Sunday, August 26, 2007

Advisory Panel Finds ‘Reading First’ Data Less Than Definitive

By Kathleen Kennedy Manzo / EdWeek
August 23, 2007

State-by-state data on the federal Reading First program were released with fanfare this past spring by the U.S. Department of Education as evidence that the initiative was fueling “tremendous progress” among students and teachers. The Reading First Advisory Committee, a federal panel that met for the first time here this week, found that information on student achievement in participating schools far less definitive and has asked for more time and technical assistance in evaluating the data from the $1 billion-a-year program.

“It seems like there are no standards anyone can interpret,” said committee member Frank R. Vellutino, a prominent reading researcher from the State University of New York at Albany.

He was referring to the variations among states in the assessments used to gauge students’ reading skills, as well as differences in cutoff scores used to determine whether children are on track toward becoming “proficient” readers.
See Also
Read more stories on the Reading First program.

“It’s going to be a job and a half to come to any conclusion whether states are meeting their responsibilities” in improving achievement in Reading First schools, Mr. Vellutino said.

The committee was given a thick binder filled with test-score data from grantees, which states are required to submit to the department each year. The information is based on the assessments and proficiency benchmarks set by each state as part of its Reading First plan.

Some Results ‘Impressive’

Members assigned to a subcommittee to review the information more closely described some of the data as “impressive.” But they also questioned whether the goals set by each state are appropriate, particularly those states that appear to have low expectations for student proficiency.

“I’ve never seen any test where 40 percent was considered proficient,” said Susan Brady, an early-reading expert at the University of Rhode Island and Haskins Laboratory in New Haven, Conn. “One needs to scrutinize what these levels are and what are the skills behind these cut scores.”

In measuring fluency, many states are consistent. Nearly 40 of the 54 states and other jurisdictions, for example, use the same test and the same cutoff score. But there are more differences in how states gauge comprehension skills. Arkansas and Idaho, for example, deem students to be proficient if they rank at the 40th percentile on the Iowa Tests of Basic Skills. Other states, such as Georgia and Michigan, define the 50th percentile as proficient.

The data include test results for 1st through 3rd graders from 2002 to 2006. Results are broken down into subgroups by race and socioeconomic status, as well as for English-language learners and students with disabilities.

The committee, which met Aug. 20-22, has asked the Education Department to assign a statistician to the task of reviewing the data and suggesting ways to analyze the information and compare results across states.

A more rigorous analysis of student achievement in participating schools commissioned by the department’s Institute of Education Sciences, is under way, and results are expected later this year.

Reading First has been under scrutiny by federal auditors, who have been responding to complaints from several commercial vendors that federal program officials and consultants favored particular reading textbooks, assessments, and approaches over others and directed states to use certain products, a level of federal prescriptiveness that the NCLB law prohibits.

The Education Department’s inspector general largely substantiated those claims, as did a separate review by the Government Accountability Office, the investigative arm of Congress. The House Education and Labor Committee held two hearings on the matter last spring. And a Senate investigation is continuing. ("Senate Report Cites ‘Reading First’ Conflicts," May 16, 2007.)

Puerto Rico Grant Review

Secretary of Education Margaret Spellings formed the advisory committee in response to the inspector general’s reports. The committee will make recommendations to the department, but they are not binding.

In other business from the panel’s discussions:

• Puerto Rico may be closer to getting funding under the Reading First program, but the commonwealth will have to make some additional changes to its proposal based on the committee’s recommendations. The self-governing U.S. territory is the last eligible jurisdiction still awaiting its grant award under the program, which was designed to improve reading instruction in the United States’ most disadvantaged school districts. The committee reviewed Puerto Rico’s latest proposal for its $35.6 million Reading First grant, which has been revised about a dozen times. An earlier version was approved in 2003, but allocation of the funding was held up when new leaders of Puerto Rico’s school system rejected some of the tenets of the plan. ("Puerto Rico Still Has No Reading First Funds," Nov. 30, 2005.)

• Targeted-assistance grants totaling more than $2 million will be split by Massachusetts and Virginia to help them expand their Reading First models beyond participating schools. The grants were given to states that had increased students’ test scores in Reading First schools in grades 1, 2, and 3, and for all subgroups. Tennessee received one of the grants last year.

• The committee agreed to meet again this fall to discuss the data further and take up any other business determined by department officials.

No comments:

Post a Comment