Observations and Conclusions
In my opinion...
Defects of Study
The study has many problems. First, it only uses two websites as its test data which were both chosen by the researchers and not the subjects. They based their choices on radio advertisements and not on subject preference. Hence, it is not clear whether the subjects considered the test sites in the same way as the researchers.
Secondly, it restricted the subjects to two groups: one set of 12 year olds and one set of 16 year olds. In fact, they only had 16 test subjects (8 male, 8 female). Morever, the test subjects they chose used the web for only 1-3 hours a week, which is relatively light usage. Finally, the test ran using Internet Explorer 3.02, which is has fallen far below 1% usage amongst web users. These limitations prevents them from drawing substantial conclusions about all children.
The researchers paid the subjects to answer four directed questions, which I assert most people just "surfing" would not task themselves in asking. This strategy would warp the normal browsing habits of the subjects. Actually, it is artificial to constrain browsers to only two websites of the whole web. Moreover, using only four data points is not enough to extrapolate information about data mining proficiency.
The researches also went into the study presuming that browsing a site is mainly an informational retrieval task, not an entertainment task. Thus, their conclusions about animations being meaningless are bogus. Of course they are meaningless when the subjects only have half an hour to answer two difficult questions. But the study says nothing about lazily viewing a website. Indeed, the study later goes on to say that the subjects found the Disney site "more 'fun to use'". It is possible the animations played a role in that.
For a another (better) study, see the StanfordPoynterProject.
I was recently exposed to Brenda Laurel and the work that she did at Purple Moon. Purple Moon made games for 8-12 year old girls. The girls played an interactive game that was heavy on character and narration, and then they were able to participate in online communities centered around the game. The games were based on about four years of intensive reseach on how girls play and socialize. The company eventually folded, but the new market that it created has survived. Its Web site had registered 290,000 users at one point. Here is an article that describes the project:  -- SteveHowell