Monday, November 21, 2016
Analysis Error and Corrections
With analysis, if the conclusion is errant, we can (and must) go back and trace the root of the error. It is a scientific process that allows for input and conclusion, repetition, input and conclusion, and so on.
If the analysis is thoroughly done, the conclusion is very likely based upon at least several points within the language. This means that the analyst must have erred in, for example, four or five separate places and there must be a different explanation for the sensitivity shown.
Analysis by top professionals is almost ways checked by other professionals.
An amateur often makes a conclusion that is not supported by the statement and the correction ends up being a great lesson. The most common error made by amateurs is a rush to judgment. This means taking a single indicator of deception and concluding deception by it. Even upon review, the indicator of deception was likely an indication of sensitivity that had an alternate explanation. Once the truth is learned, the sensitivity question is usually answered.
When the analysis conclusion is based upon, let's say 5 points of sensitivity, each one of these points must be reviewed and re-considered, but it is a lot. Would it mean that on each of these 5 points, the analyst was incorrect?
The greater the number of sensitivity (or deception) indicators, the less likely error is going to exist.
When To Dig In One's Heels
When the analysis is based upon overwhelming points of sensitivity throughout a statement or interview, I encourage the analyst to make no changes: let it stand and see what time does to it.
Here is why:
I have seen some very talented analysts say, "I must have been wrong; the subject passed the polygraph."
This comes after the analyst's work has been reviewed by another analyst and two analysts found the exact same result.
I always ask that the polygraph pre-screen interview be reviewed for contamination.
A polygraph when administered without contamination and only with the subject's own words is near impossible to beat.
I have had two cases in which analysis and polygraph results have differed. One case was extreme. In both cases, I stood to be "wrong" and plainly so.
Case One: Child Molestation
The accused wrote out a statement prior to the interview. I concluded that mother's boyfriend molested her daughter while mother was not home, and, according to the statement, it took place in the little girl's bedroom shortly after dinner.
I sent the statement to an instructor, without my analysis. He sent back his finding: the subject was deceptive, likely molested the little girl in her bedroom sometime early evening but before she went to bed.
The little girl's interview was recorded. In her own language, she showed that he had 'tickled' her chest area under her shirt, while she was playing with dolls on her bed after dinner. Mom was at work. She said her mommy's "friend", using the name "Uncle" played this game but she did not like it.
The mother was interviewed who was truthful in not knowing what had happened but stated that she did not believe it did happen. She stated that her daughter was not known for making things up and "really liked" her new boyfriend.
The suspect was interviewed and gave a lengthy account of his day, in great detail, but gave only a short account from dinner time to bed time. His interview affirmed the analysis but he denied contact.
Collateral interviews showed no concerns about boyfriend, and no concerns about child making up things.
Boyfriend took and passed his polygraph.
The analysis was plain, including the skip of time and the timeline the little girl gave. The local police dismissed the analysis, and the boyfriend was permitted to return to the home where I learned, years later, he re-offended on the victim.
He passed the polygraph because he "tickled" not "molested" his victim. The polygraph used language that was not in the internal dictionary of the suspect. This is precisely how we teach suspects to lie and to falsely pass a polygraph.
This second case is much worse:
Case Two: Murder
In this case, we have a battle, from my perspective, of two forces. One is Goliath, strong, large and overwhelming, the other is tiny David, young and weak.
On one side is analysis of a 4 minute statement; nothing else. This is David.
a. statement analysis of a 911 call
But on the other side:
a. A complete and thorough police investigation with a cumulative 20 years experience in investigation and training.
b. All known facts of the case, privy to test results, interviews (plural) with the subject,
c. collateral interviews
d. forensic evidence
e. Suspect cooperation
f. Suspect passed a polygraph
On one side is a short analysis where nothing else is known, while on the other side, there is a mountain of reasons to overrule the analysis. I, personally, hold great weight to the polygraph.
Which would you believe?
What would you do if you were told all of these things, and then were told, "statement analysis is junk science. He passed his polygraph. We collected all the evidence..."
If this were not enough, consider that in court, one of the two original investigators would testify against the conclusion of the analysis and stand upon the original clearing of the suspect.
I, and those who worked with me, including in depth review later, concluded "deception indicated", which is not enough for a murder case. In the analysis we concluded both "deception indicated" and "guilty knowledge" of the murder, but still, I was able to show the motive of the murder, plainly.
In other words, there was no other conclusion for me to follow, and there was no way of dismissing each indication of deception, sensitivity and guilt.
Yet, one short analysis versus an in-depth police investigation, a ton of material unknown to me, and a suspect who has passed his polygraph.
He was found guilty of the murders after initially being cleared by police.
When error exists in analysis, it can, and must, be traced and corrected. By using a scientific process, when in error, we may find and correct the error.
But when a statement is overwhelming, and no error can be found, even in light of various news announcements, the analysis should b left to stand.
There are such cases.
No one is willing to say "Casey Anthony was found not guilty, therefore..." and, there is other cases that have been dropped or even cleared by police in which the analysis still stands:
Several years ago, analysis showed that Lena Lunsford had guilty knowledge of her daughter's death but up until recently, the analysis appeared 'incorrect' by the standard of arrest or adjudication.
When analysis is 'overwhelming' and has been checked, rechecked, and reviewed by other trained professionals, it should be left to stand.
"Out of the abundance of the heart, the mouth speaks" has proven to be truth.
Truth is timeless.
Truth is not impacted by time, culture, or opinion.
In either of the two above cases, particularly the second, it would have been the easier path to fold the analysis but the road of less resistance is not always the correct course to take.
As analysts learn to work in depth, they learn to trust the results of the science, and from past errors, learn how to go back and trace the origin of error, but when the conclusion of the matter is plain, they must stay with it.
For formal training for you or your department or business, click Here
Knowing how to detect deception is to enter an investigation fueled with truth.