Error Correlations in a Two-Dimensional Compensatory Tracking Task

Abstract

Earlier studies utilizing two-dimensional compensatory tracking have generally computed either a single dimension average error or have taken the simple sum of the two average errors each computed separately. Results of the current study indicate that average error computed on the more highly variant of the two dimensions accounts for approximately 85 to 90% of the variance of the simple sum average error. For most experiments, the reduction in the number of computer amplifiers required may justify the use of the single dimension average error.