Take a look at this quick video. You'll see a set of six small images, arranged in a circle, for 1 second. Then the screen will go blank for 1 second. Finally, one image will reappear in the place of one of the first six pictures. Your job: indicate whether the final image is the same or different as the image that originally appeared in that same spot.
Click here to view the movie (QuickTime required)
In principle, this should be an easy task, right? Your visual working memory can hold around six items at a time, so it shouldn't be hard at all to remember if the new picture is the same as the one it replaced. So, is the final picture the same as the one it replaced?
Let's make this a poll (if you don't know, just guess):
If we're very accurate at this task, then one answer (the correct one) should dominate. If we're not very good at it, answers should be closer to 50-50.
In fact, visual memory is a bit more complicated than that. What constitutes a single item in your memory? Is it a whole face? Or does each portion of the face count as a separate item to be remembered?
In a 2007 study, Kim Curby and Isabel Gauthier found that "experts" seem to have a larger visual working memory in their area of expertise than non-experts. (Nearly all humans are experts at recognizing faces, unless the faces are upside-down). But a team led by Miranda Scolari wasn't convinced. Maybe the distinction between expert and non-expert isn't the size of working memory, but how detailed it is.
Scolari's team showed sequences like the one in the video to 12 volunteers, but with a few twists. Sometimes the faces were upright, and sometimes upside-down. Cubes and colored ovals were also shown, and sometimes a mixture of faces and cubes were shown.
Half the time, the single item that reappeared was the same as the item it replaced. What the researchers were interested in was accuracy the rest of the time, when the image changed. Here are the results:
The researchers made a simple calculation to estimate the size of working memory based on the responses. Remember, respondents were asked if the new image was the same or different from the image it replaced. If accuracy is 100 percent, then the size of working memory should be at least six. If it's 50 percent (equal to random chance), that would suggest that working memory size is zero. As you can see, when there was a big change in the items, the apparent size of working memory was roughly four, and there was no significant difference in the type of images viewers saw. But the key distinction was between upright faces and upside-down faces -- when a different upright face was substituted for an upright face, or when a different inverted face was substituted for an inverted face. In this case, memory size for upright faces was about 1.7, while it was only 0.5 for inverted faces. This difference was statistically significant.
The researchers say this means that experts do not have a larger working memory capacity than non-experts. Instead, their memory is more detailed. If experts really had a larger capacity, then they should be more accurate when the faces shift from upright to cube versus inverted to cube. They're not, even though they could spot a different upright face three times more often than a different inverted face. So experts remember more details, just not more different items.
How did you do on our little example? Did you notice that a different face was substituted? Our task was a little easier than the researchers' task, because our faces have a few extra features that make them easier to identify -- eyeglasses, different lighting, and so on. In the original study, these things were held constant.
M. SCOLARI, E. K. VOGEL, E. AWH (2008). Perceptual expertise enhances the resolution but not the number of representations in working memory Psychonomic Bulletin & Review, 15 (1), 215-222 DOI: 10.3758/PBR.15.1.215
- Log in to post comments
Interestingly, when the movie file had finished loading on my Quicktime player, it started off in the middle of the file, by showing the single upside-down face first. I then played it again from the beginning and, duh, I knew where to look! Perhaps it is to obvious to state, but I realized that my chances of being an"expert" skyrocketed when I knew in advance where to focus.
All I can say is that I thought I'd seen that face before. Where, what position... who knows?
Also, I never been considered an expert at anything.
Re-iterating and extending Ian's chirp, why Quicktime ? It is not standard and not one of the better formats. Why not Flash?
I had to guess, but I'm wondering whether you randomly decided what the result should be. I guessed "no" only because I thought that that option would be more likely.
Nash, I don't think Flash would be much better in terms of standardness and it depends on what you mean by "better". My guess is that Quicktime is simply easiest for Dave to make things in, although I agree that something else would probably be better.
I think most systems support mpeg videos since there's no licensing issues with it.
Ian/Nash:
I was having this problem with every method I tried. Quicktime seemed to work the best. You'd think it'd be a simple matter, with such a simple video, but no.
My first choice is usually Vimeo, but check out what Vimeo did to the video:
http://vimeo.com/3272495
So we're not going to replicate the study, but at least people can see what the stimuli are like.
One second was so short I only had time to look at one or two faces (and obviously not the one that was later displayed) so I had no idea whether it was the same or not (unless you are testing unconscious recall) Shouldn't the faces be displayed for a longer time so you can actually look at them all?
And like Chris I voted no.
I only had time to look at half the faces; for some reason I started by looking at the faces at the top, so I completely guessed (and I guessed no). Did the study you are discussing give the participants a longer time to look at the original pictures?
I agree, 1 sec was not long enough to even notice all the faces. I think I would've done better with simple shapes/colors than faces, as there is a lot to look at in a face; more than 1/6 sec allows anyway! I said "no" because I didn't recognize the face shown in the second frame, but I also knew that I didn't have time to look at all of them.
Yes, this is the exact timing of the original study. But in the study it was repeated many times -- of course, people were very inaccurate with inverted faces in the real study. I'm not sure why people did better with our sample. Possibly a combination of video problems and predisposition to answer "no" when unsure.
Of course, with one trial of experience, your performance will be terrible. These experiments have a practice block that allow the participants to get the flow of the experiment and, speaking from personal experience, you can do the task.
I wouldn't say these two studies present competing views. Curby and Gauthier (2007) suggestion of capacity differences for faces is based on the differences in the measure (Cowan's K is a measure of "capacity). They do not make strong claims as to how source of this difference, only that a higher "capacity" is found with faces. Scolari et al. (2008) are making sure perceptual expertise does not lead to WM capacity greater than the magical limit of 4 items. The important point for faces is that on average, more upright faces are stored compared to inverted faces. Sure, it is less than 4, but the relative difference is what applies to the benefit of expertise.
Finally, a recent study (Curby, Glazek & Gauthier, 2009) looks at visual short-term memory (VSTM) for objects of expertise besides faces. They recruited car experts and tested VSTM with car images and found a greater "capacity" for cars in experts compared to novices. They conclude that expertise allows maximum use of an inherently-limited VSTM system.
Curby, K.M., Glazek, K., & Gauthier, I. (2009). A visual short-term memory advantage for objects of expertise. Journal of Experimental Psychology: Human Perception and Performance, 35(1), 94-107.
Finally, a recent study (Curby, Glazek & Gauthier, 2009) looks at visual short-term memory (VSTM) for objects of expertise besides faces. They recruited car experts and tested VSTM with car images and found a greater "capacity" for cars in experts compared to novices. They conclude that expertise allows maximum use of an inherently-limited VSTM system.
I believe the same conclusion was/could be made for the Chase and Simon (1973) expert vs. novice chess study, where they found expert chess players could accurately replace 20+ chess pieces on a board after viewing actual games for 5-10 seconds, whereas novice chess players managed about 4. However, experts' ability fell to that of novices when the pieces were randomly placed on the board. This could be explained as a consequence of the manner in which experts chunk information in their domain of expertise (most likely connected to crystallized intelligence).
An interesting contemporary version of the Chase and Simon experiment based on Curby et al. and Scolari et al.'s research, would be to see if expert chess players are better than novices at remembering chess pieces (e.g., knight, rook, pawn) presented in specific colors in the context of actual game situations.
Huh? What's the connection with expertise?
Show pictures of things that some people are experts in, and others not, and then you can say something about expert memory.
This depends on other factors also like ethnicity etc. I am from a different ethnic origin and I had hard time even identifying the face difference! Put it in other terms, greater capacity may not always be in terms of rapidity!
I'm also wondering what's the connection with expertise. Is it because we're more expert at recognizing upright faces?
And I'm another one who answered "no" because "I have no idea!" wasn't a choice. (The only 2 faces I even vaguely saw were the top and bottom.)
I didn't get a good look at any of the faces in 1 second. I answered no because my intuition was that I hadn't seen the test face there, and that it might be the same as the upper right face.
Looking at it again, it seems likely to me that I was responding to color and lighting. The upper right face is the one most similar to the test face in color tone and lighting direction.