Do Polygraph Machines (Lie Detectors) Work?

[Englebert]

I was recently reading an article about a missing person, and the article stipulated that a possible suspect had passed a lie detector test. While reading the comments section I was astonished at how many people were basing their comments on the passing of the lie detector test. I was under the impression that most people knew polygraphs did not work. I started this topic to get an idea as to how many people think polygraphs do or do not work. I'm really not interested in the details of the workings of the test or what they are measuring or the accuracy of the measurements...I'm very famailiar with that. I'm interested in hearing how accurate and how credible the general public thinks these tests are.

[westend1]

They are obviously not 100% accurate, but they work well enough to use as a tool.

[mat]

I think they are pretty hard to beat but they are not admissible in court so they're not perfect.

[tvc184]

They are obviously not 100% accurate, but they work well enough to use as a tool.

 

And there you have it.

 

If a person blows it off of the charts then he is likely involved in whatever you are looking into. If not them he probably isn't your guy. As westend states, it is a tool. 

 

Many people will blow a polygraph so bad (and they know it) that they will confess or sometimes confess before they take it during the pre-test interview. 

 

In any case it helps law enforcement have an idea if they are likely on the right trail or not. Doing good doesn't mean that you aren't involved and doing bad doesn't mean guilt but it is accurate enough most of the time to help with the investigation. 

[Englebert]

Studies have shown that many people will confess when faced with a polygraph test. But polygraph tests also result in an enormous amount of false positives (people telling the truth but the administrator thinks they're lying) that the overall benefit is negated. That's why it has been dubbed the "fear test". The three variables that the polygraph measure have absolutely no correlation with lying. Phrenology tests are just as effective as polygraphs if the subject thinks that the test actually works.

 

The reason I was asking this question (and I shockingly found my answer even with just 3 responses) was I thought that the general public had been exposed to the wealth of information showing that polygraphs do not work. 60 minutes did a prime time show debunking polygraphs 20 years ago. Penn & Teller did an hour long show recently debunking them. A quick google search will give you all the info you need to realize these are phony tests. I am kinda amazed at how out-of-touch I was with how the general public views polygraphs.

 

And if anyone thinks they work, I can show you how to beat the machine with a simple little test. Write down a number from 1 to 25 on a sheet of paper (don't show the polygraph administrator what number you wrote down). Then answer all of the questions the administrator asks, with the provision that the administrator must ask if you wrote on the paper the number 1, the number 2, etc so that each number is asked. You answer “no” to each number. Therefore, it is a fact that you lied on one and only one of the numbers. If polygraphs work, the administrator should be able to tell you what number you wrote down.

 

When doing a study on polygraphs in college, I was able to get a polygraph specialist to administer this test to me (with a written confidentiality clause that I could use the results but could not reveal the company or administrator’s name). He failed, as would any other polygraph administrator. Later I was able to duplicate the results with 4 other subjects (subjects being the ones that the test was administered to). In each case, the polygraph administrator could not "guess" the correct number. The administrator actually thought he could figure out the correct number, and was shocked when he could not.

 

I kinda secretly hope that one day I am asked to take a polygraph test. I will only consent to taking a real test after the administrator correctly guesses my number. Not only would I not have to take the test, I would be able to expose the fraud that is the polygraph.

[tvc184]

Those silly experiments are nonsense. The polygraph depends on a person having something to lose or as you say, the fear test. I agree that it is fear that drives them but it isn't fake. Telling me that it is a fear test doesn't negate the results, it shows what causes them. 

 

A person that lies about a number on a test that he volunteers for has no consequences. I have watched some of those and they have no bearing. Have the person sign away his life savings or let him get hit with the Taser if they can discover the lie and see what the results are. 

 

A polygraph is not a lie detector but a measure of blood pressure, pulse, respiration, etc. It doesn't flash that a person is lying, it tells you questions that cause a person problems in some kind of fear of being discovered or as you say, a fear test. You can bet that there are people in this forum that might have stretched the truth a bit on how big of a fish they caught, etc. Their blood pressure didn't sky rocket out of fear of being discovered, their respiration didn't get heavy and their pulse didn't quicken. It is because there was no one there to call them on it and there was no real loss because of the lie. 

 

When a person is answering questions about a crime and you get to a question that makes his pulse rate go up, his blood pressure start rising and his breathing starts to get more labored and the previous 10 questions had no such response, there is a reason. 

 

The instrument does not detect lies. It detects body responses. A person that "passes" a test does not mean that he is telling the truth (as the goofy numbers test shows) but he is a person that has no fear of what he knows or about someone else finding out. A person that shows deception might have other reasons for the elevated responses. It does not mean that he lied but something in the question is causing a different response than the other questions.

 

Like two people have said, it is a tool. If it was ever a definitive "lie detector" then it would be admissible in court and we could do away with juries. It is not nor is it ever claimed to be.

 

Does the general public know that? Probably not but many of them also believe that the police can get DNA results in 15 minutes and plug it into a computer of every living soul and come up with a suspect. Most of the public knowledge of law enforcement, laws in general and prosecutions comes from television and movies. 

[Englebert]

Every thing you said is correct...but you emphasized just the true postives. The false positives are an enormous problem. Many innocent people are shown to be lying on a polygraph. So many so that when you get a positive result, you have no idea if the result is a true or false positive. If you administer a test to a person and the test shows that he is being untruthful (rapid heartrate, perspiration, etc.) then what are you to conclude. Well it could be that the person is just nervous or it could be that he is actually lying. The fact is that you have no idea what the real truth is after the test is administered. If you think he is lying and follow that up with more investigation then two things can happen. You can find the evidence and feel good that the tool helped you. Or you can not find any evidence because the person was nervous and you've just wasted time and resources on a guess that didn't pan out. If the first instance was the majority of the instances, then it could be construed as a good tool. If the second instance was the majority, the test could be construed as a hindrance to investigations. Many studies have shown that the first instance happens between 50-63 percent of the time...which is basically no better than flipping a coin.

 

In many of the studies on the subject, people were shown to be lying when they were actually telling the truth. When questioned why they got nervous on certain questions, many people responded that they were worried that the test would show them lying even though they knew they were telling the truth. These false positives show up more as the stakes are higher. That is, the more a person has to lose, the more false postives result, and the higher they blow up the test. When you have a guy in custody, he knows you think he has committed a crime. When the question arises on the polygraph (Did you shoot your wife?) or some other question, even though the subject did not do it, he's afraid his answer will register falsely, thus the rise in heartrate, perspiration, etc. An innocent man is now the main subject of your investigation.

 

In addition, some studies have shown a high number of false negatives. That is a person is lying and the polygraph is showing them as telling the truth (no significant baseline change in heartrate, etc.). If I remember correctly, many of these came about from people that were giving answers based on what he thought the administrator wanted to hear. Many of these people have a high need to please and will even lie to win approval. These also came about when the consequence of failing a question was something unwanted by the subject (such as being tasered).

 

The silly little numbers test positively proves that a "lie detector" does not measure lies. You and the others already knew that. I prefaced that paragraph with "for anybody that thinks that they work". I'm under the impression that the majority of the general public (not you three) believe that polygraph tests is a measurement of lies. That was the whole point of the rest of my post.

[tvc184]

You seem to have an axe to grind with polygraphs and note that they do not detect lies. That is true but they work the way they are intended. 

 

Remember that a person does not have to consent to taking one for a criminal investigation. I have known many people that "passed" in that they did not show deception so the idea that merely asking people a question of guilt will show that they are lying (false positive) because they are nervous is not correct, otherwise every person would show deception. 

 

When I took a polygraph to get hired by the police department, they asked me all kinds of sensitive questions (some which they can't ask any more) that made me very nervous as they would exclude me from being hired if they could find out evidence that I was not telling the truth. That is also why all questions have to be asked multiple times and there has to be a certain amount of time between questions. The idea that some questions make people nervous even if they are telling the truth is a known factor. 

 

I was watching an old Dragnet show last week from I think 1969 where a murder suspect was taking the polygraph (girlfriend was found dead). He showed "deception" but then Sgt. Friday asked about the interpretation of the "suspect's deception", the polygrapher stated that he did not think the guy did it. Of course it was television (claimed to be from actual cases) but when questioned the polygrapher stated that maybe the guy watched a recent movie where a similar event happened, maybe his girlfriend always did something a certain way or he could have merely guessed because of what is routinely thought of as a means of death like a shooting or strangulation. When asked specific questions about where the victim's car was located, where the body was found and other such details were asked, he showed no response. So even 45 years ago some television showed that even with a positive response in some areas, it does not show that a person lied or was guilty. As the polygrapher in the episode noted, a couple of deceptive answer did not show guilt and he thought that the cops had the wrong guy and in fact they did. The entire interview and test is the key, not a single "did you lie" question. 

 

I can believe that was taken from a real case where the guy on the hot seat was very nervous about general questions because everyone thinks if a woman is found dead, there is a good chance the boyfriend/husband did it but in that case the details cleared him. And again, it is a tool that a person does not have to submit to. I can assure you that many criminal cases have been cleared by the polygraph and with completely different factual evidence as the use of the polygraph itself cannot even be disclosed in court. It takes real evidence both testimony and physical for a conviction, not a polygraph. 

 

If the police had a lot of problems with false positives as you suggest, why would they still be using them? I never saw a point in wasting my time in an investigation and in the years that I was in detectives I only asked for a polygraph on a couple of occasions. One guy backed out and one confessed when I suggested it. 

 

There are also other methods to solve cases and get confessions like kinesic interview techniques which are basically a person to person "lie detector". They are extremely effective if the interview/interrogation last long enough but like the polygraph, it is just a tool to get a confession or to find evidence and the technique itself is not evidence. We can't give up crime fighting because hunches, polygraphs, interview techniques, etc., can't be stated as evidence. They all play a big part is solving some crimes however. 

[tvc184]

............ and if your point is that the polygraph doesn't specifically detect lies and the the public doesn't always know or understand that, I agree. There are lots of areas the public has absolutely no clue about. 

 

The way I look at that issue is, who cares? It is still an option for an investigation just like other methods. It is not an end all in an investigation. I wish it was that simple. I have seen a case that I worked around the edges about 25 years ago where a guy completely blew a polygraph and we thought that he kidnapped and killed a small child. He did not confess however and we at that point did not even have a body. A few weeks later we found her body and about 10 years later he confessed. The polygraph (probably through his fear of being discovered) was dead on target and it meant nothing as we could not use it. If the polygraph would have ended it, we might have found the little girl sooner and would not have to wait several years for his conscience to bother him. 

[Englebert]

No axe to grind. I was just surprised to recently find out that many people think polygraphs actually detect lies. I did the research many years ago, and I just assumed that the results I found 25 years ago was common knowledge today. (I also did a minor follow-up research paper around 12 years ago.) I knew law enforcement uses them (to what extent I have no idea), and am surprised by their use because the results are absolutely unreliable. I don't really care if they are used for trying to solicit a confession, but it just seems that in an investigation the overall results are going to waste as much time as they save. And there are many studies out there that show just that.

[tvc184]

No axe to grind. I was just surprised to recently find out that many people think polygraphs actually detect lies. I did the research many years ago, and I just assumed that the results I found 25 years ago was common knowledge today. (I also did a minor follow-up research paper around 12 years ago.) I knew law enforcement uses them (to what extent I have no idea), and am surprised by their use because the results are absolutely unreliable. I don't really care if they are used for trying to solicit a confession, but it just seems that in an investigation the overall results are going to waste as much time as they save. And there are many studies out there that show just that.

 

Our results must be atypical then because I don't see the waste of money and time. It costs almost nothing and I have never seen it cause overtime for the operators. They schedule the interviews during regular duty time just like any other interview/interrogation. It is not like there is an emergency called out at 3:00AM for a spur of the moment polygraph where the taxpayers have to pay the overtime. 

 

When you get a confession from a child molester or a murder suspect, I wonder how they measure that cost as a waste. I wonder if another method could have been used faster and cheaper. 

 

I can look at the time that we have had for overtime on surveillance and what the odds of success are. I imagine Beaumont PD has put in a lot of time and money (maybe thousands of dollars) into the serial rapist investigation and as of yet, has turned up nothing or at least not to getting probable cause for an arrest (I am sure that they have some leads). I guess stakeouts and task forces need to be stopped as they are not cost effective. The question then goes to, what do the police do to solve the serious crimes and what part does cost play to the public and are the other methods more cost effective than a polygraph? 

 

My guess is that they are not. 

 

Or as westend stated, it is a tool. 

[Englebert]

I'll try to equate using a polygraph to using a GPS. If the GPS sent you on the right course 60% of the time, and sent you on the wrong course 40% of the time, you will waste time and resources 40% of the time by having to reroute to the correct course once you figure out you were on the wrong course. To me, that GPS would not be a good tool to use, even though you get positive results 60% of the time. If you have a suspect in an investigation that you have a hunch "knows something", and you bring him in for a polygraph which he fails. I'm sure you will now focus more attention on this individual because of the failed test. If the test is this flawed, do you really know anything more than you did before the test?

[jv_coach]

Define Irony:Lie detecting test that lie

[tvc184]

But polygraphs do work a lot better than 60% and besides, the analogy doesn't make much sense. A GPS is dependent on almost pinpoint accuracy. A polygraph does not and only guages reaction. A hand grenade doesn't have pinpoint accuracy either but is a highly effective tool.

Simple interviewing of people isn't nearly as accurate as a GPS either so should the police quit questioning people?

[tvc184]

Define Irony:Lie detecting test that lie

But as Englebert keeps pointing out, they don't detect lies and never claim to.

[Englebert]

I was just using the GPS analogy as saying that the tool (polygraph) can lead you down the wrong path. That is, leading you to believe that a person was guilty when if fact he is not, or vice-versa.

 

A hand grenade is not an effective tool if your goal is to kill one person standing in the middle of a group on innocent people. :)  I know what you were implying though. Hand grenades can be very effective tools in some situations and very bad tools in others. And basically that is my whole point of polygraphs. My contention is that polygraphs are a bad tool in most situations. (Scaring someone into confessions has been shown to be somewhat effective through the use of polygraphs.) If the goal is to solve a case by gaining factual knowledge, I can't see where the use of a polygraph helps in that endeavor. If you ask a guy a question and he answers, what knowledge have you gained? You still don't know if he's lying or telling the truth. If the polygraph says he's lying, you still don't know if that is the actual case, due to the high number of false positives. If the polygraph says he's telling the truth, you still don't know if that was a false negative. No reliable data can be gleaned. If y'all are having success with the machine, then by all means keep using it. But I am highly skeptical of any success rates attributable to a polygraph.