The Use of Voice Stress Analysis to Expose Informers and Agents

by William Michael Kemp



For those who do not know or know of me, I will ask that you recall the summer of 1995, when the ATF was undergoing congressional scrutiny for their activities at Waco, and a major national news story concerning the ATF's extracurricular activities broke-- the Good O'Boy Roundup, where an annual weekend-long picnic of law enforcement agents was revealed to be extremely racist. The banner which often hung over the entrance to the gathering read "NIGGER CHECK POINT-- ANY NIGGERS IN THAT CAR?" The banner further offered 17 cents a pound for "field dressed and boned nigger meat." That news story was brought to light by the Gadsden Minute Men, a citizens' militia group which I founded.

Need I continue?

When the militia/constitutionalist movement awakened and gained momentum in the wake of Ruby Ridge and Waco, great strides were made in establishing networks of like-minded people in various groups around the country. Before very long, however, it became apparent that the various law enforcement entities, from local to federal, were not going to leave us unmolested, particularly not when our intent was to make government accountable for their unlawful and unconstitutional actions.

From the very first, their methods should have been apparent. After all, Randy Weaver was a crystalline example. ATF wanted Randy Weaver to be a snitch for them. He wouldn't, and we know the rest of the story. What Randy Weaver needed-- in fact, what we all need-- is a way to find the people in our associations who don't belong-- who are present under false pretenses-- who are acting for the interests of unlawful government, rather than in the interests of the sovereign people. Most of these people can be described simply as snitches, or informers, or informants.


Who has infiltrated us?

These folks, by and large, are not sworn, authorized agents of government. Occasionally, actual members of law enforcement are used, but this is relatively rare. A few are wannabe cops, but the majority are snitches, and they are, typically, criminals. They are acting at the orders of law enforcement. They are doing what the ATF wanted Randy Weaver to do-- living a lie. Randy Weaver wouldn't become a snitch, so they declared him to be a criminal, and attempted to run him through their snitch manufacturing machine, sometimes referred to as the Criminal Justice System-- which is aptly named. Randy was not a criminal, and decided not to participate. Most people faced with this choice would not have the moral integrity to resist.

Most informers are not so directly targeted-- most are convicted of one of the myriad of things which now constitute "criminal behavior," and are offered the alternative of being a snitch or going to jail. Rather than suffer whatever consequences they face, they are encouraged by various law enforcement agencies to spy upon their fellow men, and function to report on private associations, and encite, entrap, provoke, and disrupt at any and every opportunity. If they fail to generate criminal activity (and in this day and age, when simple possession of inanimate objects has been declared criminal) they will plant evidence, or otherwise cause legal problems for their targets-- US.

Failing all the above, they resort to creating dissension and division, and do everything they can to spread mistrust and suspicion. We should all be witness to the effectiveness of the effort of the snitches and their masters, the lawless "law enforcement" community.


Why have we been vulnerable to informers?

We are severely hampered in our dealings with this gutter trash. Since we are, by and large, honest and straightforward people, we have accepted folks on face value. This honesty has been taken advantage of; our honesty has been taken as naivete. We have been easy pickings for the immoral and unscrupulous methods employed against us. We do not have the resources to do full background checks, which would largely be useless anyway, resembling hens petitioning the wolf for relief from the fox. Government holds the records, and their snitches will be protected. Further, one must deal with a government employee to obtain any records, and trusting any results from such a source is a rather chancy affair.

Polygraphs (classic "lie detectors") are expensive to own, difficult to operate, and difficult to interpret.

So we find ourselves in our present circumstance. We can look around and see many of our brethren imprisoned or financially ruined, or both. But we don't even have to look that far. The level of mistrust is incredibly high; it would be called paranoia, save for the fact that it is absolutely justified. The prosecutions and financial ruin of so many of us attest to that fact.

Various leaders from various places are routinely suspected and accused of "being a fed." Even internally, trust by members of their leadership is low, and no leader can feel secure from infiltrators in any of the various groups. And, rest assured, there is at least one infiltrator in EVERY group of any size, in every group with public exposure.

We have attempted to operate publicly, for it is in the public interest that we act. We are not criminals, and wish only to be RID of criminality, most particularly OFFICIAL criminality. But, as the result of the lies and false colors of informers, we are fighting a losing battle and are on the verge of complete failure.


A Solution-- voice stress analysis

With this in mind, I offer a solution. The technique of "voice stress analysis," or "psychological stress analysis," goes back thirty years. It is founded on the premise that lying produces stress (from fear of discovery if not for moral reasons) and this stress is manifested in the voice. Properly applied and interpreted, this analysis, while not foolproof, has a fine track record. It is in regular use by various and sundry government entities around the nation, from the local cop-shop on up. I suggest that you visit the web site to see this illustrated. This company makes a commercial product based on a 166 megahertz Pentium laptop to do real-time analysis. Of course, this product is only available to law enforcement and other government entities.

Even in sparsely populated Alabama, the method is in common usage. The local newspaper, The Gadsden Times (which is a wholly- owned subsidiary of The New York Times) has run feature articles discussing the use of voice stress analysis by various law enforcement agencies, with much crowing by those agencies attesting to its effectiveness.

A former CIA intelligence officer, George O'Toole, has conducted an extensive investigation into the assassination of President John Kennedy using voice stress analysis (which he refers to as "psychological stress analysis"). He has concluded that the accused assassin, Lee Harvey Oswald, did not commit and was not involved in the killing, other than being a "patsy."

The House Select Committee on Assassinations Chief Counsel, Richard Sprague, in 1976, stated his intent to use voice stress analysis on all testimony.


Voice stress analysis-- equipment

This method has been subject to criticism, largely as a result of poor equipment, improper application, and poor analysis of the results.

The day of poor equipment is behind us. High quality tape recorders are everywhere, from hideout miniature recorders with remote microphones to be clipped to a collar to desk-top boom boxes. A video camera, switched on but with the lens cap in place makes a fine audio tape recorder, and while those present may initially react to the presence of a camera, will soon discount its presence since "the lens cap is on."

I have personally used this method twice, both times successfully: once on a District Attorney who said some things to a couple of folks privately (that a particular prosecution was "insane") that he would have never said, and would have denied, in public. The first he heard of the recording was when it was mentioned in court. He pitched a hissy fit, at which point I told him "Welcome to the NFL, kid. Sue me. Deny it. Drag it into the newspapers. Prosecute me. The law specifically allows me to tape any conversation that I'm a party to, and you people lie for a living."

The other occasion was in public with several freedom fighters in conversation with an ATF supervisor. He spoke some words in an unmistakably hostile tone that gave every appearance of a threat. I was immediately asked "did you get that?" The ATF boy did a double take at the video camera, and was devastated when he realized that he had been caught so easily. It was hilarious seeing this snitch-master trapped by such a simple device, and going away blustering with his feelings hurt. I ALMOST felt sorry for him (snicker).

Almost every person viewing this document has equipment perfectly capable of performing the necessary evaluations. A moderately fast computer, such as those equipped with a Pentium processor and a 16 bit sound card, are excellent tools. Software/freeware is readily available, and will perform analysis in near real time. Thus, the first hurdle is past. We have the tools.


Voice Stress Analysis-- its proper application

The next problem is improper application. We can't expect to analyze a bull session with a bunch of guys drinking beer and swapping tall tales. Further, we can't properly analyze a "casual" lie. By this I mean that it is not always possible to make up lies just to test the analysis. I cannot tell the tape recorder that my name is General Douglas MacArthur and I live in New Guinea and expect proper results.

I have performed experiments on friends who volunteered to "beat the machine." They told lies concerning the color of their socks and underwear and showed no signs of stress (lying) whatsoever.

There is no consequence to these lies, no hazard if the lie is discovered, no feeling of moral compunction to create stress. Meaningless lies generated strictly for the purpose of testing are therefore not a proper application. Using voice stress analysis in such situations will likely produce no valid results whatsoever and is likely the source of much of the criticism of the method.

However, in other circumstances, voice stress analysis is extremely effective. One is the candid recording, recorded by one party to a conversation. The second party to the conversation does not know that the recording is being made, and thus exhibits no "stage fright" or other external and artificial manifestations of stress. An example of this is the above-mentioned interview with the District Attorney.

Another successful method approximates the classic polygraph method. An individual poses questions to a second party. The expected answers should not necessarily be yes or no, but neither should they be lengthy. Answers of a few words, up to a couple of sentences, are fine. It is useful if the person doing the analysis intentionally poses a question to get a "true" answer as a calibration.

The most difficult speech to analyze is a rambling monologue by an individual who knows that the recording is being made and that a subsequent analysis will be done. I have performed analysis on two informers who admitted to being informers, and claimed to be "switching sides." One was recorded clandestinely in a telephone conversation, and was clearly demonstrated to be lying.

Another was recorded knowingly, and the first analysis was done on a rambling monolog. It was absolutely unreadable. Statements of known truths showed severe stress, making it impossible to separate lies from simple personal guilt over the activities, as his mind reacted to the things being stated, and about to be stated. The next attempt was the recording of his reading a statement. It was also so full of stress as to make the analysis almost impossible. The third attempt was done in the manner of a classic polygraph examiner, with the examiner asking questions and the subject responding with short answers. This proved to be much more satisfactory, and resulted in finally discovering the truth.


Voice Stress Analysis-- difficulty of interpretation

The inherent weakness of voice stress analysis lies in the fact that emotional topics can also create vocal stress. Subjects to which the speaker is sensitive (an unfaithful spouse, an embarrassing situation, guilt over past actions) often generate stress in the voice. This is a normal human reaction and does not necessarily indicate deception. Such things can greatly confuse the issue and MUST be taken into account when performing voice stress analysis.

In a rambling and unstructured monologue, the speaker's mind is racing to select words and topics, and every time the mind touches a "hot button" there will be a blip of stress in the voice. Every moment of indecision will show a stress bump. Therefore, this is a situation to be avoided for serious analysis.

Experience with the technique allows the operator to sort out occasional stress over sensitive topics, emotional issues, and such. The final analysis takes these matters into account and looks for the clear pattern, the stress on several words in a row on the same topic, which indicates deception.

I have gained many hours of experience in analyzing stress in voices. I have analyzed friends, politicians, candid recordings, staged interviews. From this experience comes my confidence in the method, and my ability to properly use it. I must state again that proper interpretation of the results is the most important ingredient. One must not assume that an indicated absence of stress is indicative of the truth, for some people's natural level of stress is quite low, and the escalation upon telling a lie can escape detection. Likewise, one must not assume that an indicated presence of stress indicates a lie. VARIATIONS in the level of stress is the critical criterion.


Voice Stress Analysis-- can proper training defeat it?

I have mentioned that most of the infiltration is done by informers rather than trained professional agents. It is entirely likely that intensive training can enable a person to lie without stress in the voice. However, I seriously doubt that this degree of sophistication can be successfully acquired by the average snitch and, as mentioned, there are simply not enough trained agents. And even in a trained individual, I seriously doubt that such training can extend to every word spoken in all situations. In short, even the trained agent will be vulnerable to the hideout tape recorder making a candid recording.


Voice Stress Analysis-- The Plan

I envision the leadership councils of the various groups voluntarily, eagerly submitting to voice stress analysis. I envision that the leaders of the various citizens' organizations shall affirm their loyalty for voice stress analysis for the benefit of their membership. I envision that the members of various citizens' groups shall voluntarily and eagerly require that all members affirm their loyalty publicly and submit the recorded results for voice stress analysis. I envision a gathering at Knob Creek where, by mutual consent, attendance is dependent upon passing such an analysis, conducted by an independent contractor (who has himself passed such a screening). I envision a very simple, non-intrusive, very-much-to-the-point standardized set of questions. An example might be:

What is your name?
My name is Mike Kemp

What group or association do you represent?
The Gadsden Minute Men of Gadsden, Alabama, and myself.

Which is the higher authority-- the Bill of Rights, or the Supreme Court?
The Bill of Rights.

Declare your intentions for attending this meeting.
I hereby affirm that my reason for attending this meeting is to exercise my individual right to insure that all government will adhere to the literal written words of the Constitution, and to therefore protect all my rights for my posterity.

Repeat this pledge-- I am not acting for any entity other than those which I mentioned. I have truthfully stated my beliefs and associations.


The individual conducting the voice stress analysis, particularly with questions and answers as described above, is not required to know anything about the subject of the analysis-- not the name, age, race, financial condition, or anything else outside the recorded questions and responses. The person conducting the analysis must be knowledgeable and skilled in its application, but does not need to know any personal information about the subject of analysis.


A Proposal

Government entities at all levels are using this technology against us. Since it so directly views the emotional condition of the subject, it can, and is, being used to analyze US to obtain psychological profiles. It is a very powerful tool, very reminiscent of George Orwell's 1984. I do not propose to use this technology, and my skills applying it, to intrude into the minds of we, the people. I DO propose to provide some defense against the swarms of infiltrators being sent to eat out our substance. Any of you reading this may use the technology. However, most of you do not have the experience to properly apply it, and obtain the best results. I have dedicated much time and precious resources to obtain my skill in its use, proper application, and analysis of the results.

I propose to provide a service. I can provide this service on an absolutely anonymous basis. I will analyze tape recordings sent to me anonymously, and post replies to anonymous e-mail addresses or to "snail mail" addresses, as requested. I will analyze computer sound files (known as "wave," or *.wav, files) sent to me from anonymous e- mail accounts (such as those provided by hotmail and juno among others) and reply to same. I can receive these messages encrypted with pgp, and my public keys for that purpose are provided below. I primarily use pgp version 2.6.2, but am capable of version 5.0 and 5.53 as well, though I view them as less secure.

I do not propose to provide this as a free service. I have invested hundreds of hours to acquire my skill at this technology and have invested scarce resources to obtain the computer hardware necessary to quickly and accurately perform a high volume of analysis.

However. I do not expect you to believe that what I say is true. I therefore offer to perform a one-time free analysis for anyone who is interested. Send a recording to me stating something which you know to be true, or know to be false, or one of each. It should not be trivial. That is, don't tell me that you are John Henry from Alaska, or that you are wearing purple socks, and expect a proper analysis. A proper statement for analysis would be "I believe in that the Bill of Rights requires government to respect my individual rights. I believe that the ATF acted properly at Waco and I believe that their representatives have told the truth about those events."

The best way to make a tape for analysis is for one person to act as a questioner and another answering, with a short set of questions which are known beforehand by both the questioner and the respondent.

The statements do not need to be exactly as stated above. In fact, I would prefer that the statements NOT be so easily analyzed. I would prefer statements which I COULD NOT KNOW OR REASONABLY EXPECT to be true or false but which are not about trivial subjects. YOU will know the truth and will know whether or not I am capable of determining the difference. As I have stated, I have no need or desire to know who you are, or whose voice is present on the recording.


The Recordings Themselves

I would expect a recording, whether on cassette or in a computer sound file, of up to thirty seconds or so duration, making a statement similar to the examples mentioned above. If you know how to make computer sound files, I would prefer that they be in "wave file" format with the following characteristics: 16 bit, mono, sampled at 11025 bits per second. Cassettes in mono are fine. I will be happy to instruct you individually on the proper method of making wave files, which are really quite simple to make using a standard tape recorder and a patch cord available from Radio Shack for a few federal reserve notes.

When I have satisfied your doubts, I will perform analyses for you on a confidential basis. For a nominal sum, I will perform an in-depth analysis and inform you and only you (unless otherwise instructed) of the results. If I am unable to clearly determine whether the speaker is being deceptive, I will inform you of that fact without charge. I will destroy the recordings and analysis so that no record remains, or return them to you without retaining copies.

I shall personally require all individuals with whom I associate to pass the "snitch test" herein described. For those who do not wish to use this service that I offer, I strongly encourage you to become proficient in the techniques of voice stress analysis (which is not a simple nor easy task) and rid yourselves of the snitches which are ruining our attempt to require accountability and lawful behavior by our governments. Personally, I'm tired of it. It's time to fish or cut bait. Or, more appropriately, to BE cut bait.

William Michael Kemp
P.O. Box 873,
Attalla, ALabama, 35954



Listed below are keys for pgp versions 2.6.2 and version 5.0

Version: 2.6.2


Version: PGP for Personal Privacy 5.0i




Editor's Commentary: I would like to say that Mr. Kemp's VSA apparatus has proven its worth already in demonstrating that a number of militia leaders are not infilltraitors. It has also proven that a number of other militia leaders, such as Ken-the-Snitch McWilliams ARE government informants. In fact, it was through the use of VSA which detected McWilliams that McWilliams had no choice other than to eventually confess, and have a number of other infilltraitors profess that the presence of infilltraitors among the militia general population was compatible with patriotism. Thus if it was not for Mr. Kemp's initial detection of Ken-the-Snitch, this regime criminal and his Tri-State Coalition/Militia friends and accomplices such as Red Mike Vanderboegh and Ed-the-Dog (misnomered Wolfe) would still be betraying the unaware militia generals on PIML.

Thus Mr. Kemp's VSA detection apparatus acts much like a monitor for the health of a particular open militia unit. It helps detect regime criminal infilltraitors and thus a security quarantine can be enacted. And from what I have heard, Mr. Kemp's moral checkup is not overly expensive -- not considering the added security that it offers.

This recommendation does not override Editor's contention that the most effective security operation lies in proper Resistance organization by cells of family members and long-time friends and allies.

Martin Lindstedt



Back to Modern Militiaman Issue #9, July 4, 1998?

Back to index?