Discrimination in loan approvals widespread.

In recent years newspaper reports have exposed lender discrimination against African-Americans an other minority groups in mortgages and other home loans. By a roughly 2-to-1 margin, this discrimination appears even when their incomes are similar to those of white applicants.

In response, bankers and conservative economists have tried to discredit the reports. According to The Wall Street Journal, they "have spent considerable time and money during the last few years" mounting a campaign to debunk the idea that racism exists in the country's lending practices.

But new studies by the Federal Reserve Bank show the newspaper reports were essentially correct.

The issue arose in the press with The Atlanta Journal-Constitution's pioneering Pulitzer Prize series in 1988, followed by a Detroit Free Press series on lending in middleclass black neighborhoods, continued with a national Wall Street Journal survey in 1992 which it updated in 1995, and The Washington Post's prize-winning 1993 series. Similar reports began appearing in many papers like the Pittsburgh PostGazette, Dallas Morning News and Cincinnati Enquirer in 1994, based on annual surveys by federal bank regulators under the Home Mortgage Disclosure Act

The major problem with all this, say bankers and conservative economists, is that family income is only one of several factors considered by banks to determine in applicant's creditworthiness, as pointed out by David Andrew Price, an attorney with the conservative Washington Legal Foundation, writing in Forbes MediaCritic (Summer 1995). Other common factors are payment-income ratios, amount of down payment, work or employment history and credit history, including episodes of slow payments and bankruptcy.

Price notes that there are no checks and balances on interpreting statistics by news reporters and editors, especially when the newspaper itself has done the research design, data collection and statistical analysis.

"Much reporting on statistics suggests that the writers put too much stock in their ability to eyeball a study and decide whether it makes sense," he wrote. "Few, if any, of these reporters seem to have consulted an econometrics expert, so it is little wonder that they present fairly crude correlations as proof."

This is what Stephen Hess decries as "social science journalism" in his book "The Washington Reporters." "Journalism and social science are not the same," Hess says. "The tools of research are different; so, too, are the standards of evidence, the framework of theory, and the criteria of judgment."

Meanwhile, The Wall Street Journal reported in July on a new Federal Reserve Bank study that showed that, even with similar credit histories, white mortgage applicants were only half as likely to be rejected for loans as were black or Hispanic applicants. But the Fed found that blacks and Hispanics with outstanding credit histories were about as likely as whites with comparable records to get bank approvals for mortgages, the Journal reported.

This study represents a research advance, in that by using credit history and other new factors, it adds to the mounting evidence of discrimination. But how much has not been made clear.

It also involved a re-examination by the Federal Reserve in Chicago of a study made public in 1992 by the Boston Fed, one that has sparked much debate in banking and economics circles. Critics faulted the Boston study for omitting the condition of the property to be mortgaged, and other factors used by lenders. Employment stability in the applicant's field was included. rather than the applicant's occupation itself, as were past bankruptcies and delinquent consumer payments to indicate credit history.

Defenders say these indicators were supported by statistical regressions, mathematical tests of significance or reliably performed on computers to shore up a rigid analysis of data. Few other than Statisticians understand them.

Instead, journalists often have to rely on the "garbage-in" approach, and such indicators might well be input problems. This is because a research analyst's input includes the assumptions made about the validity of the raw data, along with sufficiency and accuracy of the data. For instance. do an applicant's past credit problems count with lenders as much as current financial troubles do in deciding whether to extend credit? No, says David Horne of the Federal Deposit Insurance Corporation in a FDIC Banking Review article last year; the study didn't look at credit problems the way lenders do.

"The timing of credit problems is important." he wrote. "Current problems are given much more weight. And that just wasn't captured [in the study]."

However, the Journal's story -- while citing the earlier study, critical reaction to the new one and the legislative implications -- failed to suggest how the debate over the basic Boston study could affect the new study's import. For example:

Using the same 1990 data from the Boston area. the Chicago Fed study also concluded that different races with similar credit histories, overdue 60 days or longer on one or more debts. had different application outcomes. with only 9 percent of the whites disapproved compared with 18 percent of the blacks and Hispanics.

Moreover. it also found "a cultural affinity between white lending officers and white applicants and a cultural gap between white loan officers and marginal minority applicants" in approving or rejecting marginal applications, William C. Hunter, research director at the Chicago Fed, said this cultural gap appeared to explain the discriminatory treatment more than racial bias did.

The St. Louis Post-Dispatch, in an Aug. 17, editorial, condemned that notion as an attempt "to sugarcoat this discrimination ... by thinking up new ways to justify such bias. The cultural affinity argument amounts to political correctness on the right. Its aim is to make racially biased behavior seem both acceptable and normal."

But as the FDIC's Horne had observed, just one of the 70 banks covered by the study was responsible for almost half of the minorities' rejections. That bank was minority owned.

Only The New York Times reported Horne's published views. The reporter, Peter Passell, has a doctorate in economics and taught at Columbia University before the Times hired him.