The basic method of quantitative social research is the analysis of re
lations between so-called variables. This article indicates by using s
ome examples the explanatory deficiency of this method, and deals with
the question of why an action-theoretical background explanation is n
ecessary. Above all it is shown that statistical models of variable re
lations are themselves to be explained and cannot play the role of an
explanans. Finally, attention is drawn to some socio-structural condit
ions of Variable Sociology, which have recently taught us that an acti
on-theoretical foundation for Variable Sociology needs to be recognize
d.