Information Technology Reference
In-Depth Information
Table 11.3
Self-Efficacy Scores Across Time
Mean Scores for Self-Efficacy (Standard Deviation)
Pre-Training
Post 1
Post 2
Follow-Up
Group 1 (SS/WP: n
35)
General
4.11
a
(1.98)
5.28
b
(1.87)
5.47
b
(1.91)
5.25
b
(2.19)
WP
4.24
a
(2.12)
5.21
b
(1.97)
6.26
c
(1.67)
6.20
c
(2.44)
Spreadsheet
3.59
a
(1.96)
5.93
b
(1.44)
6.26
b
(1.67)
5.44
b
(2.65)
Group 2 (WP/SS: n 40)
General
4.14
a
(2.36)
5.11
b
(2.29)
5.25
b
(2.26)
5.33
b
(2.35)
WP
4.25
b
(2.33)
5.54
b
(2.22)
5.78
b
(2.43)
5.74
b
(2.31)
Spreadsheet
3.70
a
(2.31)
4.34
b
(2.35)
5.24
c
(2.33)
5.52
c
(2.61)
Source:
From Compeau (1992).
a,b,c
Means in the same row with the same subscript do not differ at the 0.001 level. No comparisons are made
across rows.
the other training had not yet occurred. Thus crossover effects, such as those observed by Agarwal
et al. (2000) seem to be more limited, perhaps moderated by proximal experiences. General com-
puter self-efficacy beliefs showed some malleability through training, but were only affected by
the first day of training (regardless of which software package was taught). This suggests that
GCSE may not be simply an average of other SCSEs, since if it were, it would be expected to
increase following both training sessions.
Summary
Based on this review of the nature of computer self-efficacy in terms of its definition, measurement,
and domain specificity, several important conclusions can be drawn. First, it remains essential that
IS researchers adapt the definition of computer self-efficacy to the task and context of study. Such
definitional clarity should ensure that the domain of the task, and therefore the self-efficacy judg-
ment, is clearly specified within the context of the research being undertaken. Understanding task
considerations may be a particularly important area to consider. This specificity in definition should
also ensure that the measure reflects a computer user's generative capacity to undertake a future
computer-based task behavior and not a component sub-skill. Finally, with respect to measure-
ment, it is essential that sample selection and measure adaptation attend to considerations of vari-
ability in response items so that underlying scales reflect the unidimensional nature of computer
self-efficacy.
INFLUENCE OF COMPUTER SELF-EFFICACY
Bandura (1986) argues that self-efficacy influences a variety of individual behaviors and emotions.
Within the framework of triadic reciprocality, computer self-efficacy is seen as influencing other
cognitions and emotions of individuals, including perceived usefulness, ease of use, and anxiety.
CSE is also seen as influencing individual behavior, including behavior choice (often operational-
ized as behavioral intention), performance, and effort/persistence.
Search WWH ::
Custom Search