Results from Evo 8/9 Transmission Failure Survey
May 30, 2008, 08:26 AM
#1
Newbie
Thread Starter
Join Date: Dec 2004
Posts: 51
Likes: 0
Received 0 Likes
on
0 Posts
Results from Evo 8/9 Transmission Failure Survey
If anyone recalls, a WHILE BACK I had created the following survey
http://www.esurveyspro.com/Survey.as...c-faff66e5e8b4
Ok, I know this wasn't the best designed survey in the world, and had left out a lot of details making this statistically invalid.
However, here are some general results nonetheless.
54 total respondents
26% 03GSR, 3% 04GSR, 9% 05GSR
0% 04RS, 1% 05RS, 3% 06RS
10% 05MR, 20% 06 MR
30% indicated no issues of any kind
21% experienced occasional grinding when shifting
the most problematic gear was 5th gear, with 8.75% of respondents having issues (I'm guessing they were 2003 owners)
1-2 people having diff issues, and 12% who have had clutch issues.
3% of failures happened at the track
9% at the dragstrip
37% on the street
For those who did have failures
13% failed within 5k miles
15% from 5k to 15k
15% from 15k to 30k
5% from 30k to 60k
1% over 60k
49% experienced no failures
With the sample size being so low, I doubt this info is all that useful. And looking at these high level numbers isn't all that informative to begin with.
If anyone here is a statistical whiz (perhaps with some SAS experience?) I can send you the raw data and perhaps you can figure out some meaningful inferences such as the likelihood for failure as HP goes up vs tranny fluid vs fluid drain interval, etc.
http://www.esurveyspro.com/Survey.as...c-faff66e5e8b4
Ok, I know this wasn't the best designed survey in the world, and had left out a lot of details making this statistically invalid.
However, here are some general results nonetheless.
54 total respondents
26% 03GSR, 3% 04GSR, 9% 05GSR
0% 04RS, 1% 05RS, 3% 06RS
10% 05MR, 20% 06 MR
30% indicated no issues of any kind
21% experienced occasional grinding when shifting
the most problematic gear was 5th gear, with 8.75% of respondents having issues (I'm guessing they were 2003 owners)
1-2 people having diff issues, and 12% who have had clutch issues.
3% of failures happened at the track
9% at the dragstrip
37% on the street
For those who did have failures
13% failed within 5k miles
15% from 5k to 15k
15% from 15k to 30k
5% from 30k to 60k
1% over 60k
49% experienced no failures
With the sample size being so low, I doubt this info is all that useful. And looking at these high level numbers isn't all that informative to begin with.
If anyone here is a statistical whiz (perhaps with some SAS experience?) I can send you the raw data and perhaps you can figure out some meaningful inferences such as the likelihood for failure as HP goes up vs tranny fluid vs fluid drain interval, etc.
May 30, 2008, 10:15 AM
#2
Evolving Member
Join Date: Jan 2005
Location: Spewgene OR
Posts: 191
Likes: 0
Received 0 Likes
on
0 Posts
With that small of a sample there is no way to get anything meaningful with all of the different data points you have collected. If one set of data had almost everyone selecting one or two responses, you might be able to make some inferences there. For example, you could probably run an analysis that compared relative risk of each model year for transmission failure, but you would then have to correct for mileage and use in racing, in the end you might end up with nothing.
The main problem you are having is selection bias, if it is a voluntary survey people who have had tranny problems are more likely to participate than those without. You need to have a moderator randomly select a pool of users here and ask them to fill out your survey. Try for 5000 users. Also try to reduce the number of selections possible for each question.
The main problem you are having is selection bias, if it is a voluntary survey people who have had tranny problems are more likely to participate than those without. You need to have a moderator randomly select a pool of users here and ask them to fill out your survey. Try for 5000 users. Also try to reduce the number of selections possible for each question.
