mkaresh

With JET's support, for which I am very grateful, I've been posting periodically about the vehicle reliability research I'm conducting.

The response from this forum has been more mixed than most. Clearly, some people here think that the information this research could provide (and is providing for dozens of other models) is not worth the hassle of a post like this once a month.

If this remains the sentiment, I'll spend my limited amount of time elsewhere, and we'll just have information on the B5 later rather than sooner.

As I see it, I put in a few thousand hours, car owners each put in one to two minutes four or so times per year, and we all get better, more timely reliability information than is available anywhere else.

To encourage participation, panel members receive full access to the results free of charge.

For the details, and to sign up to help out, click the link below.<ul><li><a href="http://www.truedelta.com/reliability.php">Vehicle reliability research - need more Audis</a></li></ul>

owlicks

How many times do people need to tell you? Too many variables, your study is not scientific

mkaresh

I've consulted with people who know far more than either of us about survey research, and they have supported my methods.

Your criticism is not uncommon, but tends to come from people with little or no experience in survey research.

That said, I'll take this as one vote for "no information is better than some information."

Casablanca Avant

value add to me. I'm not ripping on the work you do and the information you are trying to receive but you said it yourself the response from the B5 ownership is mixed. If it were me I would concentrate more on the people who are willing to sign up and provide you information over the people who just rip on you when you post. I think for us it is much easier to search our forum archives to look for information on issues we might have with our cars. That being said with the way the car market goes now a days most people do not keep a car past 5 years and they only care about its reliability for the time of ownership.

BSK

right now you only have a certain user group contributing to your data..

therefore your data is biased

mkaresh

As you might or might not have noticed, I haven't posted here in a few months. All it takes is for a few people to fuss, and it muddies the waters enough that few people sign up.

And yet dozens of B5 owners have signed up--13 for the 1998, 20 for the 1999, 32 for the 2000, and 22 for the 2001. I'm collecting data on the 2000, but the other owners are left hanging until more people sign up.

I hate to do this to them, so I'm doing what I can to get enough owners signed up that I can include these people in the research. They've done their part, now I'm trying to do mine.

mkaresh

Other methods of acquiring participants would be far more expensive, and make it impossible to provide the results for free.

That said, I agree that other methods would be necessary if my results were biased. But, looking at the results themselves and not just theory, there's no evidence of such bias.

With very few exceptions (fewer than statistical methods predict) the results have resembled those from other sources. The difference: an actual number rather than a dot, and much quicker.

The February results will include some 2008s. J.D. Power won't have info on these models until June. CR won't have it until October.

Monkeyboy

Good initiative, but IMO - you started with a flawed plan. The data is biased based on your target contributors.
I've seen it many times where user surveys are done and the data is pretty much meaningless (golf clubs, vacation sites, electronics, etc).

Hope you are not too upset that the time you put in really didn't lead to anything meaningful (unless maybe you learned something)

Silver Streakin'

<ul><li><a href="https://forums.audiworld.com/a4/msgs/2756470.phtml">https://forums.audiworld.com/a4/msgs/2756470.phtml</a</li></ul>

mkaresh

This isn't your usual shoddy survey research. The design accounts for many of the usual weaknesses of a "convenience sample." Most notably, participation isn't one time only.

As I've said elsewhere in this thread, I've spoken to many people with thorough knowledge of survey research about what I'm doing, and they see the method and results as valid.

As for my time, so far I've:

--created a price comparison tool unlike any on the Internet

--launched a Fuel Economy Survey more detailed than others that includes thousands of responses (one of the OEMs contacted me because it was the best such survey they could find)

--built a vehicle reliability research panel that includes nearly 19,000 members, and growing by over 1,000 per month

--the results of this research already include nearly 100 model/model year combinations

So, yeah, I think I'm doing pretty well for one guy. Not necessarily with the B5, but with many other groups of owners.