Colors are rarely uniform, yet little is known about how people represent color distributions. We introduce a new method for studying color ensembles based on intertrial learning in visual search. Participants looked for an oddly colored diamond among diamonds with colors taken from either uniform or Gaussian color distributions. On test trials, the targets had various distances in feature space from the mean of the preceding distractor color distribution. Targets on test trials therefore served as probes into probabilistic representations of distractor colors. Test-trial response times revealed a striking similarity between the physical distribution of colors and their internal representations. The results demonstrate that the visual system represents color ensembles in a more detailed way than previously thought, coding not only mean and variance but, most surprisingly, the actual shape (uniform or Gaussian) of the distribution of colors in the environment.

The Internet provides a convenient environment for data collection in psychology. Modern Web programming languages, such as JavaScript or Flash (ActionScript), facilitate complex experiments without the necessity of experimenter presence. Yet there is always a question of how much noise is added due to the differences between the setups used by participants and whether it is compensated for by increased ecological validity and larger sample sizes. This is especially a problem for experiments that measure response times (RTs), because they are more sensitive (and hence more susceptible to noise) than, for example, choices per se. We used a simple visual search task with different set sizes to compare laboratory performance with Web performance. The results suggest that although the locations (means) of RT distributions are different, other distribution parameters are not. Furthermore, the effect of experiment setting does not depend on set size, suggesting that task difficulty is not important in the choice of a data collection method. We also collected an additional online sample to investigate the effects of hardware and software diversity on the accuracy of RT data. We found that the high diversity of browsers, operating systems, and CPU performance may have a detrimental effect, though it can partly be compensated for by increased sample sizes and trial numbers. In sum, the findings show that Web-based experiments are an acceptable source of RT data, comparable to a common keyboard-based setup in the laboratory.

A popular model of the function of selective visual attention involves search where a single target is to be found among distractors. For many scenarios, a more realistic model involves search for multiple targets of various types, since natural tasks typically do not involve a single target. Here we present results from a novel multiple-target foraging paradigm. We compare finger foraging where observers cancel a set of predesignated targets by tapping them, to gaze foraging where observers cancel items by fixating them for 100 ms. During finger foraging, for most observers, there was a large difference between foraging based on a single feature, where observers switch easily between target types, and foraging based on a conjunction of features where observers tended to stick to one target type. The pattern was notably different during gaze foraging where these condition differences were smaller. Two conclusions follow: (a) The fact that a sizeable number of observers (in particular during gaze foraging) had little trouble switching between different target types raises challenges for many prominent theoretical accounts of visual attention and working memory. (b) While caveats must be noted for the comparison of gaze and finger foraging, the results suggest that selection mechanisms for gaze and pointing have different operational constraints.