The scandal arrived at the state's crime lab in mid-May packaged in two cardboard boxes. Inside, stuffed in manila envelopes, were the 153 answer sheets--two double-sided pages each--from the Iowa tests the Stratfield students took in January. The boxes also contained the March retests and another couple hundred tests taken by students at two other elementary schools, a sampling that would serve as a control group.
Henry Lee directed the lab's investigation, but James Streeter, a "questioned document examiner," did a lot of the legwork. Day after day, he peered through a microscope at the thousands of answer-sheet bubbles, searching for clues. Was there a pattern to the direction and length of the pencil marks? Children generally fill in test-sheet circles randomly, their scratchings varying with each answer. But adults are more consistent, often using the same strokes to darken every circle. Did the pencil marks slash through each bubble the same way? Why did the marks break the bubble's circumference on question three but not on question 23? Such were his days. And at night, Streeter says, the bubbles would float by in his dreams.
A burly ex-cop with a mustache and a bone-crunching handshake, Streeter once used his instincts and street smarts to catch crooks. Now, he works in a white lab coat and tinkers with James Bond-like gadgets to detect forgeries, uncover fraud, and identify the handwriting of bank robbers, kidnappers, extortionists, and the like. One of the instruments used on the Stratfield exams, the Electrostatic Detection Apparatus, or ESDA, is a breadbox-sized device that detects indentations on paper. If, for example, police found a note pad in the home of kidnapping suspects, the ESDA could bring to the surface the indentations made on the pad when the kidnappers wrote their ransom note.
In the Stratfield investigation, Streeter and his colleagues took the tests with the most suspicious pencil strokes or erasures and, one at a time, placed them on the ESDA. After stretching a polymer film similar to Saran Wrap over them, they would then slide a thin tube encasing a tungsten wire back and forth across the answer sheet, charging the paper electrically. Finally, they would pour tiny beads coated with a dark powder resembling photocopy toner across the polymer. If the test sheet carried indentations, the powder magically stuck to that part of the paper, much as glitter sticks to paper lined with glue.
To analyze the pencil markings on the answer sheets, the lab turned to its scanning electron microscope. The SEM makes the image of a scientist bent over the microscope seem as dated as the horse and buggy; it magnifies objects up to 5,000 times their actual size and projects these images onto three computer monitors. With the Stratfield answer sheets, technicians first used the SEM to study the paper disruption caused by the erasure marks. As with the pencil marks, consistency here could have pointed to an adult, not a child, doing the erasing. They also used the SEM to fire electrons into the pencil scratchings on the answer sheets; the resulting energy bursts were then recorded in color computer graphs. If the graphite in one answer bubble kicked off more or less energy than the graphite in another bubble, that could have suggested that the graphite was bonded by two different chemical agents--which would have suggested two pencils had been used on a single answer sheet.
After a month of investigation, however, Streeter and his colleagues could clear up little of the mystery. The ESDA test revealed no indentations on the answer sheets that amounted to anything. Nor did the SEM conclusively show whether more than one pencil had been used on any test. The lab did find scads of adult-sized fingerprints from the documents, and even some "friction ridge imprints" made by an adult palm moving across the sheet. But by the time the cardboard boxes arrived at the lab, a lot of adults had probably handled the tests, including teachers, central-office staff, and the researchers at Riverside. "It's what we call a 'neither nor,' " Streeter explains. He and his colleagues could neither rule out tampering nor prove it. "Was it an unusual case? Yeah," Streeter admits. "But very frustrating. Scientists generally like to find the answers to questions."
Given the lack of conclusive evidence in Lee's June 26 report, Fairfield school officials might have pulled the plug on their investigation. But five days before Lee reported his findings, the district received word from state officials reviewing Stratfield students' work on the Connecticut Mastery Test, the state's mandated achieve-ment exam. Harcourt Brace Educational Measurement, the publishers of the CMT, had compared erasure rates on Stratfield tests with the district's other eight elementary schools. Like Riverside, Harcourt Brace found that both the erasure rates and the percentage of answers changed from wrong to right were "considerably higher" on Stratfield's tests. The company's June 21 analysis did not mention tampering, but it concluded, "It is very unlikely that this occurred by chance."
The district now had analyses from two commercial publishers concluding that Stratfield test results were mighty fishy. What's more, there was evidence to suggest that the tampering had been going on for some time. Riverside researchers had been able to study the Iowa answer sheets from one year's test at Stratfield; the Harcourt Brace analysis, however, was based on two years' worth of tests--1994 and 1995.
And so the probe continued. By Lee's report in June, it was entering its third month. A nationally known testing expert, Susan Phillips of Michigan State University, was now running a statistical analysis. And the district had signed on a local private investigative firm, Checkers, that was interviewing school and district staff. A few weeks later, Nicholas Cioffi, a former judge and state public safety commissioner, was hired to oversee the investigation. All these investigators, district officials said, would go where the facts of the case led them.