In perceptual filtering tasks, congruency effects vary as a function of proportion congruent (PC), with smaller congruency effects when congruent trials are rare than when they are frequent. This effect is typically larger with extreme differences between high and low proportion congruent conditions (e.g., 80% congruent-20% incongruent) than with intermediate differences (e.g., 60% congruent-40% incongruent; Logan & Zbofroff, 1979; Blais & Bunge, 2010). Some authors have claimed that both PC effects can be explained in terms of the same reactive cognitive control mechanism that is responsible for sequential congruency (SC) effects (e.g., Botvinick, Braver, Barch, Carter, & Cohen, 2001). In fact, in most previous studies there was a systematic confounding between proportion congruent and the proportion of transitions involving an incongruent trial followed by another incongruent trial. In the present study we eliminated this confound and tested directly whether PC effects can still be measured in the absence of SC effects. Once confirmed, we studied the properties of this pure form of PC effect, in particular whether it is conflict-type specific or general, and whether it decreases gradually as a function of changes in proportion congruency (80% vs. 70% vs. 60%). Our results showed significant PC effects in the absence of SC effects, which replicates our previous findings (Torres-Quesada, Milliken, Lupiáñez, & Funes, 2014), and PC effects that can be conflict-type general or specific, depending on the nature of conflict type where they were produced. Importantly, the congruency effect was modulated by the level of proportion congruent, decreasing systematically as the absolute percentage of incongruent trials decreases.