COMPUTER SELF-EFFICACY - Human-Computer Interaction and Management Information Systems: Foundations - page 226

users and technical individuals highlighted two important findings. A small portion of the enterprise
system users did not view their failed computer-related performance attainments as their individual
performance; instead, they reported the performance failures as computer technology performance
issues. Individuals' definitional views of the causal attributions of successful performance of com-
puter technology varied depending on whether the interviewees were users of the enterprise system
or technical individuals. Specifically, enterprise system users considered effort and persistence as
positive causal attributions of their performance behaviors, whereas technicians perceived effort and
persistence as very negative attributions that reflected possible design problems with the computer
technology.
Enterprise system users viewed their causal attributions of successful performance as internal
and self-controlled dimensions, whereas their causal attributions of failed performance were
external and unstable dimensions. Respondents appeared to advance self-fulfilling perspectives
(i.e., took personal credit) for successful performance, and offered self-protecting viewpoints
(i.e., denied responsibility for their behavior by attributing failed performance to implementation
factors such as lack of training and computer support) for unsuccessful performance.
Several implications follow from this research. Attributions ascribed to users' enterprise sys-
tem performance and the causal dimensions are context-specific. Kelley et al. (1999) suggest that
Weiner's four achievement causes do not comprehensively represent causal perceptions in all IS
contexts. The context-specific findings reported by Kelley et al. (1999) are supported by other IS
studies. Specifically, these IS studies found that computer-related performance levels are attrib-
uted to bad luck, system quality (Hufnagel, 1990), task difficulty, effort, and instructor support
(Henry and Martinko, 1997; Henry et al., 1993).
One aspect of the research investigating users' retrospective reasoning of their new technology
performance is the investigation of the influence of the causal attributions ascribed to both suc-
cessful and failed performances on the formation of computer self-efficacy. Kelley et al. (1999) uti-
lized a causal model to test the associations of attributions ascribed to successful performance and
attributions attributed to unsuccessful performance on CSE. Their results indicated that 18.6 per-
cent of the variance in CSE was explained by causal attributions of successful and failed computer
performances. The path from attributions attributed to successful computer-related performance to
CSE was positive and significant. Moreover, causal attributions ascribed to failed computer-related
performance had a significant negative effect on CSE. These results suggest that successful causal
attributions are associated with heightened self-efficacy beliefs and that failed causal attributions
are associated with lower self-efficacy beliefs. As theorized by both Weiner (1985) and Bandura
(1997), individuals' retrospective judgments of their performance attainments predict subsequent
performance attainments and have motivational effects.
Bandura (1997) reported results from prior studies (e.g., Relich et al., 1986; Schunk and Rice,
1986) that showed that the influence of causal attributions on achievement outcomes is mediated
by changes in self-efficacy judgments. Bandura (1997) stated that “the more the arbitrary reasons
raise efficacy beliefs, the higher are the subsequent performance attainments” (p. 123). In a new
computer technology environment, Kelley et al. (1999) found significant, indirect paths between
successful and failed causal attributions and computer self-efficacy, which in turn significantly
influenced enterprise system users' perceived overall PeopleSoft satisfaction, PeopleSoft per-
formance, and PeopleSoft productivity.
Other research in this area examines the relationship between users' retrospective reasoning
about their successful and failed computer performances and other behavioral and learning outcomes.
Attribution theory has been applied in studies aimed at explaining IS adoption (e.g., Martinko
et al., 1996), individuals' reactions to IS (e.g., Rozell and Gardner, 2000), and skill acquisition