Computers in Human Behavior 55 (2016) 258e267

Contents lists available at ScienceDirect

Computers in Human Behavior

journal homepage: www.elsevier.com/locate/comphumbeh

Full length article

Computer anxiety and attitudes towards the computer and the

internet with Romanian high-school and university students
Ana-Maria Cazan a, *, Elena Cocorada

a, Ca

ta

lin Ioan Maican b
a
Faculty of Psychology and Education Sciences, Department of Psychology and Training in Education, Transilvania University of Brasov, Eroilor nr. 29,
500036, Brasov, Romania
b
Faculty of Economic Sciences and Business Administration, Department of Management and Economic Informatics, Transilvania University of Brasov,
Eroilor nr. 29, 500036, Brasov, Romania

a r t i c l e i n f o a b s t r a c t

Article history: The access of the general public and of the Romanian students to the use of the computer and the
Received 12 April 2015 internet has been possible only since the middle of the 90's, research on the topic having no object up to
Received in revised form that date. The variety of results obtained by studies conducted on different samples or at different time
22 July 2015
spans support the usefulness of the research in other cultural contexts than the ones in the Western
Accepted 9 September 2015
Available online 27 September 2015
countries. The current study examines the relationships between computer and internet anxiety, com-
puter self-efficacy and other personal characteristics in a Romanian context. A full mediated model was
tested. According to our findings, low computer self-efficacy predicts anxiety, the previous education in
Keywords:
Computer anxiety
the field of computer science has direct negative effects on computer anxiety and on the negative at-
Computer self-efficacy titudes towards the internet. The results show that there are no significant differences between the male
Core self-evaluation and the female participants concerning computer anxiety, self-efficacy and the negative attitudes to-
Internet attitude wards the internet. Significant differences between participants enrolled at different education levels and
Romanian students study programs, Science and Humanities, are obtained.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction Information Technologies equipment in the academic environment,

at work and at home (Chua, Chen, & Wong, 1999; Powell, 2013).
Some studies in the field of education and psychology have Research concerning the attitudes towards the computer and the
highlighted the prevalence of anxiety in the collection of emotions internet reported that some adults, as well as a part of the teenagers
which are associated to the learning process and to the low aca- and youngsters rate computers negatively, avoid using them or are
demic achievement (e.g., Linnenbrink-Garcia & Pekrun, 2011; less confident in their capacity to use them and make progress.
Shakir, 2014). The importance of anxiety is underlined by the They anticipate catastrophic consequences of their errors and
introduction of the umbrella-concept of academic anxiety, defined experience feelings of discomfort, frustration and stress (e.g. Chua
as a collection of anxieties lived by the students in the pre- et al., 1999; Heinssen, Glass, & Knight, 1987). Internet anxiety is
university and university environment, or related to it, such as: not a myth either, but rather a reality. Its sources are represented by
math anxiety, science anxiety, foreign language anxiety (Cassady, the possible negative impact of the internet on inter-personal re-
2010). Computer and internet anxiety can be added to these due lations, well-being and on the users’ personality. The anxiety to-
to the need, specific to the contemporary age, to effectively use the wards the computer and the dehumanizing aspects of using it can
computer and the internet as a learning means. be sources of negative attitudes towards the internet (e.g. Kalwar,
In Western countries, the study of the attitudes towards the Heikkinen, & Porras, 2013). Conversely, the confidence in oneself,
computer and the internet has been fostered in the last two de- in the Information Technologies and in their positive influence on
cades of the 20th century, due to the proliferation of the the well-being of people and society decrease the anxiety towards
the internet or computers (e.g. Thatcher, Loughry, Lim, & McKnight,
2007; Torkzadeh & Koufteros, 1994).
The proliferation of personal computers and of the internet has
* Corresponding author.
E-mail addresses: [email protected] (A.-M. Cazan), [email protected] been impressive over the last ten years, resulting in the fact that in
(E. Cocorada
), [email protected] (C.I. Maican). 2013 more than half of the total number of households in Romania

http://dx.doi.org/10.1016/j.chb.2015.09.001
0747-5632/© 2015 Elsevier Ltd. All rights reserved.
 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267 259

(55.8%) were equipped with a personal computer, the percentage compared to male ones (Chua et al., 1999; Durndell & Thomson,
being higher in the urban area (69.8%), according to the studies 1997; Kannan, Muthumanickam, & Chandrasekaran, 2012;
conducted by the National Institute of Statistics (2013a). 60.3% of Mikkelsen, Ogaard, Lindoe, & Olsen, 2002; Sam, Othman, &
the people aged 16 to 74 have used a computer at least once, the Nordin, 2005). The results are not surprising, as other studies
number of male users being higher than the one of female users: focusing on academic anxiety have reported that female students
61.8% as compared to 58.8%. Young people were at an advantage: register a higher level of anxiety as compared to male students. This
90% of the people aged 16 to 34 have used a computer. As far as the is explained by the higher level of the females’ emotional response,
internet access is concerned, the National Institute of Statistics has by their perceptions of threat in evaluative situations (Cassady,
concluded that a little more than half of the households (52.9%) are 2010) or by the roles associated with the two genders
connected to it, most of the households (73.2%) pertaining to the (Venkatesh, Morris, Sykes, & Ackerman, 2004). Nonetheless, con-
urban area (2013b). trary results were also reported: (i) male students have higher
A study conducted in the same year shows that, in Romania, the levels of anxiety or attitudes which are more negative towards the
people who hold a higher-education diploma spend 38 min of their use of the computer and of the internet (e.g Bao, Xiong, Hu, &
spare time in front of the computer, on average. Men spend 50 min Kibelloh, 2013; Brosnan & Lee, 1998; Cocorad
a, 2015) or (ii) fe-
on average while women spend 27 min on the average. Pre- males use social networks and the internet for communication
university and university students spend more than an hour in purposes more than males (Joiner et al., 2012; Kalwar et al., 2013).
front of the computer on a daily basis, almost equally shared: fe- Very recent results concerning gender differences in using location-
male students spend an average of 58 min in front of the computer, based social network services show that females pay more atten-
while male students register an average of 67 min (2013a). In tion to hedonic benefits (intrinsic motivations), while males pay
conclusion, the main differences regarding the use of computers more attention to utilitarian benefits (extrinsic motivation) than
and the internet in Romania are introduced by gender, age and level females (Sun, Wang, Shen, & Zhang, 2015).
of education. However, a number of studies signal the absence of gender
In this context of divergent results from other countries and differences among the investigated participants (e.g. Heinssen
based on the data regarding the use of the internet and computers et al., 1987; Hong & Koh, 2002; Madu, Otuka, & Adebayo, 2011).
in our country, this study aims to identify the attitudes towards the Studied centered on gender differences related to the attitudes
computer and the internet that appeared much later, as compared towards the computer and the internet used this construct as either
to western countries. a biological or psychological one. The results show that the psy-
Our study addresses the following research questions: chological gender would predict computer anxiety and technology
acceptance, while the biological gender would not, because the
1. Does computer anxiety mediates the relationship between psychological gender for some women is masculine or androgy-
personality aspects, such as core self-evaluation, computer self- nous, this modifying the results (Powell, 2013; Venkatesh et al.,
efficacy and attitudes towards the use of internet? 2004).
2. Are there associations between computer anxiety, internet
anxiety, computer self-efficacy and experience in using 1.1.2. Age and experience in using the computer
computers? Some researchers have found that the analyses based on the age
3. Are there any differences regarding computer anxiety, internet criterion show that lower levels of anxiety are present in the case of
anxiety, computer self-efficacy and experience in using com- young people (Chua et al., 1999; Kubiatko, Hala kova
, Nagyova , &
puters between the participants with different education levels, Nagy, 2011; Parasuraman & Igbaria, 1990), but other studies have
different academic performances and who study in different not found a relationship between age and computer anxiety in the
domains (Science and Humanities)? case of university students. We also found the idea that the adults'
difficulty of learning using the computer is rather a myth supported
To address the research questions, we designed a quantitative by daily life anecdotes and not by empirical evidence (Yan &
study. Fischer, 2004).
The students who have various years of experience in using a
1.1. Literature review computer and who own a PC have a more positive view on the
social and personal effects of using the computer and the internet
In today's world, the computer and the internet are used in (e.g. Liaw & Huang, 2006; Orr, Allen, & Poindexter, 2001). A study
different settings: work, services, learning, communication and from 2006 indicated that it was not computer experience or the use
socialization etc. They represent a means to get online education, of computers that had the largest significant relationship with
facilitating the access to various types of knowledge and its sharing, computer anxiety, but self-efficacy beliefs (Wilfong, 2006).
and the access to experiences with classmates in virtual classes or
virtual work groups (Zhang, de Pablos, & Xu, 2014; Zhang, de 1.1.3. Self-efficacy
Pablos, & Zhang, 2012). Self-efficacy is a good predictor of the perception regarding the
Research concerning the attitudes towards the computer and ease of computer use. According to the social role theory, some
computer anxiety was based on their relationships to gender, age, studies report that male students experience more favorable feel-
experience, ownership of a computer, geographic area and cultural ings and a higher degree of self-confidence when it comes to ac-
context, as well as to some personality features. Longitudinal or tivities involving the computer and the internet, as compared to
follow-up studies are also present. The results are divergent and female students (Gibbs, 2013; Torkzadeh & Koufteros, 1994).
seem to depend on the decade when the research was conducted or The masculine specificity of gender roles, not the biological sex,
on the samples used (Powell, 2013). was proposed as a source of these differences (Huffman, Whetten,
& Huffman, 2013; Venkatesh et al., 2004). Contrasting research
1.1.1. Gender argues that there are no gender differences as far as the internet
The research based on the gender variable identifies negative attitude or self-efficacy scores are concerned (e.g. Joyce &
emotions and perceptions concerning the use of the computer and Kirakowski, 2013; Zhou, 2014) or identifies a negative connection
of the internet, which are stronger in the case of female students as between computer anxiety and computer self-efficacy (e.g. Scott &
260 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267

Walczak, 2009). in other countries.

An important factor in determining computer self-efficacy could
be the discipline of study: computer self-efficacy is positively 1.2. Current study
correlated with the participation in ICT courses; undergraduates
studying computer related disciplines appeared to have higher self- In 2011, Romania held the last place in Europe as far as the
efficacy as regards computers and the internet (Sam et al., 2005; internet usage was concerned (Eurostat, 2011). Given this context,
Scherer & Siddiq, 2015). the studies concerning the attitude towards the computer and the
Other studies show that computer self-efficacy does not internet and the anxiety towards them have rarely been conducted
demonstrate significant relations with internet anxiety (Thatcher on the Romanian population.
et al., 2007). Ambiguous findings regarding computer self-efficacy Most of the studies conducted on college students do not report
due to gender differences could be explained by the multidimen- gender differences as far as computer anxiety is concerned, this
sional nature of the construct (Scherer & Siddiq, 2015). being explained by the high educational purposes and equivalent
learning opportunities (e.g., Hong & Koh, 2002), while other
1.1.4. Academic results research has come to opposite results. In Romania, the differences
Some studies have not found any relationship between aca- between men and women concerning the use of the computer
demic performance and computer anxiety (Williams & McCord, during leisure time are in favor of men (National Institute of
2006), while others have found a negative relationship. Other Statistics, 2013b). Thus, we propose the following:
studies show that internet anxiety moderates the relationship be-
tween academic self-efficacy and learning performance (e.g. Shakir, H1.1 Computer and internet anxiety are stronger for female than
2014). for male students.
H1.2 Computer self-efficacy is stronger for male than for female
1.1.5. Studies in Romania students.
Two studies conducted at the beginning of this millennium
(Durndell & Haag, 2002; Durndell, Haag, & Laithwaite, 2000) on The participants' experience in using the computer has been a
university students (in Physics, Law, Medicine and Sociology) in frequent criterion in analyzing the attitude towards the computer
Romania, investigated computer, internet anxiety and computer and the internet, but the results are also divergent (Kubiatko et al.,
self-efficacy. The studies were underpinned by the novelty of the 2011; Powell, 2013; Sam et al., 2005). Because the participants'
historic and cultural context. As opposed to Western countries, in experience in using the computer and the internet is operational-
the former communist countries women and men had equal access ized variously, we introduced several indicators: (1) number of
to studies in the field of engineering, mathematics and cyber- years the participants reported with respect to the study of IT&C
science. The hypothesis according to which the gender gap might subjects in school, (2) how long they have used the computer, (3)
be missing in Romania as far as computers and the internet are the education level, (4) study program (Science or Humanities).
concerned was not confirmed, the authors identifying statistically Therefore, we propose the following:
significant differences between women and men. Men were more
experienced with using the internet and registered higher self- H2.1 Computer anxiety and internet anxiety correlate directly
efficacy scores on average. Computer anxiety was stronger with with the number of years the participants reported with respect
women and the attitudes towards the internet were more favorable to the study of IT&C subjects in school and how long they have
in the case of men than in the case of women. used the computer.
A recent study conducted on pre-university and university H2.2 Computer anxiety and internet anxiety have lower levels
students confirms the differences between female and male stu- with the respondents studying Science, as compared to the ones
dents concerning computer anxiety (higher in the case of female from the Humanities, while computer self-efficacy is higher in
subjects) and self-efficacy (higher in the case of male subjects). The Science than in the Humanities.
experience with the use of computers is negatively associated with H2.3 Students with a higher computer anxiety obtain lower
the computer anxiety level and computer self-efficacy with aca- grades, while students with a higher self-efficacy obtain higher
demic results (Cocorad
a, 2014). Another study conducted on a grades.
sample of participants studying in the socio-humanities field H2.4 Participants with higher education (e.g. university stu-
(socio-humanities high-school students and university students in dents) have lower computer anxiety and internet anxiety than
the field of psychology) shows that computer anxiety is signifi- the participants with a lower education level (e.g. high school
cantly different with men and women. But on that particular students), but computer self-efficacy is higher for the first
sample, men have proven to be more anxious than women and the category.
attitude towards the internet is more favorable with women than it
is with men (Cocorada
, 2015). When the attitude towards the Previous studies used computer anxiety as a mediator or
internet is analyzed according to the school cycle, it turns out that it focused on computer anxiety in order to understand individual
is more negative in the case of high-school students as compared to reactions and attitudes towards the computer. There are studies
psychology students, as a result of the richer experience in the case which show that computer anxiety is the strongest predictor of the
of the latter, while computer self-efficacy is equivalent for the two negative attitude towards the computer among the demographic,
genders and study cycles (high-school and university). personality and cognitive style variables (Igbaria, Guimaraes, &
A research study focused on identifying the aims of internet Davis, 1995). Other researchers showed that computer anxiety
usage among teenagers revealed that less than a quarter of the and computer playfulness are significant mediators of the effect
participants (22.7%) often use the internet to find out information that the system experience has on the ease of use, the effect being
about school and a lot fewer (10.2%) use it in order to improve their fully mediated only by computer anxiety (Hachbart, Grover, & Yi,
learning skills. Female subjects use the internet as a means to study 2003). More recent studies emphasized the relationship between
more than male participants, who prefer to play various types of computer anxiety and electronic media use and their implications
video games either alone (single player) or in teams (multiplayer on the overall life satisfaction, adjustment or academic perfor-
online games) (Damean, 2008), thus confirming studies conducted mance, showing that electronic media use has a negative impact on
 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267 261

academic performance, mental health and subjective well-being, translated from English into Romanian and the back-translation
the impact being mediated by the level of anxiety (Lepp, Barkley, was accomplished by translators who fluently speak the two lan-
& Karpinski, 2014). guages. The back translation was compared with the English
The association between the study of computer science and the version of the questionnaires in order to identify any lack of fidelity
students' attitudes is mediated by the level of anxiety. Previous and to find the most appropriate equivalents for the items adapted
studies showing that the frequent use of computers and technology to the Romanian cultural context. For all the scales, the Alfa Cron-
led to decreased anxiety and to increased confidence, which in turn, bach coefficients ranged between .71 and .96, indicating an
could provide strategies to enhance computer use and positive acceptable reliability (McCrae, Kurtz, Yamagata, & Terracciano,
attitude towards computers and technology (Chu & Mastel-Smith, 2011).
2010; Werner, Carlson, Jordan-Marsh, & Clark, 2011). The CARS comprises 19 items, measured on a four point Likert
The association between computer self-efficacy and computer scale. The Alpha Cronbach coefficient for the entire scale showed a
anxiety has also been explored. Igbaria et al. (1995) found that self- good reliability (.95). The exploratory factor analysis revealed a
efficacy was positively correlated with perceived ease of use, two-factor solution explaining 63.86% of the total variance after the
perceived usefulness and usage, but negatively correlated with Varimax rotation. The meaning of the items suggested the
computer anxiety. Moreover, in the study of information systems, following labels: Confidence (13 items: Anyone can learn to use a
anxiety has been viewed as a personality variable that influences computer if they are patient and motivated, I am confident that I
system use (Agarwal & Karahanna, 2000), showing that the rela- can learn computer skills) and Fear (6 items: I hesitate to use a
tionship between anxiety and behavior is mediated by the personal computer for fear of making mistakes that I cannot correct, It scares
beliefs (Schlenker & Leary, 1982) and anxiety is incorporated as an me to think that I could cause the computer to destroy a large
antecedent to the beliefs of usefulness and ease of use (Venkatesh & amount of information by hitting the wrong key). The two di-
Davis, 2000). mensions had high alpha Cronbach coefficients, .96 for the first
The model, used in a recent meta-analysis (Powell, 2013), sug- factor (Confidence) and .71 for the second (Fear).
gests the following organization of constructs: (i) the variable to be The IAS consists of 20 items and the Likert scale is a five-point
studied (computer anxiety/the attitude towards the computer and one. The Romanian version of IAS showed a different structure, as
the internet) and its correlates (self-efficacy, perceived ease of use, compared to the original scale. After reversing the items, the alpha
perceived usefulness of the computer and satisfaction etc.), (ii) Cronbach showed that several items tended to decrease the reli-
antecedents: personal antecedents (age, gender, personality, edu- ability of the scale and to correlate negatively with the scale,
cation, profession), other types of anxiety (i.e. math anxiety, e-mail despite the logical meaning of the items. Eliminating items 8 (The
anxiety, depression, general distress), and interactions between overuse of the internet may be harmful and damaging to humans) and
individual and computer/interactional antecedents (experience, 17 (The internet will bring us into a bright new era) improved the
perceived ease of use, ownership and training) and (iii) outcomes reliability of the scale (Alfa Cronbach ¼ .88). The exploratory factor
or consequences (performances, intent to use). Following the same analysis revealed a two-factor solution explaining 58.92% of the
pattern, we propose the following: total variance after the Varimax rotation. The labels used for the
two factors were chosen according to the meaning of the items
H3 Computer anxiety mediates the relationship between per- included in each factor. Thus, the first factor included items related
sonality aspects, such as core self-evaluation, previous experi- to the perceived complexity and difficulty of the internet (such as:
ence and education regarding the use of computers, computer The Internet makes me uncomfortable because I don't understand
self-efficacy and attitudes towards the use of the internet. it, I feel intimidated by the internet). The second factor included
Thus, we expect that a high core self evaluation and a higher items related to the social impact of the internet (such as: Soon our
experience in using computers will predict lower levels of lives will be controlled by the internet, The internet is dehumanizing to
computer anxiety, which in turn, will predict a positive attitude society). The first factor was labeled Perceived complexity and diffi-
towards the internet. culty of the internet and the second factor was labeled Social impact
of the internet.
2. Materials and methods The high scores obtained are indicative of computer anxiety
(CARS) and of a negative attitude towards the internet (IAS).
2.1. Participants The CSE comprises ten items and measures self-efficacy in using
the computer, on a ten-point Likert scale. The Romanian version of
The sample included 486 participants, from different study CSE showed a good reliability, with an Alfa Cronbach coefficient of
programs, covering both Science (53.1%) and Humanities (46.9%) .81, for the entire scale. The exploratory factor analysis revealed a
and two educational levels: high school (30%) and university stu- two-factor solution explaining 54.82% of the total variance after the
dents (70%), female (70%) and male (30%), with an average age of Varimax rotation. The small number of items (3 items) included in
21.08 years. The questionnaires were administered in a paper- factor 2 led to a uni-dimensional model solution for the computer
pencil format, during courses, and the participation was volun- self-efficacy scale.
tary and unpaid. The participants were students at a university and The CSES consists of 12 items concerning self-esteem, general-
a high school in Brasov, a large city in Romania. ized self-efficacy, neuroticism and locus of control. For the Core
Self-evaluation scale, Cronbach's alpha coefficient was .79. The
2.2. Instruments and their psychometric properties interpretation of the questionnaire is the following: high sum
scores of the CSES are indicative of high core self-evaluation.
In order to verify the hypotheses, the following instruments Academic performance was measured with IT&C grades (per-
were used: the Computer Anxiety Rating Scale - CARS (Heinssen formance in the field of computer science, by grades in courses such
et al., 1987), the Internet Attitude Scale e IAS, adapted in order to as Information Technology) and GPA (grade point average at the
assess the attitude towards the internet, according to a scale built end of the previous semester).
by Nickell and Pinto (1986), the Computer Self-Efficacy Scale - CSE The demographic data include gender, age, the way in which
(Compeau & Higgins, 1995) and the Core Self-Evaluations Scale e computer literacy was achieved (at school or outside school as non-
CSES (Judge, Erez, Bono, & Thoresen, 2002). All the scales were formal or informal education), years of computer-related
262 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267

experience. Ownership of a computer and of an e-mail address was towards internet, but a higher computer self-efficacy. These cor-
also taken into consideration as an indicator of the participants' relations with the grades in revealed interesting findings: the
complexity of computer science skills. The descriptive statistics negative correlations showed that Science students with higher
showed that more than 97% of the participants owned a computer grades have a more positive attitude and do not experience com-
and an e-mail address at the date when the research was con- puter anxiety.
ducted. Therefore, we eliminated these items because of their low
discriminability. 3.3. Differences concerning the education level (H2.4)

3. Results The comparison between high school students and university

students shows significant differences and high effect size co-
The male participants were more experienced in using a com- efficients, high school students having a higher level for all the
puter (M ¼ 10.33 years, SD ¼ 2.8) as compared to the female par- variables, excluding the computer self-efficacy (Table 2).
ticipants (M ¼ 9.14 years, SD ¼ 2.97) (t ¼ 4.09, p < .001), the average There were no differences between students who learned to use
for the sample being of 9.5 years, Xmin ¼ one year, Xmax ¼ 21 years. the computer on their own and those who learned from others
For the whole sample, the GPA was 8.54 (SD ¼ .86, Xmin ¼ 5, (school, friends, parents etc).
Xmax ¼ 10) and the grades average for IT&C related disciplines was
8.49 (SD ¼ 1.82, Xmax ¼ 10 and Xmin ¼ 3). 3.4. Mediation model (H3)

3.1. Gender differences (H1.1 and H1.2) A full mediated model was tested in order to determine if
computer anxiety mediated the relationship between personality
In this paper, only the biological gender that was declared by the aspects (core self-evaluation), previous experience and education
participants was used. The t test for independent samples did not regarding the use of computers (study computer science), com-
reveal significant differences between the male and female par- puter self-efficacy and the attitudes towards the internet. The
ticipants regarding computer anxiety (t ¼ .10, p ¼ .92), negative Pearson correlation coefficients were computed in order to analyze
attitudes towards the internet (t ¼ .49, p ¼ .62) or computer self- the pattern of the relationships between the variables. These cor-
efficacy (t ¼ 1.43, p ¼ .15). Thus, the hypotheses regarding the relations showed significant associations between computer anxi-
gender differences (H1.1. and H1.2.) were not confirmed. ety, attitudes towards the internet and computer self-efficacy: the
students with a high level of anxiety also had a negative attitude
3.2. Academic results and experience with using a computer (H2.1, towards the use of the internet and a low level of computer self-
H2.2 and H2.3) efficacy. The core self-evaluation correlated only with computer
self-efficacy (Table 3). As Iacobucci (2008) explains, multi-
For the entire sample (including both study programs), the collinearity is generally expected in the mediational analysis of the
findings show that students who had studied Computer Science at mediator variable and the dependent and the independent vari-
school for a longer period of time had a lower level of anxiety and able, and therefore it cannot be avoided by the researcher.
more positive attitudes towards the internet (Table 1). The corre- Given the significant correlations between the dimensions of
lation with the grades (GPA) was positive, but small, showing that the scales, we used the overall scores in the structural model, to
students with high grades tended to be more anxious, to have avoid overlapping. The model included the overall score for the
negative attitudes towards the internet and a higher level of self- computer anxiety scale and the overall score for the negative atti-
efficacy (Table 1). Taking into consideration the study program, tude towards the internet scale as an endogenous variable (Fig. 1).
the results were different: previous experience (experience ac- The goodness-of-fit indices for the tested model showed a good fit
quired at school with using a computer) was negatively associated (Table 4). According to Hu and Bentler (1999) and Kline (2011), the
with anxiety and negative attitudes for the students in the Science goodness-of-fit criteria used in the current study acknowledged the
study program (S), but the correlations were positive for those in potential for acceptable (c2/df ratio <3, CFI and TLI >.90, SRMR <.10,
the Humanities study programs (H). RMSEA <.08) and excellent fit (c2/df ratio <2, CFI and TLI >.95, SRMR
The correlations between grades and anxiety and computer self- <.08, RMSEA <.06). Mahalanobis distance showed no evidence of
efficacy, and the attitude towards the internet were positive, multivariate outliers.
showing that the students who obtain higher grades experience The significance of the mediating effect of computer anxiety was
more anxiety towards the computer, have negative attitudes tested using the Bootstrap estimation procedure in AMOS. The

Table 1
Pearson correlation coefficients between grades, use of computer, and CARS, IAS, CSE and CSES.

How many years have you How long have you used the GPA IT&C grades
studied computer science at computer?
school?

Entire sample S H Entire sample S H Entire sample S H Entire sample S H

** ** * * **
Computer confidence (CARS) .24 .39 .09 .01 .02 .14 .14 .04 .11 .02 .55 .02
Computer fear (CARS) .16** .33** .14* .08 .07 .33** .05 .04 .03 .06 .41** .01
Computer anxiety total (CARS) .24** .39** 14 *
.03 .03 .27** .12* .04 .09 .01 .54** .01
Internet complexity (IAS) .26** .40** .02 .01 .05 .08 .13* .04 .09 .01 .55** .07
Internet social impact (IAS) .10* .19** .01 .01 .10 .09 .05 .02 .02 .01 .16* .08
Internet negative attitude-total (IAS) .25** .41** .03 .01 .07 .12 .13* .04 .09 .02 .51** .08
Computer efficacy total (CSE) .02 .01 .10 .10* .14* .07 .17** 18* .09 .20** .13 .13
Core self-evaluation (CSES) .14** .15* .14* .25** .25** .26** .04 .07 .15 .01 .06 .12

Notes. **Correlation is significant at the .01 level (2-tailed), *Correlation is significant at the .05 level (2-tailed), S ¼ Science study program (N ¼ 257), H ¼ Humanities study
program (N ¼ 228).
 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267 263

Table 2
Differences between high school and university students.

Education level N Mean Std. dev t df p Cohen's d

Computer confidence (CARS) High school 143 3.47 .25 15.93 427.15 <.001 1.27
University 340 2.55 .99
Computer fear (CARS) High school 144 3.01 .61 7.88 292.42 <.001 .77
University 339 2.52 .66
Computer anxiety total (CARS) High school 142 3.33 .29 15.12 467.85 <.001 1.25
University 339 2.54 .84
Internet complexity (IAS) High school 146 3.43 .27 17.37 453.54 <.001 1.42
University 337 2.52 .86
Internet social impact (IAS) High school 146 2.68 .51 4.06 483 <.001 .41
University 339 2.47 .50
Internet negative attitude-total (IAS) High school 146 3.07 .26 15.09 477.89 <.001 1.30
University 337 2.51 .55
Computer efficacy total (CSE) High school 145 7.70 1.2 1.41 482 .157 .14
University 339 7.52 1.33

Table 3
Pearson correlation coefficients between the CARS, IAS, CSE and CSES scales.

Mean XMin. XMax. 1 2 3 4 5 6 7 8

1. Computer confidence (CARS) 2.77 1.16 4.74 1

2. Computer fear (CARS) 2.66 1.00 7.67 .69** 1
3. Computer anxiety total (CARS) 2.77 1.16 4.74 .98** .82** 1
4. Internet complexity (IAS) 2.80 1.00 4.00 .94** .66** .93** 1
5. Internet social impact (IAS) 2.54 1.00 4.00 .28** .34** .31** .29** 1
6. Internet attitude total (IAS) 2.68 1.17 3.83 .89** .68** .89** .94** .56** 1
7. Computer efficacy total (CSE) 7.58 3.10 10 .14** .14** .14** .13** .11** .14** 1
8. Core self-evaluation (CSES) 2.96 1.58 4.17 .02 .04 .01 .04 .01 .03 .25** 1
** *
Notes. Correlation is significant at the .01 level (2-tailed), Correlation is significant at the .05 level (2-tailed).

Fig. 1. The structural equation model regarding the mediating effect of computer anxiety.

standardized path coefficients and the standardized indirect effect

and its associated 95% confidence levels of significance are dis-
Table 4 played in Table 5. The indirect effects are statistically significant,
The Goodness-of-fit measures for the mediating effect of computer anxiety. showing that there is a mediation effect (Cheung & Lau, 2008;
Model c2 (df) CFI TLI AIC RMSEA (90% CI) Iacobucci, 2008).
Testing the individual paths presented in Fig. 1 and analyzing
Mediation model 1.64 (1) p ¼ .20 .998 .992 29.64 .037 (.000.135)
the significance levels of the indirect effects (Table 5), it can be
Notes. GFI: Goodness-of-Fit Index, CFI: Comparative Fit Index, TLI: TuckereLewis concluded that the mediation hypothesis is confirmed (MacKinnon,
Index, AIC: Akaike Information Criterion, RMSEA: Root Mean Square Error of
Approximation, 90% CI: 90% confidence interval for RMSEA.
Lockwood, Hoffman, West, & Sheets, 2002).
264 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267

Table 5 answer to the question concerning the possibility of some changes

Direct, indirect and total effects (Standardized estimates) for the structural model. in people's attitudes, which would turn this decade into a one of
Variables Computer anxiety Negative attitude towards diminishing gender differences (Durndell & Thomson, 1997;
the internet Iloma€ki & Rantanen, 2007). The results indicate a beginning in
Effects Direct/Total Direct Indirect Total modifying gender roles with respect to the use of computers and
** ** the internet in Romania.
Study computer science .24 .03 .21 .24**
Computer self-efficacy .16** .14* .14*
Core self-evaluation .02 .03 .02 .05 4.2. Experience with computers (H2.1) and differences between
Computer anxiety .89** .89** Science and Humanities students (H2.2)
Note: *p < .05, **
p < .01.
The anxiety levels towards the computer and the internet are
lower with the participants who studied IT&C disciplines more
According to the results, low computer self-efficacy predicts years in school, for the whole sample and for the Science domain.
anxiety. The study of computer science (previous education) has The results are not significant for the other groups, revealing that
negative direct effects on computer anxiety and on the negative the technical and mathematical expertise in using the computer
attitudes towards the internet. These findings confirm the hy- plays a positive role in the management of beliefs about computers
pothesis of the mediated effect of computer anxiety. and internet.
The presence of significant levels of computer and internet
4. Discussion anxiety for the participants who had used the computer before
studying it at school shows that the source of these attitudes may
This article presents the results of a research study focused on be represented by learning how to use the computer. We suppose
examining computer anxiety and self-efficacy when using the that the participants, when they learned by themselves or occa-
computer and the attitude towards the internet on a sample sionally helped by friends how to use a computer, before studying it
comprising Romanian high-school and university students. The at school, used strategies based on trial and error, in contrast to
research was built from the perspective of some widely used learning in the group-class which are characterized by scaffolding
criteria, such as gender, usage experience, level of education or type (Yan & Fischer, 2004) and sharing team tacit knowledge (Zhang, de
of study program and some personality traits. Pablos, & Zhu, 2012). The two acquisition ways may have different
The expansion of computer and internet usage in Romania raises consequences over the emotions, beliefs and attitudes towards the
the question of the convergence of the participants' reaction in this computer and the internet. Building knowledge in the IT&C domain
area towards the new technologies with the ones identified by by beginner adults, as in an experimental study conducted by Yan
previous research in other geographic areas. The variety of the re- and Fischer (2004), is characterized by many fluctuations and
sults obtained by the research conducted on different samples or at regress, emphasizing the need for social and technical support.
different times explained the need of such studies. Although In- Similar microdevelopment pathways are also possible for the par-
formation Technology is global, the individuals' attitudes are ticipants who learned by themselves, but may be with more sig-
context-dependent (e.g. Li & Kirkup, 2007). Our hypotheses are nificant fluctuations and regress. We suppose that repeated failure,
mostly supported by proof. uncertainty and negative emotions experienced were generalized
as regards the computer and the internet.
4.1. Gender differences for computer and internet anxiety and For the participants who study in the Humanities, the anxiety
computer self-efficacy (H1.1 and H1.2) increases if they have used the computer for many years. In this
case, the different disciplines of study for the two sub-samples may
Stimulated by the novelty of the cultural and historic context, a explain the differences between the levels of anxiety. In other
previous study conducted on a sample of Romanian students “tends words, the participants who have studied more IT&C disciplines
to support the point of view that the emerging countries of Eastern and disciplines connected to IT&C (e.g. mathematics or technical
Europe will not be immune to the gender related effects of tech- disciplines) are less anxious. Conversely, the respondents in the
nology as found in the west” (Durndell & Haag, 2002, p. 531). Being Humanities study programs, who study fewer IT disciplines, are the
concerned with computer anxiety, our study, conducted at a dis- more anxious, the more years they report for studying computer
tance of more than a decade from the one quoted, did not reveal outside school.
significant differences between the male and female participants The process of trial-and-error learning, doubled by the activa-
on the whole sample. On the other hand, if we select the university tion of the stereotype according to which the IT domain is not
students' sub-sample, the gender differences are significant, but specific to Humanities study programs, led to higher uncertainty in
not significant for high-school students (digital natives). The dif- using the computer and the internet and to a lower confidence in
ference between male and female participants is still maintained as the students' ability for using IT&C. The results show that the
far as experience with using a computer is concerned. The new presence of an authorized support (e.g. teacher) in learning the
results subscribe to the tendency of overcoming the gender gap, the computer plays a positive role in the management of beliefs about
attitudes towards the computer and the internet being almost the computer and the internet and in the management of self-
equal for the whole sample. Our findings show that self-efficacy efficacy.
does not differ by gender, in contrast to the above mentioned
study, which indicates that males tend to report greater computer 4.3. Anxiety, self-efficacy, grades and level of education (H 2.3 and
self-efficacy (Durndell & Haag, 2002). H 2.4)
The tendency to decrease the gap between genders is probably
favored by the learning objectives and opportunities offered by the The association between grades and anxiety, attitude and
educational system. The results converge with studies about the computer self-efficacy needs also further exploration and a more
equivalence of duration in computer use in the leisure time by male detailed debate. The results showed that the higher level of
and female students in Romania, reported by the National Institute expertise in using the computers in the case of the students from
of Statistics (2013b). These results could also constitute a positive the Science study programs was associated with a lower level of
 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267 265

computer anxiety; the students from the Humanities study pro- 2014; Sam et al., 2005).
grams have a lower amount of technical knowledge and therefore Empirical evidence suggests that computer anxiety and internet
experienced higher levels of anxiety. On the other hand, because anxiety are associated, being however, distinct constructs (e.g.
the GPA includes assessments for different competences for the two Brown, Fuller, & Vician, 2004; Wu & Tsai, 2006). The strong and
study programs, an association with the anxiety levels should be significant direct correlation between computer anxiety and the
interpreted with caution. attitudes towards the internet validates the hypothesis according to
While these associations with the use of computers are not which the two constructs, computer anxiety and attitude towards
significant for the whole sample and for the Science students, the the internet, are strongly connected. Due to the fact that accessing
results showed positive associations for the Humanities sample. the internet generally requires the use of a computer, the in-
Our results are not concordant with those reported in the literature, dividuals who experience irrational feelings of fear towards the
showing that internet usage interferes with assignments or that computer could transfer them towards the internet as well
electronic media use has a negative impact on academic perfor- (Thatcher et al., 2007).
mance (Junco, 2012; Junco & Cotton, 2011; Lepp et al., 2014) and The studies which analyze the variables in a combined manner
was negatively related to overall college GPA. The significant cor- indicate the fact that the experience in using the computer sup-
relations with the GPA found in our study could be explained by the ports the self-efficacy towards it, more in the case of men than in
fact that the internet is not only a source of spare-time or distrac- the case of women, while computer anxiety may lead to a lower
tion, but also a source of information and aid for learning. Recent self-efficacy, more often in the case of women than in the case of
research suggests that it is not the amount of time that a student men (Lee & Huang, 2014).
spends online, but rather what a student does online which affects
his academic performance (Chen & Tzeng, 2010). 5. Limitations of the research and future research
The level of education is another antecedent of computer anx-
iety. In line with other studies (Bozionelos, 2001; Kubiatko et al., The main limitations of this study concern the non-randomized
2011), our research proved that university students had lower sample and the unequal status of the participants as far as the
levels of computer anxiety and higher levels of computer self- gender variable is concerned (340 female participants versus 146
efficacy than high school students. Training (previous use or edu- male participants). Due to the fact that the investigated sample only
cation on the use of computers) decreases anxiety and increases consisted of university students and high-school students, the
self-efficacy, a longer previous period dedicated to studying com- conclusions cannot be generalized to other age groups or to the
puter science having a negative effect on anxiety, and thus leading students who have advanced skills in the Information Technologies
to the decrease in computer anxiety (Orr et al., 2001). field.
Another limitation could be related to the fact that the in-
4.4. Mediation model (H3) struments used in the research measure computer anxiety and
computer self-efficacy, but the mediation model includes ‘Internet
An important reason as a basis for our central research hy- attitudes' as the predicted endogenous variable. The fact that
pothesis (the mediation hypothesis) was the classical view of accessing the internet is influenced by the attitude and the anxiety
anxiety, which shows that anxiety mediates the relationship be- towards computers, verified by the correlation coefficients in
tween beliefs and behavior (Spielberger, 1972). Applying Spiel- Table 3 as well, reduces the negative consequences.
berger's model, our study shows that computer anxiety mediates The results of this research could be negatively influenced by the
the relationship between beliefs (computer self-efficacy and core objectives and the complexity of the IT&C disciplines studied by the
self-evaluation) and attitude (attitude towards the internet, participants from different study programs. Thus, a Science student
including behavioral, cognitive and emotional dimensions) and a Humanities student with the same grades (GPA and IT&C
(Fishbein & Ajzen, 1975). grades) have different competences and different levels of learning
Our findings confirm the hypothesis of the mediated effect of outcomes. Another limit of the research could be generated by
computer anxiety: low computer self-efficacy predicted computer anxiety as a trait that was not measured.
anxiety. The study of computer science (previous education) had The results must be interpreted with caution because the
direct negative effects on computer anxiety and on negative atti- number of years related to using computers is self-reported and the
tudes. Personality traits have been considered antecedents of answer of the participants could be related to different contexts:
computer anxiety and attitudes towards the computer and the doing homework, listening to music, watching movies etc.
internet. Computer self-efficacy has a direct positive effect on The future research directions are encouraging in the social
computer anxiety and an indirect positive effect on the students' settings of Romania, where studies as the one we conducted are
attitudes. Computer anxiety tends to be the most powerful pre- few. For example, we believe that research on the following themes
dictor of negative attitudes, while core self-evaluation had no sig- is important:
nificant effects, despite its significant correlations with computer
self-efficacy and previous computer science education. Although - Diversifying the actual domains of using computers for learning,
the core self-evaluation contains three special items for the entertainment, communication etc., in relation with anxiety and
generalized self-efficacy and it is significantly associated with self-efficacy.
computer self-efficacy, the results show that computer self-efficacy, - Comparing computer anxiety to the anxiety caused by other
a specific measure, is more salient than a general measure, such as devices (smartphones, tablets).
core self-evaluation. Computer self-efficacy was also identified as - Longitudinal studies regarding the evolution of emotions and
being a predictor of computer anxiety, confirming other studies attitudes towards IT&C.
(e.g. Bao et al., 2013; Hasan, 2006; Henderson, Deane, & Ward,
1995). The use of the computer and of the internet and the study of
Being studied more after 2000, computer self-efficacy appears computer science in a formal manner could be antecedents of the
as a significant predictor of the positive attitude towards the positive image of the participants (a higher level of core self-
internet and of the acquisitions in online learning, being higher in evaluation), showing that the competence in using computers
the case of students who use effective learning strategies (Cazan, and the internet is an important aspect of self-image. The findings
266 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267

are more important in the educational context, because they may Research on Technology in Education, 35(1), 27e49.
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure
provide extra information necessary to make decisions concerning
analysis: conventional criteria versus new alternatives. Structural Equation
the teaching-learning strategies and interventions in order to in- Modelling, 6, 1e55.
crease the well-being of the learners pertaining to different cultural Huffman, A. H., Whetten, J., & Huffman, W. H. (2013). Using technology in higher
spaces. education: the influence of gender roles on technology self-efficacy. Computers
in Human Behavior, 29(4), 1779e1786.
Iacobucci, D. (2008). Mediation analysis. London: Sage Publications.
Igbaria, M., Guimaraes, T., & Davis, G. B. (1995). Testing the determinants of
References microcomputer usage via a structural equation model. Journal of Management
Information Systems, 11(4), 87e114.
Agarwal, R., & Karahanna, E. (2000). Time flies when you're having fun: cognitive Ilom€ aki, L., & Rantanen, P. (2007). Intensive use of ICT in school: developing dif-
absorption and beliefs about information technology usage. MIS Quarterly, ferences in students ICT expertise. Computers & Education, 48(1), 119e136.
24(4), 665e694. Institutul National de Statistic
a [National Institute of Statistics]. (2013a). Utilizarea
Bao, Y., Xiong, T., Hu, Z., & Kibelloh, M. (2013). Exploring gender differences on timpului in Roma ^nia [The use of time in Romania]. http://www.insse.ro/cms/files/
general and specific computer self-efficacy in mobile learning adoption. Journal statistici/comunicate/com_anuale/util_timpul/utilizarea_timpului_r13.pdf.
of Educational Computing Research, 49(1), 111e132. Institutul National de Statistica
[National Institute of Statistics]. (2013b). Accesul
Bozionelos, N. (2001). Computers anxiety: relationship with computer experience populaţiei la tehnologia informaţiilor şi comunicaţiilor în anul 2013 [People's ac-
and prevalence. Computers in Human Behavior, 17(2), 213e224. cess to technology of information and communication in 2013]. Retrieved from
Brosnan, M., & Lee, W. (1998). A cross-cultural comparison of gender differences in http://www.insse.ro/cms/files/statistici/comunicate/com_anuale/tic/tic_r2013.
computer attitudes and anxieties: the United Kingdom and Hong Kong. Com- pdf.
puters in Human Behavior, 14(4), 559e577. Joiner, R., Gavin, J., Brosnan, M., Cromby, J., Gregory, h., Guiller, J., et al. (2012).
Brown, S. A., Fuller, R. M., & Vician, C. (2004). Who's afraid of the virtual world: Gender, internet experience, internet identification, and internet anxiety: a ten-
anxiety and computer-mediated communication. Journal of the Association for year follow-up. Cyberpsychology, Behavior, and Social Networking, 15(7),
Information Systems, 5(2), 81e105. 370e372.
Cassady, J. C. (2010). Test anxiety: contemporary theories and implications for Joyce, M., & Kirakowski, J. (2013). Development of a general internet attitude scale.
learning. In J. C. Cassady (Ed.), Anxiety in schools: The causes, consequences, and In A. Marcus (Ed.), Design, user experience, and usability. Design philosophy,
solutions for academic anxieties (pp. 7e26). New York, NY: Peter Lang. methods, and tools, 250e260, Third International Conference, DUXU 2014, held as
Cazan, A. M. (2014). Self-regulated learning and academic achievement in the part of HCI International 2014, Heraklion, Crete.
context of online learning environments. In The International Scientific Confer- Judge, T. A., Erez, A., Bono, J. E., & Thoresen, C. J. (2002). Are measures of self-esteem,
ence Elearning and Software for Education 3 (pp. 90e95). Bucharest: “Carol I” neuroticism, locus of control, and generalized self-efficacy indicators of a
National Defense University. common core construct? Journal of Personality and Social Psychology, 83(3),
Chen, S. Y., & Tzeng, J. Y. (2010). College female and male heavy internet users' 693e710.
profiles of practices and their academic grades and psychosocial adjustment. Junco, R. (2012). Too much face and not enough books: the relationship between
Cyberpsychology, Behavior, and Social Networking, 13(3), 257e262. multiple indices of Facebook use and academic performance. Computers in
Cheung, G. W., & Lau, R. S. (2008). Testing mediation and suppression effects of Human Behavior, 28(1), 187e198.
latent variables: bootstrapping with structural equation models. Organizational Junco, R., & Cotton, S. R. (2011). Perceived academic effects of instant messaging use.
Research Methods, 11(2), 296e325. Computers & Education, 56, 370e378.
Chua, S. L., Chen, D. T., & Wong, A. F. L. (1999). Computer anxiety and its correlates: a Kalwar, S. K., Heikkinen, K., & Porras, J. (2013). Internet anxiety: myth or reality?
meta-analysis. Computers in Human Behavior, 15(5), 609e623. Human-computer interaction. Towards Intelligent and Implicit Interaction Lecture
Chu, A., & Mastel-Smith, B. (2010). The outcomes of anxiety, confidence, and self- Notes in Computer Science, 8008, 431e440.
efficacy with internet health information retrieval in older adults: a pilot Kannan, B., Muthumanickam, A., & Chandrasekaran, S. (2012). Computer anxiety
study. Computers Informatics Nursing, 28(4), 222e228. among higher secondary students. International Journal of Development
Cocorada
, E. (2015). The internet attitude with socio-humanities high-school and Research., 2(4), 1008e1011.
university students. Romanian Journal of Experimental Applied Psychology, 6(1). Kline, R. B. (2011). Principles and practices of structural equation modelling (3rd ed.).
Cocorada
, S. (2014). Computer anxiety, computer self-efficacy and demographic New York, NY: Guilford Press.
variables. In Proceedings of the 10th International Scientific Conference “eLearning Kubiatko, M., Hala kova, Z., Nagyova , S., & Nagy, T. (2011). Slovak high school stu-
and Software for Education” Bucharest, April 24-25, 2014. dents' attitudes toward computers. Interactive Learning Environments, 19(5),
Compeau, D. R., & Higgins, C. A. (1995). Computer self-efficacy: development of a 537e550.
measure and initial test. MIS Quarterly, 19, 189e211. Lee, C. L., & Huang, M. K. (2014). The influence of computer literacy and computer
Damean, D. (2008). Internet uses and gratifications among Romanian teenagers. In anxiety on computer self-efficacy: the moderating effect of gender. Cyberp-
M. Barbovschi, & M. Diaconescu (Eds.), Teenagers' actions and interactions online sychology, Behavior, and Social Networking, 17(3), 172e180.
in Central and Eastern Europe. Potentials and empowerment, risks and victimiza- Lepp, A., Barkley, J. E., & Karpinski, A. C. (2014). The relationship between cell phone
tion (pp. 9e28). Cluj-Napoca: Babeş-Bolyai University Press. use, academic performance, anxiety, and satisfaction with life in college stu-
Durndell, A., & Haag, Z. (2002). Computer self-efficacy, computer anxiety, attitudes dents. Computers in Human Behavior, 31, 343e350.
towards the internet and reported experience with the internet, by gender, in Liaw, S. S., & Huang, H. M. (2006). Information retrieval from the World Wide Web:
an east European sample. Computers in Human Behavior, 18, 521e525. a user-focused approach based on individual experience with search engines.
Durndell, A., Haag, Z., & Laithwaite, H. (2000). Computer self-efficacy and gender: a Computers in Human Behavior, 22, 501e517.
cross cultural study of Scotland and Romania. Personality and Individual Dif- Li, N., & Kirkup, G. (2007). Gender and cultural differences in internet use: a study of
ferences, 6(1), 1037e1044. China and the UK. Computers & Education, 48, 301e317.
Durndell, A., & Thomson, K. (1997). Gender and computing: a decade of change? Linnenbrink-Garcia, L., & Pekrun, R. (2011). Students' emotions and academic
Computers and Education, 28(1), 1e9. engagement: introduction to the special issue. Contemporary Educational Psy-
Eurostat. (2011). Internet access and use in 2011. Retreived from http://ec.europa. eu/ chology, 36, 1e3.
eurostat/documents. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002).
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An intro- A comparison of methods to test mediation and other intervening variable
duction to theory and research. Reading, MA: Addison-Wesley. effects. Psychological Methods, 7, 83e104.
Gibbs, S. (2013). Computer self-efficacy e is there a gender gap in tertiary level Madu, S. N., Otuka, C. C., & Adebayo, I. A. (2011). Attitude of male and female stu-
introductory computing classes? Journal of Applied Computing and Information dents towards the use of internet. Gender & Behaviour, 9(1), 3817e3826.
Technology, 17(1). Retrieved from http://www.citrenz.ac.nz/jacit/JACIT1701/ McCrae, R. R., Kurtz, J. E., Yamagata, S., & Terracciano, A. (2011). Internal consistency,
2013Gibbs _ComputerSelfEfficacy.html. retest reliability and their implications for personality scale validity. Personality
Hachbart, G., Grover, V., & Yi, M. Y. (2003). Computer playfulness and anxiety: and Social Psychology Review, 15(1), 28e50.
positive and negative mediators of the system experience effect on perceived Mikkelsen, A., Ogaard, T., Lindoe, P., & Olsen, O. (2002). Job characteristics and
ease of use. Information and Management, 40, 221e232. computer anxiety in the production industry. Computers in Human Behavior, 18,
Hasan, B. (2006). Effectiveness of computer training: the role of multilevel com- 223e239.
puter self-efficacy. Journal of Organizational and End User Computing, 18(1), Nickell, G. S., & Pinto, J. N. (1986). The computer attitude scale. Computers in Human
50e68. Behavior, 2, 301e306.
Heinssen, R. K., Glass, C. R., & Knight, L. A. (1987). Assessing computer anxiety: Orr, C., Allen, D., & Poindexter, S. (2001). The effect of individual differences on
development and validation of the computer anxiety rating scale. Computers in computer attitudes: an empirical study. Journal of End User Computing, 13(2),
Human Behavior, 3, 49e59. 26e40.
Henderson, R., Deane, F., & Ward, M. (1995). Occupational differences in computer Parasuraman, S., & Igbaria, M. (1990). An examination of gender differences in the
related anxiety: implications for the implementation of a computerized patient determinants of computer anxiety and attitudes toward microcomputers
management information system. Behavior & Information Technology, 14(1), among managers. International Journal of Man Machine Studies, 32(3), 327e340.
23e31. Powell, A. L. (2013). Computer anxiety: comparison of research from the 1990s and
Hong, K., & Koh, C. (2002). Computer anxiety and attitudes toward computers 2000s. Computers in Human Behavior, 29, 2337e2381.
among rural secondary school teachers: a Malaysian perspective. Journal of Sam, H. K., Othman, A. E. A., & Nordin, Z. S. (2005). Computer self-efficacy, computer
 A.-M. Cazan et al. / Computers in Human Behavior 55 (2016) 258e267 267

anxiety, and attitudes toward the internet: a study among undergraduates in construct. Journal of Applied Social Psychology, 34(3), 445e467.
Unimas. Educational Technology & Society, 8(4), 205e219. Werner, J. M., Carlson, M., Jordan-Marsh, M., & Clark, F. (2011). Predictors of com-
Scherer, R., & Siddiq, F. (2015). Revisiting teachers' computer self-efficacy: a puter use in community-dwelling ethnically diverse older adults. Human Fac-
differentiated view on gender differences. Computers in Human Behavior, 53, tors, 53(5), 431e447.
48e57. Wilfong, J. D. (2006). Computer anxiety and anger: the impact of computer use,
Schlenker, B. R., & Leary, M. R. (1982). Social anxiety and self-presentation: computer experience, and self-efficacy beliefs. Computers in Human Behavior,
conceptualization and model. Psychological Bulletin, 92(3), 641e669. 22, 1001e1011.
Scott, J. E., & Walczak, S. (2009). Cognitive engagement with a multimedia ERP Williams, J., & McCord, D. (2006). Equivalence of standard and computerized ver-
training tool: assessing computer self-efficacy and technology acceptance. In- sions of the Raven progressive matrices test. Computers in Human Behavior,
formation & Management, 46(4), 221e232. 22(5), 791e800.
Shakir, M. (2014). Academic anxiety as a correlate of academic achievement. Journal Wu, Y. T., & Tsai, C. C. (2006). University students' internet attitudes and internet
of Education and Practice, 5(10), 29e36. self-efficacy: a study at three universities in Taiwan. CyberPsychology &
Spielberger, C. D. (1972). Anxiety: Current trends in theory and research (Vol. I). New Behavior, 9(4), 441e450.
York, N.Y: Academic Press. Yan, Z., & Fischer, K. W. (2004). How children and adults learn to use computers: a
Sun, Y., Wang, N., Shen, X., & Zhang, X. (2015). Location information disclosure in developmental approach. New Directions for Child and Adolescent Development,
location-based social network services: privacy calculus, benefit structure, and 2004(105), 41e61.
gender differences. Computers in Human Behavior, 52, 278e292. Zhang, X., de Pablos, P. O., & Xu, Q. (2014). Culture effects on the knowledge sharing
Thatcher, J., Loughry, M., Lim, J., & McKnight, H. (2007). Internet anxiety: an in multi-national virtual classes: a mixed method. Computers in Human
empirical study of the effects of personality, beliefs, and social support. Infor- Behavior, 31, 491e498.
mation & Management, 44(4), 353e363. Zhang, X., de Pablos, P., & Zhang, Y. (2012). The relationship between incentives,
Torkzadeh, G., & Koufteros, X. (1994). Factorial validity of a computer self-efficacy explicit and tacit knowledge contribution in online engineering education
scale and the impact of computer training. Educational and Psychological Mea- project. International Journal of Engineering Education, 28, 1341e1346.
surement, 54(3), 813e921. Zhang, X., de Pablos, P., & Zhu, H. (2012). The impact of second life on team learning
Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology outcomes from the perspective of IT capabilities. International Journal of Engi-
acceptance model: four longitudinal field studies. Management Science, 46(2), neering Education, 28, 1388e1392.
186e204. Zhou, M. (2014). Gender difference in web search perceptions and behavior: Does it
Venkatesh, V., Morris, M., Sykes, T., & Ackerman, P. (2004). .Individual reactions to vary by task performance? Computers & Education, 78, 174e184.
new technologies in the workplace: the role of gender as a psychological