Information Technology Reference
In-Depth Information
software. However, the challenge is that as software packages become increasingly comprehensive
(e.g., MS Excel includes database management features; both Excel and PowerPoint include graph-
ing features, etc.), there is less certainty that the task is defined by the software package. As soft-
ware becomes more interrelated and interchangeable, it may become more removed from the work
task. This suggests that it is perhaps necessary to be more precise in our CSE domain specificity,
in order to reflect the tasks being performed by computer users that form the basis of their CSE
judgments.
Thus, we need to think carefully about the domain of self-efficacy judgments and the relation-
ship between task and technology in formulating our research and measurement contexts. Our
definitions of GCSE and SCSE within IS have been, perhaps, too driven by technology consider-
ations and not sufficiently driven by task considerations.
A second issue with regard to the definition of domain is the relationship between the general
and specific computer self-efficacy judgments. Marakas et al. (1998) argue, as noted above, that
GCSE can be thought of as a weighted average of a collection of SCSE judgments.
Agarwal et al. (2000) have examined the relationship between general computer self-efficacy
and specific computer self-efficacy. They report on the relationship between GCSE and SCSE
formed during computer training. They show that GCSE influences the initial development of
SCSE (i.e., self-efficacy for the first software package learned) but not for SCSE developed for
the second software package learned. They also showed the existence of cross-over effects between
the specific software efficacies.
Compeau (1992) also discussed the evolution of GCSE and SCSE judgments throughout a
classroom training program. In their program, students learned two application software packages
(a word processor and a spreadsheet). This contrasts with Agarwal et al. (2000) whose subjects
learned Windows followed by a spreadsheet, which can be thought of as a progression from begin-
ning to more advanced topics (for novice computer users). But the Compeau study looked at soft-
ware packages that did not explicitly build on one another. Their results, while similar to those of
Agarwal et al. (2000) suggest a slightly different pattern.
Compeau (1992) reported the mean of GCSE, SCSE for word processing, and SCSE for spread-
sheets at four different points in time (before training, after training in the first package, after train-
ing in the second package, and after the completion of the training program) for two groups of
subjects. The first group learned the spreadsheet followed by the word processor; the second group
learned the word processor followed by the spreadsheet. The results are reported in Table 11.3.
Repeated measures analysis of variance was conducted to see whether computer self-efficacy
had changed through the course of training. Six one-way analyses of variance were conducted, for
general self-efficacy, word processing self-efficacy, and spreadsheet self-efficacy for each of the
two training groups. The overall tests showed significant differences for all of the self-efficacy
measures. Follow-up tests were conducted using the Neuman-Keuls post hoc test.
For the group that learned the spreadsheet first, GCSE scores increased between the pre-training
measure and the first post measure (after spreadsheet training), and then stayed constant from the
first to the second post measure (after word-processing training) and to the follow-up conducted at
the conclusion of the entire session. SCSE for word processing increased both following spread-
sheet training and word processing training. It did not change from the second post measure to the
follow up. SCSE for spreadsheets increased following spreadsheet training, and then stayed con-
stant following word processing training through the follow-up measure. An equivalent pattern
was observed for the group who learned word processing first.
These patterns suggest that training in either package influenced specific computer self-efficacy
with respect to the package. SCSE regarding the other package in the study increased, but only if
Search WWH ::
Custom Search