Professional Documents
Culture Documents
2006 - JAH - Cyberloafing
2006 - JAH - Cyberloafing
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/7165535
CITATIONS READS
35 612
3 authors, including:
Gary Blau
Temple University
138 PUBLICATIONS 4,093 CITATIONS
SEE PROFILE
All content following this page was uploaded by Gary Blau on 31 January 2014.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
ORIGINAL ARTICLES
Using a primary sample of medical technologists (MTs) and Since then, cyberloafing has become an important
a second validation sample, the results of this study showed employee issue for many companies. In a survey of 224
initial support for a three-factor measure of cyberloafing. organizations in the United States, almost all of these
The three scales were labeled browsing-related, non-work- organizations had Internet access policies (IAPs) outlining
related e-mail, and interactive cyberloafing. MTs who per-
appropriate and inappropriate use of the Internet at work.7
ceived unfair treatment in their organization (i.e., lower
By contrast, in Singapore, fewer organizations have such
organizational justice) were more likely to exhibit all three
types of cyberloafing. MTs who did not care as much about IAPs.3 Despite the prevalence of these IAPs, more than
punctuality and attendance (i.e., higher time abuse) were 60% of these companies had disciplined employees for
more likely to display browsing-related and non-work- inappropriate uses such as accessing pornography, online
related e-mail cyberloafing. Finally, MTs who perceived an chatting, gaming, investing, or shopping at work.7 The pur-
inability to control their work environment (i.e., power- pose of this study is to further explore the construct of
lessness) were more likely to display interactive cyberloaf- cyberloafing2 by testing an expanded version of an online
ing. Study limitations and suggestions for future research survey cyberloafing measure initially used by Lim et al.2,3
are discussed. J Allied Health 2006; 35:9–17.
Construct Definition
9
Lim2 found a two-factor measure of cyberloafing in her misuse of company property, specifically the employee’s per-
study of 188 working adults in Singapore. She labeled factor sonal computer and related resources (e.g., Intranet). This
1 “browsing activities,” which consisted of eight items, for quadrant is more potentially serious or harmful to the
example, browsing sports-related Web sites, investment- organization than production deviance.9 Downloading
related Web sites, general news sites, and adult-oriented non–work-related information, posting messages on
(sexually explicit) Web sites. The second factor was labeled non–work-related items, or using the Internet to gain addi-
“e-mailing activities,” which consisted of three items: tional income at work are generally violations of use of
checking, sending, and receiving non–work-related e-mail. company property.8,14 Integrating the different types of
Surprisingly, however, Lim2 did not keep these factors sep- cyberloafing behavior measured across the previously cited
arate in subsequent analyses. Instead, she combined all 11 studies inductively suggests that the two more “passive”
items together into one general cyberloafing scale, which types of cyberloafing (i.e., browsing and e-mailing) both fit
had a reliability of 0.88. Lim’s2 cyberloafing scale was the into the production deviance DWB quadrant, and a third
only previously tested measure found in a literature search. more “interactive” type of cyberloafing (i.e., interactive
cyberloafing) fits into the more organizationally serious
An Expanded Cyberloafing Construct property deviance DWB quadrant.
within the DWB Typology
Relationships of Correlates to
Additional examples of cyberloafing behavior found in the Types of Cyberloafing
literature include moonlighting (using the Internet to gain
additional income), posting messages, downloading ORGANIZATIONAL JUSTICE
non–work-related information, using chat rooms, and game
playing.8,11,12 These examples fit within Lim’s2 general defi- Why would employees cyberloaf? From a theoretical per-
nition of cyberloafing (misuse of the Internet) yet were not spective, exchange theorists15,16 have suggested looking at
incorporated into her initial scale. All types of cyberloafing human interactions through the “lens” of social economics,
behaviors involve the employee sitting in a chair and star- that is, individuals being concerned about the ratio of
ing at their computer screen. However, these latter cyber- inputs they invest in a relationship relative to the outcomes
loafing items generally require more synchronous virtual or or rewards they receive. If employees perceive that they
live interaction, either with other people or the software have been treated unfairly (i.e., a lack of justice), they can
technology.13 In order to post messages, download informa- become upset and motivated to reciprocate by cyberloaf-
tion, play games, or use a chat room, the user is typically ing.2,3 Prior research suggests three types of organizational
required to have greater interaction either with the soft- justice17: distributive justice (perceived fairness in outcome
ware or other users. Such “interactiveness” typically allocations), procedural justice (perceived fairness of the
requires greater immediate effort and often takes longer to procedures used to determine outcome allocations), and
complete than browsing or a quick e-mail.13 Furthermore, interactional justice (perceived fairness of interpersonal
the user would have probably identified the Web site treatment). Adding all three types of justice together cre-
required for such interactive cyberloafing based on previous ates a general organizational justice measure. Lim2 found all
browsing-related activity. three types of justice to be equally negatively related to her
For example, if when initially browsing a non–work- general cyberloafing scale, which indicates that they can be
related Web site, an employee then decides to download combined. The general nature of this relationship is that as
information from this Web site, additional time and effort employees believe they are unfairly treated by the organiza-
are involved in doing this. Similarly, if an employee comes tion, they are more likely to cyberloaf.
across information found while browsing that he or she
then wants to post for others, additional interaction with TIME ABUSE AND POWERLESSNESS
the software is required. Online chatting and playing games
require two-way “live” or virtual interaction with others. It will be useful to examine differential relationships of cor-
Finally, using the Internet to moonlight or gain additional relates to types of cyberloafing. Finding evidence for such
income requires a higher level of interaction with either differential relationships represents a form of discriminant
the software or other users than e-mailing or browsing. E- validity.18 Such discriminant validity would further support
mailing is asynchronous or not immediately interactive, distinguishing among these three types of cyberloafing. One
because messages typically stay on one’s computer system differential correlate is time abuse, defined in this study as
and can be responded to at a later time. Browsing generally “the degree to which an employee is concerned about their
means surfing or looking and is generally one way and punctuality and attendance.” As noted earlier, punctuality
therefore not as “interactive.”13 and attendance behaviors fall in the same production
These additional cyberloafing behaviors seem to be deviance quadrant9 as general cyberloafing.2 Attitude
better captured in a different quadrant within Robinson toward one’s punctuality and attendance can be considered
and Bennett’s9 DWB typology, that is, property deviance, or a more “passive” type of work withdrawal.19 The general
loafing,” and this accounted for 9% of the variance. loafing, 0.75; non-work-related e-mail cyberloafing, 0.87;
Regarding the loadings across factors for items 8 (“play and interactive cyberloafing, 0.68. Test-retest reliabilities
online games”) and 9 (“download non–work-related infor- for the follow-up sample of 63 repeat respondents were as
mation”), it can be argued that a difference of at least 0.10 follows: browsing-related cyberloafing, 0.64; non-work-
across loadings, with the secondary loadings being mar- related e-mail cyberloafing, 0.70; and interactive cyberloaf-
ginal, can justify item classification on the “dominant” ing, 0.60. These test-retest reliabilities suggest that the self-
loading factor.30 Overall, the three-factor solution reported cyberloafing was fairly stable over the reported
explained 51% of the variance, which is consistent with 6-mo period. Overall, the results support hypothesis 1, that
other studies using exploratory factor analyses.31 These is, browsing-related, non-work-related e-mail, and interac-
items were then summed for each factor and used in subse- tive cyberloafing are three distinct types of cyberloafing.
quent analyses. Reliabilities for each scale were assessed
using Cronbach’s .28 Ideally, scale reliabilities should be at ADDITIONAL RESULTS WITH MT SAMPLE
least 0.70.28 The following ? levels were found for each
cyberloafing scale: browsing-related cyberloafing, 0.78; Table 3 shows the means, standard deviations, reliabilities,
non-work-related e-mail cyberloafing, 0.91; and interactive and correlations for study variables for MTs. Study variable
cyberloafing, 0.69. means are given based on the four-point response scale for
Confirmatory factor analyses were used on the valida- easier interpretation. Looking at the scale means for the
tion sample to test how well these 16 items “fit” the pro- three cyberloafing scales, non-work-related e-mail cyber-
posed three-factor solution compared with other factor- loafing has the highest mean of the scales (2.01), but all
solution models on a new set of data.32 The results across three have low means (browsing-related cyberloafing, 1.44;
different models are shown in Table 2. Significant model interactive cyberloafing, 1.22). For the validation sample, a
improvement is seen going from one factor (overall cyber- similar pattern of cyberloafing scale means was found, that
loafing) to two factors (non-work-related e-mail cyberloaf- is, non-work-related e-mail cyberloafing had the highest
ing and browsing-related cyberloafing) to three factors mean (2.19), followed by browsing-related cyberloafing
(non-work-related e-mail cyberloafing, browsing-related (1.64) and then interactive cyberloafing (1.43).
cyberloafing, and interactive cyberloafing). Adequate fit is Given that cyberloafing is an example of DWB, which
indicated when the fit indices are at least 0.90 and the root should have a lower frequency of occurrence,9 lower means
mean square measures are less than 0.08.33,34 Overall, these are to be expected. Scale reliabilities for study variables are
statistics indicate an acceptable fit and further support the generally good (i.e., >0.70), with only the interactive
distinctiveness of the three cyberloafing scales. The follow- cyberloafing scale being marginal (0.69). Correlations
ing scale reliabilities were found: browsing-related cyber- among the cyberloafing scales are moderate, with the high-
est being 0.50, indicating 25% (0.502) overlap between in browsing-related and non-work-related e-mail cyberloaf-
browsing-related cyberloafing and non-work-related e-mail ing, while those who perceived less control at work were
cyberloafing. A similar pattern of correlations among cyber- more likely to interactively cyberloaf. Because the time
loafing scales was found for the validation sample: r = 0.53 abuse and powerlessness data were collected 1 yr before the
(browsing with non-work-related e-mail cyberloafing), r = cyberloafing data, stronger causal inferences about these
0.51 (browsing with interactive cyberloafing), and r = 0.27 relationships can be made.18 However, these correlations
(non-work-related e-mail with interactive cyberloafing). were not strong. Several reasons may account for this.
Statistical texts28 generally use 0.80 as the correlation First, by using only a four-point frequency response scale
“cutoff” for indicating variable redundancy. These correla- for each cyberloafing item, this reduced the score range for
tional results support using each cyberloafing scale as a dis- items and subsequently formed scales. Range restriction in
tinct variable. variables can decrease correlations with other variables.28
Looking at correlational relationships, powerlessness had Future research should consider using a seven-point fre-
a positive correlation to only interactive (0.24) cyberloaf- quency response scale for measuring cyberloafing items: 1,
ing. Time abuse was positively related to both browsing- never; 2, hardly ever (about once every few months); 3,
related and non-work-related e-mail cyberloafing but not to rarely (about once a month); 4, occasionally (at least sev-
interactive cyberloafing. Organizational justice was nega- eral times per month); 5, sometimes (at least once per
tively related to all three cyberloafing scales, that is, brows- week); 6, frequently (at least once per day); 7, very fre-
ing-related cyberloafing (–0.19), non-work-related e-mail quently (at least several times per day). Greater measure-
cyberloafing (–0.16), and interactive cyberloafing (–0.13). ment precision should also help scale reliability,28 because
[Comp: set as minuses.] Organizational justice was the only the reliability of the interactive cyberloafing scale used here
variable significantly related to overall cyberloafing. These was marginal. In addition, gathering data on the amount of
correlation results provide additional support for hypothesis time spent on cyberloafing behaviors would be useful.
1, that is, that browsing-related, non-work-related e-mail, Cyberloafers who take frequent but short breaks to cyber-
and interactive cyberloafing are three distinct scales. loaf may still stay productive versus cyberloafers who spend
longer periods cyberloafing. This suggests that future
Discussion research should also examine the relationship between
cyberloafing and employee performance.
The exploratory and confirmatory factor analytic results of Another factor that may help explain the lower levels of
this study suggest that there are three related but distinct cyberloafing found and weaker correlational results may be
types of cyberloafing: browsing-related, non-work-related e- the nature of the samples used. Both the primary and vali-
mail, and interactive cyberloafing. Perhaps some might view dation samples used paper-and-pencil surveys to record
distinguishing among these types of cyberloafing as “splitting cyberloafing behavior as opposed to using an online survey,
hairs,” because they all collectively mean that an employee is as Lim2 did. Lim2 acknowledged the possibility that using
not being work productive. However, we believe that these such an online survey technique may build in a stronger
distinctions are meaningful. The first two types of cyberloaf- cyberloafing bias among online respondents. Two limita-
ing fit within Robinson and Bennett’s9 production deviance tions associated with the paper-and-pencil survey approach
DWB quadrant, while the third type of cyberloafing is better used here are eliminating respondents who did not have
placed in their property deviance DWB quadrant. To our easy work access to the Internet and respondents underesti-
knowledge, this is the first study that has found empirical mating their actual cyberloafing to show higher levels of
support for such cyberloafing scale distinctions. social desirability.35 However, results with a smaller subset
The correlational results also supported distinguishing of validation sample respondents did indicate that cyber-
between types of cyberloafing. Greater time abuse was loafing behaviors were fairly stable over a 6-mo period.
related to more browsing-related and non-work-related e- Given the large number of organizations represented by
mail cyberloafing, while greater powerlessness was related to the primary and validation sample respondents, IAPs from
more interactive cyberloafing. MTs who were more lax in such organizations were not gathered.7 [Author: Is the sen-
their attendance and punctuality were more likely to engage tence starting “Given the…” okay as edited?] Gathering
Note. N= 227. r > 0.13 significant at p < 0.05; r > 0.17 significant at p < 0.01 (two tailed). Correlations in bold indicate antecedents that are statistically
significant. [Author: Please check that the note in Table 3 is correct as edited.]
*Internal consistency scale reliabilities (in parentheses).
†Overall cyberloafing = (browsing + non-work-related e-mail + interactive cyberloafing)/3.
such policy information in future research may help to ately spreading computer viruses, or somehow sabotaging a
understand different levels of employee cyberloafing behav- company’s Intranet.8 Clearly, these hostile cyberloafing
ior, especially by using a large sample within one organiza- behaviors require higher levels of user information technol-
tion. For example, an organization’s culture may be more ogy proficiency. In addition, recent research has found that
permissive about some types of cyberloafing (e.g., non- certain personality characteristics, including loneliness/
work-related e-mail) versus other types (e.g., interactive).36 depression and diminished impulse control, are positively
Additional variance in cyberloafing behavior could have associated with problematic Internet use.40
been accounted for by incorporating more variables in the From a theoretical standpoint, these more hostile cyber-
research design. For example, Lim2 found that the “metaphor loafing behaviors would fit into another DWB quadrant
of the ledger” variable mediated the general relationship of (i.e., personal aggression) within the Robinson and Bennett
organizational justice to cyberloafing. The metaphor of the typology. Personal aggression involves “behaving in an
ledger allows employees to rationalize that they are “entitled” aggressive or hostile manner towards other individuals.”9
to cyberloaf because of their past good behaviors, which they Deliberately spreading computer viruses or somehow sabo-
can then “cash in.”37 Such a variable is part of a mental taging a company’s Intranet seems clearly intended to hurt
exchange16 that each employee develops within their own the productivity of other company employees. Specific
psychological contract with their organization. Employee noncyberloafing behaviors within this personal aggression
feelings of frustration or anger may have been useful to meas- quadrant include employees stealing from other employees,
ure, because angrier38 and more frustrated39 employees verbally abusing a customer, and endangering coworkers
exhibit greater DWB. Unfortunately, survey length con- with reckless behavior.9 Future research could test for three
straints prohibited such additional measures; however, these different levels of cyberloafing, each level fitting into a dif-
variables should be measured in future research. ferent DWB quadrant: browsing-related and non-work-
There are other ideas to think about for future research. related e-mail (production deviance), benign interactive
While browsing-related and especially non-work-related e- (property deviance), and hostile interactive (personal
mail cyberloafing are more “tightly defined and measure- aggression). Getting employees to self-report such
able,”2 the interactive cyberloafing scale is broader and is not extremely hostile cyberloafing behaviors would clearly be
as easy to measure. There may even be a kind of “pecking very challenging, if not impossible. It may be left to the
order” to different types of cyberloafing. For example, non- information technology staff vigilantly monitoring an orga-
work-related e-mail tends to focus on specific family and nization’s IAP during difficult times (e.g., downsizing-
friends. In some ways it may be the easiest to rationalize related layoffs or disciplinary action against specific
doing, as the higher means found here suggest. Browsing is employees) to “catch” such extreme employee behavior.
less personal and involves an employee deliberately going to Research on organizational aggression41 suggests that
specific non-work-related Web sites if the organization’s IAP social learning theory, or learning from others,42 can help us
allows them.7 An employee generally would have already to further understand cyberloafing as a DWB. If employees
first “browsed” a Web site to then engage in interactive choose others who exhibit deviant behaviors, specifically
cyberloafing (e.g., downloading non-work-related informa- cyberloafing behaviors, as role models, these employees are
tion, posting messages, and playing online games). Future more likely to behave in the same way.41,42 For example,
research is needed to test such a cyberloafing “pecking order.” Bacharach et al.43 found that coworkers’ permissiveness
The types of interactive cyberloafing measured here about drinking, as measured by perceived coworker drink-
were sample relevant but also fairly “benign” compared ing norms and coworker drinking behavior, was positively
with hostile interactive behaviors such as hacking, deliber- related to employee problem drinking. This suggests meas-