A privately financed scholarship program in Washington State strongly improves students' odds of attending college, two researchers announced on Monday at the annual meeting of the American Educational Research Association. The researchers said that their study of the program suggests that need-based programs could be more effective at raising college-going rates than merit-based grants such as Georgia's HOPE Scholarships.

"There's often a serious gap between low-income students' educational aspirations and what they actually expect to accomplish," said Edward P. St. John, a professor of education at Indiana University at Bloomington. "That gap was significantly reduced by the financial guarantees offered here."

The study concerns the Washington State Achievers Program, a three-year-old project that is financed by the Bill & Melinda Gates Foundation and administered by the Washington Education Foundation.

The program offers college scholarships to students in 16 public high schools located in low-income neighborhoods. Students apply for and secure the grants during the 11th grade, and then receive special college-preparatory assistance during 12th grade. They may apply for tuition grants of up to $4,000 to attend a community college, $6,500 to attend a public four-year college, or $9,000 to attend a private four-year college.

In order to receive the grants, students must enroll at an institution in Washington. The grants are renewable for up to five years, provided that the student is making satisfactory progress. Any other scholarships the students receive are deducted from their Achievers grants.

Students are selected for the program not through grade-point averages or other traditional measures of achievement, but through assessments of their "noncognitive skills," such as their ability to demonstrate leadership. That feature is important to Mr. St. John's analysis: The program has essentially no "merit" component, except insofar as the noncognitive skills serve as proxies for personal traits that might also be captured by traditional measures such as grade-point averages.

Mr. St. John and Shouping Hu, an assistant professor of education at Seton Hall University, looked at members of the program's first cohort -- students who applied for the scholarships in 2000-1 and graduated from high school in 2002 -- in 3 of the 16 participating schools. Ninety-three percent of the students who received a guarantee of college aid went on to enroll in college. Compared with students in neighboring high schools that were not part of the program, the awardees were 2.33 times as likely to attend any college, 2.81 times as likely to attend a four-year college, and 2.95 times as likely to attend a private college.

"If you let students know that they'll be able to afford college, they'll do everything they can to prepare," Mr. St. John said in an interview. "And their parents will do everything they can to help them."

The findings remained strong when the scholars statistically accounted for the socioeconomic differences between the scholarship awardees and students in the comparison schools.

The discovery that students will take advantage of such a financial-aid program might seem like a no-brainer -- of course families will gladly make use of tuition grants. Mr. St. John, however, sees the findings as a partial rebuttal to scholars who argue that the major barrier that blocks low-income Americans from attending college is not money but a general lack of preparation. Such scholars tend to argue that college-readiness programs such as TRIO and Gear Up are the most important remedies.

In Mr. St. John's view, the tasks of college preparation -- taking Advanced Placement courses, filling out applications -- are indeed crucial. But he insists that economic need is the primary factor. Students are unlikely to prepare for college if they believe they cannot afford it, he said.

Lorraine Solaegui, director of scholarship services at the Washington Education Foundation, predicted that from 66 percent to 75 percent of the program's first cohort will earn bachelor's degrees within six years of enrolling in college. A 2002 federal study, by contrast, found that only 27 percent of low-income Americans (those from the lowest income quartile) who begin college complete a bachelor's degree, she said. The Washington program's awardees are all from households below the 35th percentile of income distribution, and most but not all are from the lowest quartile.

The program "demonstrates that it is possible for students previously tagged as 'not college material' to achieve bachelor's degrees," Ms. Solaegui said in an e-mail message. "The scholarship awards are central, but their effectiveness is maximized by appropriately supporting students to make the most of the opportunities afforded by the dollars."

The Washington program and a similar project in Indiana, known as the Twenty-First Century Scholars Program, offer what is known as "comprehensive encouragement" -- that is, they provide both guaranteed tuition assistance and TRIO-style college-readiness coaching.

Mr. St. John contrasted such programs with merit-aid programs like Georgia's HOPE Scholarships, which provide tuition assistance to all students who maintain a certain grade-point average, no matter how high their family income. Mr. St. John said that such programs have increased low-income students' college-enrollment rates, but have also encouraged many low-income students with lower grades to drop out of high school.

"In Georgia, a high-school student with a 2.5 GPA" -- that is, below the 3.0 threshold for a HOPE Scholarship -- "might decide to prepare to drop out, or to join the military," Mr. St. John said. "Many middle-achieving, middle-income students are coming to believe that they can't afford college."

Georgia officials, however, strongly defended their program. Ben Scafidi, an education-policy adviser to Gov. Sonny Perdue, a Republican, wrote in an e-mail message to The Chronicle: "If the amount of HOPE Scholarships that goes to needy students were counted as need-based aid rather than merit-based aid, Georgia's need-based aid per full-time-equivalent undergraduate student would be measured at $379. ... So Georgia ranks very high in aid per student, yet we have very low tuition. If you look at the consumer cost of college paid by low-income Georgia students, we have got to be very near to (or at) the lowest out-of-pocket cost in the nation."