research papers IN statistics analysis -SEM

Information technology (IT) in Saudi Arabia:

Culture and the acceptance and use of IT

Said S. Al-Gahtani a , Geoffrey S. Hubona

b,*, Jijie Wang b

a King Khalid University, Abha, Saudi Arabia

b Department of Computer Information Systems, J. Mack Robinson College of Business Administration,

Georgia State University, Atlanta, GA 30303, USA

Received 6 September 2006; received in revised form 10 May 2007; accepted 13 September 2007

Available online 26 October 2007

www.elsevier.com/locate/im

Available online at www.sciencedirect.com

Information & Management 44 (2007) 681–691

Abstract

The unified theory of acceptance and use of technology (UTAUT), a model of the user acceptance of IT, synthesizes elements

from several prevailing user acceptance models. It has been credited with explaining a larger proportion of the variance of ‘intention

to use’ and ‘usage behavior’ than do preceding models. However, it has not been validated in non-Western cultures. Using a survey

sample collected from 722 knowledge workers using desktop computer applications on a voluntary basis in Saudi Arabia, we

examined the relative power of a modified version of UTAUT in determining ‘intention to use’ and ‘usage behavior’. We found that

the model explained 39.1% of intention to use variance, and 42.1% of usage variance. In addition, drawing on the theory of cultural

dimensions, we hypothesized and tested the similarities and differences between the North American and Saudi validations of

UTAUT in terms of cultural differences that affected the organizational acceptance of IT in the two societies.

# 2007 Elsevier B.V. All rights reserved.

Keywords: Unified theory of acceptance and use of technology (UTAUT); Technology acceptance; IT adoption; Cultural differences; Technology

social factors; Saudi Arabia

1. Introduction

In mainstream MIS research, there are many studies

that have investigated user acceptance and usage of new

IT. Of these many have used TAM [9,10] or made

changes to it [4,13]. Other models, such as the theory of

planned behavior (TPB) [1,18], and social cognitive

theory (SCT) [8] are also well known.

Venkatesh and Davis [26] introduced an extension to

TAM, TAM2, which examined the influences of select

antecedent social influence and cognitive instrumental

constructs on perceived usefulness and usage intentions.

* Corresponding author. Tel.: +1 404 413 7360.

E-mail address: [email protected] (G.S. Hubona).

0378-7206/$ – see front matter # 2007 Elsevier B.V. All rights reserved.

doi:10.1016/j.im.2007.09.002

Subsequently, they [27] synthesized the various models

into the unified theory of acceptance and use of

technology (UTAUT).

However, it has been exclusively validated in the

North American contexts. Clearly, in contexts removed

from Western nations, the impact of subjective norms

on the individual and organizational acceptance of IT

could vary markedly. Accordingly, the objectives of our

research was to: (1) empirically validate a modified

UTAUT model in a non-Western cultural context,

specifically Saudi Arabia and (2) explain anomalies

between these validations in terms of cultural differ-

ences that affect the organizational acceptance of IT. To

achieve this second objective, we draw from research on

cultural dimensions [14,15] that explain international

differences in work-related values.

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691682

2. Theory and background

2.1. Technology acceptance

TAM postulated that two belief constructs, perceived

usefulness and perceived ease of use, accounted for a

large proportion of the variance in behavioral intentions

and voluntary usage behaviors of new ITs. Empirical

validations of TAM have typically accounted for

between 15% and 45% of the variance in the ‘intention

to use’ and self-reported ‘usage’.

An extension examined the impact of social

influence and cognitive instrumental processes as

predictors of the perceived usefulness and behavioral

intention to use. The subjective norm, originally

developed from the theory of reasoned action (TRA)

predicts general behavioral intentions in strictly

voluntary contexts; it assumes that people’s perception

that others important to them think a behavior should or

should not be followed has an impact on their actions.

Additional models, including the theory of planned

behavior (TPB) [11] and the decomposed theory of

planned behavior (DTPB) [24], lead to the combined

model as illustrated in the UTAUT model of Fig. 1.

UTAUT postulates that four constructs act as

determinants of behavioral intentions and usage

behavior:

1. P

erformance expectancy: ‘‘The degree to which an

individual believes that using the system will help

him or her attain gains in job performance.’’

2. E

ffort expectancy: ‘‘The degree of ease associated

with the use of the system.’’

3. S

ocial influence: ‘‘The degree to which an individual

perceives that important others believe he or she

should use the new system. Social influence is

system- or application-specific, whereas subjective

norm relates to non-system-specific behavior.’’

4. F

acilitating conditions: ‘‘The degree to which an

individual believes that an organizational and

technical infrastructure exists to support use of the

system.’’

In addition, UTAUT also posits the role of four key

moderator variables: gender, age, experience, and

voluntariness of use.

2.2. Technology acceptance and culture

The globalization of business has highlighted the

need to understand the effectiveness of IS that span

different cultures. Multinational and trans-cultural

organizations use IT to achieve economies of scale,

coordinate operations, and facilitate collaborative work

across locations and cultures. Cultural differences have

become an important issue in the evaluation of

computer applications.

To make valid comparisons, models should be robust

across cultures. Therefore, determining whether similar

models are comparable across cultures is a first step

needed to: (1) enhance understanding of cultural effects

of IT acceptance and (2) improve the organizational

management of IT globally.

Rose and Straub [20] conducted a study of IT

adoption and use in the Arab world. Using a cross-

sectional survey of 274 knowledge workers in five Arab

nations (Egypt, Jordan, Saudi Arabia, Lebanon, and the

Sudan), they applied a modified TAM to assess the

diffusion of personal computing. Their model explained

40% of the variance of PC use in these nations.

Subsequently, Straub et al. [23] developed a cultural

influence model and suggested that Arab cultural beliefs

were a strong predictor of resistance to IT transfer. Loch

et al. [17] applied this model to examine culture-specific

enablers and impediments to the adoption and use of the

Internet in the Arab world. They showed that both social

norms and the degree of technological culturation can

impact the individual and organizational acceptance

and use of the Internet.

Different approaches have been used to study the

organizational effects of culture; one is to apply a

quantitative methodology to identify and measure

national cultural dimensions. Such studies include

Tiandis’s [25] and Hofstede’s national cultural dimen-

sions, and social identity theory [21]. Of these,

Hofstede’s are most commonly used. These facilitate

national-level analyses and allow multiple country

comparisons. Furthermore, Hofstede’s cultural dimen-

sions have been used to explore the impact of cultural

differences on technology acceptance [22]. We also

drew on Hofstede’s dimensions to describe select

cultural differences between Saudi Arabia and North

American Nations and to discuss cultural implications

of IT user acceptance. Hofstede’s dimensions are shown

in Table 1, which briefly describes his five cultural

dimensions.

Table 2 shows country scores of these dimensions for

the United States and Saudi Arabia. Thus, Saudi Arabia

ranks much higher than the US in uncertainty avoidance

and power distance; approximately the same in

masculinity; and much lower in individualism.

High uncertainty avoidance deals with tolerance for

uncertainty and ambiguity. It indicates to what extent a

person feels uncomfortable in unstructured situations.

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691 683

Fig. 1. Unified theory of acceptance and use of technology.

High power distance refers to the propensity to defer to

authority, and conform to the expectations of others in

superior social roles. Masculinity focuses on the extent

to which a society stresses achievement versus caring

and nurturing behaviors, and low individualism refers to

the extent to which individuals are integrated into

cohesive in-groups and value the protection it provides.

2.3. Research model and hypotheses

The model for our study is presented as Fig. 2. It was

derived from UTAUT, but was modified. First, as we

examined the factors that promoted the use of

computers only on a voluntary basis, we eliminated

voluntary use as a moderating construct. And second,

we substituted subjective norm for social influence; our

Table 1

Measures of cultural dimensions

Hofstede’s dimension Definition

Uncertainty avoidance (UA) Focuses on the level of tolerance for u

a structured, rule-oriented society that

amount of uncertainty

Power distance (PD) Focuses on the degree of equality, or i

indicates that inequalities of power and

Masculinity (MAS) Masculinity measures the degree to wh

and competition prevail over ‘‘feminin

relationships, service, caring, and solid

Individualism (IDV) Focuses on the degree the society rein

relationships. Low IDV typifies societi

These cultures reinforce collectives wh

Long-term orientation or

confucian dynamism (LTO)

Cultures typified by a long-term orient

and thrift, while a short-term orientatio

in particular, the respect for tradition,

target behavior related to the use of desktop computers

in general, and not to any application or system. Two of

our four usage measures explicitly related to computer

usage in many different computer applications and with

respect to performing a variety of tasks. Thus, we used

the more general subjective norm construct.

We analyzed all of the path linkages in our research

model (Fig. 2), including all direct and moderating (or

interacting) effects, forming specific hypotheses for

every path except for that from behavioral intention to

use behavior. The theoretical basis for the research

model is UTAUT, which is assumed to justify the

indicated path linkages. However, we assumed a

number of specific effects inherent that we expected

to arise from cultural differences. Since the score for

long-term orientation dimension (sometimes termed a

ncertainty and ambiguity within the society. High UA indicates

institutes rules, regulations, and controls in order to reduce the

nequality, between people in the country’s society. High PD

wealth are accepted practices and have been allowed to grow

ich ‘‘masculine’’ values like assertiveness, performance, success

e’’ values like the quality of life, maintaining warm personal

arity

forces individual or collective achievement and interpersonal

es of a more collectivist nature with close ties between individuals.

ere everyone takes responsibility for fellow members of their group

ation are oriented towards future rewards, in particular perseverance

n is characterized by values relating to both the past and present,

preservation of ‘‘face’’ and the fulfillment of social obligations

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691684

Table 2

Hofstede country scores for the USA and Saudi Arabia

Cultural dimension United States Saudi Arabia

Uncertainty avoidance 46 68

Power distance 40 80

Masculinity 62 52

Individualism 91 38

Long-term orientation 29 N/A

Confucian effect) is not available for Saudi Arabia, we

only drew on the remaining four dimensions.

There are rigid boundaries in social roles and

expectations for women compared to men in Saudi

Arabia and thus there are far fewer women in professional

knowledge worker roles. Accordingly, we expected that

women in Saudi Arabia would be less inclined than men

to expect that the use of computers would enhance their

job performances and thus advance their professional

careers. Additionally, the majority of the Saudi work-

force is young, under the age of 40 [2], and well educated.

Consequently, we hypothesized:

� H

1: Performance expectancy will have a positive influence on behavioral intentions to use computers.

� H

1a: Gender will positively moderate the influence of performance expectancy on behavioral intentions to

use computers for men.

� H

1b: Age will not moderate the influence of performance expectancy on behavioral intentions to

use computers.

Again, we expected a positive influence of effort

expectancy on behavioral intentions to use computers.

There is no reason to suspect that the effect of the ‘‘degree

Fig. 2. Researc

of ease associated with the use of the system’’ should be

influenced by Hofstede’s cultural measures. In Saudi

Arabia, increased levels of ease of using computers

should be associated with increased behavioral intentions

to use them. Furthermore, we hypothesized that men

would be more inclined than women to associate an

increased ease of use with increased intentions to use

computers. Furthermore, we expected no interaction of

age with this relationship. However, more experienced

users would tend to be less influenced by the ease of using

computers. Thus we hypothesized:

� H

h m

2: Effort expectancy will have a positive influence on behavioral intentions to use computers.

� H

2a: Gender will positively moderate the influence of effort expectancy on behavioral intentions to use

computers for men.

� H

2b: Age will not moderate the influence of effort expectancy on behavioral intentions to use computers.

� H

2c: Experience will negatively moderate the influence of effort expectancy on behavioral inten-

tions to use computers for men.

In cultures characterized by high power distance,

individuals will acquiesce to the expectations of others

who are seen as important or influential. Consequently,

in an Arabic culture, employees should exhibit a

stronger association between social influence variables

and behavioral intention, than, for example, in the US.

Furthermore, the low individualism score for Saudi

Arabia is characteristic of a culture that values

collective achievements and interpersonal relationships.

A high regard for groups suggests that the opinions of

others would impact an individual’s behavioral inten-

odel.

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691 685

tions. Consequently, the collective opinions of others

would strongly influence individual behavioral inten-

tions and result in a positive relationship between

subjective norm and behavioral influence.

In addition, younger people tend to occupy

subordinate roles and thus are likely to be influenced

by the subjective norm in a culture characterized by

high power distance. Consequently, we hypothesized:

� H

3: Subjective norm will have a positive influence on behavioral intentions to use computers.

� H

3a: Gender will positively moderate the influence of subjective norm on behavioral intentions to use

computers for men.

� H

3b: Age will negatively moderate the influence of subjective norm on behavioral intentions to use

computers.

� H

Table 3

Saudi predictor latent construct items

Performance expectancy (PE)

PE1: I find computers useful in my job

PE2: Using computers in my job enables me to accomplish tasks

more quickly

PE3: Using computers in my job increases my productivity

PE4: Using computers enhances my effectiveness on the job

Effort expectancy (EE)

EE1: My interactions with computers are clear and

3c: Experience will negatively moderate the influence of subjective norm on behavioral intentions

to use computers.

In terms of UTAUT, we argued that the relationships

between facilitating conditions and use behavior should

be strong for cultures that score high on uncertainty

avoidance. We reasoned that increasing levels of

facilitating conditions should serve to reduce uncom-

forting levels of uncertainty or ambiguity with

computers. Therefore, we expected that a direct

relationship between facilitating conditions and use

behavior would hold true in Saudi Arabia. Additionally,

we reasoned that age and experience should negatively

interact with the influence of facilitating conditions on

computer usage. Specifically, we speculated that

increasing levels of age and experience would mute

the dependence on a facilitating infrastructure to utilize

computers. Consequently, we hypothesized:

understandable

EE2: It is easy for me to become skillful using computers

EE3: I find computers easy to use

� H

EE4: Learning to use computers is easy for me

4: Facilitating conditions will have a positive influence on computer usage behavior.

� H

Subjective norm (SN) SN1: Most people who are important to me think I should use

4a: Age will negatively moderate the influence of facilitating conditions on computer usage behavior.

computers

� H

SN2: Most people who are important to me would want me to use

computers

SN3: People whose opinions I value would prefer me to use

computers

Facilitating conditions (FC)

FC1: I have the resources and the knowledge and the ability to make

use of the computer

FC2: A central support was available to help with computer

problems

FC3: Management provided most of the necessary help and

resources for computing

4b: Experience will negatively moderate the influence of facilitating conditions on computer usage

behavior.

3. Method

The data used in our study were part of the material

collected in a project financed by the Saudi government

to build a comprehensive model of the antecedents and

of the mediating, moderating, and outcome factors, that

affect the acceptance and use of computers by

knowledge workers in Saudi Arabia. A list of the

major companies in the four main provinces of Saudi

Arabia was compiled with the assistance of the

chambers of commerce in each region. The general

managers of these organizations were asked to allow

participation of knowledge workers from their organi-

zation. Those that agreed to participate were asked to

nominate a contact person to help the researchers

distribute and collect the survey instruments. The

organizations participating included banking, merchan-

dising, manufacturing, and petroleum industries.

A total of 1190 usable survey responses were

collected. Of these, 468 responders indicated that their

use of computers was mandatory. The remaining 722

survey responders indicated volitional use of compu-

ters. These responses constituted our survey sample.

All survey items, originally published in English, as

discussed later, were converted into Arabic using

Brislin’s [3] back translation method. The items were

translated between English and Arabic by several

bilingual professors and repeated until both versions

converged.

Table 3 shows the items used to estimate the Saudi

predictor latent constructs. A seven point Likert scale

with anchors of strongly disagree to strongly agree was

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691686

Table 5

Saudi predicted latent construct items

Behavioral intention (BI)

BI1: I predict I will continue to use computers on a regular

basis (seven-point Likert scale anchored with strongly

disagree to strongly agree)

BI2: What are the chances in 100 that you will continue

as a computer user? (1) Zero; (2) 1–10%; (3) 11–30%;

(4) 31–50%; (5) 51–70%; (6) 71–90%;

or (7) more than 90%

BI3: To do my work, I would use computers rather than any

other means available (seven-point Likert scale anchored

with strongly disagree to strongly agree)

Use behavior (USE)

USE1: On an average working day, how much time do you

spend using computers? (1) Almost never; (2) less than

30 min; (3) from 30 min to 1 h; (4) from 1 to 2 h;

(5) from 2 to 3 h; and (6) more than 3 h

USE2: On average, how frequently do you use computers?

(1) Less than once a month; (2) once a month; (3) a few

times a month; (4) a few times a week; (5) about once a day;

and (6) several times a day

USE3: How many different computer applications have you

worked with or used in your job? (1) None; (2) one;

(3) two; (4) three to five applications; (5) six to ten

applications; and (6) more than 10 applications

USE4: According to your job requirements, please indicate

each task you use computers to perform (count of all that

apply)? (1) Letters and memos; (2) producing reports;

(3) data storage and retrieval; (4) making decisions;

(5) analyzing trends; (6) planning and forecasting;

(7) analyzing problems and alternatives; (8) budgeting;

(9) controlling and guiding activities; (10) electronic

communications with others; and (11) others (please indicate)

used to measure each item. The performance expec-

tancy and effort expectancy constructs were items used

by Venkatesh et al. with the references to ‘‘the system’’

changed to ‘‘computers.’’ The subjective norm items

originated from TRA/TPB. Facilitating conditions,

defined by Venkatesh et al. as: ‘‘the degree to which

an individual believes that an organizational and

technical infrastructure exists to support use of the

system,’’ included references to having resources,

knowledge, and technical and management support to

use computers.

Table 4 gives the Saudi items used to measure the

moderator variables. Similar to Venkatesh et al., we

used a binary dummy variable, 0 for female, and 1 for

male, to indicate gender. There was an imbalance of

male gender representation (82% of responses), but this

was inevitable due to the cultural preponderance of

working males in Saudi Arabia. As indicated, we

measured age using five ordinal categories in response

to the question: ‘‘For how many years have you been

using computers?’’

Table 5 indicates the items used to estimate the Saudi

predicted latent constructs, behavioral intention and use

behavior. For behavioral intention, we chose items that

reflected an individual’s self-assessment of his (or her)

likelihood to continue to use computers for an indefinite

period. Our measure differs from other measures in the

literature that have used time-specific measures. We

wanted to capture a self-assessment of likely continuing

computer usage.

We used a multi-item, self-reported latent usage

construct to provide four dimensions of computer

usage: (1) amount of time spent using computers

per day; (2) frequency of using computers; (3)

number of different software applications used; and

(4) number of different business tasks supported

through computer use. These four items had been

used in a study [16] investigating the acceptance of

desktop computing by 358 users in small private firms

in New Zealand.

Table 4

Saudi moderating (interacting) variables

Gender Male (82%) or female (18%)

Age (1) Less than 20 years (1%); (2) 20–30 years (35%);

(3) 31–40 years (43%); (4) 41–50 years (19%);

and (5) above 50 years (2%)

Experience For how many years have you been using computers?

(1) Less than a year (8%); (2) 1–3 years (25%);

(3) 4–7 years (30%); (4) 8–10 years (13%);

(5) more than 10 years (24%)

4. Results

The research model of Fig. 2 was analyzed using

PLS-Graph (build 1126), a PLS structural equation

modeling tool [5]. It assesses the psychometric

properties of the measurement model, and estimates

the parameters of the structural model. This tool enables

the simultaneous analysis of up to 200 indicator

variables, allowing the examination of extensive

interactions among moderator and latent predictor

variable indicators.

4.1. The measurement model

Reliability results are given in Table 6. The data

indicates that the measures are robust in terms of their

internal consistency reliability as indexed by the

composite reliability. The composite reliabilities of

the different measures range from 0.76 to 0.95, which

exceed the recommended threshold value of 0.70 [19].

In addition, consistent with the guidelines of Fornell and

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691 687

Table 6

Assessment of the measurement model

Variable constructs The composite

reliability (internal

consistency reliability)

Average variance

extracted/explained

Performance

expectancy

0.90 0.70

Effort expectancy 0.90 0.70

Subjective norm 0.95 0.87

Facilitating

conditions

0.77 0.53

Behavioral

intention

0.76 0.52

Use behavior 0.85 0.58

Larcker [12], the average variance extracted (AVE) for

each measure exceeded 0.50. Table 7 reports the results

of testing the discriminant validity of the measure

scales. The elements in the matrix diagonals, represent-

Table 7

Discriminant validity (intercorrelations) of variable constructs

Latent variables 1 2

1. Performance expectancy 0.84 2. Effort expectancy 0.43 0.84 3. Subjective norm 0.32 0.24

4. Facilitating conditions 0.32 0.48

5. Behavioral intention 0.43 0.50

6. Use behavior 0.20 0.29

Table 8

Factor loadings (bolded) and cross loadings

Performance

expectancy

Effort

expectancy

Subjectiv

norm

PE1 0.79 0.39 0.24 PE2 0.88 0.32 0.28 PE3 0.89 0.37 0.31 PE4 0.77 0.36 0.25 EE1 0.34 0.84 0.17 EE2 0.39 0.82 0.22 EE3 0.33 0.83 0.22 EE4 0.37 0.85 0.19 SN1 0.30 0.26 0.94 SN2 0.29 0.19 0.95 SN3 0.31 0.22 0.92 FC1 0.29 0.44 0.19

FC2 0.20 0.19 0.13

FC3 0.19 0.20 0.19

BI1 0.32 0.34 0.34

BI2 0.17 0.28 0.12

BI3 0.45 0.46 0.34

USE1 0.13 0.23 0.07

USE2 0.15 0.22 0.06

USE3 0.15 0.26 0.00

USE4 0.19 0.20 �0.02

ing the square roots of the AVEs, are greater in all cases

than the off-diagonal elements in their corresponding

row and column, supporting the discriminant validity of

our scales.

We tested convergent validity using PLS-Graph by

extracting the factor and cross loadings of all indicator

items to their respective latent constructs. These results,

presented in Table 8, indicated that all items loaded: on

their respective construct from a lower bound of 0.70 to

an upper bound of 0.95; and more highly on their

respective construct than on any other. Furthermore,

each item’s factor loading on its respective construct

was highly significant ( p < 0.0001) as indicated by the T-statistics of the outer model loadings in the PLS-

Graph output. These values ranged from a low of 16 to a

high value of 121. The constructs’ items’ loadings and

cross loadings presented in Table 8, and the highly

significant T-statistic for each individual item loading

3 4 5 6

0.93 0.24 0.73 0.36 0.47 0.72 0.04 0.38 0.47 0.76

e Facilitating

conditions

Behavioral

intention

Use

behavior

0.25 0.32 0.14

0.25 0.37 0.15

0.28 0.39 0.18

0.28 0.37 0.20

0.40 0.43 0.18

0.44 0.47 0.34

0.37 0.39 0.25

0.39 0.39 0.21

0.24 0.34 0.05

0.24 0.32 0.05

0.18 0.36 0.01

0.74 0.45 0.34 0.72 0.27 0.21 0.72 0.24 0.23 0.34 0.73 0.23 0.31 0.70 0.45 0.37 0.72 0.22 0.27 0.41 0.79 0.26 0.35 0.76 0.31 0.36 0.80 0.31 0.32 0.71

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691688

Fig. 3. Structural model results (without interacting variables).

both confirm the convergent validity of these indicators

as representing distinct latent constructs.

4.2. The structural model

Fig. 3 shows the structural model results omitting the

influence of the interacting moderator variables. All

beta path coefficients are positive (i.e. in the expected

direction) and statistically significant (at p < 0.05).

Fig. 4. Structural m

To model the interaction effects, we conformed to

Chin et al. [6,7]. Interaction terms were formulated by

multiplying the corresponding indicators of the predictor

and moderator constructs. Furthermore, we followed the

hierarchical process that they recommended to construct

and compare models with and without the respective

interacting constructs. Fig. 4 shows the results of the

structural model with interaction effects. It presents the

results of the structural model with moderator variables.

odel results.

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691 689

For purposes of clarity, only statistically significant

moderator variables (e.g. age and experience) were

included. The beta values of all path coefficients are also

shown. Performance expectancy had a positive influence

(beta = 0.17, p < 0.001) on intention. Effort expectancy had a non-significant (beta = 0.12) influence on inten-

tion. Subjective norm had a positive influence

(beta = 0.45, p < 0.001) on intention. The weak influ- ence of facilitating conditions on use (beta = �0.03) was not statistically significant. Behavioral intention had a

positive influence (beta = 0.31, p < 0.001) on use. For the moderator (interacting) variables, statisti-

cally significant beta path coefficients were indicated.

Surprisingly, gender did not exhibit significant inter-

actions with any predictor latent variables. Age had a

negative (beta = �0.09, p < 0.05) interacting effect with subjective norm upon behavioral intention. Age

also exhibited a negative (beta = �0.09, p < 0.05) interacting effect with facilitating conditions on use.

Experience exhibited three interacting effects: a

negative (beta = �0.56, p < 0.001) interacting effect

Table 9

Hypotheses conclusions

Hypotheses Finding

H1: Performance expectancy will have a positive

influence on behavioral intentions to use computers

Yes: (beta

H1a: Gender will positively moderate the influence

of performance expectancy on behavioral intentions

to use computers for men

No: not s

H1b: Age will not moderate the influence of performance

expectancy on behavioral intentions to use computers

Yes: not s

H2: Effort expectancy will have a positive influence on

behavioral intentions to use computers

No: (beta

H2a: Gender will positively moderate the influence of

effort expectancy on behavioral intentions to use

computers for men

No: not s

H2b: Age will not moderate the influence of effort

expectancy on behavioral intentions to use computers

Yes: not s

H2c: Experience will negatively moderate the influence of

effort expectancy on behavioral intentions to use computers

Yes: (beta

H3: Subjective norm will have a positive influence on

behavioral intentions to use computers

Yes: (beta

H3a: Gender will positively moderate the influence of

subjective norm on behavioral intentions to use computers

for men

No: not s

H3b: Age will negatively moderate the influence of

subjective norm on behavioral intentions to use computers

Yes: (beta

H3c: Experience will negatively moderate the influence of

subjective norm on behavioral intentions to use computers

Yes: (beta

H4: Facilitating conditions will have a positive influence on

computer usage behavior

No:(beta =

H4a: Age will negatively moderate the influence of

facilitating conditions on computer usage behavior

Yes: (beta

H4b: Experience will negatively moderate the influence

of facilitating conditions on computer usage behavior

No: (beta

with effort expectancy on behavioral intention; a

negative (beta = �0.30, p < 0.05) interacting effect with subjective norm on behavioral intention; and a

strongly positive (beta = 0.53, p < 0.001) interacting effect with facilitating conditions on use.

It is important to note that the strength and direction

(i.e. positive or negative) of main path coefficients cannot

be adequately interpreted without also considering the

influences of interacting variables. However, as a basis of

comparison, the (direct only) model explains 35.3% of

the variance in behavioral intention and 25.1% of the

variance in use behavior. In contrast, by including the

effects of the interacting variables, a larger proportion of

the respective variances in behavioral intention

(R 2

= 0.391) and use (R 2

= 0.421) are accounted for.

5. Discussion

Table 9 presents the hypotheses and outcomes. The

‘CONCLUSION’ column indicates whether that

hypothesis was: (1) supported; (2) refuted; or (3) not

Conclusion Venkatesh finding

= 0.17, p < 0.001) Supported Yes: (beta = 0.18, p < 0.05)

ignificant Not supported No: not significant

ignificant Supported Yes: not significant

= 0.12, n.s.) Not supported No: (beta = 0.04, n.s.)

ignificant Not supported No: not significant

ignificant Supported Yes: not significant

= �0.56, p < 0.001) Supported No: (beta = 0.02, n.s.)

= 0.45, p < 0.001) Supported No: (beta = 0.02, n.s.)

ignificant Not supported No: not significant

= �0.09, p < 0.05) Supported No: (beta = 0.02, n.s.)

= �0.30, p < 0.05) Supported No: (beta = 0.04, n.s.)

�0.03, n.s.) Not supported No: (beta = 0.11, n.s.)

= �0.09, p < 0.05) Supported No: (beta = 0.02, n.s.)

= 0.53, p < 0.001) Refuted No: (beta = 0.00, n.s.)

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691690

supported. The ‘VENKATESH FINDING’ column

indicates the corresponding relationship presented in

the UTAUT study.

5.1. Findings

As suggested by Venkatesh et al., we found that

performance expectancy had a positive effect on

intention, but we found no interacting effect with

performance expectancy and either gender or age on

intention.

Also we found that effort expectancy did not have a

significant effect on intention in the presence of

interactions with the moderating variables. The negative

interaction between effort expectancy and experience

on intention indicated that, with increased years of

experience with computers, ease of use becomes less

important in predicting Saudi’s behavioral intentions.

In cultures characterized by a high power distance

dimension, we argued that individuals would be more

inclined to show deference to authority and conform to

the expectations of others in important or superior roles.

Consequently, we expected that higher power distance

cultures would exhibit a stronger association between

subjective norm and behavioral intention. We also

argued that the low individualism country score for

Saudi Arabia might indicate a strong relationship

between subjective norms and behavioral intentions in

the Arab world. In Fig. 4, subjective norm positively

influences intention, but negatively interacts with

increasing levels of age and experience on intention.

These results indicate that, among Saudi users,

subjective norm positively influences intention, but,

as expected, this influence is diminished by both

increasing age, and increasing years of experience using

computers.

The weak negative effect of facilitating conditions

on use was not significant in the presence of: the

negative interacting effect of increasing age with

facilitating conditions on use; and the strong positive

interacting effect of increasing experience with facil-

itating conditions on use.

5.2. Limitations

Our study was not longitudinal in design, and did not

target intentions and behaviors with respect to applica-

tion software use. Further, we substituted subjective

norm for social influence. We also hypothesized cultural

affects on the basis of Hofstede’s cultural dimensions.

Although this approach has been proven in general, it

would be more informative if measures of cultural-

specific work values were collected at the individual

level, commensurate with the survey data.

References

[1] I. Ajzen, The theory of planned behavior, Organizational Beha-

vior and Human Decision Processes 50, 1991, pp. 179–211.

[2] S.S. Al-Gahtani, Computer technology acceptance success fac-

tors in Saudi Arabia: an exploratory study, Journal of Global

Information Technology Management 7 (1), 2004, pp. 5–29.

[3] R. Brislin, The wording and translation of research instruments,

in: W. Lonner, J. Berry (Eds.), Field Methods in Cross-Cultural

Research, Sage, Beverly Hills, 1986.

[4] A. Burton-Jones, G.S. Hubona, The mediation of external vari-

ables in the technology acceptance model, Information & Man-

agement 43 (6), 2006, pp. 706–717.

[5] W.W. Chin, PLS-Graph User’s Guide, Version 3.0, 2001.

[6] W.W. Chin, B.L. Marcolin, P.R. Newsted, A partial least squares

latent variable modeling approach for measuring interaction

effects: results from a Monte Carlo simulation study and voice

mail emotion/adoption study, in: J.I. DeGross, A. Srinivasan, S.

Jarvenpaa (Eds.), in: Proceedings of the International Confer-

ence on Information Systems, Cleveland, OH, 1996, pp. 21–41.

[7] W.W. Chin, B.L. Marcolin, P.R. Newsted, 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), 2003, pp. 189–217.

[8] D. Compeau, C.A. Higgins, S. Huff, Social cognitive theory and

individual reactions to computing technology: a longitudinal

study, MIS Quarterly 23 (2), 1999, pp. 145–158.

[9] F. Davis, Perceived usefulness, perceived ease of use, and end

user acceptance of information technology, MIS Quarterly 13

(3), 1989, pp. 318–339.

[10] F.D. Davis, R.P. Bagozzi, P.R. Warshaw, User acceptance of

computer technology: a comparison of two theoretical models,

Management Science 35, 1989, pp. 982–1003.

[11] M. Fishbein, I. Ajzen, Belief, Attitude, Intention and Behavior:

An Introduction to Theory and Research, Addison-Wesley,

Reading, MA, 1975.

[12] C. Fornell, D. Larcker, Evaluating structural equation models

with unobservable variables and measurement error, Journal of

Marketing Research 18, 1981, pp. 39–50.

[13] B. Hasan, Delineating the effects of general and system-specific

computer self-efficacy beliefs on IS acceptance, Information &

Management 43 (5), 2006, pp. 565–571.

[14] G. Hofstede, Culture’s Consequences: International Differences

in Work-Related Values, Sage, Beverly Hills, CA, 1980.

[15] G. Hofstede, Culture’s Consequences: Comparing Values, Beha-

viors, Institutions and Organizations Across Nations, Sage,

Newbury Park, CA, 2001.

[16] M. Igbaria, N. Zinatelli, P. Cragg, A.L.M. Cavaye, Personal

computing acceptance factors in small firms: a structural equa-

tion model, MIS Quarterly 21 (3), 1997, pp. 279–305.

[17] K. Loch, D. Straub, S. Kamel, Diffusing the Internet in the Arab

world: the role of social norms and technological culturation,

IEEE Transactions on Engineering Management 50, 2003, pp.

45–63.

[18] K. Mathiesen, Predicting user intentions: comparing the tech-

nology acceptance model with the theory of planned behavior,

Information Systems Research 2, 1991, pp. 173–191.

S.S. Al-Gahtani et al. / Information & Management 44 (2007) 681–691 691

[19] J.C. Nunnally, Psychometric Theory, McGraw Hill, New York,

1978.

[20] G. Rose, D. Straub, Predicting general IT use: applying TAM to

the Arab world, Journal of Global Information Management 6,

1998, pp. 39–46.

[21] D. Straub, Toward a theory-based measurement of culture,

Journal of Global Information Management 10 (1), 2002, pp.

24–32.

[22] D. Straub, M. Keil, W. Brenner, Testing the technology accep-

tance model across cultures: a three country study, Information

& Management 33, 1997, pp. 1–11.

[23] D.W. Straub, K. Loch, C. Hill, Transfer of information technol-

ogy to the Arab world: a test of cultural influence modeling,

Journal of Global Information Management 9, 2001, pp. 6–28.

[24] S. Taylor, P.A. Todd, Understanding information technology

usage: a test of competing models, Information Systems

Research 6 (2), 1995, pp. 144–176.

[25] H. Triandis, Dimensions of cultural variation as parameters for

organizational theories, International Studies of Management

and Organization 12 (4), 1982, pp. 139–159.

[26] V. Venkatesh, F.D. Davis, A theoretical extension of the tech-

nology acceptance model: four longitudinal field studies, Man-

agement Science 46 (2), 2000, pp. 186–204.

[27] V. Venkatesh, M.G. Morris, G.B. Davis, F.D. Davis, User

acceptance of information technology: toward a unified view,

MIS Quarterly 27 (3), 2003, pp. 425–478.

Said S. Al-Gahtani is an associate profes-

sor of computer information systems in the

Department of Administrative Sciences at

King Khalid University, Abha, Saudi Ara-

bia. He has a BSc in systems engineering

from King Fahad University of Petroleum

& Minerals, MSc in computer sciences

from Atlanta University, Atlanta, Georgia,

and a PhD in computer-based information

systems from Lougborough University,

Loughborough, UK. His research interests include the user acceptance

of information technologies, the modeling of IT acceptance, end-user

computing, and organizational cross-cultural research. He has pub-

lished journal research articles in Information Technology & People,

Journal of Global Information Technology Management, Information

Technology for Development, Information Resources Management

Journal, and the Behaviour & Information Technology.

Geoffrey S. Hubona is an associate pro-

fessor of computer information systems in

the J. Mack Robinson College of Business

at Georgia State University in Atlanta, GA.

He has a BA in psychology from the Uni-

versity of Virginia, an MBA from George

Mason University, and an MA in economics

and a PhD in MIS from the University of

South Florida. His research interests

include the user acceptance of information

technologies, the human perception of computer visualizations, and

technology usability issues. He has published journal research articles

in Information & Management, Information Technology & People,

ACM Transactions on Computer–Human Interaction, IEEE Transac-

tions on Systems, Man and Cybernetics, Part A: Systems and Humans,

International Journal of Human–Computer Studies, The DATA BASE

for Advances in Information Systems, International Journal of Tech-

nology and Human Interaction, and the Journal of Information

Technology Management.

Jijie Wang is a PhD candidate in the

Department of Computer Information Sys-

tems at Georgia State University. She got

her bachelors degree in accounting from

Beijing University, P.R. China, and master

degree in CIS from Georgia State Univer-

sity. Her work has been presented at Amer-

icas Conference on Information Systems

and Information Resources Management

Association International Conference.

She has published papers in Decision Sciences, Information Resource

Management Journal, and the Communications of the Association for

Information Systems, and book chapters in Encyclopedia of E-Com-

merce, E-Government, and Mobile Commerce and Human-Computer

Interaction and Management Information Systems: Applications. Her

dissertation focuses on organizational controls in mobile virtual work.

Her other research interests include IT project management, open

source software community, and research methodology in information

systems research.