Search for noisy tracks camouflaged by noisy backgrounds in colour or black and white

Abstract

Four separate groups, each of 12 women members of the APU subject panel, searched for coloured target tracks. Four more groups searched without colour. A complete target track was a sloping line comprising 16 dots, but up lo 12 of the dots could be missing. In the conditions with colour, red target tracks sloped steeply down lo the right, yellow less steeply down, green sloped up to the right, blue steeply up. The target tracks were camouflaged by 1%, 5%. 15% or 35% of noise dots in the background. In the conditions with colour, 10% of the noise dots present were red, 40% yellow, 40% green, and 10% blue.

When all 4 kinds of target track had to be searched for, the average detectability as measured by the d' statistic of signal detection theory was reliably greater for the brighter target tracks, whether coloured or not. The average d' was also reliably larger with more target dots and with fewer noise dots. For a constant S/N ratio, the average i/’ was reliably larger with complete or almost complete target tracks in noisy backgrounds, than with incomplete target tracks in less noisy backgrounds, both with and without colour.

Colour reliably increased the average d' by reducing the number of noise dots in the background of the same colour as the target dots. However, the average increase in d' was reliably larger still with the colourless displays when a comparable reduction was made in the total number of noise dots in the background. For the displays without target tracks, there were only about half as many false detections with colour as without. Search for only the steep target tracks sloping down to the right gave reliably larger average d' on these target tracks, both with colour and without, than search for all 4 kinds of target tracks