COMPUTER SELF-EFFICACY - Human-Computer Interaction and Management Information Systems: Foundations - page 229

problems. Both of these processes (vicarious experience and enactive mastery) are acting together
during a given support event because when users require technical help, they can observe both how
the support resource is diagnosing and suggesting solutions to their problems and how they must
enact those solutions to determine if they solve the problem. Thus, support and how it is provided
during a given event is influential in the development of users' SCSE judgments towards resolving
the same or a similar problem on their own in the future, along with other learning outcomes.
In turn, specific computer self-efficacy judgments in this research were theorized as influencing
the thought patterns (Martocchio and Dulebohn, 1994) and emotional reactions (Compeau and
Higgins, 1995b) of users. In other words, following a support event, individuals who develop higher
levels of SCSE beliefs as a result of the support provided will be more motivated towards learning
because they judge themselves as being more able (Martocchio and Dulebohn, 1994). Similarly, the
influence of support events on user satisfaction is also seen as partially taking place through the
mediating influence of users' SCSE judgments. Individuals who develop a lower sense of SCSE as
a result of the support event are likely to judge themselves as less able and this will have a negative
influence on their reactions to support events (Bandura, 1986; Gist and Mitchell, 1992). The evi-
dence for this relationship is indicated by the study of negative emotions, such as fear, anxiety, and
stress (Bandura, 1986, 1997). However, it is expected that the influence of CSE on more positively
oriented affective reactions, such as satisfaction, will also exist. Thus SCSE is represented in the
research as partially mediating the influence between the nature of support provided during the sup-
port event and user learning and user satisfaction. A mediating influence rather than moderating
influence was judged as most accurately representing Bandura's theory of the role of self-efficacy
on individual behaviors, learning, and affective reactions (Bandura, 1986, 1997) and is consistent
with prior IS research (Agarwal et al., 2000b; Compeau and Higgins, 1995a, 1995b; Johnson and
Marakas, 2000; Marakas et al., 1998).
On the basis of this theoretical development, and focusing specifically on support provided by
organizational help desks, Haggerty (2004) developed a causal model via a detailed, qualitative
study involving critical incident interviewing of support users and detailed observation of over
one hundred support users and the help desk analysts they contacted. This facet of the research
revealed how and why support influences users' specific computer self-efficacy development and
other learning outcomes, validating and extending the theoretical stance discussed above. Further,
the work utilized this understanding to develop empirical measures and test the causal model in a
subsequent survey study of almost three hundred computer users in two different organizational
settings. The findings provide significant insight into, among other outcomes, the development of
SCSE in everyday computer usage settings, outside of formal training/study sessions.
In the qualitative study, we found that the use of a well-structured problem-solving process;
detailed, yet non-jargon-filled, verbal explanations; and aspects of service quality such as patient,
helpful, and knowledgeable analysts form the basis of effective support. Conversely, support
providers who use trial-and-error approaches to diagnosis and solution attempts take too much
time; provide little if any explanations, or explanations with too many technical terms; act impa-
tiently or inexpertly; or are unhelpful by providing an ineffective support event.
We also found evidence as to how and why these characteristics of support influenced a user's
specific computer self-efficacy. A user's SCSE—in this setting defined as his or her belief in his
or her ability to resolve the same or a similar computer problem on his or her own in the future—
was qualitatively related to attributes of the service quality demonstrated by the analyst and the
language he or she used in providing explanations to the user. By acting impatiently, seeming to
be unhelpful, failing to provide explanations or using technical jargon, the analyst could intimi-
date the user and make him or her feel “like an idiot.” This was seen as having a strong potential