SOLVING GEOSPATIAL TASKS IN 2D AND 3D VISUALIZATIONS OF MAPS - Information Visualisation: Techniques, Usability and Evaluation - page 245

times, subjects responded faster when working with the 2D visualization
( ~ =9.74 sec.) than in 3D ( ~ =11.21 sec.).
Again, the hypothesis of equal means of times in the different
visualization conditions was tested. Time observations in these kinds of
psycho-physical experiments are known not to be normally distributed,
which was confirmed to be true in experiment I.
Due to the similarity of the stimuli and task in experiment II, non-
parametric testing methods were here chosen straight off for statistical
analysis.
In subsequent analyses of response times, a Wilcoxon rank sum test
was employed on the non-transformed time observations. For the
comparison of 2D and 3D, no significant differences were obtained (n=18,
W=501, p-value=0.0992).
Learning effects
Regarding potential effects of learning between successive trials, a two-
way ANOVA was performed on the accuracy observations with
visualization and presentation order as factors. This analysis did not reveal
any significant effect for either factor, nor any interaction effect. As the
ANOVA is not applicable for the observed times, times from the groups
that started 2D or 3D, respectively, were compared using a Wilcoxon rank
sum test. This comparison also revealed no difference between trials.
Significant learning effects could, in part, be observed within both
visualization conditions in a block-wise comparison of the observations
for block pre- (stimuli 1-13) and block post- (stimuli 14-25).
Although the boxplots for accuracies in Fig. 12.9(c) suggest better
accuracy in the second block of stimuli in both visualizations, those
differences were neither significant for the 2D condition nor for the 3D
condition. However, in regard to response times, clear improvements in
the post blocks can be seen for both visualization conditions (see Fig.
12.9(d)).
The apparent reductions of times between blocks were significant both
in 2D (n 1 =13, n 2 =12 , W=238, p-value=0.01557) and in 3D (n 1 =13, n 2 =12,
W=281, p-value<0.001).
Personal abilities and skills
The answers from the questionnaire responses for different subject groups
were compared. Among 18 subjects, n 1 =12 rated their spatial vision