Abstract
A series of experiments on short-term memory is summarized. The results imply that when possible the following rules should be observed in presenting information to an operator who must keep track of a changing situation: (a) Each variable of which he must keep track should have its own exclusive set of possible states. (b) There should be few variables with many possible states, not many variables with few states, (c) A variable should not change state any more often than necessary. Three other conclusions may be useful to system designers as background information: capacity for random information is low; regularity within the sequence of states assumed by an individual variable is not particularly helpful; and orderly relations among the present states of different variables can be very helpful, at least in the extreme cases that were considered.
Keywords

This publication has 5 references indexed in Scilit: