On 1/10/2012 2:10 PM, dpb wrote:
....
OK, let's see what Titebond themselves says...publish that

TBII 3750 psi 72% wood failure
TBIII 4000 psi 57% wood failure

in the physical properties section.

... The anomaly in the
above data is that they claim a (marginally) higher strength for the
joint but there's a significantly higher fraction of failures of the
joint over the surrounding wood w/ II vis a vis II.
....

That is III vis a vis II above, of course.

I'll also note that as a "probabilistic engineer" given the marginal
difference in strengths and the apparently confounding results of the
wood failure percentages I'd wager that while the two values undoubtedly
are a mean or median of the test data, it is highly unlikely in my
estimation that the difference in the population means would be
"statistically significant" at any level of confidence. IOW, I'm
guessing there's enough spread in the measurements of strength that the
two are essentially indistinguishable in reality.

--