Tracking the Fuel Economy of Cars.com's 2013 Honda Civic, 2013 Subaru BRZ

We welcomed a 2013 Honda Civic to Cars.com's long-term test fleet in March to evaluate the unheard of changes Honda made just one year after the Civic's complete redesign. Like our 2013 Subaru BRZ, the one-year ownership experience is tracked in numerous ways including detailed gas mileage logs.

The Civic's gas mileage was not one of the numerous alterations for 2013; the 2012 Civic's EPA-rated 28/39 mpg city/highway was already competitive.

Follow Our Long-Term Fleet

We've only just begun to put the LX sedan trim level with an automatic transmission through its paces after logging 1,400 miles in just over two months. The Civic is mostly being used as a commuter car in and around Chicago with city and suburban stop-and-go driving, a style of driving reflected by the observed 28.4 mpg average. Road-trip season is approaching quickly, and we expect to rack up quite a few trips this summer to test the Civic's 39 mpg highway rating.

Cars.com $26,000 Midsize Sedan Showdown: Mileage Results

We took six of the top-selling, most fuel-efficient cars in the U.S. and pitted them against each other. How’d they do in our mileage drive?

While Nissan's redesigned 2013 Altima may have the highest EPA-estimated mileage rating among family sedans, its trip computer had the largest discrepancy with our calculations at the pump. After a 185-mile loop, the Altima's trip computer read 4.38 mpg above our calculations — far higher than the variances we observed for the 2013 Ford Fusion, 2013 Honda Accord, 2013 Hyundai Sonata, 2013 Kia Optima and 2012 Toyota Camry. The Altima won the day's trip-computer mileage with an observed 35.9 mpg, but its calculated pump mileage — 31.52 mpg — landed it a third-place finish behind the Camry and Accord.

Cars.com/USA Today/MotorWeek $26,000 Midsize Sedan Showdown

We took six contenders on a daylong mileage loop in mixed city/highway conditions. Per our usual mileage-challenge procedures, we began and ended at the same gas pump, swapped drivers over roughly equal chunks of seat time, kept windows and sunroofs closed and avoided cruise control. With cool temperatures and low humidity, we kept air conditioning off this time.

At day's end, our calculated pump mileage amounted to less than 1 mpg difference from what the trip computers read for five of the six cars, but the Altima's was off by 4.38 mpg. Why the discrepancy?

Loopholes and Fuzzy Math: The Tangled Science of MPG Ratings

In the wake of the revelations that Hyundai's and Kia's EPA mileage ratings were overstated, it's more important than ever for consumers to understand that mileage ratings are not always what they seem.

On every new-car window sticker, shoppers will see in the same fine print: "Actual results will vary for many reasons, including driving conditions and how you drive and maintain your vehicle."

That makes it sound like your gas mileage is solely your fault, and often, it is. But there's at least one other factor: The particular version of the car you buy and the features that come with it will affect the mileage of your new car no matter how passively you drive it.

Automakers determine gas-mileage ratings based on tailpipe-emissions tests, and the EPA admits that it audits only about 15% of automakers' claims for accuracy. Not every variation of every car model is tested by automakers, though, and you could easily wind up with a version that gets significantly worse mileage than what its Monroney sticker claims. Those claims take into account the engine, transmission and drivetrain, but there are many more variables.

What do we mean? Think sport trim levels, gigantic sunroofs, larger wheels, different seating configurations in large crossovers, optional storage panels ... you name it. They can all affect your gas mileage.

Hyundai, Kia to Pay 900,000 Owners for Bad MPG Estimates

Hyundai and Kia announced today they would reimburse owners of about 900,000 vehicles for misstated mileage ratings.

The news came after an investigation by the EPA into complaints from consumers that the two companies' mileage claims were inflated. As a result, the stated mpg ratings for 75 different trim levels over three model years will be changed, and window stickers of cars on sale will be replaced to reflect the new numbers.

Cars.com's Mileage Tests of Hyundai and Kia Vehicles In Line With EPA Results

The investigation began when the EPA received complaints from owners of the 2012 Hyundai Elantra compact sedan who said that they were not seeing gas mileage near the stated numbers. Once the EPA’s investigators confirmed a discrepancy, they broadened the audit to the rest of the vehicles.

Hyundai and Kia are owned by the same parent company in South Korea but operate separately in the U.S. Both have seen phenomenal growth in the past few years, not just because of improved styling and quality, but also for their exceptional fuel economy.

The changes affect the 2011-2013 model years with mileage ratings dropping from 1 to 6 mpg on the highway. Combined mileage ratings were more accurate with seven of the 43 Hyundai models and four of the 32 Kia models seeing no change in combined mileage after the audit. Twenty-four of the 43 Hyundais showed a difference of 1 mpg combined, while 11 of the 32 Kias were within 1 mpg of the original rating.

The company explained the error this way: "The fuel-economy rating discrepancies resulted from procedural errors during a process called 'coastdown' testing at the companies’ joint testing operations in Korea. Coastdown testing simulates aerodynamic drag, tire rolling resistance and drivetrain frictional losses and provides the technical data used to program the test dynamometers that generate EPA fuel economy ratings."

The two brands will announce a reimbursement program for current and former owners of the vehicles. A personalized debit card will be issued that will calculate the difference based on fuel prices and miles driven, plus 15% as a mea culpa from the automakers.

Will it be enough to satisfy consumers? That’s a question that will be asked nationwide as owners digest the news.

Below are the breakdowns of every vehicle impacted in the change, provided to Cars.com by Hyundai/Kia. We’ll have much more on this story as it develops throughout the day and coming weeks. If you have a question you’d like answered, leave it in the comments below or email us at editor@cars.com.

Daily News Briefs: July 24, 2012

How is the EPA gas mileage on that new car's window sticker determined? It's a complex system, and federal regulators verify just 15% of the ratings, the Detroit Free Press reports. At its Michigan laboratories, the EPA checks between 150 and 300 of the 1,500 to 2,000 car models rated, including every drivetrain variant, the newspaper says. Automakers get 3% wriggle room — for example, EPA tests could render 30.2 mpg on a 31-mpg car — but above that, the EPA imposes penalties and requires the automaker to use the lab results. The penalties are substantial: up to $37,500 per vehicle per day, plus jail for anyone who falsified the report, the Alliance of Automobile Manufacturers told the newspaper. But the laboratory procedures, which comprise five drivers driving at temperatures from 20 to 95 degrees, are "literally hundreds of pages" long, an AAM spokeswoman told the Free Press.

In other news:

Most-Efficient Mazda CX-5 Has Least-Accurate Trip Computer

As part of our just-released $25,000 Compact SUV Shootout, we took the six contenders on an extended drive route of roughly 200 miles to test how efficient they truly were.

In the end, it wasn't a surprise that the highest EPA-rated vehicle won. The 2013 Mazda CX-5 is rated 26/32 mpg city/highway and 29 mpg combined. In our testing of mixed city and interstate roads that slightly favored highway speeds, the CX-5 returned 32.8 mpg. The next closest as well as only other compact SUV to best 30 mpg was the Honda CR-V at 30.3 mpg.

In the past we've often used the mileage computers as the basis of our scoring with the pump calculations as backup. We've written that today's modern trip computers can be more accurate than trying to do the math on a fill-up at the pump.

This time around we were glad we had the pump calculations to fall back on because while most trip computers were off by less than 1 mpg, the Mazda CX-5 was off by 2.7 mpg. That's nearly twice the number of the next worst offender, the Hyundai Tucson, which was off by 1.4 mpg.