Biomedical Engineering Reference
In-Depth Information
statistical power, other posthoc comparisons may also have been significant, which
would have led to the same pattern of results as for navigational performance in that
experiment, and all metrics for Experiment 2 of that study. However, the contradic-
tion between the findings for distance estimates [
28
,
29
] remains unexplained (if the
findings had been the opposite way around then they could have been explained by
environment layout, which was orthogonal [
28
] versus oblique [
29
]; see previous
section).
Both Full and Trans groups have the benefit of translational body-based infor-
mation, which accounts for the significantly more accurate direction and distance
estimates made by those groups than Vis and Rot groups in both experiments
of Ruddle et al. [
17
], and the significantly better navigational performance of
the Full group in Experiment 1 of that study. The lack of an effect of transla-
tional body-based information in Experiment 2 may be because the increased (and
ecologically more valid) spatial extent meant that participants had considerably
more time to process visual information as they navigated, so body-based infor-
mation made less contribution to their development of spatial knowledge. Alterna-
tively, it is possible that the environment was not complex enough for a statisti-
cally significant effect of translational body-based information to occur (a ceiling
effect).
Lastly, Suma et al's Experiment 2 also used whole-environment acquisition
and a large-scale environment [
14
]. The Full group collided with the VE's walls
less often than the Vis group, but the difference in the distance the groups trav-
eled (greater for the Full group) in the time that each participant was given may
have been due to either an inbuilt speed restriction or insufficient practice with
the virtual travel interfaces that were provided for the Vis and Rot groups. Met-
rics involving a cognition questionnaire and a map placement test produced non-
significant results between the groups, which is common for these tests' lack of
sensitivity.
5.5 Summary and Conclusions for VE Applications
So how can the findings of experimental studies of body-based information inform
the design of navigation interfaces for VE applications? Table
5.3
summarizes the
answer to this question from a navigation perspective, taking into account the need
for users to maneuver and develop of spatial knowledge, but does not attempt to
consider other factors such as cost.
5.5.1 Model-Scale Environments
In applications that use environments which in spatial cognition terms are model-
scale then users need to be able to look around, but make only localized adjust-
