Search

Rasch [4] considered a logistic model for item analysis where the probability of a subject solving an item in a test is given by the probability δζ/(1 + δζ), where 5 characterizes the ability of the person and ζ characterizes the easiness of the item. The same model was used by Sanathanan [6] in a visual scanning set up where a "scanner" takes the role of an "item" and an "event" that of a "subject." In the visual scanning model it is further assumed that the event parameter is a random variable. Some properties of the logistic model in relation to both the item analysis and scanning problems are studied here. [ABSTRACT FROM AUTHOR]