Action research is based on an empirical process, but the steps are often rushed through, ensuring they take place while still minimizing them. Action research is great for immediate problems that have a short time span in which an answer is needed to be found, but it is not the best process to use.
Action research could be based on an empirical process which is a stochastic process that describes the proportion of objects in a system in a given state. So it depends on if you are using the action research in this way if you are then it would be considered an action state.
Yes, action research is based on an empirical process. It is used to solve an immediate problem and typically takes a problem-solving approach based on asking questions and determining causes for problems, although there is some dichotomy as to how the results are viewed. The results of this research are often placed in an organizational context.
Action research is flawed in that it rarely has a control group. The activity being observed can be effected by many external factors that the observer may not be taking into account. Isolating the stimulus that is being targeted allows for an empirical evaluation. Control groups allow for a comparison to the study group with that one feature isolated. Action research is a good indicator for a more empirical process to be performed.