Did a quick google for it and can't dig up anything easy to reach. I'm remembering stuff from first-year psych that was basically a decade ago now.

The only keywords I recall are that it was a "visual wall experiment" that was done with crawling infants (ie mainly pre-language). I'll describe the experiment in case anybody else recognises it and can point us at better references.

I can remember watching the video, where infants were placed on a glass tabletop - underneath which was an obvious drop off, visible through the glass. ie the infants weren't in actual danger of falling - but it looked (to them) as though they might. The drop-off went down about a metre and was painted with a grid-pattern so the infant had clear visual clues of what it was.

A reward (toy? food? can't remember) was placed at the other end of the table, and the infant could go get it by crawling across the glass, over the visual drop-off.

I do not recall how many infants were in the study - but it produced a clearly distinct average gender-difference in the likelihood of whether the infant would brave the scary-looking crawl to go get it.

The take-home conclusion was that males were more likely to risk more for the reward, whereas females were less likely to do so.

Ok, so now that we've got the correct key-phase ("visual cliff") I see that there's heaps of research using this apparatus, and most of the studies are on development of depth-perception, or infant reactions to maternal prompting etc...

Can't seem to find anything on the study that I remember. Sorry.