Nature 458, 147 (12 March 2009) | doi:10.1038/458147a; Published online 11 March 2009

A useful way to glean social information

See associated Correspondence: Wendl, Nature 458, 831 (April 2009)

Wendy M. Williams1 & Stephen J. Ceci1

  1. Department of Human Development, Cornell University, Ithaca, New York, USA


In his Commentary (Nature 457, 786–788; 2009), Steven Rose concludes by noting, when it comes to research on race, gender and IQ, that "there is no valid knowledge to be found in this area at all". Is the worldwide increase in IQ not valid, and should its insights into the malleability of racial differences in intelligence be ignored? Is the superiority of women's mathematical grades unimportant? Does the convergence of black and white test scores not inform our understanding of how to achieve a more just society? All of these findings stemmed from research that Rose claims is invalid.

One-party science may appease some now, but will eventually stymie scientific understanding. The recent renaissance of research on Lamarckian modes of genetic transmission shows that not all widely acknowledged 'dead' theories are 100% wrong — Lamarck was dismissed years ago, but today his ideas spur valuable science.

Rose argues that a single g-loaded coefficient, such as an IQ score, cannot capture all of intelligence. We agree. However, the limitations of IQ are not a reason to muzzle researchers who believe that, albeit imperfect, it nevertheless provides important predictive information.

Rose argues that the dubious biological underpinnings of race render scientific analyses meaningless. Although we agree that the terms white, black, Caucasian or Asian do not define distinct races, this misses the point: self-defined race often provides the largest coefficient in models of achievement. Even if race is a proxy for a chain of social and environmental events that are not genetic, this does not minimize the scientific value of studying it. No one who self-defines as black will deny that this identity is omnipresent in their own and others' consciousness, influencing social interactions and stirring identity threat. Most governments require reporting of racial gaps; numerous studies target narrowing them. Such factors justify studying race, even if it is entirely socially constructed.

The same goes for gender, which — as Rose points out — is highly influenced by socialization but nonetheless continues to offer important insights as a category. Empirical research reveals cognitive differences between the sexes in 4-month-old babies. Even if one argues that these infants' environment has already caused sex differences, this does not gainsay their empirical basis nor their scientific validity.

No one endorses pseudoscience, but Rose has prejudged what qualifies.

See also:
The belief that genes cannot be changed is now outdated
Identifying adaptive differences could provide insight
The arrogance of trying to sum up abilities in a number
Is poverty better explained by history of colonialism?
Would you wish the research undone?
Measured intelligence is a product of social processes
Don't fan the flames of a dead debate

Further contributions are welcome online at