I think I meant photos are high quantity of data (1 photo=1000 words).
But high quantity doesn't equal high quality. High quality (high data density? high signal?) is found in a single, 10-word succinct sentence more often than it's found in a photo. In that sense the signal in a photo is more diluted by more noise.
This definition is in line with what I said about elite scientists. They prefer photos less than average scientists do, because the signal is low. It's interesting that this difference isn't as big though. Maybe it's harder in general to find the signal in high noise (the few good words in the 1000 words of a photo), irrespective of being elite or average.
Maybe that's where painting comes in. It isn't as fixed of a medium as a photo, which may make it easier for the elite to capture the few good words a photo may not. Although this is probably connected with time too. You can continue working on a painting; paintings are abandoned. Whereas a photo is a still in time, often completely uncapturable at other times.
I wish the study had also captured physical exercise. That'd be more data on kinetic thinking.
A thought on music, which is off the charts for elites. Music is patterns and paradoxically patterns compress very well. Literally, think .mpeg, deflate, arithmetic coding and dictionary compression that can yield 100x benefit. Music is probably an uncompressed spider web that reels us in to some small, powerful data function. Music is the macroexpanded source code and its data expression is the macro.