Daniela Zambarbieri Universit di Pavia

1
Daniela ZambarbieriUniversità di Pavia
http//e-tracking.unipv.it
2
E-TRACKING Eye tracking analysis in the
evaluation of e-learning systems

• OBJECTIVES
• To define and validate a quantitative method,
  based on eye tracking analysis, for the
  evaluation of the level of functionality,
  usability and acceptability of e-learning systems
• To identify guidelines for the development of
  functional, usable and acceptable e-learning
  systems
• To develop optimised prototypes of e-learning
  systems based on the identified guidelines

3
E-learning

• Underlying assumptions
• the user actually looks at the screen in the
  region where the information of interest is
  presented
• that he remains in this region for a time long
  enough to acquire all relevant information
• that he explores all the parts of the screen
  where information is displayed

4
Eye Movements

• bring the image on the fovea
• maintain it stable
• saccades fixations scanpath
• WHERE WHAT FOR HOW LONG

5
From my position paper

• Concerning data recording, to make the state of
  the art on the available commercial
  instrumentation would represent a useful starting
  point. Moreover it could encourage producers of
  such instruments not only to improve the
  performances of their products but also to try to
  be competitive from a financial point of view.
• The second problem is related to data analysis.
• Each research group and industry usually develop
  their own SW (as we actually did at the
  University of Pavia).
• It should be very useful to critically discuss
  data analysis in terms of algorithms and
  parameters and their validity for the purpose of
  eye tracking applications.
• As a matter of fact, even if standardisation is a
  very ambitious goal to reach, at least a sort of
  quality assurance of the results provided should
  be a valuable result to reach among people
  interested in this kind of research and
  application.

6
Acquisition

• Eye movement
• Mouse
• Keyboard
• Scrolling

EyeGaze LC Technologies
7
Scanpath
8
Regions of Interest (RoIs)
9
Data Analysis

• pages
• RoIs

• permanence time
• mean fixation duration
• number of fixations
• scanpath length

• sequence of access to the RoIs

10
Analysis of accesses

• Sequence of accesses
• The first RoI accessed by subjects for each page
• The total number of accesses to the RoIs

11
2
3 - 5
1
4 6 - 8
7 - 11
9
10
12
Levenshtein distance

• Levenshtein distance or edit distance between two
  strings is given by the minimum number of
  operations needed to transform one string into
  the other, where an operation is an insertion,
  deletion, or substitution.
• Levenshtein distance (LD) is a measure of the
  similarity between two strings

13
First access
78
14
Total number of accesses
4/16
4/16
15
Conclusions

• Access analysis can add more insight in the study
  of exploration strategies
• In fact quantitative parameters such as
  permanence time, number of fixations etc. are
  static measures
• Sequence of accesses, first access and repetition
  of access are dynamic measures