Talk:Voice stress analysis/Archive 1

Factual accuracy
This seems like complete bunk to me (just like the polygraph). Does anyone have any references to any academic studies? —Preceding unsigned comment added by 220.233.50.147 (talk • contribs)
 * I agree that it's a garbage technique, but that doesn't mean that an article shouldn't be here — notability, not factual accuracy, is the determinant (hence articles like Phrenology). However, if you do have some referenced academic criticism to add, by all means, please do — it would improve the article.  Tijuana Brass E@ 09:51, 28 May 2006 (UTC)

YOU agree that it's a garbage technique??? What are you talking about??? WE ARE USING VSA in our departments - AND it is a great technique if you know how to use it!!

Added the following:

 Note - then you must actually refer to such studies by way of a link, or traceable data
 * A sentence to introduction indicating that multiple studies have concluded that VSA's accuracy is no better than chance


 * A table of contents

''The APA's own research was debunked by NAS in 2003 - APA is not a worldwide authority on  all matters DOD - in fact, the APA's 'so called research' on VSA is hearsay. Please dont use independent and competing researcg studies. its like saying pepsi says coca-cola is crap.)''
 * A section on accuracy with a link to the American Polygraph Association's VSA research section

please. Unless you publish your full name and academic qualifications, you are NOT qualified to label anything as psuedo or fictional. To do so borders on slander.
 * Category references to pseudo-science and measuring instruments - keep personal opinions out

VSA appears to be complete bunk, and I think that the article should reflect that more strongly.

''What you 'THINK' is irrelevant. What you know by peer reviewed research that you can refer to specifically, is relevant. Otherwise your personal opinions have no place here.'' StraussianNeocon 10:16, 28 May 2006 (UTC)

VSA accuracy can definitely be affected by the questions asked.

''I recently failed a VSA. I did not attempt to deceive, but was worried that I had left something out. The questions were of the nature "Other than what we talked about, have you ever (engaged in illegal behavior - insert here - i.e. used illegal drugs etc.)?''

''The problem? I had disclosed as much as I could of my sordid past but had I disclosed everything? Of course not. Growing up in the 60s, 70s and 80s it would take more than a two hour interview to disclose everything I had done that might be considered "shady."''

Naturally, my test results showed "deception"  JohnWillTenney (talk) 13:49, 18 December 2008 (UTC)

''The whole point of a VSA is to get you to tell the whole truth before you start the test. You normally fill out a very long questionnaire which gives you ample room to tell them everything. Then they will ask you, "Other than what we discussed, have you ever (insert interrogative here)?". They want you to pass the interrogative portion because they want you to have already listed all of your offenses. You also discuss the entire form before you actually answer the question. The technique is very different than that of a polygraph (I have been given both) and the VSA is very conversational in tone as opposed to the polygraph. They want you to review everything first before asking you the question of "other than what we discussed." If there was more to discuss, you should have discussed it. Schlice (talk)'' — Preceding undated comment added 00:20, 6 January 2015 (UTC)

This article is highly biased and lacking many credible sources. For example, the Chapman report is mentioned but the link leads to a source which isn't even in English. The study in question is a key piece of scientific evidence in support of VSA:  TN7742 (talk) 19:26, 30 January 2017 (UTC)

This actually is the original source. And, just wondering, do all US government officials publish their key pieces of scientific evidence in Ukrainian Journals that are nowhere to be found? — Preceding unsigned comment added by Stringfellows (talk • contribs) 17:33, 12 June 2017 (UTC)

Skype
Skype users to get lie detectors KishKish VSA (presumably the product in question)

Not sure if this should be added, or what exactly should be said about it. JulesH 17:32, 15 December 2006 (UTC)

Kishkish is a party trick, based on LVA. LVA has no place in the serious world of DOD (Deception Of Detection) John Elder (see bottom page) is also referring to LVA. Ctka 15:58, 17 October 2007 (UTC)Ctka - 16 October 2007

Amplitude? or Frequency?
I admit that I know nothing about this subject. But, a tremor, as any waveform, has an amplitude and a frequency. Since the article later says that the stress level is measured in Hertz, it sounds like we're talking about the frequency of the tremor, not the amplitude. Am I wrong? Can somebody clarify the text? Appljax123 15:09, 9 January 2007 (UTC)


 * There is plenty of misinformation on the net about this - not in the sense of falsehoods, but total misunderstanding by people who have written articles. This article needs to be strongly edited - the answer appears to be that the integrated power of the tremor signal is compared with the integrated power spectrum of the entire signal. If the two do not track, it indicates that the vocal cords are stressed, which, proponents claim, indicates a lie. I should dig up enough info to write this up clearly in the main article, but don't have time right now. John Elder 16:46, 18 January 2007 (UTC)

WELL, YOU ARE INCORRECT JOHN. VSA CANNOT BE REDUCED TO AN ELECTRONIC EQUATION. VSA RECORDS THE DIFFERENCE BETWEN UTTERED RESPONSES, COMPARING THE NORMAL VOICEPRINT IN Hz OF THE VARIOUS RESPONSES. UNDER PSYCHOLOGICAL STRESS, THE VOICE RESONATES AT LOWER Hz THAN WHEN UNSTRESSED. —Preceding unsigned comment added by Ctka (talk • contribs) 15:52, 17 October 2007 (UTC)

''McQuiston-Ford based VSA (not LVA) measures the Lippold Tremor in Frequency. If you require a detailed and tech explanation - drop me a line via Wiki channels.''

Is there anything besides lying that might cause the lower frequency? David —Preceding unsigned comment added by 81.132.1.151 (talk) 23:59, 29 January 2008 (UTC)


 * Yes, low frequency resonances can be achieved in any number of ways, including slight shifts in vowel pronunciation, any kind of change in the shape of the mouth (or unusual muscle contraction), or even just a higher or lower amount of saliva. I was hoping this article might have the actual circuit algorithm. The problem with analog lie detection or any of these old-school techniques to produce discrete results is that the signal they claim they are looking for is going to be so small that any tiny variability can throw the results off completely. Then to make the signal readable, they have to digest the data through analog filters which by necessity reduces the signal to noise ratio. That's why analog tests that have been scrutinized heavily have very specific procedures and use a lot of contact gels and require patients to sit perfectly still (like polygraph and EEG). But as you may find out, by clenching your buttcheeks you basically make a polygraph unreadable (but you have to clench for the entire test, otherwise they'll know you're clenching). Digital signal processing, by contrast, can be very effective in finding precise signals in a sea of noise, but you have to know what you're looking for (see the Emotiv gaming headset for an example of this, where a very precise EEG is set up without regard for greasy hair, geometry, and long baseline signals). 99.142.95.19 (talk) 17:42, 1 September 2009 (UTC)

Use to prevent fraud
This is being used to prevent welfare benefits fraud in the UK. Secretlondon 03:53, 12 September 2007 (UTC)

The relevance of Lippold
Has anyone read the article by Eriksson and Lacerda on VSA? They claim to pick appart any suggestion that LVA is based on scientific studies by stating that none of Lippold findings have anything to do with microtremor that can be detected through the human voice. According to the article, all the microtremor studies are of completely different muscle groups. It's a pretty serious challenge to the claims of VSA-based lie detectors. Has anyone actually read any of these studies? Do they actually make any claims about the presence of microtremors in the human voice? Peter Isotalo 19:37, 3 February 2009 (UTC)
 * LVA is not microtremor based - LVA detects changes in pitch and tone. Lowering pitch and tone produces 'Truthful' results with LVA.
 * The tremor of Lippold is well documented. Every mammal has a natural tremor in all muscle groups.
 * Drill down and ascertain who paid for the Eriksson Lacerda study. ?? — Preceding unsigned comment added by Ctka (talk • contribs) 11:15, 8 February 2012 (UTC)


 * I've edited the article based on the summary of microtremor research described in Eriksson and Lacerda. Their article is critical to the point of being polemic, but they do seem to have good points. I've tried to use their conclusion in a neutral manner. If anyone can support the claim that there is independent confirmation of the presence of microtremor in the 8-12 Hz frequency in human speech, please supply reliable support for this.
 * Peter Isotalo 10:23, 7 February 2009 (UTC)
 * Your edits were not applicable. Please google LIppold Tremor and take your time. The Lippold tremor not only exists, it is used in speech recognition technology, whereby every persons spectrograph (unique combination of microtremors) is uniquely different. Voice recognition is a forensic science, accepted in legal systems worldwide. — Preceding unsigned comment added by Ctka (talk • contribs) 11:15, 8 February 2012 (UTC)

Stress out of fear one might be detected as lying?
The article suggests that the fear of being detected as lying influences the reading, and thus might give as a result that one is indeed lying, without that being the case. But if one is told and shown that the technique does not detect the truth as a lie, and only shows lies as such, that fear of being evaluated erroneously will greatly diminish. After that, and knowing beforehand that what one is going to say is true, the fear of being caught lying will be absent. — Preceding unsigned comment added by L. Langeweg (talk • contribs) 09:23, 3 April 2012 (UTC)

Section "Criticism" Not NPOV
That Section mentions a handful of positive or neutral reports, not academic studies. It lacks citations for most of it's claims. It doesn't mention a single study that found any problems with the technique.

Saying it is as accurate as Polygraphy isn't saying much - remember Aldrich Ames of the CIA?

Supposedly his KGB controller told him to get plenty of sleep before taking the test, to cultivate a rappor with the operator, and to never, never admit - even to himself - the slightest discreditable or off-color fact, no matter how minor. He passed the test. — Preceding unsigned comment added by 72.70.200.65 (talk) 21:27, 26 June 2012 (UTC)

Agreed. There are several messages that seem to try to 'explain away' each criticism, instead of just acknowledging them. This would be fine if it was cited rebuttals to those criticisms, but this does not seem to be the case. This section needs a re-write bad, and reflects and obvious bias. 121.98.140.124 (talk) 04:00, 27 June 2012 (UTC)

"In Anders Eriksson and Francisco Lacerda's article "Charlatanry in forensic speech science:A problem to be taken seriously" voice stress analysis is described as charlatanry, and that analysis of studies shows that these methods preform at chance levels. They argue that "there are serious ethical and security reasons to demand that responsible authorities and institutions should not get involved in such practices."  This article meet it's own criticisms from manufactures of voice stress analysis machines, leading the International Journal of Speech, Language and the Law International Association of Forensic Linguists to withdraw it." Maybe something along these lines needs to enter the article? http://www.cs.columbia.edu/~julia/papers/eriksson&lacerda07.pdf that is the link. I don't have the time to edit this properly currently. If someone else could clean this up it would be appreciated. 121.98.140.124 (talk) 04:32, 27 June 2012 (UTC)

"they are illegal in Califorina"
Wrong. That's not even what the linked penal code citations say! Come on now 64.134.233.52 (talk) 21:08, 1 June 2014 (UTC)

Merge with Voice Risk Analysis?
I'm not sure if Voice Risk Analysis refers to a specific voice stress analysis device/company or if it's just the british word for the same thing. Both articles are fairly low quality, but if Voice Risk Analysis is a specific product, that should be made clear in that article (I don't think it is), otherwise I suggest we merge and redirect that article here. 0x0077BE [talk/contrib] 21:26, 8 October 2014 (UTC)

Redirect
Why does Vipre redirect here? The term doesn't appear in the article, and Vipre is a fairly well-known anti-virus software package - see Comparison of antivirus software. My addition of a hatnote to reflect this was recently reverted. Tevildo (talk) 18:09, 16 October 2015 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 2 one external links on Voice stress analysis. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive https://web.archive.org/20090123064356/http://diogenescompany.com:80/vsaprogram.html to http://www.diogenescompany.com/vsaprogram.html
 * Added archive https://web.archive.org/20071105222550/http://www.nitv1.com:80/Agenciesusing.htm to http://www.nitv1.com/Agenciesusing.htm

When you have finished reviewing my changes, please set the checked parameter below to true to let others know.

Cheers.—cyberbot II  Talk to my owner :Online 15:23, 1 February 2016 (UTC)

Article in need of MAJOR work!
I have heavily tagged this article as being in need of improvement. Hopefully, this will attract some thorough editorial attention. At the very least, it needs non-commercial references, with citations for statements. Halwahalf 07:59, 21 October 2007 (UTC)

I agree. I changed some and added some more. I have done little to the actual article other than remove the uncited conspiracy passages and the meaningless citation that went along with it. Many references reference nonexistent web sites as well. This article is in a very pitiful state; indeed it is one of the worst I have ever come across on wikipedia. Before anyone argues with me on my edits, please read WP:V, WP:FRINGE, WP:CITE, and WP:OR.

208.81.93.114 (talk) 06:21, 24 March 2008 (UTC)


 * The references are a mess. Fuzzform (talk) 03:13, 7 November 2008 (UTC)


 * Agree that this needs much more work. The technique is of questionable validity, has been characterised as pseudoscience, and a good bit of the article read as a sales pitch. Guy (Help!) 23:47, 5 December 2008 (UTC)

This sentence: "In a study published December 7, 2013, the International Journal of Electrical, Electronics and Computer Engineering (IJEECE) found that Voice Stress Analysis (VSA) technology can identify emotional stress better than polygraph" does not make sense. A journal cannot "find", but can *publish* a finding, presumably by a researcher or investigator. — Preceding unsigned comment added by 108.65.32.10 (talk) 03:29, 9 November 2015 (UTC)

I'm sorry I can't offer more help to improve the article than this, but I just came here to get some quick answers and was disappointed with the state of the article. My reasoning for flagging this is to make it more obvious to laypeople that as it stands the article is not reliable. -AC 2A00:2381:1614:100:0:0:0:2 (talk) 10:40, 24 May 2016 (UTC)
 * This article still needs work. I am about to flag it for neutrality check. Some of my specific problems (to begin with):
 * Biased in favour of the support argument.
 * The references! For example, the very first paper listed (Patil, V. P., Nayak, K. K., & Saxena, M) is from a journal I have never heard of. Why not lead with something from IEEE? Additionally, while that paper does indeed state "we conclude that VSA technology can identify stress better than polygraph", there is no evidence in that paper for that statement! The science is poor: The experiments performed did not include the use of a polygraph, so how could the VSA results be compared!? The methodology applied did not seem robust to me. There were no statistics for the results. The mention "using multiple users", but don't state how many. I am not an expert in this field (just have a MSc in ElecEng, with a number of years experience in telecommunications related digital signal processing).

This article is heavily biased towards VSA. All references to quality studies published in peer reviewed journals are deleted, and obscure references for Indian and Ukranian journals are listed. How to make this more balanced without one used simply deleting everything?

One other flaw I would like to point out: The last section references some Guardian articles about VSA, but it is talking about the Nemesyco system, which is different. In fact, their website specifically states that they are NOT affiliated with VSA:  Therefore, it is faulty and inaccurate to use Nemesyco as an example in this article. TN7742 (talk) 19:25, 30 January 2017 (UTC)

Pseudoscience
Voice Stress Analysis can be regarded as pseudoscience. This entry illustrates this. All meaningful, peer reviewed, publications are ignored (and deleted if I enter them). The evidence listed comes from 1) an article in an Indian open access Journal of which is it completely unclear how they reach their conclusions 2) An article that appeared in an Ukranian Journal. It took me some time to track it down, but I found it. So a U.S. government official publishes U.S. field data in a Ukranian Journal a year after he dies? How does that work exactly? In any case, if the data is real, it suffers from sampling bias, overestimating the accuracy. 3) a conference abstract that actually shows that whereas VSA technology correctly classified 73% of deceptive examinees, for truth tellers - with 52% accurate classification - it performed as well as a coin flip 4) a government report (not a peer reviewed journal article) that tested VSA on material of 2 murder suspects and actually concludes that 'Although these systems state they detect deception, this was not proven'. References to peer reviewed articles, including quotations from the National Academy of Sciences are deleted because they are 'deleterious statements' and 'negative and onesided comments'  — Preceding unsigned comment added by 84.30.65.146 (talk) 15:01, 25 September 2016 (UTC)

For some reason, United States vs Scheffer (1998) is discussed, as it allegedly ruled that polygraphs are "no more accurate than coin flip.". First, this quote is nowhere to be found in the ruling (full text here: https://supreme.justia.com/cases/federal/us/523/303/case.html). Second, a supreme court ruling on polygraphs bears no relevance to VSA, so this should be deleted. — Preceding unsigned comment added by Stringfellows (talk • contribs) 09:08, 30 September 2016 (UTC)

There is no 'Scientific Journal Criminalistics and Court Expertise'. It does not exist, so this should not be listed as a reliable source. Stringfellows (talk) 12:58, 9 October 2016 (UTC)

In line with WP policy, I have removed references to primary sources, leaving the conclusion by the National Research Council. I also removed unreferenced or improperly references sentences, for example that VSA works in high stake field setting (VSA is not mentioned in the referenced article). Stringfellows (talk) 16:13, 18 June 2017 (UTC)

Factual accuracy
I would like to propose to remove the 'factual accuracy is disputed' message. In the current version the claims that this message referred to have been removed. Stringfellows (talk) 11:53, 26 June 2017 (UTC)

needs better sourcing
The following is WP:OR and needs reliable secondary sources


 * Principle and origins

Lippold discerned that an 8–12 Hz range physiological tremor was associated with the contraction of human muscles indicated a connection between psychological stress and its effect on the human body. On September 5, 1972, three retired U.S. Army officers—Allan D. Bell Jr., Wilson H. Ford, and Charles R. McQuiston—filed a U.S. patent under their corporation name, Dektor Counterintelligence and Security, Inc. for their “physiological response analysis method and apparatus.”. Bell's expertise was in counterintelligence, Ford's was in electronics, and McQuiston's was in polygraphy. Ford had invented an electronic device that utilized the theory of Lippold, Halliday, and Redfearn in which he tape-recorded the human voice, slowed it down to one-third or one-fourth its normal rate, and fed it through several low pass filters which then fed the signal into an EKG strip chart recorder. The strip chart recorder then made chart tracings on heat-sensitive paper. Their Psychological Stress Evaluator (PSE) was the first commercially available VSA. According to Allan Bell Enterprises, "All lie-detection examinations or evaluations are predicated upon the fact that telling a significant lie will produce some degree of psychological stress. Psychological stress, in turn, causes a number of physiological changes.".

-- 02:42, 31 August 2017 (UTC)

Pseudoscience
Well researched citations with titles like "Charlatanry in forensic speech science" and "Voice stress analyses: Science and pseudoscience" substantiate the label of pseudoscience, as does the Department of Justice report that explicitly states "It is clearly unlikely that a single measure such as that based on the computerized voice stress analyzer could be universally successful in assessing stress". ☆ Bri (talk) 20:23, 30 April 2018 (UTC)