Cause

From EvaluationWiki

The cause C (of an effect E) is a factor whose occurrence or presence in a particular context, explains the occurrence or presence of E. In many cases, this means that, given the circumstances, its presence is a sufficient condition for E; in many cases, this means that if C had not occurred, E would not have occurred, that is, C was a necessary condition for the occurrence of E. But, and it's a very big but, neither of the remarks in the last sentence are true in all cases, so they are no use as definitions or partial definitions of causation, though useful in explaining the notion. Effects are often over-determined and hence would have occurred even if the factor that did cause them to occur had not been present, so causes are not necessary conditions for their effects. And something that is the cause of E in a particular case could have occurred a hundred times in apparently identical circumstances without E occurring, so causes are not sufficient conditions for their effects. Many attempts have been made to tidy up these analyses but none work, and the matter is not just of philosophical interest, it intimately affects experimental design and evaluation designs. Cause is a truly basic concept, not fully reducible to other basic concepts, and is rooted in the psychomotor experience of manipulation acquired by the child before speech; and it can and is established throughout every person's life by direct observation in particular cases (by an eye witness to a shooting, for example), or by half a dozen non-experimental research investigations (autopsy, for example), or from theory-based analysis (cosmogony) and of course by 'true experiments' with control groups to which subjects are randomly allocated (RCTs). For more details, see "Causes, Connections, and Conditions in History" (Scriven) in the Handbook of Historical Methods in the Social Sciences (Sage, 2006).