UK's most unreliable car/........Which

John_

facade said:

John_ said:

facade said:

Mini is a surprise, based on my own empirical data, maybe they have improved a lot recently.

so these surveys, what % of the owner numbers are surveyed? it would need to be 100% for the replies to be meaningful.

No, that’s not how statistics work at all.

I know how statistics work, and you can get pretty much whatever result you wanted out of them.

You need big numbers for the results to be significant.

If you only asked me, and the people that I know have had problems with minis, then you would conclude that minis are rubbish.

That’s a bit of a moving of the goalposts, but it’s still not right. As long as the sample is representative then a sample of 2,000 will give a margin of error of 3%, irrespective of the population size.

Where are you getting the idea from that you need to include a large fraction of all owners to get an accurate result?

WaywardDriver

John_ said:

facade said:

John_ said:

facade said:

Mini is a surprise, based on my own empirical data, maybe they have improved a lot recently.

so these surveys, what % of the owner numbers are surveyed? it would need to be 100% for the replies to be meaningful.

No, that’s not how statistics work at all.

I know how statistics work, and you can get pretty much whatever result you wanted out of them.

You need big numbers for the results to be significant.

If you only asked me, and the people that I know have had problems with minis, then you would conclude that minis are rubbish.

That’s a bit of a moving of the goalposts, but it’s still not right. As long as the sample is representative then a sample of 2,000 will give a margin of error of 3%, irrespective of the population size.

Where are you getting the idea from that you need to include a large fraction of all owners to get an accurate result?

The key word is representative. So as long as we can assume Which readers are representative of the population of car owners, and the sample size is large enough, the results should be robust. 

Mercdriver

Some studies weight it according to the cost of repairs which is partly why the premium marques are often near the bottom.

facade

John_ said:

facade said:

John_ said:

facade said:

Mini is a surprise, based on my own empirical data, maybe they have improved a lot recently.

so these surveys, what % of the owner numbers are surveyed? it would need to be 100% for the replies to be meaningful.

No, that’s not how statistics work at all.

I know how statistics work, and you can get pretty much whatever result you wanted out of them.

You need big numbers for the results to be significant.

If you only asked me, and the people that I know have had problems with minis, then you would conclude that minis are rubbish.

That’s a bit of a moving of the goalposts, but it’s still not right. As long as the sample is representative then a sample of 2,000 will give a margin of error of 3%, irrespective of the population size.

Where are you getting the idea from that you need to include a large fraction of all owners to get an accurate result?

My own cynicism.

As Waywarddriver says, the sample must be representative. I don't believe we can assume that with something like a Which? survey.

You need to ask your 2000 people at as close to random as you can, evenly distributed over the population by age, sex, location, ethnicity, politics, socio-economic group etc., and force them to answer.

onlyfoolsandparking

I think the clue as to how factual any survey could be is in the word 'representative' and for me I will conclude 'not very well' 

Apintplease

sweetsand said:

Dr_Crypto said:

I was thinking about buying a new RR and discussed it with my mates. 4 of them had owned or do own RRs. 2 had no issues and recommended the car. The other 2 had lots of problems and told me to stay away, which I did. 
RR/LR has been known to be unreliable for some time. 

We weare aware of Landrover woes but the RR is in a class of it's own so we had to go for it and experiences as per my previous post. Plus we have other cars so just in case it broke down.  Anyways, we always sell before their warranty runs out so no drama.

So mate what car you go for then, the Bentley or the Rolls SUV or the more reailable ones like the Rav 4 or whatever its called or korean, Japan maker

Seriously you buy a car, then buy others just in case the car of choice breaks down?

Seriously, why are you on a money saving site?????

Mercdriver

I think they live in motoring utopia where nothing ever goes wrong and where all the occupants of the house, but mainly one are all geniuses, and blessed in every way.  We should feel honoured to be in such esteemed company...

Grumpy_chap

sweetsand said:

We weare aware of Landrover woes but the RR is in a class of it's own

Range Rover (assuming that is what the abbreviation RR is used for here) is hardly "in a class of it's own".  Even at list price, a new Range Rover can be had for £36k.  That is no more expensive than many a rep-mobile.

Mercdriver

I wonder if Walter Mitty is aware that Land Rover is the owner of Range Rover and they are just wearing a posher badge??

Scrapit

Mercdriver said:

I wonder if Walter Mitty is aware that Land Rover is the owner of Range Rover and they are just wearing a posher badge??

They still have a landrover badge.