4 ms·
There used to be a belief that if you were exposed to am image, say someone drinking a prominently labeled Coca-Cola, it would make you want a Coke even if the
by tzs 5y ago
There used to be a belief that if you were exposed to am image, say someone drinking a prominently labeled Coca-Cola, it would make you want a Coke even if the image only flashed by so fast that you were not even consciously aware you had seen it such as if the image was a single frame inserted in a scene in a movie.
This was debunked.
There is, however, something sort of in the same ballpark that has been reproduced and does show that your brain does have an astounding capacity to at least partially remember images seen only briefly.
Take a large collection of images of a variety of ordinary things. Pictures of things like a bunch of bananas, a dog catching a Frisbee, a cat sunning itself near a window, kids waiting for a school bus, a cat pouncing on a mouse, someone eating a banana, yadda yadda.
Divide the collection randomly into two equal sets, A and B.
Show test subjects the images from set A serially, letting them see each image for a hundred milliseconds or two, with occasionally rest breaks if the image set is large (I believe this has been with image sets up to 10K in the total collection, so it can take quite a while to go through set A).
If you then tested people on how well they could recall these images, they would suck at it. Ask them to describe all the dog images they saw, for instance, and they will only recall a small fraction of them.
Where it gets interesting is if instead the test you do is at a later time you randomly pair up each image from set A with an image from set B, and serially show the subjects these pairs in random order and with the random left/right placement, and ask the subjects to identify which of the two images is one they have seen before.
People get it right some astoundingly high percentage of the time, like 90% after several days. They might not really "remember" most of the A set in the sense that they can purposefully recall them, but apparently we do automatically and effortlessly store something that is good enough to let us tell which of two images is the one we've seen before.
I wonder if this effect would apply with ads? You see a banner ad in passing for brand X molten boron, not paying attention to it. This doesn't make you subconsciously want to go buy the product--that was the old debunked subliminal advertising theory.
But later, you actually need some molten boron and so head out to the store. They have brands X and Y. I wonder if because of the banner ad you saw a few days ago, you'd recognize brand X as one you have seen before and be more likely to choose it over brand Y.
- mattkrause 5y agoThe classic paper on this is Standing’s 1973 “Learning 10,000 pictures.” The title really says it all: subjects shown up to 10,000 pictures (once, for a five seconds apiece) remembered a substantial fraction of them two days later. This is somewhat specific to images (and even moreso for “vivid” ones); people did worse with dictionary words. However, one important detail is that subjects were told to pay careful attention to the stimuli; they were repeatedly reminded that a memory test was coming up. PDF: https://www.sas.upenn.edu/psych/rust-lab/publications/standing_73.pdf https://www.sas.upenn.edu/psych/rust-lab/publications/standi... Journal Page: https://www.tandfonline.com/doi/abs/10.1080/14640747308400340 https://www.tandfonline.com/doi/abs/10.1080/1464074730840034...
- Tarsul 5y agoyeah and then there was this other study where people had to count how often people would pass the ball in a video and no one noticed that there was a guy in a gorilla custom going around, because no one cared about that (they were distracted by counting). I think this is similar to how we usually try to avoid looking at ads in the internet: it's not what we are interested in at that moment! https://www.livescience.com/6727-invisible-gorilla-test-shows-notice.html https://www.livescience.com/6727-invisible-gorilla-test-show...