View Single Post
Old 12-29-2010, 11:43 AM   #55 (permalink)
CapriRacer
Tire Geek
 
CapriRacer's Avatar
 
Join Date: May 2008
Location: Let's just say I'm in the US
Posts: 796
Thanks: 4
Thanked 393 Times in 240 Posts
Quote:
Originally Posted by roflwaffle View Post
Someone would need more than just another brand/model. Many brands/models for many years would be needed. Even then, no where in the paper is the relationship bigger = better mentioned, you were the only one who mentioned it...........
It is true that the presentation doesn't say larger = better, and it is also true that I have drawn that conculusion, but I am not the only person to have done so.

Quote:
Originally Posted by roflwaffle View Post
.......In the end the author states...
Quote:
The lack of quality linear correlations between rolling resistances and the basic parameters investigated suggested that if the researcher is investigating manufacturer/tire design differences within a tire size, other more complex aspects of the tire will need to be considered.
So what they are stating is entirely reasonable given the data, and there were no objections because of that. You were the one who stated bigger = better, and that is where the debate is.......
True - and perhaps I've overstated this. However, I think it is painfully obvious that larger sized tires give better RRC values than smaller sized tires - and if that were wrong, I feel confident someone would have pointed this out, particularly considering that there are regulations being written and this would be a very important thing to get right.

Quote:
Originally Posted by roflwaffle View Post
.......In terms of a regression analysis, that's not exactly kosher. One of the assumptions of a regression analysis is that......
Quote:
The sample is representative of the population for the inference prediction
So w/o data for larger sizes you still can't reasonably state that bigger = better even for Goodyear Integritys, and certainly not for tires in general.
But the data IS representative. They are all the same make and model - and you don't have to have EVERY data point to draw a conclusion.

But you are right to point out that there are risks associated with extrapolating data.

Side note: The way science works is that a phenomenon is observed and reported and some conclusions are drawn from those observations - just like I did. Then the data is examined, the analysis critiqued, other studies run, etc. - all with the idea of trying to refute the conclusion. If the conclusion stands up to scrutiny, then it is accepted - until further data comes along. We are in the critique phase - and I have no problem with anyone critiquing, questioning, etc.
  Reply With Quote