Full Idea
What makes p a cause-factor relative to the effect-factor q (rather than vice versa) is the fact that by manipulating p, producing changes in it 'at will', we could bring about changes in q.
Gist of Idea
p is a cause and q an effect (not vice versa) if manipulations of p change q
Source
G.H. von Wright (Logic and Epistemology of Causal Relations [1973], §8)
Book Reference
'Causation', ed/tr. Sosa,E. /Tooley,M. [OUP 1993], p.118
A Reaction
As a solution to the direction-of-causation problem, I suspect that this proposal is begging the question. Will a causal explanation be offered of the action of manipulation? If he mistook his manipulation for a cause when it is actually an effect...