Dear all,
I'm currently working on a study focusing on the effects of employees' perceptions of a specific attribute of their leaders on their performance. In most research, I see that demographic variables or constructs like job satisfaction have been used as control variables to prevent any alternative explanations between the IVs and DVs. In some research though, I see that those control variables are also regressed on IVs, in my case employee perceptions of leadership, and the research then includes the residuals to the regression with the DVs with the same control variables also controlling the DV besides the IVs.
It seems a bit confusing in here. Though I understand the logic of regressing the control variables to the IVs, I am still unclear on whether controlling the DVs only would not be sufficient to prevent the effects of control variables in overall?
What should be the general procedure? Is it always critical to control IVs and DVs separately?