To detect an effect of landscape context on an ecological response one must measure the landscape variable at the appropriate spatial extent around the response, i.e. at its 'scale of effect'. However, it is not clear what determines this scale or if we can predict it a priori. One hypothesis is that the scale of effect increases with the temporal scale regulating the response. We tested this, comparing scales of effect of road density and forest amount on wood frog fecundity, abundance, and occurrence estimated from egg mass surveys of 34 ponds. We predicted the following order for scales of effect: fecundity < abundance < occurrence. Scales of effect were different for the three responses, but did not vary in the predicted order. This suggests that scales of effect cannot be predicted from the temporal scale regulating different responses and should thus be estimated empirically, rather than 'guestimated' a priori.