It has long been clear that attending college is a paying proposition. Research studies and personal experience alike confirm that college graduates on average earn far more over their lifetimes than workers with only a high school diploma, and in recent years that gap has been growing steadily wider.

But will any institution do, or do some schools convey a greater economic benefit than others? And--more to the point for many parents now facing fall tuition bills--are expensive, high-prestige colleges worth the whopping extra cost, compared with in-state public colleges?

The democratic traditions of the United States encourage us to believe that graduates of State U. are just as well educated and face prospects just as rosy as Ivy League alumni. And with Ivy League tuition and other costs running between $30,000 and $35,000 a year--$120,000 or more for four years--while the in-state tab at public institutions is often half or even a third of that, families have a practical reason to embrace this egalitarian outlook.

New research, though, indicates that, as laudable as these sentiments may be, they are out of touch with reality. In economic terms, at least, elitism pays.

Drawing on data from long-term "longitudinal" surveys of sets of college students dating back to 1960, new studies by Harvard economics professor Caroline M. Hoxby have found that lifetime earnings of graduates of more selective colleges are likely to be greater than those of less-selective institutions, and that the difference is enough, on average, to pay back the extra tuition cost many times over. Often the cost difference is repaid in two to three years of work, she noted.

Hoxby doesn't dispute the notion that bright, motivated kids are likely to be successful no matter where they go. But she finds that they do even better if they go to a top school.

"It is perfectly true that people of high aptitude do well" regardless of their college, she said. "My question is, what is the optimal investment to make" for economic success after college, she added.

Using a system that divided colleges into eight tiers of selectivity, with Rank 1 being the most selective, Hoxby found that moving up two tiers almost always pays off.

Thus, even if a student were to choose an expensive Rank 1 college, such as Dartmouth, over a good and much less expensive public college in Rank 3--the University of Virginia, for example--the student would recover the tuition difference sixfold over the course of a career.

And if the student paid the average tuition at both institutions, rather than the "list price," the recovery would jump to 30-fold (because top private schools tend to give bigger scholarships, thus reducing the price differential).

Put another way, a graduate of a more selective college will generally recover the extra cost within a decade.

Rank 1 schools on the list Hoxby used included such institutions as Stanford, MIT, Cal Tech and Johns Hopkins, along with well-known small colleges such as Amherst, Swarthmore and Williams, and the entire Ivy League.

At Rank 2 were the Unversity of California at Berkeley, Northwestern, the University of Chicago and Tufts, among others.

And at Rank 3 were Virginia, William & Mary, Davidson, Middlebury, Georgetown, Lehigh, Lafayette and others.

Hoxby cautioned that her study was not meant to compare specific schools, such as Dartmouth to U.Va., but only general groupings. Her findings would apply for moving from any Rank 3 school to any Rank 1 school, and the less the cost difference, the greater the payoff.

Thus, moving from an expensive private school in Rank 3 to an expensive private school in Rank 1 would have a large payback because earnings would rise but tuition would be about the same.

Moreover, the payback of a two-level move was great enough in most cases to offset even the switch from a public to a private school or from a scholarship to a non-scholarship situation.

Only "in a few cases where if you were to get free ride at a college that wasn't much less selective" would the less-selective school be a better deal, Hoxby said.

She found similar paybacks result from movement from lower ranks as well, such as from Rank 5 to Rank 3.

Of course, if two schools are in the same rank of selectivity, the less-expensive one is the better deal.

And with a school that is both more selective and cheaper, the payback is astronomical. Thus, top state schools such as Virginia and the University of California at Berkeley show a hefty advantage compared with lower-ranked but higher-cost private schools.

Hoxby also noted that her data on incomes looked at students who had reached age 32. Thus, she used a selectivity ranking from 1982, when these students would have been entering college. Individual schools may have risen or fallen since then, she said, but the general relationship of selectivity to economic return should not be affected.

Her paper, "The Return to Attending a More Selective College: 1960 to the Present," has not yet been published.

There are, of course, many ways of ranking colleges. In its September issue, Kiplinger's Personal Finance magazine ranks private colleges based on what students pay for what they get while there. The magazine balances selectivity, graduation rates, endowment and the like against tuition and other costs.

Based on that, Kiplinger's concludes that Rice University is the best deal among the nation's private colleges--a conclusion, interestingly, that would not be out of line with Hoxby's findings, as Rice is in Rank 1 for selectivity and is one of the cheapest.

But Kiplinger's also finds Princeton tied with the University of Richmond, mainly because Richmond is cheaper. By Hoxby's analysis, students attending Princeton could expect to make up such a difference many times over during their working lives.

After correcting for family income, aptitude (as measured by SAT scores and grades) and demographic factors, she concludes that there is a clear benefit to bright students from going to better schools--not merely that top schools are full of bright students who would do well anyway.

That is, very good students who go to top schools do, on average, better than students of similar ability who go to less selective schools.

"We may state that a student of given aptitude earns more . . . if he experiences a college that has a higher concentration of high-aptitude peers and higher per-student expenditures," she wrote in another of her papers, "Explaining Rising Income and Wage Inequality Among the College Educated," published earlier this year by the National Bureau of Economic Research in Cambridge, Mass.

Hoxby acknowledged that she cannot be sure whether the effect comes about because enrolling with other bright students really improves a student, i.e., that it "generate[s] actual human capital," or whether it represents "an elaborate signaling mechanism" that alerts employers and others, thus giving graduates of selective schools a leg up.

But whatever the mechanism, the payoff not only exists but has been growing in recent years, Hoxby said.

She noted that income inequality among workers who have gone to college has been increasing sharply. Among men with at least some college, the difference between the income of a worker in the 90th percentile of incomes and that of a worker in the 10th percentile climbed from $13,275 in 1975 to about $49,000 in 1995 (both in 1995 dollars).

Looking at the reasons for this growing differential, she and colleague Bridget Terry, also of Harvard, note that in general the American economy is bestowing increasingly disproportionate rewards on people of high intellectual ability.

At the same time, they found that since 1960 "colleges have become increasingly segregated on the basis of students' aptitude, so that aptitude differentials within each college are falling and aptitude differentials are rising."

This segregation results from students' increasing awareness of their options and improvements in transportation that make attendance at a wider geographical range of colleges practical. As a result, it has become less likely that a very bright student from a rural area or small town will automatically go to the state university or nearby small college.

This student today may apply instead to a highly selective college, get in, and in the process push out the child of an alumnus or an applicant from a northeastern suburb who otherwise would have been admitted.

Thus, while socioeconomic status still confers a benefit on students, the magnitude of that benefit is declining.

By the same token, applicants' broader horizon is depriving many schools of a traditional source of high-quality students. Hoxby said this has been particularly hard on some small liberal arts colleges that in the past could count on being a top choice of most high school graduates in the vicinity.

State universities, of course, report that exploding private school tuitions, as well as the "echo boom" in the population, have brought them more high-quality applicants, including many they previously would have lost to, say, an Ivy League school. Officials of some of these schools say this trend more than offsets the loss of bright rural and small-town kids, and many say they are taking advantage of it to raise standards.

Moving Up, Breaking Even

Lifetime earnings of graduates of more selective colleges are likely to be greater than those of less selective institutions, according to recent studies by Harvard economics professor Caroline M. Hoxby. This chart shows the years of earnings needed on average to break even on moving up two selectivity levels.

Moving between Moving from public

two private colleges to private college

Full tuition/Average Full tuition/Average

tuition paid tuition paid

From Rank 3 to

Rank 1 0/0 5.4/1.1

From Rank 4

to Rank 2 0.5/1.9 5.8/2.0

From Rank 5

to Rank 3 11.4/11.9 8.6/3.3

From Rank 6

to Rank 4 6.8/5.9 9.1/3.6

From Rank 7

to Rank 5 4.3/4.2 6.0/2.2

From Rank 8

to Rank 6 9.6/8.2 6.9/2.8