# Collecting Analysis Data

Status
Not open for further replies.

#### VeeDubb

I have enjoyed learning about oils from experts and reading all the analyses posted on this site. However, it seems that we don't really have a systematic way to "test" for the best oil out there; analyses seem to vary a lot by factors such as vehicle, driving conditions, etc. even for the same brand. I would love to do a statistical analysis, like a regression, that controls for all these factors to figure out which oil has the greatest impact on wear rates. However, I don't have enough time to go through all the posts and assemble the data. I would be willing to perform the regression analysis if somebody is willing to collect the data. Are there any takers? The data can be put into an Excel spreadsheet in the following format: Column 1 **Column 2**Column 3** Column4***Etc. Analysis ***OilBrand***Iron****** %Hwy 1 2 3 . . . N where N = number of analysis we have (of course this is growing by the day). The columns contain the key variables we are interested and the rows are the observations (the analyses). You should include as many important variables as available. So for example, if the car is a turbo, create a column with the heading "turbo". Then write either "yes" or "no" into the row matching the analysis depending on whether it's a turbo or not. Another variable to consider is %Hwy driving. Then enter 75 if the analysis was done for a car that was driven 75% highway over the oil run interval. Feel free to add as many variables as you think important. Let me know if anybody is interested. I think this would give us much more info than reading each analysis one by one which is only moderately better than anecdotal evidence since the human mind can't control all the variables.

You could start off with the spreadsheet from Maxima.org, which has oil analysis from Maximas and non-Maximas. Although there are many variables involved (making the study non-definitive), finding which oils have the greatest influence on wear rates (best and worst) is an interesting metric to start out with.

This is an excellent idea, but in going to www.maxima.org, I was unable to find the spread sheet that was mentioned. Where is it located?

quote:
Originally posted by segfault: You could start off with the spreadsheet from Maxima.org, which has oil analysis from Maximas and non-Maximas. Although there are many variables involved (making the study non-definitive), finding which oils have the greatest influence on wear rates (best and worst) is an interesting metric to start out with.
Thanks for the suggesting segfault. I checked out the Maxima spreadsheet and it has some good info but there aren't enough observations (analyses) to do a good regression analysis with many control variables. Plus, the way the data is formatted facilitates readability but not statistical analysis. I think if all the data there were combined with the data on this site, we probably have several hundred observations, which would be better. I think we could get some interesting comparisons between Mobil 1 and Amsoil because we have repeated observations on these two brands across different vehicles and different driving conditions. The more observations we have per brand across different conditions, the better.

Well, since there is no standard and agreed to protocol on which filters to use, how many miles to drive on each oil, driving conditions, pre test condition of vehicles (basically no controls) it appears that what you are trying to accomplish is not feasible. It could yield some indicators but a true statistically significant finding seems unrealistic.

quote:
Originally posted by Spector: Well, since there is no standard and agreed to protocol on which filters to use, how many miles to drive on each oil, driving conditions, pre test condition of vehicles (basically no controls) it appears that what you are trying to accomplish is not feasible. It could yield some indicators but a true statistically significant finding seems unrealistic.
I think most people post the number of miles driven on each oil, driving conditions (e.g. 75% hwy, 25% street), and oil filter do they not? People also give info on what oil was run before the test in most cases. These may not be perfect measures of the variables you mention, but using them beats the eyeball method of evaluating oils.

VeeDub, I have quite a bit of oil analysis data collected over the past eight years and you're welcome to all of it. TooSlick

Status
Not open for further replies.