posted 02-13-2013 08:38 AM            

The company you represent monopolizes 90 percent of the iceberg.(Lafayette advertizes this statistic)

Do you seriously think that diverting attention to the start-up company that feeds on the crumbs of this Industry helps your case?

posted 02-13-2013 08:42 AM         

quote:

---

Just to the south of us in Massachusetts, there's a colossal mess because a state employee in the crime lab facility falsified what may be thousands of test results. Cases are being thrown out, and hundreds of inmates have already been released.

---

posted 02-13-2013 09:09 AM         

quote:

---

Thinking that I, as an individual participant in this forum, can respond to threats of corporate liability is an exercise in deludedness. I cannot.

You'll have to call Lafayette. I'm probably the only person at Lafayette that reads or participates in this forum. I'll convey your message. That is all I can do. Unless you have another suggestion, let's move back to something useful.

---

Do I have another suggestion? Indeed.

LIC might want to take a pro-active approach. It might be easier than being in damage-control mode after an article appears in 30 syndicated newspapers.

posted 02-13-2013 11:59 AM            

I do not use Lafayette (the decision was made by my dept), but could someone explain to my why a test using a Lafayette system with a reported integral GSR/GSC deficiency would necessarily negate the results?

I'm just a practitioner, but I thought test results are based upon relative changes made by a comparison of intra-test physiology.

What effects one would certainly effect the others.... or am I missing something here?

The ethical points of concern, as alleged by Dan, are a separate discussion.

posted 02-13-2013 12:31 PM         

It's kinda complicated. Or not.

Go to Ray's newer, separate thread -- "EDA" -- and take a close look the the chart excerpts Ray posted as well as his comments.

Speaking as a practitioner, it appears to me that there's sufficient weirdness to warrant concern, especially if one is also consulting computerized scoring algorithms which derive half of their decision data from EDA alone.

Then, go back to the memo from LIC's Mark Lane that Jamie posted on page 2 of this thread and give that document a very careful read.

My guess is that Mr. Lane sent the memo to some bigs in federal government polygraph circles.

I can see why LIC would want to bury this thing.

If -- and I underscore IF -- the problem is real enough, the ripple effect could be enormous.

Dan

posted 02-13-2013 01:38 PM            

Its only weird if you expect perfection.

If you accept the reality that EDA data are not perfect - that compromises are made in terms of increasing the usability of the data - then maybe it is not weird. Maybe that weirdness is a normal and expected proportion of random variance (s#$%! happens).

The goal of the developers will be to do as little as possible to the data, and that means tolerating some issues that could be managed more assertively - like descending EDA. I'm sure the folks at LIMESTONE could completely remove all descending noise. We could do it. So can they. But there would undoubtedly be a price to pay in terms of increasing the frequency that you'd see score disagreements between the Auto and Raw EDA.

Of course, the graphics I posted are based on LIMESTONE's technology, so perhaps they should provide their own answer to why those scores disagree, why those data are descending, and why those data are so noisy.

r

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

posted 02-13-2013 02:29 PM            

Limestone Technologies is not the company that repeatedly tells examiners that we've fixed our EDA.

JB

posted 02-13-2013 02:44 PM            

You have been attempting to leverage a lot of issues lately. You attack Lafayette for its larger size and market share. You attack Lafayette because of your smaller size and market share. You attack Lafayette with accusations of defective hardware. You criticize EDA quality. You criticize for making improvement in response to customer feedback, and you raise questions about integrity and ethics around the use of the Lafayette polygraph.

First, there are both advantages and disadvantages to being large and small. But Lafayette did not become successful by attacking you or oppressing you or your company. Lafayette was successful before Limestone was in the polygraph marketplace. I understand the value, from your perspective, in attacking Lafayette to gain market share – but I am afraid it will do more damage than good in the long run by causing division and misunderstanding among people who need to work together to learn what are the best decisions to plan the most effective future – in terms of testing and technology, and in terms of policy development.

Regarding the EDA data quality. The images tell the story. Perceived deficiencies include data acquisition problems, descending EDA data and messy/ugly EDA data. My answer is that these are known issues inherent to EDA signal processing. These perceived "deficiencies" seem to be issues that are known to your own instrument. To me that is a kind of glass-house arrogance. EDA technology works most of the time under most circumstances with most people. At some point we should get to a more complete and accurate discussion about EDA data quality, but for now I'll leave it to you to explain those descending charts and noisy charts. I've offered my perspective and what little knowledge and wisdom I have on the matter.

As for defective hardware, that data acquisition stoppage has got to be concerning. Did you or your engineers ever correct that? Or did someone else? Or was it a software programming issue? Will you tell us what exactly was defective about the circuitry? Is this an issue of negligence and liability – faulty circuitry? How long before it was corrected? Was there a general recall?

There are ways to study the EDA and learn to make improvements. We can converse in more detail about filtering solutions and EDA signal processing when this discussion settles down to a more sensible level.

Perhaps you could set for us an example of professional accountability, and tell us what issues of defectiveness or deficiency were identified that had to be corrected with the new and improved Limestone model? And how long those issues were known about? How many defective boxes were sold and how many real people were affected by those defective old devices? I'm not so naïve that I'm going to hold my breath for an answer on all this.

I'll continue to study the EDA data, but it takes time and effort. Obviously I'm not going to have all the answers. Nobody has all the answers. But it seems to be part of my role right not to try move things along a little bit, and it should be part of your role too. Mostly, I think it is misleading to the profession for you to continue to disseminate the notion that Lafayette's products and technologies are defective or inferior. Your product seems to have a number of known deficiencies or issues of imperfection.

I believe that our biggest concern needs to be for the profession. Causing misunderstanding of these technology issues does not help the profession. Damaging the profession will damage both Limestone and Lafayette in the end.

The polygraph profession needs us to communicate more clearly and effectively on this – with less confusion and less misunderstanding - so that we can continue to make whatever incremental improvements are possible in terms of polygraph accuracy and effectiveness. Confusion and misinformation will only make if more difficult for our profession to think through these issues in any intelligent way. Aggressive marketing with misinformation will not lead to long term success for the profession, but will more likely contribute to division and stasis within the profession. We've been there. Its time to move forward.

Some people seem to like the Limestone product. That is fine. I liked using it. That is no secret. Right now, I like the Lafayette better. Do they pay me to say that? Sure. They also pay me to help make it better. I believe I have done that, and I will continue to do that. Part of making it better is making the product better it to make sure the profession is also becoming smarter and better informed - which requires less confusion and less misinformation. Whatever is best for the integrity of the profession will, in the long term, be good for all involved - Lafayette, Limestone and all examiners everywhere.

Good competition is OK, and can even be beneficial. Misinformation is damaging to the profession.

Bottom line is this: we should all just focus on selling our strengths. We should all dispense with inaccurate criticisms, attacks, insults, threats and character assassination. Good people don't like those things, and they like it when we focus on our strengths, accurate information, respect, and the kind of progress that benefits all who work in the profession. They'll like us all better, and everyone will be better to evaluate the real advantages of each different technology solution. The profession will be less divided and less vulnerable, and we will all learn more improve more and be more successful in the end.

.02

r

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

posted 02-13-2013 04:16 PM            

We'll be in touch.

JB

posted 02-13-2013 04:36 PM            

quote:

---

What effects one would certainly effect the others.... or am I missing something here?

---

That's correct. The question of whether imperfect filtering (and all the instruments filter the data) would result in FPs requires a lot of assumptions. Unless the filtering only impacts CQs or RQs in some special way, which is unlikely, any effects should be random.

If Lafayette made a mistake, it was giving examiners the ability to see what their filters are doing to the raw data. If they just filtered and that's all you got to see, nobody would ask any questions. (We should see it, and we should ask questions.)

Filtering is designed to make scoring easier. The folks from Utah used filters for that reason, and we know, as a general rule, examiners correctly classify cases when using filters (on data that would otherwise be tough to score). One problem is that you can filter garbage and make it look like something worth scoring. (That's why Stoelting and Lafayette let you look at both, so you can make an informed decision.)

When you read the scientific literature, you'll note study authors list the means of collecting EDA (and other data). What you ultimately see has to do with where you place the electrodes, the size of the electrodes, the type of electrodes, the sampling rate, etc., etc., etc. All those are in addition to the circuitry, filtering, software enhancements, etc.

What you see in the research literature on the EDA is that it is very robust. It seems whatever you do, you do okay with it (but nothing is perfect). We can easily see increases and decreases in skin conductance or resistance, and that's all we want to know.

The assumptions here, among others, are that Lafayette has an issue with its EDA and that issue - whatever it is - is unique to them. Additionally, people are assuming that the possible issue adds to the error rate.

When we come up with predictive models, we are trying to account for which variables (for example, pneumo data) explain the "variance" in another variable, in our case ground truth. Think of it this way: Let's say that with every inch increase in height, a person tends to weigh 10 more pounds. Thus, there is a linear relationship between height and weight. Let's say the correlation is .80. Thus, we'd say that height explains 64% (.8 X .8) of the variance of the variable "weight." (The same is true in reverse.) Notice 36% of the variance is unexplained (100 - 64 = 36). There is always unexplained variance, meaning that some other variable(s) are also at play.

Algorithms just do a lot of math to figure out the best weights to give each variable to optimize correct predictions / classifications. Our scoring systems are designed to do something similar. (The 7-position scoring system - or any scoring system - is really just another algorithm.)

The assumption here is that the Lafayette's supposedly faulty EDA somehow explains less of the variance than the other instruments. That's what I tested the other day. I could not find evidence that that was the case. (Lafayette's EDA explained just as much of the variance as Stoelting's.) In other words, all of the drama is just that: drama.

We know, for example, in a single issue test approximately 13% of the time (according to the meta-analysis, the NAS and others) classifications will be wrong. (If our models could account for 100% of the variance, we wouldn't have mistakes.) Given all the different studies using a variety of EDA recording and filtering methods, we should feel pretty confident that the different methods of collecting and presenting the data do little to explain any more (or less) variance. That is, any differences are probably already accounted for in our known error rates.

posted 02-13-2013 04:40 PM            

quote:

---

I'll take this opportunity to remind Raymond Nelson of the Non Disclosure Agreement that he signed 7-17-07.

---

posted 02-13-2013 06:51 PM            

quote:

---

Is that a joke? The topic always was when I heard law professors discuss them.

---

Maybe someday soon you will have the opportunity to apply your knowledge of law.

FYI... I did try Bill.

From: James S. Brown [mailto:brownjs@limestonetech.com]
Sent: Tuesday, May 27, 2008 9:52 AM
To: 'Raymond Nelson'
Subject: Lafayette EDA

Hi Ray,

I sent this to Barry today and I thought you might want a copy as well.

Hi Barry,

It was good seeing you in Jacksonville. Hope you enjoyed the conference.

I was reading the posts on the Polygraphplace and I noticed that you had made a pretty significant observation regarding Lafayette’s EDA. This has been noticed by many examiners and it unfortunately gets swept under the rug. I’m sure you can appreciate the far reaching impact of this flaw.

Regards,

--
James S. Brown
President
Limestone Technologies Inc.

posted 02-16-2013 10:57 AM            

She seems to be contacting examiners and supervision programs around the country surveying the types of instruments that people are using and seeking commentary on a document describing EDA Deficiencies.

The document is a .pdf of a powerpoint that includes information from this forum.

Whomever made the powerpoint appears to have been intending to stir controversy around the Lafayette EDA system.

Whomever provided the powerpoint to Ms. Taylor may have had that same intention, or may have been attempting to stir controversy around the polygraph in general.

Ms. Taylor's intention may have more to do with drawing up controversy around polygraph in general, using the PCSOT arena, and ethical concerns about human liberties decisions that are influenced by polygraph results for which there are uncertainties about about decision accuracy.

The EDA issue is a fictitious issue for several reasons. Not saying the EDA is perfect. Nothing is perfect. What I'm saying is that concerns about decision errors resulting from the EDA are ficticious and misguided. They appear to be the result of someone intending to stir controversy.

First, any differences in numerical scores is within the range of normally expected error variance. The EDA alone is not a complete diagnostic indicator of truth or deception. What that means is the EDA is already known to be a good but imperfect signal that works better when combined with other good but imperfect signals - the pneumo and cardio. The result is a better diagnostic indicator that also remains imperfect. In the end everything is a probability. The probability of a correct decision is high - significantly greater than chance. In fact data at this time suggest that accuracy for single issue exams is significantly greater than .83 with a mean of .89.

Second, a great deal of the criterion accuracy studies published over the years - including those used in the 2011 meta-analysis - were conducted using Auto EDA.

Finally, a lot of information in the .ppt seems to old and inaccurate from 2007 and earlier.

Drama and reactivity seem to occur only when we engage in naive expectations of perfection - and only when someone is intent on stirring drama and reactivity.

Argh.

r

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

posted 02-16-2013 12:16 PM            

But I gotta food chain too, and I don't yet have all the answers to the questions I'd anticipate.

And maybe I'm not the right person.

r

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

posted 02-16-2013 01:37 PM            

quote:

---

A little birdie just tweeted that you're talkin' shit about me on the forum today. I have never heard from this person before, but if what he said is true, I'm hurt. I did NOT leak that package to Taylor or anyone else. However, I am speaking candidly to Taylor about various polygraph issues. Happy hunting.

---

Could be someone is still trying to stir this.

r

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

posted 02-16-2013 02:19 PM            

Letter regarding McClatchy series on polygraphs
Stay Connected

McClatchy Newspapers
Editor McClatchy Newspapers, Inc.

The McClatchy’s Newspapers, December 2012, series of articles on the federal government’s polygraph screening programs presented a misleading depiction of how the federal government uses polygraph for personnel security. The thesis of the series is that the government is expanding its use of polygraph in a rapid uncontrolled manner, the tests are abusive, and the results are often inaccurate. The unstated, but clear, premise is that government officials responsible for the top secret security clearance process, as well as the congressional oversight committees are determined to employ a personnel screening method that is not only ineffective but traumatizes a large segment of the personnel in the Intelligence Community and some law enforcement agencies. The successful performance of our nation’s intelligence and security services requires a workforce of superior education, motivation, and integrity. The suggestion that government officials would deliberately conduct business in a manner that would hinder the development and preservation of such a workforce is absurd.

To support the claim that “many” are traumatized by the polygraph exams the articles cite a handful of individuals who claim to have been questioned in an inappropriate manner. Minimal evidence is provided to support those claims of abuse. In fact, Mark Zaid, an attorney representing a complainant acknowledges that he found no abuse when he was able to review an audio recording of his client’s exam. The fact is that for many years a number of federal agencies have surveyed the attitudes of the individuals they test. Those surveys consistently demonstrate that the vast majority of those who had been tested believe the exams to be fair, professionally administered, and support the use of polygraph testing for security screening.

The McClatchy articles also claim that polygraph examiners torment the examinees by delving in personal issues that are irrelevant to the security clearance process. It is suggested that polygraph examiners have great leeway in selecting topics for the exam. In fact, the issues addressed in federal screening exams are carefully selected by high level officials who manage the government’s security programs, not by the polygraph examiners. Political and/or religious affiliations or beliefs are completely outside the bounds of examinations. Sexual activity is never addressed unless it is criminal in nature. Applicants and employees who submit to a federal screening should expect the exam to address at least some of the following issues; espionage, sabotage, hidden associations with governments and citizens of foreign countries, compromise of classified information, terrorist activity, major criminal activity, recent illegal drug use, and falsification of application materials. Thoughtful individuals would agree that exploring such issues with those seeking or holding sensitive intelligence and law-enforcement positions is not only reasonable, but that it would be negligent not to address those issues.

The articles allege that a very significant and poorly managed expansion of federal polygraph testing occurred in the decade since the September 11th terrorists’ attacks. Yet no data is offered to support that premise. McClatchy even acknowledges that many federal agencies, including two of the largest (CIA and FBI), decline to provide any information regarding how many exams they administer. The American Polygraph Association (APA) informed McClatchy that it had not experienced a significant increase in membership by federal examiners since 2001. How McClatchy concluded that such a dramatic expansion occurred is never explained.

Finally, the articles suggest that the federal polygraph testing is devoid of accuracy. It intimates that each year thousands of truthful individuals are labeled as deceptive by their polygraph test. This claim is made despite the fact McClatchy has no idea whether or not a thousand be individuals are even are deemed deceptive each year or how many of those judged deceptive acknowledge their deceptive statements when informed of the test results. The APA provided McClatchy one of several methods federal officials might use to accommodate for errors in the polygraph results without harming the career prospects of those examinees. But those methods were not mentioned in the articles.

The APA is disappointed and disturbed that a national news organization published such an incomplete, poorly researched series of article on an issue that directly impacts the security of the United States.

Robert Peters Vice President Government American Polygraph Association www.polygraph.org