Public Release: 

New Study Casts Doubt On 'Bell Curve' Theories

University of Wisconsin-Madison

A new study examining verbal ability and socioeconomic success casts doubt on theories advanced in the controversial 1994 book The Bell Curve.

The study, published in the September 1997 issue of Social Science Research, found that cognitive ability - or a person's academic ability - has not created growing differences among socioeconomic classes in the United States, as argued by The Bell Curve authors, Charles Murray and the late Richard Herrnstein.

Murray and Herrnstein contend in The Bell Curve that a person's IQ largely determines their socioeconomic status, that IQ differences in race are partly genetic and that African-Americans generally have lower IQs than whites or Asians. A growing number of scholars are disputing their findings.

"We would not for a moment deny cognitive ability an important place in the stratification process, but that place appears to be limited mainly to its role in determining how far people go in school, and that role appears to have been pretty much the same throughout this century," write the study's authors, Robert M. Hauser and Min-Hsiung Huang of the Center for Demography and Ecology at the University of Wisconsin-Madison.

"Our findings suggest that, if there is a key variable in the American class system, it is educational attainment, not cognitive ability," add Hauser, a professor of sociology at UW-Madison, and Huang, of the Institute for European and American Studies, Academia Sinica, Taipei, Taiwan. The authors say the distribution of schooling has become more equal throughout this century.

Hauser and Huang examined the claims made in The Bell Curve along with data from a short verbal test administered to about 12,500 adults nearly every year between 1974 and 1994 as part of the General Social Survey of the National Opinion Research Center.

Using the verbal test data, they analyzed how social background and educational attainment affect verbal ability and how verbal ability affects occupational status and income. In all cases, they found little or no evidence that the effects of verbal ability on socioeconomic outcomes have increased in the past two decades.

"Herrnstein and Murray have offered precious little evidence to support their story line, and we find equally little support in the trend data from the General Social Survey," Hauser and Huang write.

The authors discovered that there has been almost no change in how social background influences verbal ability, except for a declining negative effect on those born in the South or on farms. And they found few differences in verbal ability between high school graduates and college graduates born since the Great Depression, which they say reflects a combination of larger postsecondary enrollments and more relaxed college admission standards.

Hauser and Huang also determined that there were no changes in the effects of verbal ability on occupational status between the 1970s and the 1990s, except for small decreases among African-American men, white men younger than 45 and middle-aged white women. Effects of ability on occupational status did increase some for older white women, they found.

There were no changes in the effects of verbal ability on earnings as well, according to the study. Hauser and Huang say their analyses show that verbal ability affects a person's income primarily through their level of education.

The authors admit that data from the General Social Survey does have weaknesses, including the narrow content of the verbal test, that it has no measure of childhood ability and that it doesn't represent either very wealthy or very poor sections of the American population in substantial numbers.

But they emphasize that the survey's verbal test offers consistent data because it was administered regularly over a 20-year period. And they add that the survey obtained standardized, measurable information on social background and socioeconomic outcomes.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.