The nature and measurement of conspicuity

B L Cole, S E Jenkins · 1980

Although conspicuity has a well understood colloquial meaning there is no agreed operational definition, nor an agreed method of measurement. This paper proposes an operational definition. A conspicuous object is defined as one that will, for a given background, be seen with certainty within a short observation time (ie, without search) regardless of the location of the object in relation to the line of sight. This definition implies that the primary measure of conspicuity should be the probability of seeing of the object when observation time is short (250 ms) and the angle between the direction of the object and the direction of fixation is specified. This angle is known as the eccentricity. Alternative measures of conspicuity are considered. Reaction time and search time are shown to be equivalent measures at least for conspicuous targets and offer the advantage of distinguishing between the conspicuity of objects that have a probability of seeing close to 1.0. However, their use complicates the definition of conspicuity by implying an additional criterion (minimum response time) and the end points of reaction/ search time scales are uncertain. Analysis of eye movement behaviour offers the prospect of an objective measure but was shown to have a poor correlation with the primary measure of conspicuity. Another alternative measure is conspicuity area. This measure is derivative of the primary measure of probability of seeing and in effect uses the angle of eccentricity as the scale, such that conspicuity is specified as the eccentricity beyond which the probability of seeing is less than some specified value. This measure relates to the operational definition proposed, has greater generality than the primary measure which is specific to a given eccentricity, has an easily conceived meaning and offers a scale that, unlike reaction and search time, has clearly defined end points. Its use is recommended for the investigation of conspicuity.

Read the paper · More papers on PaperTik