Cognitive Psychology · Simplified
Visual Search Laboratory
Some targets are found in the same time however many things surround them. Others cost you more for every item added.
Practice
Find the target, then say which half it is in
Every display contains exactly one target. Find it, then press Left or Right to say which half of the panel it was in. Six untimed practice trials first, with feedback.
What you are looking for. It changes between blocks, and you are told before each one. Nothing is ever timed out: the display stays until you answer. About four minutes.
Practice trial 1 of 6.
Your two slopes
A slope, not a speed
Each point is your mean for that many items. The line through each set is the best-fitting straight line, and its steepness is the number that matters: how much every extra item cost you.
All the values behind this chart
| Search | Items | Mean reaction time | Trials used | Accuracy |
|---|
Key idea: The interesting number is not how fast you were, it is how much slower each extra item made you. When the target differs from everything else in a single feature it tends to be found in about the same time whether there are eight items or thirty-two, so the line is nearly flat. When the target is only defined by two features in combination, and every distractor shares one of them, each added item costs you time and the line climbs. Describing those two functions is solid. Explaining them is where the argument starts, and a flat slope is evidence about the shape of the search, not proof that everything was examined at once.
One person, eight trials per point, in a browser. A slope estimated from three points and this few trials each is a rough figure and can easily come out negative by chance. Real experiments also include trials with no target at all, where the slopes are steeper still and usually about twice the target-present ones, and leaving those out is the largest simplification here. Browser timing adds noise of its own. A result that comes out flat for both searches, or steeper for the feature search, is a result about the block rather than about you, and nothing here measures anyone's attention or ability.
The longer version adds target-absent trials and the slope ratio they permit, an untimed display inspection with the target revealed, a practice mode with feedback on every trial and a transfer challenge. It is at Visual Search Laboratory in the main collection.