Perhaps you should start a new topic as we are off this one.  Until then, I'm not sure where to begin to address the errors in your post.

The Utah system is the most researched system in the world.  To say it's not peer-reviewed is just wrong.  It's not "their" invention - whoever "their" are.

The system was built upon Backster's; however, scientists found his system to be lacking in some areas, so they set out to come up with a scoring system that is empirically based, which is what we have today.  You can read about it in the works of Raskin, Honts, Kircher, Barland, Rovner, Horowitz, Ginton, Horvath, Iacono, Patrick etc, etc, etc....  You'll note the last two are not friends to the CQT.  (Sorry, friends if I missed your name, but I've got to order dinner and pay attention to my kids.)  DACA, based on a study by John Hopkins University has recently adopted the same criteria as the Utah criteria.

Just because research is done at a single university by separate researchers doesn't mean the research wasn't done independently.  (In fact, some of the "Utah" researchers did their research at DACA!)  Also, it is common (actually necessary) in science to set aside a portion of a data set in order to validate findings from the first portion of data to make sure they generalize.  I'm working on a project now that involves both Drs. Kircher and Honts, but I've never been to Utah.  Will this not be independant? Of course it is.

I believe the CPS (computer) scoring was published in a peer-reviewed journal.  I'd have to check.  It was written in the 80's if you want to hunt it down yourself.

What errors?  That's a philosophical question, and it really doesn't matter what I think.  It matters what my employer thinks.  

With that said, it depends.  In a criminal test where there is no real risk of "failing" (at least in my state), then an error only means you wait two hours for the interrogation that was going to happen anyhow.  In a screening situation (and that's what seems to irk you), you're dealing with a different animal.  I have the freedom to take all the time I need with candidates, and as a result I can get almost everybody through - eventually.  That doesn't mean they get hired.  It just means we can get to the point of truthfulness.

For those agencies that have 100 people in line, 99% of whom are more than qualified for the job, then you can set the bar high and only accept those that can make it through a test extremely biased against the truthful (which means better at catching liars).  The cost of errors there is that you lose good people (probably), but you replace them with equally qualified people.

So you see this isn't really a polygraph question.  It's philosophical, and most examiners aren't in positions to make those decisions.

I'm sorry for the brevity, but I've got a child tapping on me for my attention.

Perhaps more later in a new post.