The reason it works ya'll is that our brains work by using simplified mechanisms called heuristics in order to form judgements and make decisions. One of the heuristic we use a lot is called the "representativeness heuristic," which is the technique used to make quick judgments in response to new stimulus/events. At its simplest: The human brain picks out the salient features of a new object/event in order to compare those with our previous experiences to arrive to a quick conclusion. It's how we know an object we've never seen before is a chair. Our brains immediately pick out the distinguishing features of this object -- flat, hard, two 90 degree planes ... -- and "Oh! Bingo!" that looks like a chair! It's what you might also call "pattern recognition."
... and our brains do are almost able to do this work instantaneously. You take person and put them in front of a picture slide show with a category in mind, say chairs, and they'll be able to confirm that the picture in front of them is a chair within tenths of a second. It's all working with categories and similarities. Suddenly throw in a picture of an apple, and whoops, the brain has to throw red flags and is forced to recalculate.
Off this psychologists can perform tests. These tests are replicable and predictable. People can, in a quantifiable manner, better identify apples to apples and chairs to chairs, than they could identify apples with chairs. The can easily identify the color red to red objects, instead of
failing at the Stroop effect test; they can identify coffee pictures to coffee scents; and can compare cute things to other cute things. People can also better identify nice words to nice pictures, bad words to bad pictures, nice words to white people, and bad words to black people ... which as you could imagine, is a bit of a problem.