Be Careful in Interpreting Experiments

This is from today's NYTimes: "Three recent experiments show that even the youngest children have sophisticated and powerful learning abilities. Last year, Fei Xu and Vashti Garcia at the University of British Columbia proved that babies could understand probabilities. Eight-month-old babies were shown a box full of mixed-up Ping-Pong balls: mostly white but with some red ones mixed in. The babies were more surprised, and looked longer and more intently at the experimenter when four red balls and one white ball were taken out of the box — a possible, yet improbable outcome — than when four white balls and a red one were produced. "

I find this a weird explanation. If I reached into a pile of mostly white balls and pulled out three red ones, my toddler would be pretty interested, but not because she understood this to be an improbable event - rather, because it would seem like magic. The experimenters assume that the child noticed the ratio of red/white in the box and compared it to the sample. Is that a fair assumption? My child notices red balls more than white ones, period. I also am not sure how this study can be double-blind: doesn't the experimenter who is inviting the baby's gaze know what she is doing, and isn't she looking for a reaction?

I wanted to blog about this because it's the kind of crappy science reporting I see in newspapers all the time. Either the study was much better than the reporter makes it out to be, or the reporter should have been more skeptical of the study, or (likely) both.