posted 05-24-2004 08:31 AM            

as a practitioner of polygraph, not an academic, I will tell you how I approach that.

First, the prosecutor needs to give you the lattitude to explain the over polygraph process to the panel, the same way you would an examinee. Then the process of dividing the test charts into zones for scoring, then numerical evaluation (+1/0/-1) or 7 scale if used.

Then explain the NDI range (ie, +2 in each for Bi Zone, +6 overall = NDI; -3 in either spot = DI and anything else is a buffer of "safety zone" to preclude a false outcome decision, like a gray area).

Inc's can be caused by a multitude of factors, not setting CQ properly, outside concerns (related to, or not related to) the issue, fatigue, excitability, dysfunctional nervous tension, etc.

A "gray area" decision requires a retest.

This is a simple way to explain it and most people can understand.

Good Luck,

Jim

posted 05-24-2004 12:16 PM            

Did you run three charts or four? The cut-offs for three charts is +/-4, and four charts is +/-6; although some people don't know the research indicates that the two scores are about the same in terms of reliability.

In theory, cut-off scores are based on studies showing the average distribution of scores for known liars and truth tellers. We then look at the probability of a person achieving a particular score and decide if it's within the limits of error we can live with. (Remember those bell curves from statistics?) If the truth be known though, most are just made up based on somebody's best guess of where the cut-offs should be. That's not to say our cut-off scores are bad as the research does support (to some extent) what we've been doing. It all depends on what you want to accomplish.

In other words (and I'm just making these numbers up to make a point), let's say a person scores a +8 on any test (it doesn't matter which since this is made up). There is a possibility the person is DI because we know from researching the particular technique we are using that 1% of the confirmed DI suspects score that score or greater, BUT 95% (don't forget I'm still making this up) of NDI subjects score this score or greater, so it it likely this person is in fact NDI. The INCs are those scores that are too close to call, i.e., too many of the guilty and innocent alike achieve those (inconclusive) scores. (Remember the two bell curves - the DI bell curve and the NDI bell curve - overlap.)

I hope you can figure out from that gibberish what I was trying to say. It's a lot easier in person, especially if I were to draw the bell curves and explain as I go. (The explaination is the same regrdless of the score and final call. I've had to explain an NDI - and the chances of a false negative - in a case in which the State had taken a child from his parent for abuse.)

Good luck!