Positioning graphical objects on computer screens: a three-phase model

This experiment identifies and models phases during the positioning of graphical objects (called cursors in this article) on computer displays.

The human computer-interaction community has traditionally used Fitts' law to model selection in graphical user interfaces, whereas human factors experiments have found the single-component Fitts' law inadequate to model positioning of real objects.

Participants (N=145) repeatedly positioned variably sized square cursors within variably sized rectangular targets using computer mice. The times for the cursor to just touch the target, for the cursor to enter the target, and for participants to indicate positioning completion were observed. The positioning tolerances were varied from very precise and difficult to imprecise and easy.

The time for the cursor to touch the target was proportional to the initial cursor-target distance. The time for the cursor to completely enter the target after touching was proportional to the logarithms of cursor size divided by target tolerances. The time for participants to indicate positioning after entering was inversely proportional to the tolerance.

A three-phase model defined by regions--distant, proximate, and inside the target--was proposed and could model the positioning tasks.

The three-phase model provides a framework for ergonomists to evaluate new positioning techniques and can explain their deficiencies. The model provides a means to analyze tasks and enhance interaction during positioning.