There are two ways to look at this; anecdotal and statistical.
As far as anecdotal observations go, there is possibililty that an increased rate of failure events is occuring.
As far as statistic analysis goes, there is a yet unknown probability value that is waiting to be discovered.
As a statistical process quality control engineer, I have to agree with Wilhelm and Hyde. There are too many variables that are not understood, and far too many open-ended quantifications that are not bound.
What we see is that there are media failures that are popping up in Puro and Wix filters. But we don't know what the sample rate or sample direction is. Once an event happens, it draws focus to that direction. Folks start cutting open more Puros because they are LOOKING for Puro failures; it predisposes the rate because of bias.
There's been a suggestion to do a "standard deviation probability test" ... Here's what I see as issues to overcome:
- I, for one, have ZERO idea of the market density of each brand; how many filters for any given application does Fram, Puro, Wix, Champ, etc sell? Until I would know that, I have no ability to decide what a statistical sample quantity would be.
- I, for one, have ZERO idea of the BITOG filter density of each brand; how many of each brand does the average member consume?
- I, for one, have ZERO idea of how the BITOG population relates to the overall vehicle market population; how many BITOGers have access to direct filter autopsy analysis versus the open market? Are we even going to define the open market as the USA, North America, the World, or what?????
- I, for one, have ZERO idea of the sampled failure rate; while the failures seem to be prominent, how does this factor in with the population of BITOGers that cut filters to begin with?
- I, for one, have ZERO idea of the sampled success rate; how many are cut open and found to be fine, and so they go unreported?
- I, for one, do not have the time or funds to expend to capture the needed data; it would be prohibitively expensive for me to do this alone. I doubt I would be given access to the raw data for sales, distribution, nor the reports of failures gleaned by OEMs. Further, even voluntary cooperative data (member data) would need to be scrutinzed to a very deep level. Too much effort; too little reward. It's not like anyone is going to pay me to do this here ...
I agree that an ANECDOTAL notation could infer some previous (perhaps ongoing) failure rate has escalated in some brands. But I disagree that we have anything substantial to base a conclusion on. And we likely never will.
What each member is left with is the ability to make decisions based upon anecdotal views, because true data is missing for any logical and rational conclusion. In the absence of true data, that's all we're left with. And that's fine for many. But it's false to state otherwise with an expectation of impunity .
.
