I've just completed some fascinating reading regarding the current state of CIT research. As most have probably gathered, I'm not entrenched on either side; I find the debate intellectually stimulating in many dimensions. My impression of the polygraph's history reveals a chasm between the scientists and the practitioners over the years; the latter being more from a law enforcement / legal genre of professionals. The major beef with the polygraph is the inadequacies of the CQT. While practitioners point to criterion validity, scholars note the lack of construct validity as well as the biased feedback from the field. This is the root of the "pseudoscience" tag.

In walks the CIT (or GKT out of respect to Lykken). Current research lends support to both criterion and construct validity (Sokolov's Orienting Reflex), as well as being unbiased towards the innocent in contrast the to CQT.

So, my question is thus:

Does the CIT potentially steal the fire from the antipolygraph camp? Or are they split; some accepting the CIT as a viable approach to PDD (or more appropriately Memory Detection) and others who are totally against any kind of external intrusion into the private domain of the mind.

Comments?

Dan, Lykken did address this by having each test contain only one crime relevant key item, and running several tests with other crime relevant items. For example if 10 tests were run, each with a unique crime relevant item along with 4 controls, the chances of a person lacking guilty knowledge responding to all (or most) crime relevant items would statistically be less than 1%. Interestingly, way back in 1960, Lykken did a countermeasure study on the GKT and found that diagnostic information regarding guilty knowledge was none-the-less discernible.

In contrast to the POT where phasic responses take a back seat to tonic during analysis, the CIT is numerically scored with no Inconclusive available. Also, the POT uses only one key, whereas the CIT suggests many (accuracy increases proportionally).

Putting this all together creates a scenario whereby the examiner is woven into an investigative team to implement strategies to identify salient crime related keys and prevent them from leaking. I am not sure if the average examiner has the prowess (or gumption) and critical thinking skills to make the CIT effective.

Drew Richardson wrote on Nov 28^th, 2015 at 6:09pm:


Doc,
While pondering the idea if an OR could somehow be utilized in screening, I came up empty every time. Then, while reading a book from a noted Lithuanian polygraph examiner Vitas Saldžiūnas, I was surprised to see where he affirms he and his colleague developed what they call an Event Knowledge Test (EKT). They even called it an "improvement" over the CIT. I gather that it has its roots in the KGB counterpart of the old DODpi. It strikes me as having elements of both a CIT and an RI.

I noted that he and his colleague authored the attached article in the 2012 APA journal. Are you familiar with this proffered EKT? Intuitively I want to ask to see the studies as well as an explanation of its construct validity. But, I am humbled by his credentials. 

His book is written with his own ESL skills and reads as such. But, if only I could do so well in Lithuanian. Here's the book:

http://www.amazon.com/gp/product/3659741922?psc=1&redirect=true&ref_=oh_aui_deta...

Ark,

How do I get to the attachment you refer to?  Amazon's book offering will not be practical in the short term...

So, obviously I have not yet read the work you refer to and am not familiar with the event knowledge test, so I can't comment upon the specifics of the notion or the details of the article.  That having been said...

An RI lie detection test is the lowest of the low in the world of lie detection and combining it with a CIT would likely not improve the RI but altogether ruin a CIT exam.

Going back to my previous statement about CIT exams not being applicable to screening, I need to offer a differentiation and an explanation.  

A CIT is most definitely not useful in a typical screening test, e.g., applicant testing/employee reinvestigation/sex offender testing types of screening exams because all of the preceding are merely fishing expeditions with no developed information available from which to probe examinee memories.

There does exist a different and a general screening exam in which one screens a pool of individuals/potential examinees (or even one individual) to see who in that pool has  association with some specific group of interest (Mafia, Al Qaeda, etc) or who possesses some specialized knowledge or skill set (e.g., bomb maker, medical training, etc).  

This latter type of testing is clearly a viable use of a CIT for a generalized (group association) type of screening, but, as always, depends upon the development of relevant privileged information. This might well involve the use of  confidential sources, informants, etc. with terrorist or organized crime related exams. Examples of useful information that might be obtained are training methods, locations, occupants (specific names and demographics of) in the hierarchy of some group, etc.

My colleague, Dr. Larry Farwell, and I have demonstrated the viability of this latter type of testing several times over the last couple of decades, e.g.,

http://journal.frontiersin.org/article/10.3389/fnins.2014.00410/abstract

Ark,

Thank you for getting me to the article you referenced.   

I'm sorry, but without spending additional and considerably more time than I am prepared to devote to tracking down and reading additional foundational articles (i.e., Saldžiūnas & Kovalaenka, 2008a,b,c, 2009a,b,c,d, 2010, 2011) referenced in that paper, I will not have any fully developed opinions about the EKT.

But again, anything that commingles lie detection and CIT exams would be of no interest to me and likely of no value to anyone.

Ark, where did you go to polygraph school?

I doubt that an engineer of virtually any stripe would have much faith in the "test."

What would it take to make you a believer? Evidently more than the APA's own research...