To the legion of loyal Toyota, Rabbit and Honda owners, the latest crash ratings from the Department of Transportation must seem like an overtly political sneak attack.

What better way to give American cars a badly needed sales boost than to give the most popular small imported cars a black eye?

A closer look at the crash tests on a dozen 1980 small cars shows some dramatic differences in protection provided by the cars, at least as measured by the National Highway Traffic Safety Administration, the DOT agency that conducted the tests on the Chevrolet Chevette and the Fiat Strada passed.

The test: Each car was equipped with two human-form dummies strapped in with seat belts and wired with electronic sensores to register the collision forces on the head, chest and legs as the car struck a wall at 35 miles an hour.

The most critical measurements deal with the forces on the head as it strikes the dashboard and windshield and the blow to the driver's chest as it hits the steering column or dash, said Kennerly Digges, director of passenger vehicle research for NHTSA.

In the tests, Japanese-made minicompacts came out far worse than their American-built competitors.

NHTSA calculates the combined collision forces on the head into a single composite number called the head injury coefficient. A number of well below 1,000 is necessary if the occupant is to avoid serious injury or death, NHTSA says.

In the Honda Prelude tested by NHTSA, the head injury numbers were 2,904 for the "driver" and 1,759 for the "passenger." The corresponding numbers for a Ford Mustang were 819 for the "driver" and 567 for the "passenger"; for a Chevrolet Chevette, 668 and 669; for a Honda Civic, 2,626 and 1,506.

"Thus," said NHTSA's researchers, "it is likely that an occupant of a Honda Civic or Prelude in this crash test would have died. It is likely that the same occupant of a Ford Mustang or a Chevrolet Chevette would have survived without serious injury."

The ratings for the other 1980 cars tested by NHTSA were: Toyota Tercel -- 1,218 and 1,179 (driver and passenger); Datsun 310 -- 1,059 2,019; VW Rabbit -- 1,328 and; Subaru -- 1,087 and 2,837; Toyota Corolla -- 838 and 1,162; Fiat Strade -- 790 and 962; Audi -- 1,322 and 1,428; Datsun 200SX -- 1,091 and 1,032; Mazda -- 1,435 and 2,206; Chevrolet Citation -- 845 and 623.

The Citation was given passing marks in results announced in February.

NHTSA's tests have been challenged by manufacturers on serveral grounds.

First, some automakers object to the sweeping prnouncements on vehicle safety based on a crash test of a single vehicle.

NHTSA have issued a disclaimer that the results "are not a rating system per se, since the tests involve single crashes and do not cover a large number of makes and models." But its officials point to the large disparities between the "best" and "worst" performers in the tests. Despite the disclaimer, the results were broadcast far and wide by NHTSA at a press conference, garnished by displays of the battered cars in the DOT headquarters courtyard.

Secondly, automakers now aren't required to meet the test, NHTSA's critics complain. Toyota noted yesterday that its car passed the current NHTSA safety standards, which involve a 30 mile-an-hour crash without instrumented dummies and which serves to check the strength of selected car parts like the gasoline tank. A 35 -mile-and-hour test will not be imposed on large cars until 1982 and on small cars until 1984.

Joan Claybrook, administrator of NHTSA, concedes the more demanding tests with the instrumented dummies are made precisely to dramatize the agency's contention that small cars are particularly dangerous -- notably the popular Japanese models.

Finally, NHTSA is accused of singling out small cars for criticism. What about the safety of large cars?

There is little doubt that in a crash between a large car and a small car, the occupants of the small car are in much greater risk, say Digges. They will rebound from the collision at a greater speed, adding to the impact forces they suffer. Moreover, in 1979, 51 percent of car occupants killed in single vehicle crashes were in small cars, although small cars make up only 38 percent of the total American car fleet, says Digges -- a disparity that supports NHTSA's warnings.

Nhtsa did perform the crash tests on large 1979 cars, however, and most of the American models and all of the foreign models failed. (Those that passed the occupant protection tests were the Dodge Magnum, Chrysler Cordoba and Lincoln Continental. Those that failed this test were the Chevrolet Impala, Pontiac Catalina, Chevrolet Malibu, Olds 98 , Buick Regency, Buick Rivera, Ford LTD, Mercury Marquis, Dodge St. Regis and Chrysler Nerwport.)

Some of the worst scores in the 1980 tests were due to complete or partial seat belt failure permitting the dummies to strike the front of the car compartment with particularly severe impact, Digges said. Both buckles on the Subaru released during the crash, while the passenger buckle released on the Datsun 310.

NHTSA said the equipment failure was not an isolated event that unfairly branded these cars. The design of Japanese seat belts is particularly poor, the agency claims.

Other poor scores can be explained by structural weaknesses in the Japanese cars, said Claybrook. A key difference is the ability of a car to absorb a front-end blow and keep the engine from crunching into the occupant compartment. The tests show that the Toyota, Honda and Datsun minicompacts and the VW rabbit provide substantially less protection than most U.S. competitors.

The reason, Digges claimed, is that American manufacturers have put more engineering effort into safety design than the Japanese automakers, who appear to have emphasized appearance and driving reliability instead.