Abstract: To reveal the type and characteristics of attention quality of controllers, a 20 min continuous cancellation test was designed and carried out. The test data were analyzed by using the clustering analysis method. The error rate of reaction, the mean value of reaction times, and the variance of reaction times were chosen as the measurement parameters of attention span, attention stability, attention switching and distribution. The types of attention quality of controllers were classified by using the k-means clustering analysis method. The characteristics of attention quality of controllers were described and the differences between controllers and controller students were revealed according to the types of attention quality. The results show that 181 testees (controllers and controller students) could be classified into four types according to attention quality, that these outstanding controllers show large attention span, strong attention stability, rational attention switching and improper attention distribution, and that there is a significant difference between controllers and controller students in attention quality.

Key words: controller, attention quality, cancellation test, k-means clustering, cluster analysis