Estimates of trend direction/ vagueness and bias in map reading.

A test of map users' identifications of regional trend directions documents the imprecision and bias of directional estimates. Fewer than 40% of the azimuths marked on circular test maps by the subjects were accurate within 5^o. No relationship was readily apparent between accuracy and the complexity of the test surfaces symbolized as continuous-tone choropleth maps. Estimates of trend direction were biased toward the closest of 8 prototype cardinal and diagonal directions when the difference between correct and rptotyped directions was 10^o or less. Generalization to one of these 8 principal directions is an example in map reading of the vagueness common to numerous cognitive classifications. When long-term memory intervenes in map analysis and interpretation, this directional generalization might lead to further inaccuracy in the conceptualization of geographic relationships.-Author