posted 07-07-2009 03:05 PM            

------------------
Ben

posted 07-07-2009 03:24 PM            

------------------
Ralph Hilliard
PolygraphPlace Owner & Operator

Be sure to visit our new store for all things Polygraph Related
http://store.polygraphplace.com

posted 07-08-2009 07:24 AM            

------------------
Ben

posted 07-08-2009 02:34 PM            

an applicant is given several opportunties to be truthful prior to meeting with me (original application, update opportunities, background investigation, etc )and now has a conditional offer.

They lie about withholding drug information.

They get ONE opportunity to clear it up with a Bi-Zone/U-Phase exam.

If they fail, does it really matter regarding the particulars?

My opinion is with pre-employments, we should not get into the weeds (pardon the pun) over what they're lying about when we know for sure they are lying or showing significant response.

At that point in the process it is about integrity, not exact truthfulness.

My $.02

Jim

posted 07-08-2009 02:41 PM            

I appreciate that perspective. I treat polygraphs at our agency as a tool for both detection AND disclosure. That is why I am interested in exploring this further. Most of the time, I have enough to disqualify the applicant anyway, but for those that occur a couple of times a year, where I don't have enough to disqualify them, I want something else to assist in getting more post-test admissions. Our agency, rightfully so, will not disqualify someone based on polgyraph SR results alone, unless it includes countermeasures.

------------------
Ben

posted 07-09-2009 04:27 PM            

quote:

---

It's not a SPOT: it's an R/R test. What if there are two areas?

---

quote:

---

If they go DI or Inconclusive, then I'll go into a series of searching POT's as to what they form they have hidden the assets in (cash, jewelry, property, negotiable bonds, drugs, etc) and then where they might be hidden; e.g., offshore accounts, domestic accounts, safe deposit boxes, secure storage units, etc.

---

What if there are two areas in this or similar cases? Is this SPOT or R/R or something else?
For example:
Regarding...
...
3. in a cash?
4. in a jewelry?
5. in a property?
...

posted 07-09-2009 04:39 PM            

Where did you quote your second quote from?

------------------
Ben

posted 07-10-2009 03:11 AM            

quote:

---

It would just be a list of RQs, thus an RR test.

---

posted 07-10-2009 07:31 AM            

For example:
"Regarding the location of that property" - preparatoty phrase
"is it..." - prefix phrase
"in Moscow" - key or padding or false key

Regarding the location of that property is it in London?
...
is it in Moscow?
is it in New York?
...

In this case every question in the POT/SPOT, for example, "is it in Moscow?" is not like relevant questions in tests of other formats, for example, in R/I or ZCT.
It is just my observation, I do not make some conclusion.

posted 07-10-2009 10:54 AM            

The problem is that RR test lacks a clear numerical hand-scoring approach, and is evaluated "globally." Rank scoring systems are sometimes used, and ranking is interesting for obvious reasons that are related to the fundamental strengths of nonparametrics (they allow us to work numerically with crappy data that cannot otherwise be worked with). So nonparametrics are great in that way, because they achieve a dimmentionless transformation of non-normal data. But nonparametrics and rank systems do that at the expense of discarding data on between question variance - meaning that tiny differences are sometimes over-emphasized, and major differences are lost. So non-parametrics are mathematically limited in their potential power - compared to parametric methods, for which dimmensionless transformations like ratios and z-scores do not discard valuable data pertaining to between item variance - which is what we need to use to evaluate a multi-issue test (breathe).

I now have a model algorithm that can score RI or RR exams using parametric inferential methods, but it would put you to sleep if you tried to do the math by hand - that's what 'puters are for.

The practical aspect of an RR test as a follow-up exam is that the test can have good by-case sensitivity, but may or may not provide good by-question sensitivity - that is another problem, especially when their is a possiblity of deception to more than one item. An RR test, like an RI test, can be expected to have weak specificity to truthfulness - and we want a follow-up method to provide increased specificity, not decreased specificity.

Global scoring is not really scoring. Its evaluation. Scoring means numbers and decision models. Scientific minded persons (including scientific minded opponents of the polygraph) will not be impressed with something that is not based on measurements and numerical scores. Global scoring is simply an answer to the question: how do we feel about how the date look. Scoring is numbers, decision models, and decision thresholds - all of which are specified in advance of testing, or we're makin' it up as we go, and that ain't how science works.

What will be unimpressive to our critics and scientific minded outsiders is to over-rely on confessions and information disclosure as a justification for polygraph testing. Yes, its comfortable for us. Confession/disclosure based responses to questions or challenges regarding accuracy are unimpressive to outsiders because they look like of a circular and self-justifying process. Circular arguments, and claims that outsiders cannot understand, are a virtual invitation to accusations of pseudo-science. Science is explainable to scientific minded outsiders.

Not saying disclosures and confessions are unimportant. It's just that broader polygraph acceptance by scientific minded experts and critics will be achieved by incorporating good scientific principles into our processes and procedures, and ensuring a test that can prove its accuracy independent of confessions and remain consistent with scientific and ethical principles inherent to the behavioral sciences and testing.

I say "depending on your agency needs" because there may be possible times when a follow up test that is highly sensitive is desired. For example: if you have far more applicants than jobs, and the work is high-priority, then you could use the added sensitivity to increase the NPV (negative predictive value) of passed exams. This will, of course, be only a small portion of the tests, but with far more applicants than jobs, and a high-stakes situation, it might be desirable to send most people away, and reduce FN errors to very low levels at the same time. The cost will be a number of frustrated people complaining of FPs and an unfair process.

For many situations, it might make more sense to select a follow-up test with increased specificity, after an initial screening test with high sensitivity.

These decisions are policy matters that are best made with an understanding of both the capabilities and limitations of the science, and the agencies goals and needs. Do this well and we will start to deflate the energy of our critics and will begin to win the hearts and minds of open-minded scientists.

.02

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)