Feature Integration Theory (Treisman & Sato, 1990) assumes that targets defined by multiple features (conjunction) increase attentional demands at the perceptual level of processing (perceptual load) due to a feature integration process, relative to targets defined by a single feature. Consistent with this, Lavie (1995) showed that interference by a response incompatible distracter (RID) was less when discriminating the presence of a conjunction compared to a single-feature, indicating that the former demanded more attentional resources and hence reduced the resources available to process the distracter. However, a failure to equate discriminability between the dimensions defining the conjunction and single-feature targets may account for these results. We compared RID between conjunction and single-feature targets while equating discriminability between the dimensions defining each target type. Displays consisted of probe and distracters that varied in color (red, green) and shape (H, K), with shape being less discriminable. Observers judged the presence/absence of target features (color, shape, or color-shape) within the probe and ignored adjacent distracters. RID differed from the probe on one or more target dimensions. Target “present/absent” reaction time and accuracy (joystick movement left/right) were recorded. Results showed equivalent RID interference for conjunction (color-shape) and single-feature targets (shape) when each required identification of the less discriminable, shape dimension (target dimension discriminability equated). RID interference was greater for conjunction targets only if the conjunction alone required identification of the less discriminable, shape dimension (target dimension discriminability not equated). Thus, discriminating the presence of a feature conjunction does not increase perceptual load due to a feature integration process.