Sancho:

I have read Croddy, and it is my understand the claim was dismissed because the court ruled that the applicants could not claim a constitutional violation because they had no constitutionally protected property right to employment.  IN otherwords, no "standing".

Sancho and Mr. Nelson:

Regarding the Honts study, I was not criticizing Professor Honts himself nor his background or qualifications, just his work in this particular study.  I fully admit I have not read all his published work.

I also fully admit I am not a math whiz, and it has been 25 years+ since my undergraduate course in psychological statistics.

The Honts study has too many variables, the worse being:

College students self-reporting that they read some or all of TLBLD.  Of course, they had to say that to get extra credit. The amount of actual study of the material, or the attempt to actually decieve the polygrapher is not quantifiable, and thus unable to be replacated.

Good scientific study is by necessity, able to be replacated, in order to eliminate the possible bias of the researcher.   

Additionally, each sub group tested only had 10 individuals.  To even start to be statistically valid, the number should be 20 or more.   

Lastly, because the results of that study indicated only about a 70 percent acccuracy rate, that also is a serious flaw in making any determination about the ability to use countermeasures, because the data itself had a 30% error rate, (approx).

If questioning the study means I am arrogant, then guilty as charged.  I am in good company though, as expressed in USA v. Williams by Judge King:

"Also, although Dr. Honts struggled to avoid admitting the obvious, a recent report on polygraph examination reliability and validity issued by the National Academy of Science (“NAS”) questioned the quality of all of the available studies regarding polygraph accuracy - including those in which Dr. Honts participated or on which he relies."

The entire ruling is available here:

https://antipolygraph.org/litigation/williams/us-v-williams-daubert-report.pdf

Quote:


That's not true. You just pay a higher price, statistically - so to speak - with a smaller sample.  (You use a different statistic, which requires "more" evidence to make a determination that something is (statistically) significant. 

The reason you break studies down like this and do ANOVAs is to avoid needing an astronomical sample size.  It's very common.

I think you should write the journal editor and peer reviewers and let them know your concerns.  See if your critique can pass the peer-review process.

Quote:


You might want to read the study again.  You've missed a little something.

Quote:


And the New Mexico Supreme Court used that same report to support the acceptance of polygraph in their courts.

Let's not forget that they still found single-issue tests worked.  The biggest reservations seemed to be the then believed possibility of countermeasures, and that fear appears to have very much colored the authors' conclusions and speculations, but recent research has demonstrated that fear to be unfounded.

Do you know what they said in that case?  I've read the entire opinion.  They didn't say polygraph didn't work if that's what you're implying.