Archived here are older posts which are no longer relevant or were redundant.
1 post • Page 1 of 1
Are you compiling the performance statistics of each participant as a function of total number of movies viewed? I would expect that the sensitivity score of a participant would get better and better as they view more and more movies and "learn" from their errors. This assumes that the participants are keeping track of their scores and are reviewing their errors. Perhaps you could "weight" a participant's identified events with a parameter that is dependent upon where in the total number of movies viewed by that participant the event was identified.