COLOR AND SHAPE CODING EFFECTS ON PROCESS CONTROL TASKS

NEWSLETTER ARTICLE

Creating a CRT-based display system for a processing unit can be a formidable task, even to those who are "experts" at doing so. Problems are often found when operators try to use the display system. One frequently occurring problem is that the operator cannot quickly find information or has difficulty discriminating between different types of information. To alleviate the problem, display designers resort to coding similar information in a like manner. Obviously, there exist numerous ways in which information can be coded in a display system. Recent research has found that some coding strategies work better than others.

One recent study, "Coding effects on a process control task with uniparameter and multiparameter displays", Human Factors, 1990, 32(8), 287-297, investigated the effects of color and shape coding on search and identification tasks. In the study, subjects monitored parameters, such as temperature and oil level, which were coded following four different strategies. Examples of each strategy are as follows:

1) shape coding: high temperature = triangles, low temperature = circles

2) color coding: high pressure = red, low pressure = blue

3) color + shape coding: green circles = energized, brown diamonds = not energized

4) partially redundant color coding: red and orange circles = high flow, blue and gray squares = low flow

In the experiments, investigators tracked subjects' performance on search and identification tasks. Search tasks were characterized by providing subjects with general statements such as, "Temperature is normal on all three indicators," shortly after which subjects were presented a display containing indicators showing different temperature indications. After reading the statements and viewing the display, the subjects would respond by typing a T or F for true or false. Identification tasks, on the other hand, presented subjects with statements such as "Temperature is high on indicators D and L." Subjects would then respond to the question in the same manner.

Results of the study indicated that for search tasks, subjects could find objects more quickly when color + shape or color coding was used versus shape coding or partially redundant color coding. The strategies of color and color + shape coding were found to mitigate the detrimental effects of increased display loading in search tasks. Coding, however, had no effect on the subjects' performance for identification tasks. Subjects could not perform identification tasks any more quickly when the displays were color and/or shape coded.

Other studies have corroborated the importance of color coding for search tasks over other forms of coding. One study, Zwaga and Duijnhouwer (1984) reported finding that color was a more important coding attribute than shape, and that adding shape coding to an object that was already color coded provided no significant performance improvement.

One result of the experiments sheds light as to how humans process information. It was found that partially redundant color coding and shape coding produced slower reaction times for search tasks than color or color + shape coding. This finding can be explained in the way people mentally process color information versus shape information. It is a fairly well established fact that colors are mentally processed in parallel, whereas shapes are processed in series. For color + shape and color coding, subjects could rely on the color to determine the state of the object (shorter reaction time). For partially redundant color coding, it is suggested that subjects first processed the color and then processed the shapes in a series within the target color (longer reaction time). Likewise, for shape coding, processing the shapes in series lengthened the time to identify the states of the process.

The study results hold important implications for the chemical processing industry. Annunciator panel design is an area that could benefit from application of color coding. Adding a color coding scheme could greatly aid in reducing operator reaction time when obtaining process status.

Copyright 1991 Beville Operator Performance Specialists, Inc., All Rights Reserved