For more than 60 years we've been answering the objective questions—How fast? How quick? How much grip?—comprehensively, and with an authority based in experience that our readers have come to rely on. Our testing started back in 1956, just as the interstate highway system was coming into existence. At that time we went by the name Sports Cars Illustrated, and we tested the old-fashioned way: with a handheld stopwatch and not the high-precision GPS test gear we use today. More recently, we've added a panoply of static tests to measure cargo space, interior stowage pockets, infotainment response time, and the size of blind spots, among other things. While the Mustang fanboy may be upset his performance numbers don't one-up the Camaro's and, conversely, enjoys confirmation that the Camaro's visibility figures are far worse, we hope all readers can appreciate our transparency and objectivity when it comes to test results. There is no question as to whether or not our test results are comparable because we follow the same procedures with all cars, without exception. "What are those procedures," you ask? Read on for the full details on how we collect more than 200 data points on the roughly 400 vehicles that we test every year.
Our dynamic testing is performed on a closed test track. What we do there can be compartmentalized into three basic categories: straight-line performance, cornering/handling, and top speed. The heart of our test equipment is the Racelogic VBOX GPS data logger. VBOX uses the U.S. government's GPS satellite constellation to record speed, position, and acceleration. We have various models of this data logger in our fleet, ranging from 10-Hz units (that's 10 points of data per second) to 100 Hz, and one of them even uses the Russian GLONASS satellite system in conjunction with GPS to deliver speed accuracy within 0.1 mph and positional accuracy within about six feet. Piggybacking it with a GPS base station (a device used to correct GPS positional inaccuracy) and a VBOX 3iSL (100Hz) can deliver positional accuracy to within one inch. The VBOX is what we use to measure acceleration times, braking, and top speed. Our VBOX 3i units (we have four of them) also have the ability to log vehicle data such as steering angle, engine speed, and throttle position through the vehicle's CAN communication interface.
Straight-line acceleration consists of three different tests: the standing start (from which we pull all the zero-to-speed times), the 5-to-60-mph rolling start, and two top-gear acceleration tests (30 to 50 mph and 50 to 70 mph). The rolling start is a C/D creation, in which we creep along at 5 mph and then accelerate as hard as possible. This test illuminates differences in powertrain flexibility. The larger the difference between a 5-to-60-mph and a zero-to-60-mph run, the more lag an engine has; this is particularly relevant today with the flurry of turbocharged engines. Top-gear acceleration, in a manual-transmission car, where we simply goose the throttle and don't downshift, highlights midrange power. In a vehicle with an automatic, the transmission downshifts (and the times are much quicker), so this metric represents a combination of transmission and engine responsiveness. And that means the times between vehicles equipped with manual and automatic transmissions clearly aren't comparable.
Standing start. The quarter-mile. A race from A to B. No matter what you call it, this is the test that most people care about. We test in street conditions, so launch traction is low and not the level of stick you'd find at a local drag strip. We also do not power shift, which is keeping your right foot pinned while completing a shift. It is up to the tester to determine the best launch technique, and this process can mean that some cars (for example, a launch-control-equipped Porsche 911) require just two or three launches to get the best possible time. Conventional automatics may only require five launches. High-power, rear-wheel-drive cars equipped with manual transmissions are the most time-consuming, and finding the sweet spot of balancing wheelspin and clutch engagement (usually in the 3000-to-4000-rpm range, but it varies depending on the car) may take 10 runs or more.
All of our straight-line acceleration results are the average of the best run in opposite directions, to account for wind. Ambient weather conditions—we record absolute barometric pressure and wet- and dry-bulb temperatures trackside—determine how much power an engine makes. Because of that, we also correct acceleration results to 60 degrees Fahrenheit at sea level. Cooler air is denser and contains more oxygen, allowing the engine to burn more fuel and make more power. Similarly, high barometric pressure produces more power than low pressure, and dry air has more oxygen than moist air. All of our standing-start acceleration times also include rollout, a short period of time (typically about 0.3 second) that we subtract from the acceleration figures. It's a phenomenon that stems from the physics of the timing lights at a drag strip, where a car can be rolling for 12 inches or more before the clock starts. We recently changed our procedure to now use the industry standard 1-foot rollout.
When possible, we also measure a vehicle's top speed. We often hit an electronic limiter during the straight-line testing, but some cars' speeds are drag limited, meaning their top speed is limited because of air resistance. Fewer cars are redline limited, meaning their top speed is reached at redline in a gear—upshift and the car can't go as fast. We don't test the top speed of every car because cars have gotten ludicrously fast in the last 20 years and we don't always have access to a safe place to do it.
Chassis performance testing answers two essential questions: how short can a car stop, and how hard can it turn. Our standard braking testing consists of six stops from 70 mph to zero. Five of them are done in close succession, with the sixth stop coming after approximately a mile of cooling so that we can roughly determine how well the brakes shed heat, which is otherwise known as "brake fade." Stopping from exactly 70.0 mph is, obviously, a very difficult thing to do. So, we stop from between 70.0 and 70.5 mph, using a tape switch on the brake pedal so we know exactly when the brake pedal is first touched. Then we correct the distance to a true 70.0-mph start based on the average deceleration from that stop. To avoid any issues with a one-off accomplishment, we report the second-best stop from the group of six as our 70-mph-to-zero distance. On high-performance vehicles, we also measure 100-mph-to-zero distance. The best sports cars wearing high-performance summer tires can stop from 70 mph in the 140-foot range (we measured the new mid-engine Corvette at 149 feet), while heavier trucks wearing knobby off-road tires, such as the Jeep Gladiator Rubicon, require nearly 200 feet. When you need to stop in a hurry, those additional four car lengths it takes to come to a halt can easily be the difference between an elevated heart rate and a significant collision.
Maximum Cornering (Skidpad)
Timing the lap of a car on a flat circle—we usually use a 300-foot-diameter circle, but we occasionally use a smaller circle out of necessity—allows us to calculate the average lateral acceleration of a car. We straddle the painted line that defines the circle with the vehicle's tires in both directions and average the results, which we report as roadholding in g-force. Fun fact: Almost every car turns a faster lap in the counterclockwise direction, because the position of a driver on the inside of a car shifts less load to the stressed outside tires. Cornering results range from 0.61 g for a Mercedes-Benz G-class to nearly double that for the grippiest sports cars.
Just like any other researchers conducting a controlled case study, we use a specific procedure for our test vehicles. Before a vehicle can hit the test track, it must undergo numerous prepping protocols, ensuring that every vehicle's performance is measured on a level playing field. By taking these measures, we can consistently yield accurate real-world results and confidently compare data from a recent test to one from years back.
From start to finish, the prep process is very thorough. All the information is recorded on a template called a track sheet. When the test is completed, the track sheet's data is stored in our database where it will live on for eternity as reference material. First, the vehicle is carefully topped off with fuel and weighed using our Intercomp wireless scales. The corresponding weight of each corner is then recorded on the track sheet. This is where the total weight, as well as front and rear weight distribution, will be calculated and recorded.
Next, a technical assistant scrutinizes every detail of the vehicle, both inside and out. During this inspection, numerous data points are captured and recorded on the track sheet, such as engine layout (front, mid, or rear), driven wheels (front, rear, or all-wheel drive), transmission type, steering wheel turns from lock to lock, and tire specifications. Correctly recording the tire information is a crucial aspect of prepping a car for the track because tires influence nearly all performance metrics, including cornering grip, braking, and launch traction. Not only must the name, size, and any manufacturer-specific markings be noted, but the tires must also be set to the manufacturer's recommended cold-tire-pressure specification, which is located on the vehicle's door placard or in the owner's manual. Finally, the engine oil level is checked to ensure that it is at the recommended level. When all procedures have been completed, the vehicle is finally track ready. With a rich history rooted in instrumented testing, we take great pride in publishing test data that is honest, accurate, and dependable.
Interior Sound Level
While at the test track measuring performance, we also use a Brüel & Kjær 2250-L Class 1 sound meter to measure the sound-pressure level in each car under three different conditions: at idle, while accelerating at wide-open-throttle to 70 mph, and at a steady 70-mph cruise. Each car is tested on the same section of road to ensure the results are comparable, since the road surface has a significant impact on the noise level inside a vehicle.
Fuel Economy and Driving Range
All light-duty vehicles are required by law to have their fuel-economy estimates certified by the U.S. Environmental Protection Agency (EPA). These city, highway, and combined ratings are boldly listed on new vehicles' window stickers and often used by manufacturers as advertisement fodder. Plug-in hybrids and electric vehicles also receive estimates for electric operation. Expressed in MPGe, these estimates are intended to be an easy way to compare the efficiency of an electric to a gasoline-powered car. But there is a drawback to using EPA numbers that few people realize: the agency actually does very few of its own tests. Surprise! The EPA lists ratings that are mostly self-reported by auto manufacturers. Whether the testing is performed by the automaker or the EPA, they are done inside on a sort-of treadmill for vehicles that eliminates variables such as temperature and traffic. These scientific methods provide the best way to directly compare two vehicles. However, the EPA tests are not necessarily indicative of how people drive in the real world, and the test cycles don't include speeds as low as what's experienced in areas of dense traffic or high as those that tend to be driven on U.S. highways. That's why we created our own uniform highway fuel-economy test.
Highway Fuel-Economy Test
We run all our tests at a GPS-verified 75 mph on the same 200-mile out-and-back loop on Michigan's I-94 highway. Our consistent procedure includes a methodical fill-up process, following a specific route, using cruise control, and setting the climate control to the same temperature (72 degrees auto). We also correct for odometer error, and we don't test in heavy wind or rain or with extra passengers. In the event we encounter too much traffic or unusual conditions, we abort the run and try again later.
We follow the same procedure for electric vehicles and plug-in hybrids, except for these, we have additional steps that include making sure the battery is fully charged before starting and recording the kilowatt-hours (kWh) needed to fill the battery after the drive loop. Plug-in hybrids also get a highway EV range and MPGe economy for those miles. MPGe is calculated just like miles per gallon of gas only using the EPA's equivalence factor of 1 gallon = 33.7 kWh of electricity to arrive at the result. For plug-ins that can't hit 75 mph in electric mode, we instead first drain the battery and then start the test in charge-sustaining (hybrid) mode. Since those plug-ins don't use any electricity, their results are in miles per gallon rather than MPGe. Likewise, we have to shorten our route for EVs that don't have the range to complete the entire loop. We still give them an MPGe number, though.
Highway Driving Range
The highway range figure we report is the maximum distance that a vehicle can travel at 75 mph on a full tank of gas. We take the fuel economy from our highway test and multiply it by the vehicle's fuel-tank capacity. For example, the Honda Accord 2.0T automatic averaged 35 mpg on our fuel loop and has a 14.8-gallon tank. This equates to an impressive 518 miles of range, but we round down to the nearest 10-mile increment and publish it as 510 miles. That's because when it comes to something that can strand you by the side of the road, we believe it's better to publish conservative figures rather than distances that are more difficult to achieve. A range figure under about 400 miles is the threshold where fill ups can become annoyingly frequent.
Our process is different for electric vehicles and plug-in hybrids. For plug-ins, we simply note how many miles we get into our loop before the battery runs out of juice and the vehicle switches on the internal-combustion engine. EVs are more complicated, because as the battery charge gets really low they generally can't maintain highway speed and tend to go into a low-speed limp mode. (Plus, then we'd be stranded on the side of the highway.) And we also can't calculate range based on the energy put back into the pack after a test, because that would include the inefficiencies of the charging process. So we note the estimated range and battery state of charge from the trip computer every five miles. We then plot all of those points and fit a curve to project out to our range figure, again rounding down to the nearest 10-mile increment.
Observed Fuel Economy
To give consumers an idea of how efficient a vehicle is in mixed driving conditions, we track all fill-ups and mileage on our test vehicles. We do the same with electric vehicles and plug-in hybrids, except for those we track electrical energy (kWh) instead of gallons of fuel. This information is documented for every model that is part of a comparison, long-term, or instrumented test. However, we eliminate the miles recorded during track testing and during our highway fuel loop. We also ensure that every odometer reading is accurate to create a level playing field for all the cars we test.
The observed fuel-economy number we publish has variables such as driving style (our staffers have heavier feet than most consumers, and some more than others) and distance traveled. This means that comparing the economy of one tested vehicle to another can be imperfect except for in our comparison tests, for which all the cars are driven the same distances and in the same conditions. So we consider our observed mpg as supplementary to the EPA estimates and the results of our real-world highway fuel-economy test.
Cargo Space and Storage
The car-shopping process is both rewarding and exhausting, sometimes even frustrating. People spend countless hours between dealerships and websites narrowing down body styles, drivetrains, brands, and features. While only you can decide between black and beige leather, we can take the heavy lifting out of your car-shopping experience—literally. We measure every nook and cranny inside the vehicle so that we can compare cargo and storage space with its peers and so that you know which vehicle is going to fit the most hockey bags, tool boxes, or Costco pallets before you even set foot on a dealer lot.
Manufacturers provide scientifically measured cargo-volume numbers that adhere to engineering standards, but those figures can vary depending on which version of that standard the automaker is measuring to. These numbers are also difficult to translate into real-world practicality. If Chevrolet states that the Cruze hatchback has 25 cubic feet of space in its trunk, how do you know if that's enough space to get your in-laws and their stuff home from the airport? C/D's testing showed that the Cruze hatchback can fit five pieces of standard carry-on luggage with all seats up, a figure that we think is easier to visualize.
To perform this test, we use cardboard boxes measuring 9.0 by 14.0 by 22.0 inches, the maximum dimensions for carry-on luggage used by major U.S. airlines. Starting from the front row, we begin by moving the front seats to the minimum comfortable setting for a person that's 5'10" tall. Without removing anything we deem necessary for safe travel (headrests, spare tires, first aid kits), we fill the rearmost cargo area (either a sedan's trunk or the cargo area behind an SUV's or minivan's second or third row) with carry-on boxes, attempting to fit as many boxes as possible in the space. After reaching a maximum, we close the space, ensuring that the trunk or liftgate closes without interference—we will not force the door closed or bend any boxes. After the maximum capacity has been recorded, we fold the second and third rows and repeat. If seats fold in multiple ways (some cars have stowable seat bottoms, for example), we will determine the configuration that allows the maximum number of boxes. As with the trunk, all doors must close without interference. For models with various trims, we only retest vehicles with substantial differences in cargo area—such as those with a hybrid battery or a different seating configuration. For pickup trucks, we only test the enclosed space. Our current champ is a Ford Transit Cargo Van with 188 boxes, while a BMW i8 holds just one.
Have you ever situated yourself in a new car only find out that there isn't enough space for your phone, sunglasses, wallet, makeup, chewing gum, purse, and napkins? It's difficult to estimate how one car's interior storage space compares to another's. This is why we devised our ping-pong-ball test. By filling interior storage pockets with ping-pong balls, we can objectively show how vehicles measure up to their competitors.
To perform this test, we remove everything from the cubby, such as owner's manuals or removable storage trays, as these could reasonably be relocated should an owner need to maximize a particular storage space. For open bins, we add ping-pong balls arranged at random until the bin or cubby is full. An open cubby is considered full when the balls reach a level where no ball is more than halfway above the top edge. For closed compartments such as the glovebox or center console, the enclosure must be able to latch closed with no resistance from the ping-pong balls. At that point, the total number of balls is recorded. Every defined storage space in the vehicle is counted, including door pockets and underfloor spaces but excluding seatback map pockets and door handles that fit less than six balls.
While high seating positions and ride heights are all the rage, they often bring high cargo openings. We measure this height—to the nearest tenth of an inch—to show how high an object must be lifted in order to place it in a trunk or cargo area. For sedans this usually means the middle of the trunk lip, and for SUVs and hatchbacks this typically means the carpeted part of the load floor. If a vehicle has an adjustable suspension with a loading or parked setting, we will test at both heights. When testing pickup trucks, we measure to the surface of the open tailgate.
Visibility and Seating Height
In an automotive market experiencing the rapid proliferation of the crossover as well as the resurgence of trucks and SUVs, seating height has never been a more scrutinized metric. As people continue gravitating toward larger vehicles with higher seating heights, this measurement—which we determine by measuring the distance from the driver's hip-point (or H-point) to the ground—will undoubtedly become increasingly important to prospective buyers.
Simply put, the H-point is the theoretical location of an occupant's hip joint in a vehicle's seat, and it's determined using the SAE International engineering organization's aptly named H-point machine (HPM). The HPM is a plastic and steel human-shaped device designed to mimic a 50th-percentile male (69.1 inches tall and weighing 172 pounds). The H-point could be considered the starting point when designing an interior, because it influences many aspects such as roof height, seating height, collision performance, outward visibility, interior packaging, and even the door apertures.
To find the H-point of a car, we set the driver's seat to the middle of its vertical and horizontal travel, giving us a consistent location from which to start each of our tests. Next, the HPM is assembled in the seat. Once the HPM is settled and level, it indicates the H-point with crosshairs located on the side of the device. We verify seating height using a floor jack and a laser level, by lining up the laser with the crosshairs and a level yardstick. It is interesting to note that the H-points of a segment usually fall within a fairly tight range—usually within one to three inches of each other,—except for the SUV and truck segments, where it can be as much as five to 10 inches.
The HPM is a highly versatile instrument; besides its obvious use for locating the H-point, we also use it to measure outward visibility. By mounting a laser to the "head" of the HPM, we measure the horizontal obstruction of each roof pillar in relation to the driver's sightline, measured in degrees. We subtract the obstructed portions from 360 degrees, allowing us to calculate the amount of unobstructed outward visibility.
We also measure how much of the roadway in front of and behind the vehicle is obscured by the car itself. First we measure the distance in feet that is blocked by the hood, then we make the same measurement in feet for how much of the roadway can't be seen when looking over the trunk or hatch through the rearview mirror. The results aren't always intuitive and are heavily dependent on the vehicle's styling. For example, the 2019 Honda Accord has an abnormally long obscured rear view of 138 feet (the Camry's is half as much), while the 2020 Kia Soul has only 21 feet blocked to the rear because it is more upright and doesn't have a trunk protruding off the back of the vehicle.
A car's center of gravity (CG) is the hypothetical point in a vehicle that is equivalent to the average location of all of the masses of the individual components, and it's important because it directly influences a car's dynamic traits. The lower it is, the better. A low CG reduces the load transfer when cornering and thus limits body roll while improving transient behavior. Beyond the implications for ride quality and handling, it also reduces the possibility of a rollover incident. Low-slung sports cars, which are painstakingly engineered to carry the majority of their weight very low to the ground, generally have a very stable and confident feeling through corners, whereas SUVs and tall trucks tend to exhibit more body roll and can feel tippy in the same situations.
To calculate the CG, we first measure the height of the wheel centers and weigh the vehicle using four individual scales (one per wheel) on a perfectly level floor. We calculate the longitudinal location of the center of gravity using the static weight distribution and the wheelbase (the distance between the front and rear axles). The heaviest axle of the car is then lifted and positioned on blocks, which are 17.9 inches tall, and the non-lifted axle is weighed again.
The increase in the weight on the static axle provides the critical variable to solving a trigonometric equation that determines the height of the vehicle's center of gravity. For the mathematically minded, we've detailed this in more advanced language here.
It should be noted that we don’t measure the center-of-gravity height for every vehicle we review. We generally perform this test on performance-oriented cars, as this type of data is of greater interest for them than for mainstream sedans, crossovers, and pickup trucks.
One of the critical characteristics that influences whether or not an infotainment system is a chore to use is how much lag it has to inputs from the touchscreen, control knob, touchpad, or other method of control. To measure this, we use a GoPro camera to record interactions with the infotainment system, measuring the time it takes for the system to respond to a number of commands, such as from the home screen to each individual menu (e.g., navigation, radio, media, settings). We then average all of those times together to arrive at an overall response-time figure for an infotainment system. They vary widely; the best, such as Chrysler's Uconnect, are below 0.3 seconds while the most laggy systems, such as Mazda's, take more than a second to respond. As automakers are constantly tweaking the hardware and software running the infotainment, we record the software version of each car we test.
In addition to measuring response time and noting the many features that make up an infotainment system—including the increasingly popular Apple CarPlay and Android Auto—we also measure the output of each USB port in a vehicle, as more juice means quicker charging time for the various electronic devices we can't live without today. Here, too, the results vary considerably, from 2 amps of current or more for the best ones to under 1 amp for the worst, which can be barely enough to maintain a phone's state of charge if the screen is on and being used to run a navigation app.
Safety and Warranty
Safety is one of the few areas in which C/D relies on outside testing to inform our ratings. Two agencies, the National Highway Traffic Safety Administration (NHTSA) and the Insurance Institute for Highway Safety (IIHS), perform crash tests on cars in the United States. The ratings those agencies produce, which are found on window stickers in dealerships across the country, form the backbone of our safety ratings. When a car hasn't yet been tested by either agency, we don't assign it a safety rating.
That's not to say that we don't perform our own subjective evaluations of available safety equipment. When cars are equipped with active-safety technologies, we leave them engaged during our test drives and note whether they perform well or are intrusive. We also keep a rear-facing infant seat handy to test out the ease of child-seat installation in any given car. We're checking to see how easy it is to access LATCH anchors, whether there's space for a rear-facing seat behind a comfortably adjusted front seat, and whether the child seat sits level without the aid of bolsters or other extra equipment. Finally, we collect data on the availability of a spare tire and evaluate backup-camera operation.
Taken with the crash-test data, all of this information helps to provide a picture not only of a car's safety credentials, but also of how easy it is for buyers to opt into a model with desirable features. Safety is one of only a few categories in which our ratings are given on an absolute scale rather than determined in relation to a vehicle's direct competitors. If a compact sedan needs nine seconds to reach 60 mph but tracks closely with its key rivals, we won't knock it in our ratings. But a poor crash test result is bad news, even if the competition performs similarly.
Most of the cars we test come to us as loans from automakers, so there's not much room to test the practical benefits of any given warranty. However, our long-term test cars, which stay in our stables for 40,000 miles, provide an opportunity to become acquainted with the coverage and service offered by various brands. To rate warranties, we compare the lengths of coverage periods and note whether extras such as regularly scheduled maintenance are included or whether there is a hybrid-specific warranty. We typically award the highest ratings to cars with the longest coverage periods and deduct points when a company falls short of its direct competitors or fails to offer a feature that others in its class provide.
Lest you think our editors simply take a spin around the block before sounding off about a car's build quality, comfort, and driving dynamics, we've created a step-by-step process for capturing and organizing our impressions. We call it the editor's observation sheet (EOS), and this form allows our editors to perform exterior and interior walkarounds while collecting more than 170 individual observations using an iPad or laptop.
Upon starting a new EOS, the editor starts by recording the year, make, and model of vehicle that is under scrutiny as well as its trim and powertrain specifics. The exterior walkaround is next, and paint quality is noted, as are exterior features and panel fit. Next, the editor moves inside to evaluate the cabin, which is where the majority of the observations are made. Some items on the EOS are simply for information gathering, such as cupholder and USB count and port locations, but categories such as material quality, ergonomics, and infotainment usability are rated on a five-point scale, with notes made for each item that will later help the editor when writing the review. Additionally, a rear-facing infant seat is installed in the back seat and the editor evaluates the ease of this process. Then, on a thorough test drive, the editor makes notes in the performance and driving section, including subjective observations about the steering, handling, braking, and powertrain.
Once complete, this EOS evaluation is uploaded to our database and saved for future use, both for the review of this subject vehicle and for subsequent reviews of its competitors in which this vehicle can be used for reference.
When compiling our reviews, we assign ratings on a ten-point scale to every car, truck, and SUV. In general, we determine the ratings based on the car's results in our extensive testing, but also based on our subjective evaluations after driving and spending time in the car.
We evaluate a car only in comparison to its competitors. We won't knock a minivan for needing more time to complete a quarter-mile than a supercar does, for instance. A hybrid that earns 40 mpg on the highway may be derided, while a crossover will earn praise for topping 30 mpg. In certain instances, though, we rate cars based on the overall market. Poor crash-test scores, for instance, are not overlooked simply because the rest of the class has also struggled. Because opinions can vary among our staff, we discuss (argue) before a core team finalizes them, to ensure that no rogue opinions can unfairly raise or tank a vehicle's score.
Dave VanderWerp has spent more than 20 years in the automotive industry, in varied roles from engineering to product consulting, and now leading Car and Driver's vehicle-testing efforts. Dave got his very lucky start at C/D by happening to submit an unsolicited resume at just the right time to land a part-time road warrior job when he was a student at the University of Michigan, where he immediately became enthralled with the world of automotive journalism.