When members of the Eugenics Board of North Carolina were trying to decide whom to sterilize, they almost always relied on IQ scores as the one objective measure they could count on among a number of other considerations.

But even 50 years ago, researchers had doubts about what intelligence-quotient tests - including the Stanford-Binet IQ test used by the eugenics board - could accurately tell them. Scientists knew that environment, nutrition, early-childhood enrichment and even cultural bias on the test played a role in determining IQ.

The test has undergone radical changes since its inception in the early 1900s. Despite the doubts, people - including members of the eugenics board - still continued to see the test as something that could reveal all about a person's potential.

IQ tests can't meet that expectation, said Bruce Bracken, the president of the International Testing Commission and a professor of education at the College of William & Mary in Williamsburg, Va.

"We have no measures of potential.... The tests we use merely measure a person's present capacity. They may have previously worked at a different level," he said. "The best use of intelligence tests is just to assess what the person's current ability is."

A difference in IQ scores between the races was noted in a 1948 study of Winston-Salem schoolchildren conducted by A.M. Jordan, a professor at the University of North Carolina at Chapel Hill.

Using money from Winston-Salem textile magnate James G. Hanes, Jordan tested the IQs of 95 percent of the elementary-school students - about 10,000 children - in the city. Jordan's team decided to use an IQ score of 60 as the dividing line between retarded and normal instead of the standard 70 used in other studies.

Even with the lower dividing line, the results were not "entire successful," Jordan wrote, because researchers found that the children tested in Winston-Salem had lower average IQs than their counterparts in other parts of the country. Black children tested in the city had lower scores than white students.

Individual tests confirmed that about 60 percent of the children who tested below 70 actually had IQs that low, Jordan wrote. The rest of the children had been scored too low.

In addition to compiling information about the children of Winston-Salem, Jordan was also evaluating the use of group tests to determine which children needed more help.

But he may have been doing something more. Hanes and Dr. Clarence Gamble, a wealthy supporter of the eugenics movement, may have also been trying to use the tests to promote further sterilizations, according to a thesis by William Van Essendelft, a graduate student at the University of Minnesota who was studying a group called the Voluntary Sterilization League.

"In Winston-Salem and in Orange County, North Carolina, the field committee had participated in testing projects to identify school age children who should be considered for sterilization," Van Essendelft wrote.

Doctors with the Bowman Gray School of Medicine may have been involved in the testing and may have tried to use the results to increase public interest in sterilization, Van Essendelft wrote.

"The medical school has a long history of interest in eugenics and had compiled extensive histories of families carrying inheritable disease. In 1946, Dr. C. Nash Herndon ... made a statement to the press on the use of sterilization to prevent the spread of inheritable diseases. That release coincided with the testing program and was part of an effort to use the tests to stimulate interest in sterilization," Van Essendelft wrote.

In the end, seven children were identified as candidates for sterilization, a number that disappointed Hanes and Gamble. It is not clear whether the seven children were ever sterilized.

The group apparently didn't use typical eugenic board standards - which made anyone with an IQ under 70 a candidate for sterilization - to decide who should have the operation. If the results of Jordan's study were the only factors used to determine who should be sterilized among the 10,000 children tested in 1948, 218 white children and 526 black children would have been sterilized.

Jordan saw the wide discrepancy between black and white children as a sign of a problem with the test.

"One gains immediately the impression that the test was not fair to them or that the fact that the tester was white could have made some difference," he wrote.

Jordan blamed the low performance on the fact that most of the black children's parents had less education than white parents, and that there was a limit on the expectations people had for black children - two theories that are still considered valid today by experts in the field.

But other modern researchers think that the difference is caused by a number of other factors, not the least of which is how the test is constructed.

The IQ test was made up by whites and tested on white children until an update in 1972. So the version that the eugenics board used was based on the experiences and abilities of white administrators and children.

This could have been part of the problem, said Frank Wood, a professor of neuropsychology at Wake Forest University Baptist Medical Center. IQ tests have always been criticized for posing questions that the majority culture would have an easier time answering.

"Culture-fair items are supposed to be items that are just as regularly solved by people in one culture as another. People who grew up in a vegetarian culture might not have learned very much about cooking meat, for example," Wood said.

Intelligence-quotient tests have been around since the early 1900s. They were developed by French doctors to separate developmentally disabled children from other children who were entering school.

The Stanford-Binet is a one-on-one test that has to be administered by a trained professional. It takes more than an hour to give, and the scoring takes longer yet.

The test was created in 1905, and modified in 1916 when it was given the name Stanford-Binet.

There have been four revisions done since then, with a fifth scheduled to come out in 2003.

The Stanford-Binet has a series of questions in different subject areas. Some deal with reason and comprehension. A typical question asks children to name the similarities and differences between an orange and baseball.

Others test general knowledge by asking such questions as naming the days of the week.

Some test spatial intelligence by looking at a child's ability to draw a picture of an object or a design placed in front of him or her.

"When you add it up across these different ways of testing it you get to feel that you are getting a more general sense than you would any other way," Wood said.

Even if IQ research has led to some insight during the past 100 years, several questions still remain, such as what creates intelligence in a person.

"People tried to look at it to see if it was inherited. That turned out to be only minorly interesting. Yes, it runs in families, but familial connections, including genetics, have never been the whole of IQ," Wood said.

Most researchers agree that such factors as early-childhood enrichment and nutrition play a role in the formulation of IQ. But they disagree on many other points.

During the past 100 years, IQ scores have increased by about 15 points. The average score is still 100, but every time the average is reset, a new scale also has to be recalculated to take the difference into account.

Though test experts cannot pinpoint a reason for the increase, some point to better schooling and the effects of living in a more complex society as possible explanations.

"Our overall ability to do that as a population has increased slowly but steadily over the past 100 years. That's one reason you have to restandardize the test every few years. If you've got an IQ of a certain level on a test you took 20 years ago, that same level of skill would get you a lower score now," Wood said.

One of the most controversial aspects of the test has always been the difference in scores between black and white children.

Blacks score an average of 15 points below whites on the tests. That difference was the subject of a 1994 book, The Bell Curve, which suggested that the discrepancy was due to genetic differences between the races.

People still argue over the book's ideas. Bracken thinks that the reason for the difference is more complex than a simple genetic difference.

"If you look at the nature-nurture issue, some people believe that the differences are genetically related, some believe the differences are almost purely environmental. Most people agree there is a combination between nature and nurture and we also have the effects of cumulative deficiency. When you compound socioeconomic deficit over generations, you get a compounding effect," Bracken said.

No matter how they were constructed, IQ tests should not be used to decide who could have children, Bracken said.

"I think that's despicable. There are many good reasons to use IQ tests and there are obviously extensions that are inappropriate," he said. "Deciding who should reproduce should not be one of those uses."

• Danielle Deaver can be reached at 727-7279 or at ddeaver@wsjournal.com