Reliability 2.0 vs 2.5

[brnpttmn]

Anything over 30 works with statistics, anything less than 30 is considered small n.

Lol. no.

Anyway, sample size isn't CR's biggest issue. Sampling frame is. You can have all the statistical power in the world, but without a random sample you can still get killed by sampling error.

Sponsored

 

[Fotomoto]

It turns out Toyota has been using square wire for years, so +1 to nothing but mature technologies here.

I swapped out the engine and motor in a 2012 Honda CR-Z hybrid and its pancake EV motor had square windings too. Very cool looking if nothing else!

 

[vap0rtranz]

CR & JDPower will always be criticized because of their methods.

I've never met a CR subscriber. Have you? But I've also never met a Nielsen family. I watch whatever TV shows that I want, including ones that get cancelled after 1 Season And if you ever watch Family Feud, you'll know that the sample really skews the survey. So believe what you will when it comes to surveys.

carcomplaints.com is a source for technical detail to get away from simple surveys. CarComplaints website name sounds negative but it let's you quickly look up TSBs and cross-reference the complaints against the fixes. None of the other sources online let you quickly get at the tech detail behind complaints.

 

[Rkbrumbelow]

Lol. no.

Anyway, sample size isn't CR's biggest issue. Sampling frame is. You can have all the statistical power in the world, but without a random sample you can still get killed by sampling error.

Umm yes, anything over 30 is valid statistically. That’s the cutoff point for big n vs small n. I mention there are a number of other factors but sampling frame is a known variable and there are automatic corrections done for it.

come on this is basic statistics. Want me to go into the actual calculus behind the statistical methods? If you do, this is not the place for it because about 8 people will get it, assuming you are one of those who gets it otherwise probably 7. Scientists, medical persons who have done research, mathematicians, actuaries, statisticians and big data/ad personnel are about it and only 1 cares.

 

[brnpttmn]

[/QUOTE]

Umm yes, anything over 30 is valid statistically. That’s the cutoff point for big n vs small n. I mention there are a number of other factors but sampling frame is a known variable and there are automatic corrections done for it.

So you're saying that n=30 is a valid sample for a survey of car reliability?

 

[Rkbrumbelow]

So you're saying that n=30 is a valid sample for a survey of car reliability?
[/QUOTE]
It is valid for statistical analysis, I did not say it would not lead to a large margin of error.
For a very basic run through: https://www.statisticshowto.com/probability-and-statistics/hypothesis-testing/margin-of-error/

 

[Down]

So you're saying that n=30 is a valid sample for a survey of car reliability?

I think what he's trying to say is that 30 is typically considered the bare minimum that one needs to hit before they can even reasonably expect something approaching a normal distribution. I believe the rule of 30 was really constructed to simply determine whether a z test or t test was required but tons of folks seem to misinterpret this as meaning "a sample size of 30 is always sufficient". The fact he ends up conceding that this sample size, in this situation, could lead to a huge margin of error shows that he also believes this is not a practical solution here; so really? He agrees with you.

I'd guess this is more of an attempt to act super smart based off of a couple Googled ideas, while not actually bringing anything valuable to the discussion, than a valid counterpoint.

Anyway, good luck with this one! I rarely shy away from a nice Internet argument but over statistical theory? Even I have better things to do than that.

 

[brnpttmn]

I think what he's trying to say is that 30 is typically considered the bare minimum that one needs to hit before they can even reasonably expect something approaching a normal distribution. I believe the rule of 30 was really constructed to simply determine whether a z test or t test was required but tons of folks seem to misinterpret this as meaning "a sample size of 30 is always sufficient". The fact he ends up conceding that this sample size, in this situation, could lead to a huge margin of error shows that he also believes this is not a practical solution here; so really? He agrees with you.

Exactly. The rule of 30 is specifically for running specific statistical tests (and it's value in that is debated). I've never seen it used to actual determine sample size (because it's not). A sample of 30 in any population over about 100 is essentially meaningless because you're getting into the range of 20 point confidence intervals. My response was mostly because he was seemingly unequivocally saying that a sample of 30 was enough to determine the reliability of the 2020 Ford Escape. That's absurd.

 

[stoptothink]

Its a bit off topic when it comes to reliability, but there was a great video I watched a little while back with one of those auto channels comparing FWD with all seasons, AWD with all seasons, FWD with winter tires, and AWD with winter tires on the same vehicle. What seemed to surprise a lot of people was that winter tires had a MUCH bigger impact than AWD, particularly in all the safety related tests such as the emergency swerve and the braking test for what should be obvious reasons.

So if you are on a limited budget and had to pick between a second set of wheels/tires for winter on FWD and year round all seasons on AWD, it might not be intuitive, but the the FWD option would drastically outperform the AWD in all the ways that really matter. And, to no one's surprise, the AWD with a set of winter wheels/tires to swap on is the best albeit most expensive option. A lot of people do end up picking one or the other though to keep costs down.

You would think this would be common sense, the huge majority of weather-related accidents are caused by issues which having more drive wheels simply won't help with (braking, steering). The difference in handling in the snow between my mom's AWD outback on all-seasons and my FWD VW with snow tires is massive. My n=1, wife and I with a combined ~25yrs of driving in a snowy mountainous region, never having owned an AWD or 4WD vehicle, have zero snow-related accidents; my mom, brother, and sister (all with Subarus, all-seasons) all got into at least one snow-related accident their first winter in Utah.

Countless videos and articles out there (with data) like this: https://shift.com/articles/all-wheel-drive-versus-front-wheel-drive-in-snow-which-one-is-better

"When it comes to steering, stopping and performing in winter conditions in general, experts tend to agree that there's actually little difference between AWD-equipped vehicles and their standard front-wheel drive counterparts...All-season or winter tires, on the other hand, can improve a vehicle's traction anywhere from 25 to 50 percent..."

If snow, not off-road capability, is the primary concern, it's a better value proposition to get a set of snows for a FWD vehicle than AWD with all-seasons. The large majority of drivers are fine with FWD in snowy mountainous regions. If you have a steep driveway or are commonly hitting the slopes, I'd probably consider AWD.

Me, I don't want to deal with the minor headaches of AWD (replacing all 4 tires at the same time, more maintenance, slightly less efficiency).

 

[fbov]

something approaching a normal distribution

N=30 is a common sample size because it's the minimum number that lets you ask: "What is the distribution of the data." Not all distributions are Normal, only the distribution of averages from repeated random samples is guaranteed to be Normal.

You would think this would be common sense,

+1 Snow driving doesn't need equipment near as much as it needs this!!

 

[WasChops]

I think what we are witnessing here is simply a matter of confirmation bias, and you see the same thing on forums when CR publishes some good news about a particular vehicle the forum specializes in.

Someone is telling you something that you aren't predisposed to want to believe, such as a car you are excited about tempered by a poor rating, and so you're trying hard to dismiss it by making assumptions pulled out of thin air like their sample size being small based on nothing other than the fact that you hope it is.

The Ford Escape is a high sales volume vehicle, and CR indicates that for the 2020 model year they received ~329,000 survey responses, making it the largest independent survey size from a company that does not allow any manufacturer to advertise or compensate them in any way (they buy all their test vehicles, don't take ads, etc). Then can also enhance survey data by combining it when applicable, such as adding survey results for a certain powertrain's reliability across a wide variety of vehicles since manufacturers often parts share across their lineups.

The surveys are also not essay format where people can discuss subjective feelings, but about checking boxes of what broke and then showing all that breakdown in a later issue and giving different weight to different issues to determine their impact, for example:

So this is not a "I didn't like that BMW oil changes were more expensive" when determining reliability. That the suspension was too stiff however would be a complaint that is worthy of noting, and overall satisfaction level or buyers remorse is also important because if ten thousand people answering the survey mostly had buyer's remorse about the vehicle in question, chances are you will too.

I don't work for CR, but I can't think of any other resource that is as trustworthy in their business model. JD Powers for example? Their customers are literally the manufacturers, and they sell their rating accordingly. Heck, GM seemed to pretty much brag about the fact that they paid JD Powers to give all their vehicles awards:

No argument here... I will never proclaim I "know it all"...

 

[bombast]

Good to hear that from a New Englander. Many here in the south somehow deem AWD necessary for winter in the off-chance we even get a decent snow.

When I lived in Texas, I refused to drive in the snow, no matter how light it was. Half an inch of slush and the highway turns into a bumper car rally, with giant AWD and off-roading megatrucks spinning like tops.

I know people down there don't get much practice and all, but seriously, need stop slamming on the brakes and overcorrecting whenever they encounter the slightest slip. No drive system can handle that.

 

[DryHeat]

I've never met a CR subscriber. Have you?

Please allow me to introduce myself. I'm a man of wealth and taste. (h/t Rolling Stones)

 

[Grandalf57]

You would think this would be common sense, the huge majority of weather-related accidents are caused by issues which having more drive wheels simply won't help with (braking, steering). The difference in handling in the snow between my mom's AWD outback on all-seasons and my FWD VW with snow tires is massive. My n=1, wife and I with a combined ~25yrs of driving in a snowy mountainous region, never having owned an AWD or 4WD vehicle, have zero snow-related accidents; my mom, brother, and sister (all with Subarus, all-seasons) all got into at least one snow-related accident their first winter in Utah.

Countless videos and articles out there (with data) like this: https://shift.com/articles/all-wheel-drive-versus-front-wheel-drive-in-snow-which-one-is-better

"When it comes to steering, stopping and performing in winter conditions in general, experts tend to agree that there's actually little difference between AWD-equipped vehicles and their standard front-wheel drive counterparts...All-season or winter tires, on the other hand, can improve a vehicle's traction anywhere from 25 to 50 percent..."

If snow, not off-road capability, is the primary concern, it's a better value proposition to get a set of snows for a FWD vehicle than AWD with all-seasons. The large majority of drivers are fine with FWD in snowy mountainous regions. If you have a steep driveway or are commonly hitting the slopes, I'd probably consider AWD.

Me, I don't want to deal with the minor headaches of AWD (replacing all 4 tires at the same time, more maintenance, slightly less efficiency).

How about AWD and Falken Wildpeak AT3s?

 

[stoptothink]

Falken Wildpeak AT3s

What about them? The Wildpeaks AT3s are 3PMSF, so while maybe not quite as solid as a true dedicated snow tire, they should be great.

Sponsored