Geography Reference
In-Depth Information
Universit¨t Berlin. Raw data (in ASCII) from different types of eye trackers can be
imported and processed based on user defined settings regarding the duration and
dispersion of a fixation. The tools available in OGAMA to visualise and analyse the
eye tracking data are similar as in BeGaze, but the software has the advantage that it
does not require a licence key.
The structure of the raw eye tracking data exported by BeGaze does not
correspond to the requested structure of that data by OGAMA. Consequently, the
authors have written a tool in JAVA that can do this conversion automatically:
SMI2OGAMA
. The statistical analyses are executed in a specialised software
package: SPSS. Because the (difference in) users
'
attentive behaviour is the main
interest of this user test, a fixation report will be exported from OGAMA. This text
file lists all fixation from all participant underneath each other, with its character-
istics (e.g. duration, x-position, y-position) on the same line.
In order to be able to visualise the values of the fixations in a grid of AOIs (see
section “Creating the Gridded Visualisation”), the authors have written a second
tool in JAVA,
fr2grid
, that takes the fixation report as input, places the fixations in
the corresponding cell in the grid (based on its x and y position) and calculates the
total number of fixations and total duration of fixations in each grid cells. This is
done separately for the twelve stimuli. The resulting grids are visualised in
(Microsoft Office) Excel, based on the
tool.
However, before analysing the recorded eye movements, the data had to be
verified and subsequently selected. This selection was based on the accuracy of the
calibration, the tracking ratio of the eye tracker and a visual verification of the
location of the fixations. This is explained in more detail in the next section.
conditional formatting
'
'
Data Selection
Not all data could be included in the analyses due to some problems in the
calibration phase, which could result in a too low accuracy, too low tracking rate,
or a shift in the data.
Calibration Accuracy and Tracking Ratio
When using the eye tracker in a standard set-up (attached to the 22 in. screen), the
acceptable accuracy is typically limited to 0.5
. Because this study places the eye
tracking device in a stand-alone setting, in front of a much larger screen and with
different relative distances, higher deviations in the calibration will still be
accepted. For this study the limit is placed on 1
. The justification of this values
takes the viewing distance and the task into consideration. The distance between the
user and the screen was 116 cm. A deviation of 1
in the participant
s calibration
would consequently result in a deviation of 2.02 cm on the screen. As the task is to
locate a label on the map, this can be verified within a 2.02 cm range. All deviation
'
