How Good Are People at Counting?

[Andrew Mason]

You make an assumption that the 100 viewers of the film did not talk to each other.

No. I make the assumption that the five did not all agree to lie in the same way.  Even if some or all of the five had talked to each other, it is unlikely they would enter into a conspiracy to lie, especially absent  some reason to lie about what they had seen.  It is much, much, much more likely that at least two of them actually recalled seeing what they said they saw.

But how could you know that? How do you know the witnesses at Dealey Plaza did not influence each other?

I don't know if they were all free from outside influence. It appears unlikely that they received information about the shot pattern but they could have heard about the number of shots from others.   But receiving information from others does not mean they are not independent - that they were not reporting what they actually recalled observing. Studies show (Loftus) that while information from others may affect some witnesses' recollection, it does not affect most and is much less of an influence when the witness is interviewed soon after the events, as most of the witnesses were (in relation to the shot pattern and the number of shots. )

You admit that sometimes, this ?Find the Truth by seeing what the majority thinks? breaks down.

I have never said that the what the majority thinks is a reliable method of determining what happened.  It is a matter of the statistical significance of specific observations by a number of observers.  What the majority thinks may or may not be statistically significant.
 

But it doesn?t matter, because you can use your judgement to tell when it did, or it likely did, break down. You can tell that if 95 people don?t recall a man in a gorilla suit, but 5 do, that this is a special exception.

It is not a matter of judgement. It is a matter of probability of independent observers reporting the same observation. That does not happen by random chance.

Even if this is true, this Great Principle is not a great principle, if the majority of the witnesses can be way off in some cases. How can we tell which examples violate this great principle? Can we rely on your intuition?

It is not about intuition. It is about mathematics.  The majority of witnesses are rarely way off in their observations of salient facts.  Where large numbers of witnesses have been wrong is where their ability to observe or compare is limited and they are effectively giving opinions of what they observed rather than simple describing what they observed.  This is particularly so with respect to facial comparisons (eg. witness tries to identify a person that they do not know from a photo line) or a witness tries to compare a sound to something they have heard before.

Let?s take the gorilla film example. You say you can tell what happened, because if 95 people don?t recall a man in a gorilla suit, but 5 do, you can logically conclude that there must have been a man in a gorilla suit. Because it is unlikely that 5 people would see something so off the wall and all saw the same thing.

So at least you agree with me on the gorilla observation. That's a start....

But does this work in general? Suppose there wasn?t a man in a gorilla suit. The film showed two teams, one in white basketball uniforms, the other in black basketball uniforms. In the middle of the practice, a man in some black clothes with a black hat walked across. If 5 people reported that one of the men was not wearing a basketball uniform, would this be so off the wall that we could expect you to conclude that the 95 witnesses were wrong and the 5 witnesses were right?

The short answer is : if the observations were independent, yes. Let's assume that in the group of 100 people there were 5 liars.  A man walking through with a hat is not the only thing a person would make up. But let's say there was a 1/10 chance that a liar who wanted to make up a story would make up a story about a man wearing a hat.  If they are independent, the probability that 4 liars would independently make up the same story as the first liar is 1 in 10,000.  The conclusion has to be that it is extremely unlikely that there was not a man with a hat in that video.

Or let?s go with the original gorilla film. 95 views report nothing unusual but 5 report there was a black gorilla. Can we really deduce there must have been a man in a gorilla suit? Maybe one the views thought he saw a gorilla in the last second of the film. And he told other people that he saw a man in a gorilla suit. He could be so convincing, that some other people might believe him. In interviews taken down later that day, 5 people might report seeing a man in a gorilla suit. What you assumed was 5 independent events weren?t really independent events at all. The error of all 5 people was caused by the error of just one man.

Then they would not be independent.  If 5 people said there was a gorilla in the film and there really wasn't, you would know that the witnesses were not independent.  If they were really independent, it is virtually impossible that even two witnesses would report seeing a gorilla if there was no gorilla.

And in the Dealey Plaza witnesses, you conclude the majority in a list of witnesses reporting the limousine stopped or almost stopped, you just claim the list is faulty. Without providing us with a better list.

Not true. Here is my analysis of all the witnesses who observed the limo. 8 said it stopped. Most said it slowed or almost stopped.  This is more an illustration of witnesses who were not in a good position to observe whether it actually stopped providing an inference rather than an actual observation.

It seems that whenever a failure of the ?Majority must be right? witness theory, you do special pleading with each example. The gorilla film is not a problem because you can use your special analysis to spot a special exception. The majority of the witnesses report the limousine stopping or almost stopping, contradicting the Zapruder films, and all other films, is just the case of a bad list. That is what your judgment tells you. While your judgment tells you that the list of the ?Spacing of the Shots? witnesses is a good list and should be trusted.

All I am assuming is that at many of the witnesses(I don't have to assume all) were providing independent honest recollections of what they recalled hearing.   That is all.

No. I would not be comfortable with going with either the 95 ?non-gorilla? witnesses nor the 5 ?gorilla? witnesses. I do not trust witnesses.

Fine. You could decline to draw the conclusion that there was a gorilla in it.  But you can't conclude that the 5 witnesses were wrong unless you have evidence that the 5 were not independent.  Not just a possibility but actual evidence that they all colluded.  There are only two reasonable possibilities: the five were right or they all colluded.

But a man who relies on the ?Majority opinion of the witnesses? would be wrong. Unless he had some special intuition that told him when to apply this principle and when not to.



The observers of the gorilla film failed to do what one would think they should easily do, spot the gorilla. Because they were concentrating on something else, counting the number of passes the white team made to each other. Similarly, the Dealey Plaza witnesses may have failed to do what one would think they should easily do, tell us how many shots were fired, the spacing of the shots, the speed of the limousine. Because they were concentrating on something else. Hoping to catch the eye of the President or the First Lady. Trying to remember their likely one and only close up view of a President and First Lady.

As a final aside, how could the shot spacing witnesses be explained? Perhaps the witnesses were distracted, but not during the entire event. After the fatal headshot, they realized something terrible happened. The remembered the previous 5 seconds pretty well, from 5 to 10 seconds not so well, and over 10 seconds not well at all. The could have forgetting the first ?backfire? or ?firecracker?, remembered the second shot vaguely, the last shot rather well, with it?s ?Crack-Thump?. Hence becoming a ?3 shot ? last two shots close together? witness.

That is speculation.  One  cannot draw conclusions from speculation.  Without evidence that such influence actually occurred and was widespread, the likelihood that they were all influenced to provide the wrong pattern of shots is much, much, much smaller than the probability that they actually observed the 1......2...3 pattern. 

[Bill Chapman]

Apparently not if you think McDonald said in that video that Oswald tried to shoot him.

Sorry, but "Oswald carrying his rifle to the crime scene in a brown paper bag" is the thing you're supposed to be proving.  You don't just get to state it as a given.  The reason I LOL is because you do that all the time.  "Oswald's gun bag", "Oswald's rifle', "Oswald's ammunition", ad nauseum.

Are you the same person who whined when somebody called him "Mutton"?



Those who can make coherent arguments do.  Those who cannot make up silly nicknames and try to change the subject.

The reason I LOL is because you do that all the time.  "Oswald's gun bag", "Oswald's rifle', "Oswald's ammunition", ad nauseum.

Will your 'random guy' be more acceptable?
How about 'Anyone But Oswald' ('RandomMan')?

[John Iacoletti]

Will your 'random guy' be more acceptable?

What do you mean my random guy?  I never said anything about a random guy.

[Jerry Organ]

It is not about intuition. It is about mathematics.  The majority of witnesses are rarely way off in their observations of salient facts.

Three (if that's what it was) loud noises became increasing "salient" as the sounds unfolded. The first few loud noises weren't "salient" to everyone. Some dismissed the first loud report as a backfire or firecracker. They were also in a distraction setting with peak concentration on the motorcade at the time of the first shot.

While it's true they could not fail to hear three shots (if there were, in fact, three shots), it doesn't mean it was stored in memory equally. For some of the witnesses, their perception and retention of initial events (some termed the first shot as a backfire or firecracker) could be affected by a greater concentration on the latter shots and things occurring visually up to and including the shock of the head shot and the dramatic Jackie/Clint potential tragedy.

Spectators in the stands at the Boston Marathon Bombing barely react until the second bomb goes off.

[Bill Chapman]

What do you mean my random guy?  I never said anything about a random guy.

It was a description you used to describe Oswald in relation to the scene around the TT

[Joe Elliott]

No. I make the assumption that the five did not all agree to lie in the same way.  Even if some or all of the five had talked to each other, it is unlikely they would enter into a conspiracy to lie, especially absent  some reason to lie about what they had seen.  It is much, much, much more likely that at least two of them actually recalled seeing what they said they saw.

No. I?m not assuming five decided to lie. I am postulating that one person was mistaken and influenced four others, before they could be interviewed.

I have never said that the what the majority thinks is a reliable method of determining what happened.  It is a matter of the statistical significance of specific observations by a number of observers.  What the majority thinks may or may not be statistically significant.

 It is not a matter of judgement. It is a matter of probability of independent observers reporting the same observation. That does not happen by random chance.

This is real helpful. To paraphrase what you are saying:

What the majority thinks may have happened is a matter of statistical significance. Except when it isn?t.



You can tell when the statistical method can be used using, what appears to me, by intuition. If the majority reports a certain shot pattern, but a minority reports a different pattern, you can tell, that in this case, the majority is right. But in the case of the gorilla film, if the majority reports no gorilla, but a minority reports a gorilla, you can tell, that in this case, the statistical method cannot be used, because the minority, in this case, are probably right.

It is not about intuition. It is about mathematics.  The majority of witnesses are rarely way off in their observations of salient facts.  Where large numbers of witnesses have been wrong is where their ability to observe or compare is limited and they are effectively giving opinions of what they observed rather than simple describing what they observed.  This is particularly so with respect to facial comparisons (eg. witness tries to identify a person that they do not know from a photo line) or a witness tries to compare a sound to something they have heard before.

It?s not about mathematics. It will only be mathematics if the witness errors are random. We cannot assume that.

Not true. Here is my analysis of all the witnesses who observed the limo. 8 said it stopped. Most said it slowed or almost stopped.  This is more an illustration of witnesses who were not in a good position to observe whether it actually stopped providing an inference rather than an actual observation.

The bottom line is, using the statistical method to determine the approximate ?Speed of the Limousine? using witnesses fails. We would not know this is we did not have the films of the assassination.

As you acknowledge, the statistical method fails in this case. Probably, indeed, because most witnesses were not in a clear position to see the limousine, only the follow up cars which did (probably) stop. But that did not prevent them from giving a confident opinion. Something we should remember about with the ?Spacing of the Shots? witnesses.

This is a classic case as to why we cannot rely on the statistical method. Because the errors witnesses made were not random. We cannot assume these errors are always going to be random. Which is what we need if the statistical method is going to be dependable.



Could the ?Shot Spacing? witnesses be an example of widespread witness error, non-random witness error, just like the ?Speed of the Limousine? witnesses? Easily. Below is a possible theory:


Most witnesses were distracted. Concentrating on the President and the First Lady. They may have ignored, possibly forgot what they assumed were backfires or firecrackers. When the realized shots were fired, they may have remembered the more recent shots better than the first. Hence:

First shot: Had forgotten about 15 seconds later, it never really got stored in their permanent memory.

Second shot: Remembered, but not very well.

Third shot: Remembered it very well, as a ?Crack-Thump?.

Hence, many may have remembered ?Bang <pause> Bang Bang?.

As I recall, some witnesses thought all the shots occurred in pairs, ?Bang Bang <pause> Bang Bang?.


My theory may be false. But we cannot assume it is false. There may have been not just significant but non-random witness errors, in the ?Shot Pattern? witnesses, just was there was in the ?Speed of the Limousine? witnesses.

[Joe Elliott]

Three (if that's what it was) loud noises became increasing "salient" as the sounds unfolded. The first few loud noises weren't "salient" to everyone. Some dismissed the first loud report as a backfire or firecracker. They were also in a distraction setting with peak concentration on the motorcade at the time of the first shot.

While it's true they could not fail to hear three shots (if there were, in fact, three shots), it doesn't mean it was stored in memory equally. For some of the witnesses, their perception and retention of initial events (some termed the first shot as a backfire or firecracker) could be affected by a greater concentration on the latter shots and things occurring visually up to and including the shock of the head shot and the dramatic Jackie/Clint potential tragedy.

This is correct. They may have a better memory of the previous 5 seconds, when they first realized that shots had been fired, than they did of 5 to 10 seconds earlier and worse yet for 10 to 15 seconds earlier. If certain events are not remembered within a few seconds, they may be, likely will be, forgotten altogether.

Spectators in the stands at the Boston Marathon Bombing barely react until the second bomb goes off.

Yes, an example of distracted witnesses becoming aware witnesses.

[Andrew Mason]

No. I?m not assuming five decided to lie. I am postulating that one person was mistaken and influenced four others, before they could be interviewed.

But if the 5 witnesses said that they saw a gorilla and, in fact, did not (and, of course, they would know they did not) they would be lying in saying that they saw one.  If they were truthful, they would say something like: "I don't remember but I think there may have been a gorilla because Bob told me".

This is real helpful. To paraphrase what you are saying:

What the majority thinks may have happened is a matter of statistical significance. Except when it isn?t.

No. That is not what I am saying. If 51% said there were 3 shots and 49% said there were 4, I would not be able to conclude whether there were 3 or 4 shots.  I would conclude that witnesses had difficulty observing the number of shots and were confused by something. This is because 51-49 is not statistically significant.  Even 60-40 may not be enough, especially if the quality of the recollections was poor (eg. " I am not sure but I would say there were three, maybe four, no i think three"). You cannot simply draw conclusions based on what the majority observed.

You can tell when the statistical method can be used using, what appears to me, by intuition. If the majority reports a certain shot pattern, but a minority reports a different pattern, you can tell, that in this case, the majority is right. But in the case of the gorilla film, if the majority reports no gorilla, but a minority reports a gorilla, you can tell, that in this case, the statistical method cannot be used, because the minority, in this case, are probably right.

No. The point of the video is that most people are not observing anything other than the white players so if they noticed nothing, their failure to notice is not significant.  But for those who did notice, the similarity of their observations cannot be explained by anything other than 1. they saw the gorilla or 2. they were lying and colluding.

It?s not about mathematics. It will only be mathematics if the witness errors are random. We cannot assume that.

Again: two possibilities: they saw a gorilla or they were colluding and lying about it.  If they were not colluding you can conclude that any errors would be random. 

The bottom line is, using the statistical method to determine the approximate ?Speed of the Limousine? using witnesses fails. We would not know this is we did not have the films of the assassination.

As you acknowledge, the statistical method fails in this case. Probably, indeed, because most witnesses were not in a clear position to see the limousine, only the follow up cars which did (probably) stop. But that did not prevent them from giving a confident opinion. Something we should remember about with the ?Spacing of the Shots? witnesses.

You can tell by the distribution of witnesses to the limo slowing down that the ability of many to observe whether it actually stopped was restricted. They saw the brake lights and a motorcycle stop (it actually did) and other cars in the motorcade ( not seen in the zfilm) may have actually stopped.  They formed the opinion that it stopped. This was an inference based on what they saw.  We can see this in some of the witnesses who initially said it stopped and then had to admit that they weren't sure it stopped, just that it slowed eg. Dallas police officer Earle Brown

This is a classic case as to why we cannot rely on the statistical method. Because the errors witnesses made were not random. We cannot assume these errors are always going to be random. Which is what we need if the statistical method is going to be dependable.

There are two possibilities: errors are random or they are caused by a common factor that induces the error (such as collusion/lying).  If the circumstances permit a possible explanation for error other than collusion/lying (such as the brake lights combined with slow speed combined with the motorcycle stopping and the officer getting off and giving people the same impression that the car had actually stopped) you can look at that.  But, invariably, in such cases you rarely get statistically significant distributions - as we see in the "limo stopped" witnesses.  There were 8 who said it stopped and 19 who said it slowed or that the motorcade stopped without specifying whether the limo stopped.  We can't be sure if any cars in the motorcade actually stopped but we certainly cannot say that none stopped.

Could the ?Shot Spacing? witnesses be an example of widespread witness error, non-random witness error, just like the ?Speed of the Limousine? witnesses? Easily. Below is a possible theory:


Most witnesses were distracted. Concentrating on the President and the First Lady. They may have ignored, possibly forgot what they assumed were backfires or firecrackers. When the realized shots were fired, they may have remembered the more recent shots better than the first. Hence:

First shot: Had forgotten about 15 seconds later, it never really got stored in their permanent memory.

Second shot: Remembered, but not very well.

Third shot: Remembered it very well, as a ?Crack-Thump?.

Hence, many may have remembered ?Bang <pause> Bang Bang?.

As I recall, some witnesses thought all the shots occurred in pairs, ?Bang Bang <pause> Bang Bang?.


My theory may be false. But we cannot assume it is false. There may have been not just significant but non-random witness errors, in the ?Shot Pattern? witnesses, just was there was in the ?Speed of the Limousine? witnesses.

The question is whether it is true.  The only way you can determine if it is a true theory is by comparing it to the evidence. Not a single witness said they had difficulty recalling the 1......2....3 shot pattern.  Ask Robert Jackson or Mary Woodward what the shot pattern was. They still remember it.  Ask them what they think of your theory.