Clarifying the validity of eye movement measures from various eye tracker types

Eye data quality has a large effect on choice of event detection algorithm, which in turn effects all measures of eye movements recorded for a study. Quality is effected by system used as well as by the individual characteristics and behaviours of the individual participant. A clear evaluation of algorithms and their thresholds across typical eye data quality profiles is lacking. Here we present first results from a larger study on eye data quality, selecting 20 participants, 4 eye tracking systems, and three commonly implemented event detection algorithms. We assess how eye movement measures from the same eye behaviour are effected by system noise characteristics, sample rate, and choice of algorithm. RESULTS