BY one estimate, 50 percent of all births in the United States will be out of wedlock by 2003.
Can federal welfare policy be shaped to curb the trend?
That question is a focal point of the debate in the Senate this week as lawmakers take up changes to the welfare system that would fundamentally alter the way America helps its poor.
Sen. Phil Gramm (R) of Texas thinks denying additional benefits to mothers on welfare who have more children will discourage out-of-wedlock births. And he has some prestigious social-policy analysts to back him up.
Not so, says Senate majority leader Bob Dole (R) of Kansas. Far ahead of Senator Gramm in the race for the Republican presidential nomination, he opposes a so-called ''family cap'' on welfare benefits.
The schism represents more than just Dole-Gramm presidential wrangling. It highlights differences among conservatives in general, reflecting the variations in nuance and emphasis among the so-called pro-family groups that usually work together.
For Gramm, the issue of out-of-wedlock births is at the root of the nation's welfare crisis. In Senate debate, he said: ''The most serious problem with the bill before us is that it does not deal with illegitimacy.''
He is pushing amendments to the Dole bill that would ban additional payments to welfare mothers who give birth again (the ''family cap''), bar cash benefits to unwed teen mothers, and deny benefits to mothers who won't identify the father of their child.
But is there evidence that these measures would make a serious dent in the problem? For now, New Jersey is the only state giving analysts anything to go on as they look at the effectiveness of family caps. New Jersey's family-cap pilot program went into effect in October 1992 and began limiting benefits for children born after July 1993.
But what happened after July 1993 has become a subject of intense debate among experts. Researchers at Rutgers University in New Brunswick, N.J., contracted by the state to evaluate the pilot program, have issued a preliminary finding that from August 1993 through July 1994, ''there is not a statistically significant difference between the birth rates'' of groups that faced capped benefits (6.9 percent) and those that did not (6.7 percent).
In a June letter to a state official, Michael Camasso, the principal investigator on the study, acknowledged that earlier analyses by other researchers had found that the family cap did reduce out-of-wedlock births, but he attributed that to ''lengthy reporting delays of births by clients,'' which resulted in incomplete data.
Mr. Camasso called his own results preliminary, and indicated that more recent data, covering August 1984 through July 1995, would allow for more complete analysis.
Robert Rector, a senior policy analyst at the Heritage Foundation and an influential figure in the welfare debate, disputes the Rutgers results.
Instead of looking just at limited samples of the New Jersey welfare population, he looked at the state's total population of women receiving Aid to Families With Dependent Children (AFDC) and compared their birthrates before and after August 1993.
By that measure, Mr. Rector found an average monthly decline of 9 percent in the AFDC birthrate since August 1993.
He also cites the research of June O'Neill, now director of the Congressional Budget Office, who looked at the limited samples and found declines in births of 20 percent to 29 percent.
These competing studies may only point to one firm conclusion - that it's too soon to say definitively that family caps bring down illegitimate births. But that does not mean Gramm should hold off on his amendment, says Rector.
At first, he says, the New Jersey cap was intended as a moral statement. ''I felt it was important to say we don't support people who have babies they can't take care of,'' he says. ''We didn't believe it would have a big effect.''
After all, he points out, the additional AFDC payment for an added child to a New Jersey welfare family is just $64 a month.
Senator Dole, in agreement with some conservative (and liberal) groups, argues that family caps and other ''child-exclusion'' proposals would encourage abortion and are therefore unacceptable.
His approach is to allow states to decide for themselves how, if at all, to stem the tide of out-of-wedlock births. Dole's bill would set no federal requirements along those lines, while Gramm's version contains eight anti-illegitimacy features.
Jennifer E. Marshall, a policy analyst at the conservative Family Research Council, counters the notion that family caps will seriously hurt children. ''There's nothing more cruel than the current system,'' she says.
And since the family cap bars only an increase in the cash benefit a new baby would bring - while still allowing increases in other forms of aid, such as Medicaid and nutrition programs - it is debatable how severe a burden the family cap would be.
But, as with other experiments in welfare reform, it's too soon to tell.