Reliability study- Who likes stats?

This site may earn a commission from merchant affiliate
links, including eBay, Amazon, Skimlinks, and others.

Nottajeep

Nottajeep

SILVER Star
10 Year 15 Year
Joined
Aug 17, 2008
Threads
102
Messages
3,422
Location
BFE
From the numbers in the link, the short answer is that the LX and the LC are the same, statistically speaking, in terms of power train failures at least. They don't give the exact formula (that I could find at a quick glance on my phone) for their "quality index", but interestingly enough, it's one of the few studies that published their raw data. It also didn't mention what model years were included in the survey, and if they were proportional for all models.

Anyway, on to the data for the LX v LC...

For the LX, they reported a 3.48% Power Train failure rate, out of 230 sampled. This works out to 8 failures. The LC had a 5.46% failure rate out of 183 sampled, working out to 10 failures.

Start with the assumption (hypothesis) that the failure rates are equal between the LX and the LC. Then use the data collected - 8/230 failures for LX and 10/183 failures for the LC - to test that hypothesis. Using a two sample binomial (pass/fail) test of proportions, we find out that the data collected DOES NOT provide enough evidence to reject our initial hypothesis that the failure rates are the same, at least at the 95% confidence level (alpha = 0.05). The p-value for this test is 0.327. Therefore, we must conclude that it is possible that the sample results (LX better than LC) "might" be a function of simple sampling variation due to the relatively low sample sizes. In other words, if a different sample of 230-ish LX owners and a different sample of 183-ish LC were obtained, the results might be reversed, or at least more similar, or even slightly worse, but would still not refute our initial assumption of equal power train failure rates...

In all, the best way to really answer the question "Is the LX power train failure rate significantly lower than the LC power train failure rate" would be to get a larger sample size (size dependent mainly on what you mean by "significantly lower". 1% better? 2% better, 0.5% better? etc.), and also to make sure the sample was randomly obtained, and representative of the population of LC/LX's. For instance, in the early years, 100-series LC's outsold the LX's by a wide margin. After 2001's introduction of the Sequoia, the LX began to outsell the LC by a large margin, therefore the sample of cars in the study, if not controlled, may skew towards older LC's and newer LX's, but that split is not available on the website (that I could see), which might impact mileage-related failures.

So, enough rehash of "that class" for you??? :lol::lol:
 
re_guderian said:
From the numbers in the link, the short answer is that the LX and the LC are the same, statistically speaking, in terms of power train failures at least.
Click to expand...

Exactly. Well played Rob! Another example of why journalists should not handle statistics.
 
Only on this site would anyone complain about a journalist screwing up the comparison of the LC to the Lexus.

All I know is the LC was at the top!(statistically speaking).
 
re_guderian said:
From the numbers in the link, the short answer is that the LX and the LC are the same, statistically speaking, in terms of power train failures at least. They don't give the exact formula (that I could find at a quick glance on my phone) for their "quality index", but interestingly enough, it's one of the few studies that published their raw data. It also didn't mention what model years were included in the survey, and if they were proportional for all models.

Anyway, on to the data for the LX v LC...

For the LX, they reported a 3.48% Power Train failure rate, out of 230 sampled. This works out to 8 failures. The LC had a 5.46% failure rate out of 183 sampled, working out to 10 failures.

Start with the assumption (hypothesis) that the failure rates are equal between the LX and the LC. Then use the data collected - 8/230 failures for LX and 10/183 failures for the LC - to test that hypothesis. Using a two sample binomial (pass/fail) test of proportions, we find out that the data collected DOES NOT provide enough evidence to reject our initial hypothesis that the failure rates are the same, at least at the 95% confidence level (alpha = 0.05). The p-value for this test is 0.327. Therefore, we must conclude that it is possible that the sample results (LX better than LC) "might" be a function of simple sampling variation due to the relatively low sample sizes. In other words, if a different sample of 230-ish LX owners and a different sample of 183-ish LC were obtained, the results might be reversed, or at least more similar, or even slightly worse, but would still not refute our initial assumption of equal power train failure rates...

In all, the best way to really answer the question "Is the LX power train failure rate significantly lower than the LC power train failure rate" would be to get a larger sample size (size dependent mainly on what you mean by "significantly lower". 1% better? 2% better, 0.5% better? etc.), and also to make sure the sample was randomly obtained, and representative of the population of LC/LX's. For instance, in the early years, 100-series LC's outsold the LX's by a wide margin. After 2001's introduction of the Sequoia, the LX began to outsell the LC by a large margin, therefore the sample of cars in the study, if not controlled, may skew towards older LC's and newer LX's, but that split is not available on the website (that I could see), which might impact mileage-related failures.

So, enough rehash of "that class" for you??? :lol::lol:
Click to expand...

I think you just gave me a headache :)
 
I actually think the failure rates are higher in LCs. They both have the same build, but people who buy a new LX are less likely to go off road and are more likely to have their cars serviced at Lexus at the proper intervals. They are also less likely to tow things, etc. These are a couple of reasons that I would rather buy a used LX than a used LC.
 
You must log in or register to reply here.

Share this page

Back
Top Bottom