Subtracting the "blind", audiophiles essentially ABX everything they try in their system. While the hometheaterhifi experiment has problems affecting its validity, these problems are not inherent to ABX. Two factors that apparently people seem to be implied with ABX are time and stress. Time does not have to be a parameter, really. It could be a factor with this experiment - I think the training was too short, but it does not have to be. Although stress is typically not the significant contributor people think it is to cognitive tests, it too can be varied (and tested if one wants). What I would want with an ABX test for PCs is to mimic home trials, without revealing the pc models. That has practical difficulties, but...
Another problem with this experiment is that there is no real hypothesis and barely a conclusion. This subject is difficult to discuss when people are coming up with wrong, overly broad and multiple conclusions. There is only one conclusion per variable which includes the conditions of experiment, and there should have been a distinct hypothesis to eliminate the BS. The hypothesis determines the test procedure...there is no general best way. While the experiment was a good exercise, it was too flawed, imo, to support whether or not "pcs make a difference". Plus no one experiment can yield such a sweeping conclusion. Other factors can be manipulated that could affect the results (that is why I stated the conclusions must include test conditions).
PS -
"Some people can hear the differences and repeatedly get them right, while some people can't."
"There is so much psychological stuff at play here that we can't tell if the differences are real or imagined"
These are not factors after decent sampling, and statistical analysis.
Another problem with this experiment is that there is no real hypothesis and barely a conclusion. This subject is difficult to discuss when people are coming up with wrong, overly broad and multiple conclusions. There is only one conclusion per variable which includes the conditions of experiment, and there should have been a distinct hypothesis to eliminate the BS. The hypothesis determines the test procedure...there is no general best way. While the experiment was a good exercise, it was too flawed, imo, to support whether or not "pcs make a difference". Plus no one experiment can yield such a sweeping conclusion. Other factors can be manipulated that could affect the results (that is why I stated the conclusions must include test conditions).
PS -
"Some people can hear the differences and repeatedly get them right, while some people can't."
"There is so much psychological stuff at play here that we can't tell if the differences are real or imagined"
These are not factors after decent sampling, and statistical analysis.