Open brandonhackney opened 7 months ago
Statistical test would be a correlation of time-on-target to the 1-5 rating of how well it was understood
Statistical test would be a correlation of time-on-target to the 1-5 rating of how well it was understood
analysis
already performs such a correlation, as long as the metric is a single number. So we just need to implement it in selectMetric
: the percentage of the video's duration (i.e. not the raw time on target) the eye falls on ANY target, without discrimination. But also allow the eye to be within a certain radius of the character, to allow for some measurement error.
Next question: what should that radius be?
We NEED to include some sort of analysis that utilizes the 'actor' XY-coordinate data from the TriCOPA videos. e.g. measuring total (or average) duration the eyes were on (or within a specific area around) any (or a specific) character in each video.
selectMetric
that compares eye position (from EDF) to ROI position (from above)