When Sen. Edward M. Kennedy's car plunged off a bridge on Massachusetts' Chappaquiddick Island 10 1/2 years ago, he was driving at a much faster speed than he reported to police, according to a new scientific study of the accident.

The study, commissioned by the Reader's Digest and said to have been based on sophisticated analytical techniques, concluded that Kennedy was driving at about 34 miles an hour just before he slammed on his brakes.

Kennedy told police and later testified at the inquest into the death of his woman passenger, Mary Jo Kopechne, that he had been going approximately 20 miles an hour. He said he didn't see the narrow, hump-backed bridge ahead of him "until the split second before I was on it."

By contrast, the Digest study, summed up in an article in the magazine's February issue, said that Kennedy saw the bridge when he was at least 50 feet away from it, engaged in "panic braking" that locked the front wheels, and was still traveling at 22 to 28 miles an hour when the car sailed into the air.

A spokesman for Kennedy said early yesterday afternoon that the senator had no comment on this or other aspects of the article. The Digest account labeled Kennedy's account of the accident "false" in several significant respects including his story of swimming back to Edgartown after the accident.

With just a week left before the crucial Iowa precinct caucuses in the Democratic presidential race, however, the Kennedy forces changed signals and staged a full-dress press conference last evening, the first ever about Chappaquiddick for Kennedy campaign manager Stephen Smith. He asserted that the Digest had not given Kennedy a fair chance to respond before publication and maintained that the story was "seriously in error." But he and several experts he brought with him, including two admiralty lawyers from Boston, spent most of their time in an effort to substantiate Kennedy's swim back to Edgartown.

As for Kennedy's speed at the time of the accident, the senator's press secretary, Tom Southwick, said: "He's testified to that. His testimony is accurate."

The Digest said it repeatedly asked Kennedy for an interview during the preparation of its story but "he would not agree to meet with us."

The Massachusetts judge who conducted the inquest, Edgartown magistrate James A. Boyle, held that "a speed of even 20 miles per hour, as Kennedy testified to, operating a car as large as [his 1967] Oldsmobile, would at least be negligent, and possibly reckless."

Actually, Kennedy appears to have been uncertain from the outset of what his speed had been.

In his original accident report to Edgartown police on July 23, 1969, five days after the accident, Kennedy put his speed at 20 miles an hour, but the figure was typed over an erasure. The erased figure appeared to have been 15 mph when the report was inspected by a Washington Post reporter in the summer of 1969. Another reporter thought it was 35 mph.

Asked about the erasure in an interview last fall, Kennedy said he could not recall making any change. He said 20 mph was his "best estimate" of his speed. Kennedy noted that a study conducted for him in 1969 by Arthur D. Little Inc. agreed that "the approach to the bridge was made at approximately 20 mph."

According to the Digest article, written by senior editor John Barron, the study of the accident was conducted by Raymond R. McHenry, "one of the nation's foremost experts in automobile-accident analysis." A copy of McHenry's summary report was made public yesterday, along with the text of the article. He is a staff scientist for Research Engineers Inc. a consulting engineering firm in Research Triangle Park, N.C.

Summarizing his conclusions, the Digest said McHenry found that:

"Driving on the wrong [left] side of the road, Kennedy approached the bridge at approximately 34 mph." Abiding by scientific standards, McHenry stipulates that his speed computations could be in error by plus or minus 4 mph. So the car would have been traveling at a minimum of 30 mph and could have been going as fast as 38 mph.

"Kennedy saw the bridge when he was at least 50 feet away from it, probably from farther away. At least 17 feet from the bridge, he slammed the brakes down hard, 'panic braking,' which locked the front wheels. Propelled by the high speed, the car skidded 17 feet along the road, about another 25 feet up the bridge, jumped a 5 1/2-inch-high rub rail, and hurled approximately 35 more feet into the water. Despite Kennedy's braking effort, the car was still traveling between 22 mph and 28 mph when it shot out over the pond."

The Digest said that McHenry used "sophisticated analytical techniques, validated by the Department of Transportation, and accepted in numerous legal cases."

The magazine said McHenry fed masses of data, including the weight and wheel base of the car, the elevation of the road, and the geometic features of the bridge, into an IBM computer. Then, after "repeated computer runs, he mathematically re-created the movements of the car."

A spokesman for the National Highway Traffic Safety Administration, Bob Boaz, confirmed last night that McHenry "is looked on as one of the country's foremost men in accident reconstruction."

At the news conference, Smith devoted more attention to the dispute over whether Kennedy swam back to Edgartown around 1:30 a.m. on July 19, 1969, after attempting to rescue Kopechne. The Digest had commissioned another scientific study, which concluded that far from being swept northward and almost drowning, as he testified, Kennedy, if he "encountered any current at all," would have been swept southward, "in exactly the opposite direction."

One of the experts hired by Kennedy, Lawrence J. Hoch, an admiralty lawyer from Boston, said the current did not turn slack until 1:36 a.m. and that Kennedy would have encountered a 1/4-knot pull to the north six minutes earlier. But Hoch acknowledged that Kennedy would have felt a much stronger pull to the north if he had jumped into the water much earlier -- without having tried to rescue Kopechne -- as an Associated Press study several years ago suggested.

Apparently relying largely on the testimony of Kennedy's cousin, Joseph Gargan, the Digest estimated that Kennedy jumped into the water to swim back to Edgartown sometime between 1:35 and 1:45 a.m. on July 19. By then, the magazine reported, the current was flowing "southward with increasing velocity."