Of course, I just made it up based on anecdotal experience with friends and family. For all I know, my friends and family could just be weird. However, I wonder if you encountered significant selection bias by polling people that are technical enough to participate in Mechanical Turk. Have you done anything other than an MT survey? I bet Flickr has some really good data on how many images get viewed at the default medium resolution and what percentage of people click through to the high resolution versions. Even then, Flickr may skew towards artistic images, rather than just people sharing photos of social events.
Of course, I just made it up based on anecdotal experience with friends and family. For all I know, my friends and family could just be weird. However, I wonder if you encountered significant selection bias by polling people that are technical enough to participate in Mechanical Turk. Have you done anything other than an MT survey? I bet Flickr has some really good data on how many images get viewed at the default medium resolution and what percentage of people click through to the high resolution versions. Even then, Flickr may skew towards artistic images, rather than just people sharing photos of social events.