A modification of the compound stimuli paradigm has been used to measure the impact of a certain single element on the local-to-global effect and to compare the measured impacts of central and non-central elements matched on diagnosticity. In addition to global letters made of identical response-associated elements, some global letters comprised of only one response-associated element at a specific location (with all other ones being response-neutral), and in some other global letters that critical element was rather response-neutral (with all other ones being response-associated). Experiment 1 showed that the contribution of a central element that served as a distinctive feature was as large as the joint contribution of all other elements. Experiment 2 (as well as Experiment 4) showed that, in contrast, a non-central element that served as a distinctive feature did not contribute at all to the effect. Experiment 3 showed that the contribution of a central element was still as large as the joint contribution of all other elements even when it was completely irrelevant for selecting the response.