Information Technology Reference
In-Depth Information
Table 2.
Results for Hypothesis 2 (Influence of PST-Distance)
PST-distance cases percentage of correct answers perceived diculty
Mean
SD Mean
SD
1
19 80.2%
20.62
2.89
0.31
2
13 79.5%
10.35
3.30
0.26
3
15 76.9%
15.06
3.42
0.27
4
5 76.4%
76.38
3.47
0.54
5
6 64.9%
64.85
3.55
0.29
6
4 74.8%
74.80
3.47
0.36
7
2 79.6%
79.64
3.54
0.28
5.2 Results for Hypothesis 2
As expected, the covariate PST-distance (element interactivity) has an influence
on the percentage of correct answers (F
1,55
= 4.32,
p
=0
.
042). Additionally
there is a highly significant effect on perceived diculty (F
1,55
= 22.04,
p<
0
.
001). Table 2 shows the average percentages of correct answers and the average
perceived diculties across different PST-distances. Hypothesis 2 predicted that
PST-distance will be positively associated with cognitive diculty. Therefore,
Hypothesis 2 is supported.
5.3 Results for Hypothesis 3
79.9% of the questions about two activities with a cut-vertex between them
have been answered correctly, compared to 75.8% of the questions about two
activities without a cut-vertex. Although the difference between means shows in
the expected direction, the results of the ANCOVA indicate that this difference
is not statistically significant and that there is also no significant influence on
perceived diculty. Moreover, there are also no interaction effects of “element
separateness” (cut vertex) with “relations between elements”.
Therefore, there is not enough evidence to support Hypothesis 3, which ex-
pected that the presence of a cut-vertex makesiteasiertoansweraquestion.
6 Discussion
This study provides empirical results on the influence of different relation types
of elements, their interactivity and their separateness on cognitive diculty of
understanding the relation between elements.
In line with our predictions in Hypothesis 1, we found that different control
structures in a BPM (like order or concurrency) differ according to their diculty
to be understood. Our results are in line with [18]. However, results are not
directly comparable, as we used different wordings of possible understandability
questions based on possible issues concerning ambiguousness (see [24] for details)
and consistently addressed two model elements in the questions.
