Hi,

AFAIR - yes.

I have participated and actually conducted a number of DB Tests. These included cables. The results invariably suffer from a fairly small sample size, which makes statistical analysis difficult. That said, when using not a singular criteria like a .05 Significance but more advanced statistical methodes I seem to be doing okay, similar to the old tests conducted.

So a 75 - 80% statistical certainty that my correct identifications of items are due to actual audible differences instead of chance. My score increases significantly in a blind preference test (that is not an ABX test, but a blind test between two random items where preference is marked instead of attempting a "same/different" identification.

Such tests however do not lend themselves to the simplistic "different/same" with a 95% certainty that any "different" identifications are not due to chance (and BTW, as a result of this certainty a equally near certainty that small differences will be missed due to small sample size) so beloved by "Debunkers".

Hence I have so far not bothered to formally publish much, at very small sample sizes and small (if subjectively relevant or significant) differences (such as passive components or cables are arguably limited to) we, if we practice honest and equitable statistics, simply must conclude that we lack sufficient data to satisfy traditional scientific rigor of proof.

Neverless, we may (and do) let such results influence our choices of parts or cables, both personally or in our case for our products.

Kind regards T