**Exclusion Criteria** Subject data will be excluded on the following grounds: * Subjects reported being under 18 years of age. * Subjects reported problems with experimental playback, such as stuttering, freezing (pre-determined options) or another issue specified in a free-response. * Subjects reported needing correction to their vision but not wearing it during the experiment. * Subjects reported counts that err by more than 50% in either direction more than once. * Subjects report any abnormalities with their color vision. * Subjects misidentify the number in Ishihara Plate 9. * Subjects reported having performed a similar task before, wherein they tracked multiple objects and something unexpected appeared. * Subjects failed to answer any question during the course of the experiment. **Measures** <br> After each trial in which no unexpected object appears and after the critical trial, we will collect subjects' counts of how many times the attended objects bounced off the edge of the frame. On the critical trial, we will ask whether an extra object was present during the trial, and then ask them to describe the shape and color from a menu of pre-determined options. Subjects will be counted as having noticed the object only if they affirm having seen something new and correctly identify either its shape, color, or both. <br> We will also collect a set of demographic data for exclusion purposes, as well as characterizing our sample. We will verify that there are no systematic differences in the data based on country of origin, and we will verify that the two conditions (attend white vs. attend nonwhite) have similar error rates to ensure that one version of the task is not more difficult than the other. **Analysis** <br> Three different data possibilities emerge from this design. 1. If attention sets are always category-based, regardless of working memory load, then all colors should be noticed at equal, high rates when subjects attend to nonwhite shapes, and at equally low rates when they ignore nonwhite shapes. 2. If attention sets only become category-based when working memory is overloaded, then only the unexpected object that matches the current color should be suppressed. The other unexpected objects, with their radically different features, should be noticed at high rates. 3. Attention sets may depend on having encountered the color already during the course of the experiment. In that case, when ignoring colors, only the unique color (green) should escape suppression. For each condition (current color, previous color, or new color), we will determine the difference in noticing rates (notice when attending white - notice when attending color) and compute a 95% confidence interval around that difference. In all conditions, we expect a large difference for the current color; it should be noticed frequently when attending to the color, and noticed infrequently when ignoring it. If we also see a difference in the other two colors, hypothesis 1 is supported; if we see only a difference in the previously seen colors, hypothesis 3 is supported. We will also conduct exploratory analyses to investigate possible effects of presenting a previously seen color from the first trial versus the second trial in the previous-color condition.