2009 Annual Auto Reliability Survey: Domestics vs. the imports

Asian vehicles continue to lead in car reliability, but a number of domestic and European automakers are showing improvements. Of the 48 models with top reliability scores, 36 are Asian—Toyota accounts for 18; Honda, eight; Nissan, four; and Hyundai/Kia and Subaru, three each. The rest come from Ford with eight; Mercedes-Benz, one; Volkswagen, two; and Volvo, one.

European brands continue to improve in reliability with a significant rebound from Mercedes-Benz and improvements from Volkswagen and Audi. Most Volvo and Porsche models have also done well in our survey.

Ford leads the domestics with about 90 percent of models scoring average or better in our survey. Overall 20 of the 48 GM models have average reliability, but only the Chevrolet Malibu V6 has above average reliability in our survey. Last year, we couldn’t recommend any Chrysler product due to low test scores and poor reliability, but this year the 4WD Dodge Ram did well in our tests and has average reliability. However, more than one-third of Chrysler products are much worse than average in reliability.

The graphs below show the percentage of models that rate above average, average, and below average for Asian, European, and American brands.

 

For more details on our 2009 Annual Auto Survey including the full list of most and least reliable new cars by vehicle type, see our reliability report. Also, check out our guide to car reliability for more details on new and used car reliability as well as owner satisfaction.

—Liza Barth 

Comments

[Sjones]

What constitutes "average" here? If the average car has a 3% problem rate does that make a less than average car "unreliable"? Also, since we don't know what percentage of respondents drive the most reliable brands we have no idea if the "average" is representative of the mix of vehicles purchased by Americans as a whole. If 75% of CR respondents drive Toyotas and Hondas that would skew the average reliability rating which would make any of the lesser brands look worse than they actually are. This is why random sampling is so important.

[Jeff Bartlett]

The data reveals the models that are better than average, average, and worse than average. While true that most vehicles prove to be relatively reliable, this valuable insight can steer consumers toward models that are less likely to give them troubles.

The survey is conducted among our 7 million print and online subscribers. While it is not a nationwide, random sample, the sample size is massive. No vehicle is included in the final report that does not have a large enough sample to be statistically valid. Our survey teams also conducts additional validation surveys to ensure the findings are fair and accurate.

For more information on how the reliability survey is performed, read our FAQ:
http://www.consumerreports.org/cro/cars/new-cars/auto-test/consumer-reports-car-reliability-faq-8-06/overview/0608_consumer-reports-carreliability-faq_ov.htm

[Jeff]

Jeff Bartlett,

I find your comment "While true that most vehicles prove to be relatively reliable, this valuable insight can steer consumers toward models that are less likely to give them troubles" quite interesting. Indeed, many knowledgeable people in the automotive industry or who track the automotive industry would say you are correct about today's car's reliability.

With very few exceptions, most new cars are reliable. If you would break down the numbers in your survey between a car with average reliability or below average reliability, you would find that the differences are statistically insignificant.

As to your other point about the size of the sample. Consider me unimpressed. First, how many people out of those 7 million responded? I say this because the Literary Digest survey of who would win the 1936 election between the incumbent President Democrat Franklin D. Roosevelt and the Republican challenger, Kansas Governor Alfred Landon proves the size of the survey does not matter if the sample is bad. The Literary Digest survey (which had a reputation of being quite accurate going into the 36 election) showed that Landon would defeat Roosevelt fairly handily. The survey went out to 10 million people and nearly 2.5 million people responded. However, as we now know, FDR handily defeated Landon.

The survey was found to have a sample bias (one that certainly can be directed at CR), and despite the large number of responses, overall the response rate was good, but not great. I could bring up several other issues with CR's survey from a methdology standpoint, but personally I don't think CR cares. It is a marketing tool that helps sell CR magazines and online subscriptions. It is not really about what is a reliable car. You admitted the truth in your response.

Also, the Gallup poll of the 1936 election got it right. They predicted a FDR landslide win. The survey used rigorous statistical methods. The number of people surveyed: 50,000. A far cry from Literary Digest's 2.5 million respondents. They got it right, however, because the survey was conducted of a cross section of likely voters that represented percentage wise the various demographics that existed at that time.

[Jeff Bartlett - Consumer Reports]

As our numerous reports and blogs list, this year's data is based on 1.4 million vehicles. Great efforts are taken by our staff to ensure the survey data is fair, and for decades our revised methodology has proven an excellent tracker of past performance and predictor of future reliability.

Learn more about how we conduct the surveys and the latest results at:
http://www.consumerreports.org/cro/cars/reliability/index.htm