Parallel search ("pop-out") is pretty straightfoward. In these situations, the relevant aspect of the target can be detected quickly regardless of how many distractors there are.
Limited capacity search seems straightforward at first. The time required to detect a target (or to reject that a target is present) increases with the number of distractors. In addition, the time required to say that no target is present increases faster than the time required to say that the target is present.
This pattern of data is consistent with a serial search. In a serial search, you look at one item at a time, and respond as soon as you see a target. Because you have to search all of the items to say that the target is absent, but you can stop as soon as you find the target if it is present, you would expect the time to increase more for target absent trials than for target present trials. So far so good.
What keeps us from concluding that there is a serial search when we get the pattern of data with response times that increase with display size? The problem is that there is another process that would also yield the same pattern of response times that involves a parallel search.
Imagine that processing takes place in steps, and that the more steps you carry out the longer the response time. In this model, processing involves allocating points to objects in order to recognize them. A given object requires paying a certain number of points before you can recognize it. In our example, imagine that you have to pay an object 240 points to recognize it. Finally, imagine that you can process by allocating all of your points to a single object (that is a tight focus of attention), or you can spread your attention to more than one object at the cost of giving each object fewer points on any given time step. Focusing on one object at a time, and giving each one 240 points is like doing a serial search. If you spread your attention over 4 objects, you could give each one 60 points on each time step, but it would take 4 time steps until you recognized all of the objects. If there were 8 objects and you processed them all in parallel (by paying 30 points to each object) it would take 8 time steps until you recognized the objects. Thus, it would take longer to respond as the display size increases.
Because it is possible to get increases in the response time with display size with this limited capacity parallel search, we cannot conclude that this pattern of data reflects a serial search.
Return to the PSY 355 Question Corner.
This page last modified on <INSERT_MODIFIED.
You are visitor <INSERT_COUNT Q355F0005 to this page.