EAI schools' test scores fall short

Baltimore's school system overstated gains last spring when it released preliminary test scores for the eight elementary schools run by for-profit Education Alternatives Inc., newly revised results released yesterday show.

Yesterday's new data also show that reading and math scores have declined at EAI-run schools over the past two years while rising districtwide.


At a June news conference, city school leaders had touted overall average increases between last spring and the previous year in reading comprehension scores on the Comprehensive Test of Basic Skills for the eight "Tesseract" elementary schools run by EAI. In reality, the average scores in reading comprehension for the combined eight schools declined, new results show.

The school system reported in June that reading comprehension scores had increased at five of the eight schools. But they actually rose at only two of them and declined at the other six EAI-run schools, the new results show.


The new data also show that the math concepts and applications scores increased only about half as much as the school system officials reported they had in June.

Superintendent Walter G. Amprey and other top school system officials could not explain the discrepancies at a news conference yesterday when they released citywide and EAI test results.

Dr. Amprey said in a statement last night that he did not know about the discrepancies until yesterday. "The discrepancies -- though small -- are particularly disappointing to me because we've made every effort to respond accurately to the intense public scrutiny of the Tesseract program," he added.

At yesterday morning's news conference, Dr. Amprey said: "I'm not sure I can answer the question about last spring or why it's different from now. I'd be happy to examine that. . . . I would rather spend the time focusing on whether [the revised data] is accurate and whether it's an appropriate reflection of what we're doing with young people."

Dr. Amprey's new chief of educational accountability, L'Tanya Sloan, said: "I think you were given incorrect data in the spring. . . . I don't know what you got previously, but you can rest assured that this is correct."

EAI skeptics say the discrepancies heighten doubts about the validity of school system claims purporting success of the privatization experiment, now in the third of a five-year city contract worth more than $140 million.

The harshest critics, including the 8,500-member Baltimore Teachers Union, suggest that Dr. Amprey and his administration have failed in the crucial oversight role and instead acted more as an advocate for EAI.

Speaking of the overstated gains, Linda Prudente, Baltimore Teachers Union spokeswoman, said: "It sends up a lot of red flags to us because we have long been concerned with the lack of oversight for these schools. There's nobody minding the store; there's nobody watching so [EAI] can do anything they want."


EAI vied for contract

The June news conference came as EAI, a Minnesota-based company, was in the midst of negotiating a contact to run the entire Hartford, Conn., school district.

Hartford's school board approved a five-year contract this month giving EAI management of the district's 32 schools and $200 million annual budget.

Dr. Amprey denied critics' suggestion that Hartford's approval had anything to do with Baltimore's releasing the faulty preliminary test data.

City Councilman Carl Stokes, who heads the council's Education and Human Resources Committee, questioned whether EAI should be allowed to continue working in the city, based on misleading claims and a lack of promised results.

"It seems to me we ought to make hard decisions about either pulling the contract or placing severe modifications in the contract," said the 2nd District Democrat, who called for more stringent oversight and a cut in city spending on EAI.


How results are computed

The district's test results are measured on what is known as "normal curve equivalents" -- roughly equivalent to a percentile scale -- ranging from 1 to 99, with 50 being the national average.

On its reading tests -- composed of tests in reading comprehension and vocabulary -- the district average for Baltimore's 121 elementary schools rose from 42.3 in 1992 to 44.8 last spring.

Districtwide total math scores -- based on tests in math computations and concepts and applications -- rose from 44.4 in 1992 to 47.7 last spring.

At the EAI-run schools, by contrast, the overall average reading scores fell from 39.8 in 1992 to 37 last spring; and math scores fell from 41.7 to 40 last spring.

Several school system sources describe tension-filled meetings leading up to the June news conference in which administrators fretted about extremely discouraging results of the privatization experiment and made clear the urgency to present test scores in the best possible light.


"Everybody knew the results were disastrous, and the idea was to go public and put the school system's spin on it before anybody else got a hold of any results," said one source, who asked not to be identified.

Controversial test scores

EAI found itself at the center of a controversy over Baltimore's test scores this summer, when the company admitted overstating the academic progress of students.

It had reported in August 1993 that 4,800 students had advanced an average of 0.88 grade levels in three months based on the company's computerized tests. But, responding to published reports, EAI admitted that the level of academic progress reported was accurate for only 954 underachieving students in five of the nine schools.

Test scores released last summer for another EAI-run public school also raised concerns about its ability to improve student performance. A three-year comparative study showed no significant advancement for South Pointe School in Dade County, Fla.

The study tracked the test scores of the students who were in the company's program for three years and found that they advanced at about the same rate as similar students at a regular elementary school.


Measuring progress

Baltimore's school system also had selected "control" schools by which to measure the progress of the Tesseract schools, based on such factors as demographics, number of special education students, promotion rates and test scores.

But the school system said last week that three of the original DTC schools have been replaced "in an effort to improve the selection process" -- a move regarded as unconventional by some statisticians.

District's attendance is up

Also yesterday, the school system released attendance figures showing gains in average daily attendance for the district, from 86.5 percent in 1991 to 89.4 percent in 1994. While releasing separate attendance figures for elementary, middle and high schools, the school system produced no corresponding figures for EAI-run schools, despite The Sun's repeated requests for that information.

EAI could not immediately be reached for comment.