Geography Reference
In-Depth Information
4.4.1 Cognitively adequate animation design
We still know very little about how effective novel interactive graphical data depictions and
geovisualization tools are for knowledge discovery, learning and sense-making of dynamic,
multidimensional processes. Cognitive scientists have attempted to tackle the fundamental
research question of how externalised visual representations (e.g. statistical graphs, organi-
zational charts, maps, animations etc.) interact with people's internal visualization capabili-
ties (Tversky, 1981; Hegarty, 1992). Experimental research in psychology suggests that static
graphics can facilitate comprehension, learning, memorization, communication of complex
phenomena and inference from the depiction of dynamic processes (Hegarty, 1992; Hegarty,
Kriz and Cate, 2003). However, in a series of publications surveying the cognitive literature
on research on animated graphics (that did not include map animations), Tversky and
colleagues claim they failed to find benefits to animation for conveying complex processes
(Betrancourt, Morrison and Tversky, 2000; Betrancourt and Tversky, 2000; Morrison, Tver-
sky and Betrancourt, 2000; Tversky, Morrison and Betrancourt, 2002). In the cartographic
literature, results on animation experiments are not conclusive in part because they depend
on how 'success' is measured. For example, in think-aloud experiments comparing passive,
interactive and inference-based animations for knowledge discovery (e.g. differing in their
interactivity levels), Ogao (2002) found that animations did play a crucial role in facilitating
the visualization of geospatial data. Different animation types are used at specific levels of
the exploratory process, with passive animation being useful at earlier observatory stages
of the exploration process, and inference-based animation playing a crucial role at later
stages of discovery, such as in the interpretation and explanation of the phenomenon under
study. Similarly, participants in a study by Slocum
et al
. (2004) suggested in a qualitative
assessment that map animations and small multiples are best used for different tasks, the
former being more useful for inspecting the overall trend in a time series dataset, the latter
for comparisons of various stages at different time steps.
In studies reviewed by Tversky and colleagues, typical 'success' measures are either re-
sponse time, also known as
completion time
(a measure of efficiency) or
accuracy of response
(a measure of quality). In some experiments comparing map animations with static small
multiple displays participants answer more quickly, but not more accurately, with anima-
tions (Koussoulakou and Kraak, 1992); in other experiments they take longer, and answer
fewer questions more accurately (Cutler, 1998), or the time it takes to answer the question
does not matter for accuracy at all (Griffin
et al.
, 2006).
There is a fundamental problem with these kinds of comparative studies. To precisely
identify differences in the measures of interest, the designs of the animation and the small
multiples to be compared require tight experimental control to the extent that it might make a
comparison meaningless. Animations are inherently different from small multiples. Making
an animation equivalent in information content to a small multiple display to achieve good
experimental control may actually mean degrading its potential power. Animations are not
simply a sequence of small multiples. Good animations are specifically designed to achieve
more than just the sum of their display pieces.
As mentioned earlier, there are many design issues to consider for the construction of
potentially useful map animations. Dynamic geographic phenomena may not only change in
position or behaviour over time, but also in their visual properties (e.g. attributes). Moreover,
the observer or camera location may change in position, distance and angle in relationship
to the observed event. Lighting conditions that illuminate the scene and dynamic events may
Search WWH ::
Custom Search