Explore chapters and articles related to this topic
Inferential Control and Model Identification
Published in Brian Roffel, Patrick Chin, Computer Control in the Process Industries, 2017
The main reason for controlling an inferential variable is that the variable that should be controlled cannot be measured directly or there is not sufficient economic justification for its measurement. Analyzers, for example, are still relatively expensive and require a considerable amount of maintenance, which may force us to infer the required measurement. Also process conditions, e.g., high temperature and high pressure, may be such that actual process variable measurement is difficult and it is easier to infer the measurement.
Working safety and workloads of Chinese delivery riders: the role of work pressure
Published in International Journal of Occupational Safety and Ergonomics, 2023
Qi Zheng, Jing Zhan, Xiliang Feng
Based on the traditional approach for testing mediating variables proposed by Baron and Kenny [41], the regression equations are established as follows: where Pressurei = mediating variable; Zi = control variable; c1 = total effect of the quantity of weekly orders delivered on the occurrence of occupational injuries; α1 = effect of the quantity of weekly orders delivered on the mediating variable (work pressure); b1 = effect of the mediating variable on the choice of the occurrence of an occupational injury after controlling the other variables; c1′ = quantity of weekly orders delivered on the occurrence of an occupational injury after controlling the mediating variable. The mediation effect is usually tested using three methods, i.e., the stepwise method, the Sobel test and the bootstrapping method. The stepwise method is adopted in this study, after which the bootstrapping method is also used to test the significance of mediation effects.