TY  - JOUR
PY  - 1937//
TI  - Illusory trends in the observations of bar graphs
JO  - Journal of experimental psychology
A1  - Graham, J. L.
SP  - 597
EP  - 608
VL  - 20
IS  - 6
N2  - In this bar graph analysis, six variables were analyzed: horizontal or vertical mode of presentation, coarseness or fineness of the scale units, spacing distance between the bars, width of bars, bar length, and nearness or remoteness from the scale. The factors compared were equated either in identical comparisons or in groups where the factors and degrees were equated (but not in identical combinations of degrees). The probability of an illusory trend holding consistently was expressed in terms of its ratio to the sigma of the difference. Results reveal that the factors which should give the maximum amount of overestimation in combination are short bars, presented with wide spacing distances, coarse scale units, and narrow bars viewed vertically if the most important comparison is a long bar, and horizontally if it is short. The bar positions remote from the scale, also, give greater overestimation. (PsycINFO Database Record (c) 2012 APA, all rights reserved)<p /> 
LA  - 
SN  - 0022-1015
UR  - http://dx.doi.org/10.1037/h0060831
ID  - ref1
ER  -