This is the approach taken by the testers at Automobile magazine:
TESTING PROCEDURES AND PHILOSOPHY
Equipment: VBOX II GPS-based system manufactured by Racelogic, Buckingham, England. (You can verify our results by buying your own at
www.racelogic.co.uk for $11,800.) A Doppler shift applied to radio signals received from the US Department of Defense Global Positioning System satellites yields velocity, position, and other information which is recorded for analysis.
Vehicle state: Driver (no passengers), full fuel tank, break-in mileage.
Driver: In the interests of consistency and comparability, all formal testing is conducted by Don Sherman, Automobile Magazine's technical editor. He's been doing this for 34 years.
Acceleration test: Standing start with aggressive clutch engagement (vehicles equipped with manual transmission) or brake torque (vehicles equipped with automatic transmission). Traction and stability control systems are disabled when possible. An appropriate amount of wheel spin is encouraged to obtain peak performance. Lift-throttle up-shifts are rapidly executed with redlines observed. Drag strip "roll out" time and distance are NOT subtracted from published results.
Passing acceleration: With the transmission in third gear (manual-shift vehicles) or drive (automatic-equipped vehicles), and the vehicle cruising at 30mph, the time needed to accelerate to 70mph is measured. Corrections applied to all acceleration figures adjust results to standard weather conditions (60 degrees F, 29.92 in. mercury barometric pressure).
Braking: A pressure switch located on the brake pedal initiates recording of activation speed, deceleration rate, and distance information. Braking distance is mathematically adjusted to a 70-mph initiation speed. The deceleration rate reported is the peak g's measured during a stop.
Cornering: The test car is accelerated to the adhesion limit on a 400-foot-diameter skid pad circle. The peak g's reported are the averages observed during one-second intervals in each direction.
Speed in gears: The maximum observed velocity without exceeding the engine's redline rpm.
Philosophy: Automobile tests cars for a variety of reasons. While many makers offer a few snippets of acceleration, top-speed, and gas-mileage information, many do not. Conducting our own tests fills in gaps and adds trustworthy information about passing ability, cornering grip, and stopping performance. Having accurate performance profiles helps us pass judgment when we compare one contender to the next or draw conclusions at the end of a Four Seasons evaluation. Since there is no industry-wide standard for car testing, every maker has its own pet procedures. Some test with less than a full tank of fuel, some with two passengers and luggage aboard. Doing our own tests is the only means of leveling the playing field. Some publications strive for the quickest, fastest, or most spectacular results. We make no attempt to emulate quarter-mile drag strip results by subtracting the roll-out portion (the 0.3-0.4 seconds required to move the first foot) of the acceleration run. Drag strips divulge nothing but the speed achieved near the end of the quarter mile and the time required to accelerate that distance. Since the strip reports no other time-to-speed information, subtracting roll-out from 0-60 mph results is never warranted. Many magazines do adjust all of their acceleration results by subtracting the roll-out. We do not because, even though that yields quicker, more tantalizing performance figures, it presents a less accurate picture of the car's abilities.