap resampling procedure with an iteration of 100 sub-sample
extracted from the initial sample (322 Atlantic Canadian people). The
results are presented in Table 5.
As shown in Table 5, PLS analysis indicates that all square roots
of AVE (boldfaced elements on the diagonal of the correlation matrix)
are higher than the correlations with other variables of the model. In
other words, each variable shares more variance with its measures than
it shares with other variables of the model. Consequently, discriminant
validity is verified. Finally, as supposed previously, we can see in
Table 5 that PLS analysis showed high composite reliability indexes for
all variables of the theoretical research model. The variables have
therefore a high internal consistency, with composite reliability
indexes ranging from 0.74 to 0.99.
Hypothesis Testing
First, to get the significant variables in the study and the
percentage of variance explained ([R.sup.2] coefficient) by all the
variables of the theoretical research model, we developed a PLS model
similar to those of Fillion (2005), Fillion and Le Dinh (2008), Fillion
et al. (2010a), Fillion and Booto Ekionea (2010b), and Yoo and Alavi
(2001). And to ensure the stability of the model, we used the PLS
bootstrap resampling procedure with an iteration of 100 sub-sample
extracted from the initial sample (322 Atlantic Canadian people). The
PLS model is depicted in Figure 2.
As we can see in Figure 2, all of the variables of our theoretical
research model, used as independent variables, are explaining 49.2% of
the variance on the dependant variable user satisfaction. And about half
of these variables are significant, that is, they are determining
factors in satisfaction of using high speed Internet by people in
household. More specifically, the five more significant variables (in
order of significance) are mobility (t = 5.177, beta = 0.238, p <
0.001), cost (t = 3.839, beta = -0.177, p < 0.001), applications for
fun (t = 3.504, beta = 0.218, p < 0.001), age (t = 3.009, beta =
0.245, p < 0.001), and perceived ease of use (t = 2.800, beta =
0.288, p < 0.001). And two other variables are significant at the
level of significance required in this study, that is, p [less than or
equal to] 0.05. They are fear of technological advances (t = 1.908, beta
= -0.101, p < 0.05) as well as self-efficacy (t = 1.646, beta =
0.163, p < 0.05). As shown in Figure 2, our new variable mobility is
by far the more significant variable in the global PLS structural model.
So, in this study, the fact that high speed Internet allows using only
this technology to perform all personal and professional activities is
by far the more satisfying factor for Atlantic Canadian people when they
choose to get access to high speed Internet from Internet services
providers (ISP).
[FIGURE 2 OMITTED]
Finally, to measure interaction effect of moderator variables (the
life cycle stage characteristics: income (I), marital status (MS), age
(A), and child's age (CA)) in order to verify hypotheses 1 through
8, we used the PLS procedure proposed by Chin et al. (2003) (see the
paper for more details). On the other hand, in a review of 26 papers
assessing interaction effect of moderator variables published between
1991 and 2000 in information systems (IS) journals, Carte and Russell
(2003) found nine errors frequently committed by researchers when they
estimate such an effect, and provided solutions (see their paper for
more details). We tried to avoid these nine errors in applying their
solutions to test hypotheses 1 through 8. Indeed, among others, in the
verification of hypotheses 1 through 8 that follows, interaction effect
of a moderator variable is significant if, and only if, the path between
the latent variable (the multiplication of items of independent and
moderator variables forming interaction effect) and the dependent
variable is significant, as well as if the change in [R.sup.2]
coefficient (the difference between the [R.sup.2] calculated before the
addition of interaction effect and those calculated after the addition
of interaction effect, that is, A[R.sup.2] (named delta [R.sup.2])) is
greater than 0.
For a matter of space, as the test of hypotheses 1 through 8
required the development of several PLS structural equation models (two
models per hypothesis, that is, 16 models), we summarize PLS analyses to
test each hypothesis. And, as for the PLS model developed to get the
significant variables in the study and the percentage of variance
explained by all the variables of the theoretical research model
previously, for each PLS model developed, we used the PLS bootstrap
resampling procedure with an iteration of 100 sub-sample extracted from
the initial sample (322 Atlantic Canadian people) to ensure the
stability of the model.
Concerning hypothesis 1 related to the independent variable
applications for personal use (APU), the path from the latent variable
APU*MS*A to the dependent variable user satisfaction is not significant
(t = 0.882, beta = -0.181), but there is a substantial change in
[R.sup.2] (^[R.sup.2] = 0.006). So, contrary to our expectations, the
moderator variables marital status and age have not an influence on the
relationship between applications for personal use and satisfaction of
using high speed Internet by people in household. Hypothesis 1 is
therefore not supported. The scenario is similar for hypothesis 2
related to the independent variable utility for children (UC). The path
from the latent variable UC*CA to the dependent variable user
satisfaction is not significant (t = 0.219, beta = 0.039) and there is
no change in [R.sup.2] (^[R.sup.2] = 0.000). Also, contrary to what we
expected, the moderator variable child's age has not an influence
on the relationship between utility for children and satisfaction of
using high speed Internet by people in household. As a result,
hypothesis 2 is not supported. For hypothesis 3 related to the
independent variable utility for work-related use (UWRU), the path from
the latent variable UWRU*A to the dependent variable user satisfaction
is significant (t = 1.646, beta = 0.457, p < 0.05) and there is a
huge change in [R.sup.2] (^[R.sup.2] = 0.015). Thus, as we expected, the
moderator variable age has an influence on the relationship between
utility for work-related use and satisfaction of using high speed
Internet by people in household. Hypothesis 3 is therefore supported.
The scenario is similar for hypothesis 4 related to the independent
variable applications for fun (AF), the path from the latent variable
AF*A to the dependent variable user satisfaction is significant (t =
1.695, beta = -0.334, p < 0.05) and there is a huge change in
[R.sup.2] (^[R.sup.2] = 0.014). Thus, as we expected, the moderator
variable age has an influence on the relationship between applications
for fun and satisfaction of using high speed Internet by people in
household. As a result, hypothesis 4 is also supported. And the scenario
is still similar regarding hypothesis 5 related to the independent
variable status gains (SG), the path from the latent variable SG*A to
the dependent variable user satisfaction is significant (t = 1.712, beta
= -0.339, p < 0.05) and there is a huge change in [R.sup.2]
(^[R.sup.2] = 0.011). Thus, as we expected, the moderator variable age
has an influence on the relationship between status gains and
satisfaction of using high speed Internet by people in household.
Consequently, hypothesis 5 is also supported.
In the case of hypothesis 6a related to the independent variable
friends and family influences (FFI), the path from the latent variable
FFI*MS*A*I to the dependent variable user satisfaction is not
significant (t = 0.894, beta = -0.122), but there is a change in
[R.sup.2] (^[R.sup.2] = 0.004). Also, contrary to our expectations, the
moderator variables marital status, age, and income have not an
influence on the relationship between friends and family influences and
satisfaction of using high speed Internet by people in household.
Hypothesis 6a is then not supported. The scenario is similar for
hypothesis 6b related to the independent variable secondary
sources' influences (SSI), the path from the latent variable
SSI*MS*A*I to the dependent variable user satisfaction is not
significant (t = 0.263, beta = 0.041) and there is no change in
[R.sup.2] (^[R.sup.2] = 0.000). Contrary to what we expected, the
moderator variables marital status, age, and income have not an
influence on the relationship between secondary sources' influences
and satisfaction of using high speed Internet by people in household. As
a result, hypothesis 6b is not supported. The scenario is also similar
concerning hypothesis 6c related to the independent variable workplace
referents' influences (WRI), the path from the latent variable
WRI*MS*A*I to the dependent variable user satisfaction is not
significant (t = 0.327, beta = 0.042) and there is no change in
[R.sup.2] (^[R.sup.2] = 0.000). Contrary to our expectations, the
moderator variables marital status, age, and income have not an
influence on the relationship between workplace referents'
influences and satisfaction of using high speed Internet by people in
household. Hypothesis 6c is therefore not supported.
As for hypothesis 7a related to the independent variable fear of
technological advances (FTA), the path from the latent variable FTA*A*I
to the dependent variable user satisfaction is not significant (t =
0.888, beta = -0.109), but there is a change in [R.sup.2] (^[R.sup.2] =
0.003). Contrary to our expectations, the moderator variables age and
income have not an influence on the relationship between fear of
technological advances and satisfaction of using high speed Internet by
people in household. Hypothesis 7a is therefore not supported. The
scenario is similar for hypothesis 7b related to the independent
variable declining cost (DC), the path from the latent variable DC*A*I
to the dependent variable user satisfaction is not significant (t =
0.434, beta = 0.068) and there is no change in [R.sup.2] (^[R.sup.2] =
0.000). Also, contrary to what we expected, the moderator variables age
and income have not an influence on the relationship between declining
cost and satisfaction of using high speed Internet by people in
household. Consequently, hypothesis 7b is not supported. And the
scenario is also similar for hypothesis 7c related to the independent
variable cost (C), the path from the latent variable C*A*I to the
dependent variable user satisfaction is not significant (t = 0.021, beta
= 0.003) and there is no change in [R.sup.2] (^[R.sup.2] = 0.000). Thus,
contrary to our expectations, the moderator variables age and income
have not an influence on the relationship between cost and satisfaction
of using high speed Internet by people in household. As a result,
hypothesis 7c is not supported.
Finally, concerning hypothesis 8a related to the independent
variable perceived ease of use (PEU), the path from the latent variable
PEU*A to the dependent variable user satisfaction is not significant (t
= 0.801, beta = -0.324), but there is a change in [R.sup.2] (^[R.sup.2]
= 0.003). Thus, contrary to our expectations, the moderator variable age
has not an influence on the relationship between perceived ease of use
and satisfaction of using high speed Internet by people in household. As
a result, hypothesis 8a is not supported. The scenario is different for
hypothesis 8b related to the independent variable self-efficacy (SE),
the path from the latent variable SE*A to the dependent variable user
satisfaction is significant (t = 2.412, beta = -1.286, p < 0.01) and
there is a huge change in [R.sup.2] (A[R.sup.2] = 0.033). So, as we
expected, the moderator variable age has an influence on the
relationship between self-efficacy and satisfaction of using high speed
Internet by people in household. Consequently, hypothesis 8b is
supported. In order to provide the reader with an overall view of the
test of hypotheses, Table 6 presents a summary.
In summary, as shown in Table 6, four hypotheses have been
supported in our study, that is, H3, H4, H5, and H8b. Thus, the
moderator variable age had several moderating effects in this study. As
for moderator variables marital status, income, and child's age,
these ones had not a significant moderating effect on the relations
between the independent and dependent variables involved. Hence
hypotheses H1, H2, H6a, H6b, H6c, H7a, H7b, and H7c were not supported.
And the moderator variable age had not a significant moderating effect
on the relation between perceived ease of use and satisfaction of using
high speed Internet at home. Hence hypothesis H8a was not supported.
In the next and last section of the paper, we discuss about the
more important findings of the study, the theoretical and practical
implications, the limitations, and the future directions.
DISCUSSION AND CONCLUSIONS
This last section is devoted to a discussion about the findings of
the study and some conclusions. First, to support our discussion and
conclusions, we provide the reader with a more detailed view of the PLS
structural equation model developed to get the significant variables in
the study, including the percentages of variance explained of variables
(see Table 7).
As shown in Table 7 (and Figure 2), the eighteen independent
variables examined in the study explained 49.2 percent ([R.sup.2] =
0.492) of the variance in satisfaction of using high speed Internet at
home. And we can also see in Table 7 that the seven variables who showed
to be significant (see also the significant beta path coefficients in
Figure 2), that is, mobility, cost, applications for fun, age, perceived
ease of use, fear of technological advances, and self-efficacy explained
alone 45.9 percent of the variance in satisfaction of using high speed
Internet at home. Thus, these seven variables are assuredly very
important factors to take into account in future studies on high speed
Internet and on the part of high speed Internet providers, and more
particularly self-efficacy and perceived ease of use which explained
alone 33 percent of this variance (see Table 7). It is very interesting
and surprising here to see that the new variable that we added to the
Brown and Venkatesh's (2005) theoretical research model, that is
mobility, showed to be the more significant (t = 5.177, beta = 0.238, p
< 0.001; see Table 7) in satisfaction of using high speed Internet by
people in household. Indeed, the present study showed that people are,
to some extent, using high speed Internet for a matter of mobility
(e.g., high speed Internet provides them with the possibility to use
only this technology to perform all their personal and professional
activities). So, here is a new variable that we can now assuredly
include in the integrated research model of MATH and household life
cycle characteristics suggested by Brown and Venkatesh (2005) as well as
Brown et al. (2006) to test in future studies. In fact, we included this
new variable mobility in the integrated model of MATH and household life
cycle characteristics in several different studies (see Fillion &
Berthelot, 2007; Fillion & Le Dinh, 2008; Fillion & Booto
Ekionea, 2010b) and it always showed a very significant effect on the
dependent variables involved. Of course, its inclusion in the integrated
model will depend on its relevance to the technologies examined in the
studies. For example, mobility can be included in studies on mobile
phone, high speed Internet, or PC, but it cannot be integrated in
studies on e-government services, e-learning, or course management
software. On the practical point of view, this new variable mobility can
be included in the sales marketing plan of high speed Internet
providers.
In the large-scale study in which Brown and Venkatesh (2005)
integrated MATH and some household life cycle characteristics (as
moderating variables), the integrated model explained 74 percent of the
variance in intention to adopt a PC for home use, a substantial increase
of 24 percent over baseline MATH that explained 50 percent of the
variance. In the present study, we used the integrated model proposed by
Brown and Venkatesh (2005). We also added a new independent variable to
the model, that is, mobility. And we also used the household life cycle
variables as moderating variables in our research model as did Brown and
Venkatesh (2005). Finally, as we investigated the perceptions of people
already using high speed Internet at home instead of those having the
intention to adopt high speed Internet, as did Brown and Venkatesh
(2005) for the PC, then we used the dependent variable user satisfaction
instead of behavioral intention. And the model explained 49.2 percent of
the variance in satisfaction of using high speed Internet by people in
household (see Table 7 and Figure 2). Thus, in this study, using a
different dependent variable than did Brown and Venkatesh (2005), that
is user satisfaction instead of behavioral intention, our research model
explained the same percentage of variance than those explained by MATH
alone (e.g., without the household life cycle characteristics and using
behavioral intention as dependent variable).
Further, in a previous study in which we investigated the intention
to buy a mobile phone by people in household (see Fillion &
Berthelot, 2007), we also used the theoretical research model suggested
by Brown and Venkatesh (2005) to which we added the same independent
variable mobility than we included in the present study in which we
investigated satisfaction of using high speed Internet at home. And our
model explained the same percentage of variance in intention to buy a
mobile phone than in the present study in satisfaction of using high
speed Internet, that is, 50 percent. According to this finding, we can
then see that the variable user satisfaction is as much appropriate as
dependent variable in the research model proposed by Brown and Venkatesh
(2005) than is behavioral intention. And this finding is also consistent
with what is argued by Brown et al. (2006), that is, the model is
expected to generalize to other IT products and systems in the household
context. However, when the dependent variable of the model is
interchanged (e.g., user satisfaction instead of behavioral intention)
as did Fillion and Le Dinh (2008) and Fillion and Booto Ekionea (2010b)
in studies examining the determining factors in the use of mobile phone,
it seems that high speed Internet (the technology involved in the
present study) is a more appropriate technology to examine than is
mobile phone, since the amount of variance explained by the model in the
present study is largely superior, that is, 50 percent comparatively to
32 percent and 35 percent respectively for the two studies on the mobile
phone quoted above. Besides, it is to be noted that, in the model we
used in this study, less independent variables showed to be good
predictors in satisfaction of using high speed Internet by people in
household than in the two studies quoted above examining the predictors
in satisfaction of using mobile phone by people in household. So, this
study brings several interesting findings which contribute to the
technology adoption and use literature by offering key insights
regarding the differences between adoption, use, and satisfaction of
using technology at home.
First, our main findings regarding user satisfaction of a certain
technology are consistent with those got in Tao et al.'s (2009)
study in the sense that several variables have a significant effect on
user satisfaction, but other variables need an improvement on some
elements, for example, consumer service, transmission line and
connection stability need to be improved in Tao et al.'s (2009)
study, while other's usage influences, utility for work-related use
as well as applications for personal use need to be improved in our
study. Second, we found seven very important variables that seem to be
good predictors in satisfaction of using high speed Internet at home,
and more particularly perceived ease of use, cost, applications for fun,
age, and the new variable that we added to the Brown and
Venkatesh's (2005) model, mobility (see Table 7). And the fact that
the moderator variable age has been found a very significant predictor
(taken as independent variable) in satisfaction of using high speed
Internet and a very significant influencing factor (taken as moderator
variable) in all hypotheses supported in the study provides additional
evidence concerning the importance of integrating household life cycle
stage in research examining household technology adoption and use. These
seven variables are also important to take into account by high speed
Internet providers in order to improve actual services, to offer new
services still better adapted to people's needs, as well as to
perform their sales marketing. Third, we found that people are, to some
extent, using high speed Internet for a matter of mobility given our new
variable mobility showed to be the more significant in the study (see
Table 7). Fourth, we found that, depending on the technology studied,
the dependent variables behavioral intention and user satisfaction might
be interchanged in the model proposed by Brown and Venkatesh (2005)
given the amount of variance explained by the models are quite varying
across technologies and dependent variables observed. The dependent
variable use behavior proposed by Thompson et al. (1991) and the
dependent variable user satisfaction (examined in the present study)
conceptualized in the work of Cyert and March (1963), and initially
developed by Ives et al. (1983), may also be further tested in future
studies. And, finally, we suggest the test of new independent variables
which may explain a greater amount of variance in satisfaction of using
high speed Internet by people in household in future studies. To that
end, we recommend three new independent variables in the next paragraph.
Indeed, depending on the technology examined, it would be
interesting in future studies to add a variable such as utility for
security (in utilitarian outcomes) to the theoretical research model
suggested by Brown and Venkatesh (2005) augmented with the new variable
mobility that we tested in this study. This variable has been found very
significant in the case of mobile phone technology in the studies
conducted by Fillion and Berthelot (2007), Fillion and Le Dinh (2008),
as well as Fillion and Booto Ekionea (2010b). Who knows, people might be
also using high speed Internet for a matter of their own security and
those of their family given this technology allows to rapidly
communicate with helping people or organisms everywhere in the world.
The variable social norm might be also added in social outcomes. Who
knows, people might be using high speed Internet just to do as
everybody. And the variable provider support might be added in external
control beliefs. People might be according a great importance to the
quality of support offered by the high speed Internet provider. It would
be also interesting to test the actual model in other situations and
with other populations. For example, with colleagues from Brasil
(University of Lavras) and Cameroon (University of Yaounde I), we are
now testing the actual model with people who are using a mobile phone at
home. As in this study, we used the dependent variable user satisfaction
since the respondents are already using a mobile phone. The results of
these studies will follow in subsequent papers. It will be interesting
to see whether the results remain the same as those got from people who
are using high speed Internet in household.
Regarding the limitations of this study, as pointed out by Brown
and Venkatesh (2005), the primary limitation is the reliance on a single
informant. It is possible that other members of the household would have
provided different responses concerning the motivations of using high
speed Internet at home. Future research in household use of technology
should incorporate responses from multiple members of the household to
truly assess the nature of household use. A second limitation of the
study is that it was conducted in only one area in Atlantic Canada. If
the study would have been carried out in the whole Atlantic Canada, its
results would be of a higher level of generalization. But the fact that
the sample of the study was a randomized sample allows a high level of
generalization of its results. Another limitation of the study is the
administration of the survey instrument over the telephone. Some
respondents might have not very well understood some items of the survey
instrument over the telephone and then provided more or less precise
ratings on these items, introducing the possibility of some response
bias. But the method we privileged in this study to administer the
survey instrument is not an exception to the rule. Each method has its
own limitations.
To conclude, much more research will be needed on the use of
technology in households in order to better understand its impacts on
people's daily life. The research will allow, among others, at
least to minimize, if not to remove, some negative impacts of technology
in people's daily life in the future and to develop new
technologies still better adapted to people's needs. So, rest
assured that we will continue to inquire into this new and exciting
field.
ACKNOWLEDGMENTS
The authors would sincerely like to thank professor Wynne W. Chin
(University of Houston at Texas) who kindly offered to us a license of
the last version of his structural equation modeling software PLS to
perform the data analysis of this study. We are also grateful to the
Faculte des Etudes Superieures et de la Recherche (FESR) at the
University of Moncton for its financial contribution to this study.
REFERENCES
Al-Omoush, K.S. & A.A. Shaqrah (2010). An empirical study of
household Internet continuance adoption among Jordanian users.
International Journal of Computer Science and Network Security, 10(1),
32-44.
Anderson, B. (2008). The social impact of broadband household
Internet access. Information, Communication, & Society, 11(1), 5-24.
Barclay, D., C. Higgins & R. Thompson (1995). The partial least
squares (PLS) approach to causal modeling, personal computer adoption
and use as an illustration. Technology Studies, 2(2), 285-309.
Brown, S.A., V. Venkatesh & H. Bala (2006). Household
technology use: Integrating household life cycle and the model of
adoption of technology in households. The Information Society, 22,
205-218.
Brown, S.A. & V. Venkatesh (2005). Model of adoption of
technology in households: A baseline model test and extension
incorporating household life cycle. MIS Quarterly, 29(3), 399-426.
Cambini, C. & Y. Jiang (2009). Broadband investment and
regulation: A literature review. Telecommunications Policy, 33, 559-574.
Carte, T.A. & C.J. Russell (2003). In pursuit of moderation:
Nine common errors and their solutions. MIS Quarterly, 27(3), 479-501.
Chin, W.W. (1998). The partial least squares approach to structural
equation modeling. In G.A. Marcoulides (Ed.), Modern Methods for
Business Research (pp. 295-336), Mahwah, NJ: Lawrence Erlbaum
Associates.
Chin, W.W., B.L. Marcolin & P.R. Newsted (2003). A partial
least squares latent variable modeling approach for measuring
interaction effects: Results from a Monte Carlo simulation study and an
electronic-mail emotion/adoption study. Information Systems Research,
14(2), 189-217.
Compeau, D.R. & C.A. Higgins (1995a). Application of social
cognitive theory to training for computer skills. Information Systems
Research, 6(2), 118-143.
Compeau, D.R. & C.A. Higgins (1995b). Computer self-efficacy:
Development of a measure and initial test. MIS Quarterly, 19(2),
189-211.
CRACIN (2005). Written Submission to Telecommunications Policy
Review Panel. Toronto: Canadian Research Alliance for Community
Innovation & Networking.
Cyert, R.M. & J.G. March (1963). A Behavioral Theory of the
Firm. Englewood Cliffs, NJ: Prentice-Hall.
Danko, W.D. & C.M. Schaninger (1990). An empirical evaluation
of the Gilly-Enis updated household life cycle model. Journal of
Business Research, 21, 39-57.
Davis, F.D. (1989). Perceived usefulness, perceived ease of use and
user acceptance of information technology. MIS Quarterly, 13(3),
319-340.
DeVellis, R.F. (2003). Scale Development: Theory and Applications
(2nd ed.). Thousand Oaks, CA: Sage Publications.
Digital Home (2009). Canada ranked fifth in broadband penetration.
Retrieved August 6, 2010, from http://www.digitalhome.ca/
2009/09/canada-ranked-fifth-in-broadband- penetration/.
Dumitru, R.C., T. Burkle, S. Potapov, B. Lausen, B. Wiese &
H.-U. Prokosch (2007). Use and perception of Internet for health related
purposes in Germany: Results of a national survey. International Journal
of Public Health, 52, 275-285.
Dunn, J. (2010). Digital home thoughts: High-speed Internet access
in Canada: It's expensive and slow. Digital Home Thoughts,
http://www.digitalhomethoughts.com/news/show/97187/high-
speed-internet-access-incanada. Retrieved August 9, 2010.
Efron, B. & R.J. Tibshirani (1993). An Introduction to the
Bootstrap. New York: Chapman and Hall.
Fillion, G. (2005). L'integration des TIC dans la formation
universitaire : une etude des resultats educationnels des etudiants dans
les contextes de presence et de non presence en classe. Doctoral Thesis
(Ph.D.), Faculty of Administration, Laval University, Quebec.
Fillion, G. (2004). Publishing in the organizational sciences: An
extended literature review on the dimensions and elements of an
hypothetico-deductive scientific research, and some guidelines on
"how" and "when" they should be integrated. Academy
of Information and Management Sciences Journal, 7(1), 81-114.
Fillion, G., M. Limayem, T. Laferriere & R. Mantha (2010a).
Onsite and online students' and professors' perceptions of ICT
use in higher education. In N. Karacapilidis (Ed.), Novel Developments
in Web-Based Learning Technologies: Tools for Modern Teaching (Chapter
6), Hershey, PA: IGI Global Publishing.
Fillion, G. & J.-P. Booto Ekionea (2010b). Testing a
moderator-type research model on the use of mobile phone. Forthcoming in
Academy of Information and Management Sciences Journal, 13.
Fillion, G. & T. Le Dinh (2008). An extended model of adoption
of technology in households: A model test on people using a mobile
phone. Management Review: An International Journal, 3(1), 58-91.
Fillion, G. & S. Berthelot (2007). An extended model of
adoption of technology in households: A model test on people's
intention to adopt a mobile phone. Management Review: An International
Journal, 2(2), 4-36.
Fuhr, J.P. & S.B. Pociask (2007). Broadband services: Economic
and environmental benefits. Retrieved August 9, 2010, from,
http://www.theamericanconsumer.org/2007/10/31/broadband-services-
economic-andenvironmental-benefits/.
Gill, K.E. (2010). What does 'high speed Internet' mean
exactly?. Retrieved August 6, 2010, from http://wiredpen.com/2010/
03/26/what-does-high-speed-internet-mean-exactly/.
Government of Canada (1999). Speech from the throne to open the
second session of the 36th parliament of Canada.
http://www.pco-bcp.gc.ca/
default.asp?Language=E&Page=InformationResources&sub=sftddt&doc= sftddt1999_e.htm. Retrieved August 6, 2010, from
Government On-Line Advisory Panel (2003). Connecting with
Canadians: Pursuing Service Transformation. Ottawa: Treasury Board of
Canada.
Helsper, E.(2010). Gendered Internet use across generations and
life stages. Communication Research, 37(3), 352-374.
Hobbs, V.M. & D.D. Osburn (1989). Distance Learning Evaluation
Study Report II: A Study of North Dakota and Missouri Schools
Implementing German I by Satellite. ERIC ED 317 195.
Horrigan, J. (2009). Home Broadband Adoption 2009. Report for the
Pew Internet & American Life Project, Retrieved August 6, 2010, from
http://pewinternet.org/Reports/2009/10-Home-Broadband-Adoption2009.aspx.
[Accessed 6 August 2010]
Howard, P.N. & N. Mahazeri (2009). Telecommunications reform,
Internet use and mobile phone adoption in the developing world. World
Development, 37(7), 1159-1169.
Ida, T. & K. Sakahira (2008). Broadband migration and lock-in
effects: Mixed logit model analysis of Japan's high speed Internet
access services. Telecommunications Policy, 32, 615-625.
Ives, B., M.H. Olson & J.J. Baroudi (1983). The measurement of
user information satisfaction. Communications of the ACM, 26(10),
785-793.
Jones, S. & S. Fox (2009). Generations Online in 2009. Report
for the Pew Internet & American Life Project, January 28, 2009.
Kwon, H.S. & L. Chidambaram (2000). A test of the technology
acceptance model: The case of cellular telephone adoption. Proceedings
of HICSS-34, Hawaii, January 3-6.
Labriet-Gross, H. (2007). High speed Internet: Can broadband save
the planet?. L'Atelier North America, Retrieved August 9, 2010,
from http://www.atelier-us.com/internet-
usage/article/high-speed-internet-can-broadbandsave-the-planet.
Majumdar, S.K. (2008). Broadband adoption, jobs and wages in the US
telecommunications industry. Telecommunications Policy, 32, 587-599.
Matthews, D. & L. Schrum (2003). High-speed Internet use and
academic gratifications in the college residence. The Internet and
Higher Education, 6(2), 125-144.
Middleton, C.A. & J. Ellison (2006). All Broadband Households
Are Not the Same: Why Scope and Intensity of Use Matter. Report for
Ryerson University, Toronto.
Moore, G.C. & I. Benbasat (1991). Development of an instrument
to measure the perceptions of adopting an information technology
innovation. Information Systems Research, 2(3), 192-222.
National Broadband Task Force (2001). The New National Dream:
Networking the Nation for Broadband Access. Ottawa: Industry Canada.
Orazem, P.F. (2005). The Impact of High-Speed Internet Access on
Local Economic Growth. Research Report Prepared for University of Kansas
School of Business, Topeka, Kansas.
Perry, T.T., L.A. Perry & K. Hosack-Curlin (1998). Internet use
by university students: An interdisciplinary study on three campuses.
Internet Research, 8(2), 136-141.
Platt, R.G., W.B. Carper & M. McCool (2010). Outsourcing a high
speed Internet access project: An information technology class case
study in three parts. Journal of Information Systems Education, 21(1),
15-25.
Rains, S.A. (2008). Health at high speed: Broadband Internet
access, health communication, and the digital divide. Communication
Research, 35(3), 283-297.
Rosston, G., S.J. Savage & D.M. Waldman (2010). Household
Demand for Broadband Internet Service. Final Report for the
Broadband.gov Task Force, Federal Communications Commission (FCC).
Selouani, S.. & H. Hamam (2007). Social impact of broadband
Internet: A case study in the Shippagan area, a rural zone in Atlantic
Canada. Journal of Information, Information Technology, and
Organizations, 2, 79-94.
Straub, D.W. (1989). Validating instruments in MIS research. MIS
Quarterly, 13(2), 147-169.
Tao, C.J., S.C. Chen & L. Chang (2009). Apply 6-Sigma
methodology in measuring the competition quality of satisfaction
performance--An example of ISP industry. Quality & Quantity, 43,
677-694.
Taylor, S. & P.A. Todd (1995). Understanding information
technology usage: A test of competing models. Information Systems
Research, 6(2), 144-176.
Telecommunications Policy Review Panel (2006). Telecommunications
Policy Review Panel--Final Report 2006. Ottawa: Industry Canada.
Thompson, R.L., C.A. Higgins & J.M. Howell (1991). Personal
computing: Toward a conceptual model of utilization. MIS Quarterly,
15(1), 124-143.
Varon, E. (2010). Testing the benefits of high-speed Internet
access. Appeared on NetworkWorld, Retrieved August 9, 2010, from
http://www.netorkworld.com/news/2010/051710-
testing-the-benefits-of-high-speed.html.
Venkatesh, V. & S.A. Brown (2001). A longitudinal investigation
of personal computers in homes: Adoption determinants and emerging
challenges. MIS Quarterly, 25(1), 71-102.
Wagner, J. & S. Hanna (1983). The effectiveness of family life
cycle variables in consumer expenditure research. Journal of Consumer
Research, 10, 281-291.
Webster, J. & J.J. Martocchio (1993). Turning work into play:
Implications for microcomputer software training. Journal of Management,
19(1), 127-146.
Webster, J. & J.J. Martocchio (1992). Microcomputer
playfulness: Development of a measure with work place implications. MIS
Quarterly, 16(2), 201-226.
Windhausen Jr., J. (2008). A Blueprint for Big Broadband. Report
for EDUCAUSE, Washington D.C., Retrieved August 9, 2010, from
http://www.educause.edu.
Yoo, Y. & M. Alavi (2001). Media and group cohesion: Relative
influences on social presence, task participation, and group consensus.
MIS Quarterly, 25(3), 371-390.
Gerard Fillion, University of Moncton
Jean-Pierre Booto Ekionea, University of Moncton
Table 1: Generations Explained
(adapted from Jones & Fox, 2009, p. 1)
Generation Birth Years, % of Total % of
Names Ages in 2009 Adult Internet-Using
Population Population
Gen Y Born 1977-1990, 26 30
(millennials) Ages 18-32
Gen X Born 1965-1976, 20 23
Ages 33-44
Younger Born 1955-1964, 20 22
Boomers Ages 45-54
Older Boomers Born 1946-1954, 13 13
Ages 55-63
Silent Born 1937-1945, 9 7
Generation Ages 64-72
G.I. Born -1936, 9 4
Generation Age 73+
Source: Pew Research Center's Internet & American
Life Project December 2008 survey. N = 2,253 total
adults, and margin of error is [+ or -] 2%. N =
1,650 total Internet users, and margin of error is
[+ or -] 3%.
Table 2: Main High Speed or Broadband Technologies
(adapted from Gill, 2010, pp. 3-4)
Technologies Speeds (1)
Cable Basic: 4 Mbps to 6 Mbps
High End: 12 Mbps to 16 Mbps and faster
DSL Basic: 768 Kbps to 1.5 Mbps
High End: 3 Mbps to 7 Mbps
Fiber Optic 15 Mbps to 25 Mbps
Cable
Mobile--EDGE Up to 58 Kbps, average 22 Kbps
Mobile--3G AT&T: Download: 700 Kbps to 1.7 Mbps;
Upload: 500 Kbps to 1.2 Mbps
Sprint: Download: 600 Kbps to 1.4 Mbps
Verizon: 600 Kbps to 1.4 Mbps
Mobile--4G Download: 3 to 6 Mbps
Satellite 10 to 20 Kbps
WiMax Download: 3 to 6 Mbps
(like Clear)
South Korea 1 Gbps (2012)
Japan Average: 93.6 Mbps (2007)
France Average: 44.1 Mbps (2007)
(1) One kilobit per second (Kbps) is 1,000 bits
per second (bps). One megabit per second (Mbps) is
1,000 Kbps or 1,000,000 bps. One gigabit per
second (Gbps) is 1,000 Mbps or 1,000,000 Kbps or
1,000,000,000 bps.
Table 3: Related Literature Survey
Research Area References
High speed Internet social impact and Orazem (2005)
economic growth. Selouani & Hamam (2007)
Anderson (2008)
High speed Internet and wages and Majumdar (2008)
employment.
High speed Internet and health. Dumitru et al. (2007)
Rains (2008)
High speed Internet and regulation. Cambini & Jiang (2009)
Howard & Mahazeri (2009)
High speed Internet migration, Ida & Sakahira (2008)
implementation, and support. Platt et al. (2010)
High speed Internet adoption and use. Perry et al. (1998)
Matthews & Schrum (2003)
Middleton & Ellison (2006)
Dumitru et al. (2007)
Windhausen Jr. (2008)
Horrigan (2009)
Howard & Mahazeri (2009)
Al-Omoush & Shaqrah (2010)
Helsper (2010)
Rosston et al. (2010)
High speed Internet (e.g., ISP) user Tao et al. (2009)
satisfaction.
Table 4: Variables and Definitions
Beliefs and Variables Definitions
Characteristics
Dependent User Satisfaction According to Cyert and March
Variable (1963, p. 126), an information
system or information
technology which meets the
needs of its user will
reinforce satisfaction with
that system or technology. If
the system or technology does
not provide the needed
information or service, the
user will become dissatisfied
and look elsewhere.
Attitudinal Applications for The extent to which using high
Beliefs Personal Use speed Internet enhances the
(independent effectiveness of household
variables) activities (Venkatesh & Brown,
2001).
Utility for The extent to which using high
Children speed Internet enhances the
children's effectiveness in
their activities (Venkatesh &
Brown, 2001).
Utility for The extent to which using high
Work-Related Use speed Internet enhances the
effectiveness of performing
work-related activities
(Venkatesh & Brown, 2001).
Mobility The extent to which high speed
Internet allows using only
this technology to perform all
personal and professional
activities.
Applications for The pleasure derived from high
Fun speed Internet use (Venkatesh
& Brown, 2001). These are
specific to high speed
Internet usage, rather than
general traits (Brown &
Venkatesh, 2005; see Webster &
Martocchio, 1992, 1993).
Status Gains The increase in prestige that
coincides with the purchase of
high speed Internet access for
home use (Venkatesh & Brown,
2001).
Normative Friends and Family "The extent to which the
Beliefs Influences members of a social network
(independent influence one another's
variables) behavior" (Venkatesh & Brown,
Secondary Sources' 2001, p. 82). In this case,
Influences the members are friends and
family (Brown & Venkatesh,
2005). The extent to which
information from TV,
newspaper, and other secondary
sources influences behavior
(Venkatesh & Brown, 2001).
Workplace The extent to which coworkers
Referents' influence behavior (Brown &
Influences Venkatesh, 2005; see Taylor &
Todd, 1995).
Control Beliefs Fear of The extent to which rapidly
(independent Technological changing technology is
variables) Advances associated with fear of
obsolescence or apprehension
regarding high speed Internet
access purchase (Venkatesh &
Brown, 2001).
Declining Cost The extent to which the cost
of high speed Internet access
is decreasing in such a way
that it inhibits adoption
(Venkatesh & Brown, 2001).
Cost The extent to which the
current cost of high speed
Internet access is too high
(Venkatesh & Brown, 2001).
Perceived Ease The degree to which using high
of Use speed Internet is free from
effort (Davis, 1989; see also
Venkatesh & Brown, 2001).
Self-Efficacy The individual's belief that
(or Requisite he-she has the knowledge
Knowledge) necessary to use high speed
Internet. This is closely tied
to computer self-efficacy
(Compeau & Higgins, 1995a,
1995b; see also Venkatesh &
Brown, 2001).
Life Cycle Income The individual's year gross
Characteristics income (see Wagner & Hanna,
(moderator 1983).
variables)
Marital Status The individual's family status
(married, single, divorced,
widowed, etc.) (see Danko &
Schaninger, 1990).
Age The individual's age (see
Danko & Schaninger, 1990). In
this case, age is calculated
from the individual's birth
date.
Child's Age The age of the individual's
youngest child (see Danko &
Schaninger, 1990). In this
case, age is represented by a
numeral.
Table 5: Means, Standard Deviations, Composite
Reliability Indexes, Correlations, and Average
Variance Extracted of Variables
Variables M SD Reliability Correlations
Index and
Average
Variance
Extracted (e)
1 2
1. Applications 5.17 1.87 0.82 0.78
for Personal
Use
2. Utility fo 3.07 3.00 0.99 .13 0.98
Children
3. Utility for
Work-Related 4.47 2.63 0.89 .21 .12
Use
4. Mobility 4.70 2.63 0.89 .11 .17
5. Applications 4.83 1.95 0.86 .29 .09
For Fun
6. Status Gains 2.88 2.08 0.94 .08 .06
7. Friends and
Family 3.75 2.48 0.93 .06 .01
Influences
8. Secondary
Sources' 3.73 2.20 0.85 .12 .09
Influences
9. Workplace
Referents' 3.15 3.00 0.91 .15 .07
Influences
10. Fear of
Technological 2.87 2.07 0.87 .03 .11
Advances
11. Declining 3.93 2.07 0.86 .16 .08
Cost
12. Cost 4.77 1.93 0.74 -.22 -.07
13. Perceived
Ease 5.73 1.45 0.83 .30 -.04
of Use
14. Self-Efficacy 6.37 1.05 0.93 .29 .01
15. Income (a) NA NA NA .10 .25
16. Marital NA NA NA -.07 -.29
Status (b)
17. Age (c) 40.00 13.70 NA -.05 .08
18. Child's 14.91 8.89 NA -.06 -.20
Age (d)
19. User 5.72 1.38 0.90 .28 .06
Satisfaction
Variables Correlations
and
Average
Variance
Extracted (e)
3 4 5 6 7 8
1. Applications
for Personal
Use
2. Utility fo
Children
3. Utility for
Work-Related 0.85
Use
4. Mobility .28 0.85
5. Applications -.02 .30 0.78
For Fun
6. Status Gains .11 .17 .29 0.92
7. Friends and
Family .10 .20 .27 .36 0.87
Influences
8. Secondary
Sources' .05 .20 .37 .26 .39 0.81
Influences
9. Workplace
Referents' .29 .19 .11 .21 .42 .27
Influences
10. Fear of
Technological .09 -.04 .01 .13 .16 .11
Advances
11. Declining .20 .22 .13 .16 .14 .10
Cost
12. Cost -.19 -.11 -.11 -.12 -.15 -.12
13. Perceived
Ease .12 .25 .26 .10 .07 .17
of Use
14. Self-Efficacy .11 .24 .20 .07 .04 .11
15. Income (a) .13 .01 -.20 -.11 -.10 -.09
16. Marital -.01 .16 .13 .05 .09 .11
Status (b)
17. Age (c) -.07 -.28 -.22 -.05 -.08 -.18
18. Child's -.12 -.20 -.07 -.01 .01 -.08
Age (d)
19. User .13 .38 .39 .18 .15 .17
Satisfaction
Variables Correlations
and
Average
Variance
Extracted (e)
9 10 11 12 13 14
1. Applications
for Personal
Use
2. Utility fo
Children
3. Utility for
Work-Related
Use
4. Mobility
5. Applications
For Fun
6. Status Gains
7. Friends and
Family
Influences
8. Secondary
Sources'
Influences
9. Workplace
Referents' 0.92
Influences
10. Fear of
Technological .06 0.83
Advances
11. Declining .15 -.01 0.83
Cost
12. Cost -.16 .18 -.29 0.70
13. Perceived
Ease .14 -.26 .13 -.23 0.76
of Use
14. Self-Efficacy .08 -.25 .13 -.18 .47 0.90
15. Income (a) .04 -.12 .03 -.06 .04 .05
16. Marital .14 -.13 -.12 -.05 .12 .06
Status (b)
17. Age (c) -.30 .19 -.06 .15 -.22 -.26
18. Child's -.20 .07 -.08 .11 -.13 -.20
Age (d)
19. User .11 -.20 .20 -.32 .56 .49
Satisfaction
Variables Correlations
and
Average
Variance
Extracted (e)
15 16 17 18 19
1. Applications
for Personal
Use
2. Utility fo
Children
3. Utility for
Work-Related
Use
4. Mobility
5. Applications
For Fun
6. Status Gains
7. Friends and
Family
Influences
8. Secondary
Sources'
Influences
9. Workplace
Referents'
Influences
10. Fear of
Technological
Advances
11. Declining
Cost
12. Cost
13. Perceived
Ease
of Use
14. Self-Efficacy
15. Income (a) NA
16. Marital -.26 NA
Status (b)
17. Age (c) .28 -.27 NA
18. Child's -.02 -.05 .34 NA
Age (d)
19. User .03 .02 -.08 -.07 0.77
Satisfaction
(a) This variable was coded as an ordinal
variable. It was measured in terms of non
quantified distinct ordered categories.
(b) This variable was coded as a nominal variable.
It was measured in terms of non quantified
distinct categories.
(c) This variable was coded as a continuous
variable. It was measured using the respondents'
birth date.
(d) This variable was coded as a numeral. It was
measured using the age of the respondents'
youngest child.
(e) Boldfaced elements on the diagonal of the
correlation matrix represent the square root of
the average variance extracted (AVE).
For an adequate discriminant validity, the
elements in each row and column should be smaller
than the boldfaced element in that row or column.
NA: Not applicable.
Table 6: Summary of the Test of Hypotheses
Hypotheses Results Software
(beta sig.)
H1-Marital status and age will Not supported PLS (-0.181)
moderate the relationship
between applications for
personal use and satisfaction
of using high speed Internet
at home.
H2-Child's age will moderate Not supported PLS (0.039)
the relationship between
utility for children and
satisfaction of using high
speed Internet at home.
H3-Age will moderate the Supported PLS (0.457*)
relationship between utility
for work-related use and
satisfaction of using high
speed Internet at home.
H4-Age will moderate the Supported PLS (-0.334*)
relationship between
applications for fun and
satisfaction of using high
speed Internet at home.
H5-Age will moderate the Supported PLS (-0.339*)
relationship between status
gains and satisfaction of
using high speed Internet at
home.
H6-Age, marital status, and a- Not supported PLS (-0.122)
income will moderate the b- Not supported PLS (0.041)
relationship between the c- Not supported PLS (0.042)
normative beliefs ((a)
friends and family influences;
(b) secondary sources'
influences; and (c) workplace
referents' influences) and
satisfaction of using high
speed Internet at home.
H7-Age and income will a- Not supported PLS (-0.109)
moderate the relationship b- Not supported PLS (0.068)
between the external control c- Not supported PLS (0.003)
beliefs ((a) fear of
technological advances; (b)
declining cost; and (c) cost)
and satisfaction of using high
speed Internet at home.
H8-Age will moderate the a- Not supported PLS (-0.324)
relationship between the b- Supported PLS (-1.286**)
internal control beliefs ((a)
perceived ease of use; and (b)
self-efficacy) and
satisfaction of using high
speed Internet at home.
* p < 0.05; ** p < 0.01.
Table 7: Beta Path Coefficients, T-Values, and
Percentages of Variance Explained of Variables
Variables Beta t-values [R.sup.2]
Coefficients (one-tail)
Applications for Personal 0.032 0.566 0.001
Use
Utility for Children -0.031 0.729 0.001
Utility for Work-Related -0.016 0.317 0.000
Use
Mobility 0 238 **** 5.177 0.037
Applications for Fun 0.218 **** 3.504 0.047
Status Gains 0.036 0.823 0.004
Friends and Family 0.005 0.117 0.004
Influences
Secondary Sources' -0.017 0.380 0.005
Influences
Workplace Referents' 0.028 0.588 0.002
Influences
Fear of Technological -0.101* 1.908 0.003
Advances
Declining Cost 0.012 0.258 0.005
Cost -0 177 **** 3.839 0.038
Perceived Ease of Use 0.288 **** 2.800 0.081
Self-Efficacy 0.163 * 1.646 0.249
Income -0.041 0.819 0.006
Marital Status -0.016 0.300 0.000
Age 0 245 **** 3.009 0.004
Child's Age -0.066 0.903 0.005
* p < 0.05; **** p < 0.001.
