Thursday, April 15, 2010

100 / 40 = 2.5 ?

From a report on Williams syndrome, excerpted and then discussed by Steve:
Despite their lack of racial bias, children with Williams syndrome hold gender stereotypes just as strongly as normal children, the study found. That is, 99 percent of the 40 children studied pointed to pictures of girls when asked who played with dolls and chose boys when asked, say, who likes toy cars.
Solomon must have had his way with one of those kids!


Anonymous said...

Clearly, the last kid was black, following the 3/5 rule of the constitution...

Sparky said...

Sailer is quoting a news report that, in typical fashion, completely misunderstood something written in the paper. Reactions to gender were reported as being indistinguishable between WS and normal children and given a p value as follows:

"Conversely, sex-role bias was pronounced and identical in the two groups (p > 0.99), indicating that absent racial stereotypes were not due to an overall feature attribution impairment."

This was the only occurrence of 0.99 or 99% in the paper itself. Moreover, it is evident from a glance at the graph in Figure 1 that sex-role bias was not 99% in either the normal or WS groups. Bad science reporting strikes again.