Bigfoot Caught on Camera in PA Woods

[stereologist]

But still the estimate is really only as good as your sample. How do those doing the sample know it is representative?

This is looking at the problem the wrong way. Trying to find a so-called representative sample leads to problems, ie the wrong answer. Use a fair sample, not a representative sample. In this situation it appears that you have used a sampling technique to examine a small population. Next you describe samples with a high variance. Then you decide to cook the data by tossing out outliers. All of this is bad, bad, bad.

Finally you conclude with an invalid statement.

These estimates are built on estimate, which are built on estimates.

Clearly you don't understand sampling and estimating. The goal of sampling is to get the right answer. The problem is that the work to get the accurate answer, the actual number of bears down to the last bear is:

1. Too costly in terms of money, people required, and time

2. There is no need to know a number that precisely. Does it matter if there are 13,000 bears or 13,002? Does it matter if there are 13,000 or 14,000 bears?

So now in your example you've decided that there are only 4 samples and they are 5, 2, 4, and 11. That's pretty bad since the ratio of the standard deviation to the mean is 0.7. That tells us that you have done a terrible job at setting up the sampling protocol. If this part of the work is done poorly, then the answer is going to lack precision.

There is a proof that the only unbiased linear estimator is to divide the sampled values by the sample fraction to obtain an unbiased estimate of the bear population. Each sample is 1/4 of the population. Dividing by that fraction is the same as multiplying by 4.

Pick one of the 4 samples at random and multiply that population by 4.

The population estimates are: 20, 8, 16, and 44. The actual population is 22.

Now you are going to say "That's worthless." Fine, I'm not the one that did the bad job of sampling.

[stereologist]

The brown bear/black bear issue you mentioned simply means finding a way to count each population separately. This is called stratifying the sampling. It is done all of the time.

[stereologist]

They admited that their estimates were low. They as much as admit that if no one had done a study, that estimate would still be low and some 20000 plus bears would be roaming about unofficially. I am supposed to believe that a state agency that misses more then half the bears can successfully track a possible population consisting of only possibly several hundred (statewide). Are we to assume that they could be expected to track half of every population as a minimum? The smaller the population the easier it is to be ignored.

Clearly you don't understand sampling. They not only admitted they were low, but stated why they were low and took corrective action. They are not tracking bears. They are estimating populations. very different.

[stereologist]

A precise estimate should be accurate

A precise estimate does not mean an accurate estimate. These are very different concepts.

[stereologist]

Back tot he main topic here. The fact that an estimation technique can be flawed is not new. The problems with sampling are well understood. The issue here is that sampling does not apply to big foot. To prove the existence of big foot you don't sample. You simply find one and you are done.

[+DieChecker]

No, the discrepancy was picked up within one year. In other words a years study picked up the disparity. I suspect it would probably take near that to accurately count 20,000 Bears? I would not expect them to breed all year, and then line up for a count on December 1 lol. That would be funny though.

Oh. I understand your statement now.

meaning that in only one year they accurately extrapolated the current state of the species. Perhaps it has been as much as 10 years since the last such in depth count was done. It's a big call as they are claiming to have an accurate count right now, saying they have counted 20,000 Bears means no Bear has escaped? I think that is vigilant and would have required quite a large and co-ordinated team.

All the meantime these Bears are under close scrutiny. Now even more so to see why the disparity to begin with. Miscount? Population Explosion? Migrational Patterns? Computer Error?

I too would think 4-6 years more likely for that large an expansion.

I have to agree here. Your logic is too clear to fight.

Only problem, mate, that is somewhat of a biggie!

A bigger problem is people like Meldrum are looking for him, and claim to have found footprints. He has had more than enough time to to come forward. If there was a body to find, Meldrum or similar would have found it by now. That is what they claim to do, but Only Tom Biscardi has produced a real specimen to date, shame they all be hoaxes, but I think that is also somewhat telling?

Old Tom has had a live BF or a body part, what, four or five times. Once in a cave on Mount Shasta, if I remember right. He is a piece of work that Tom.

This is looking at the problem the wrong way. Trying to find a so-called representative sample leads to problems, ie the wrong answer. Use a fair sample, not a representative sample. In this situation it appears that you have used a sampling technique to examine a small population. Next you describe samples with a high variance. Then you decide to cook the data by tossing out outliers. All of this is bad, bad, bad.

Finally you conclude with an invalid statement.

I took Statistics 1 and 2 in College, but that was ten... no... seventeen years ago. You are right that I do not know what you are getting at. How is a "fair" sample derived. I would think sampling would be to randomly determine positions of a fixed size of a known habitat and count the entire population. Bam! you cordon off ten square miles and you count the bears. What more can there be? Repeat enough to get a good average and mean. Then pop that into your computer that has the habitat info in it and again, Bam!, you get an estimate. If I am truely ignorant, perhaps you can link me to a site or other resource to enlighten me further. That is one of the things I am here for, to Learn, not just to enjoy discussion.

Clearly you don't understand sampling and estimating. The goal of sampling is to get the right answer. The problem is that the work to get the accurate answer, the actual number of bears down to the last bear is:

1. Too costly in terms of money, people required, and time

2. There is no need to know a number that precisely. Does it matter if there are 13,000 bears or 13,002? Does it matter if there are 13,000 or 14,000 bears?

No, but it matters if there are 13,000 or 34,000 bears. Economically, environmentally and for public safety reasons.

So now in your example you've decided that there are only 4 samples and they are 5, 2, 4, and 11. That's pretty bad since the ratio of the standard deviation to the mean is 0.7. That tells us that you have done a terrible job at setting up the sampling protocol. If this part of the work is done poorly, then the answer is going to lack precision.

What does that mean to have a poor sampling protocol? How can you set protocol for something that should be random? Or are you implying that you measure from the same locations each time? So as to get the changes from year to year. Maybe that is where I am confused. I think of them as taking random samples, and you are saying that they are not random, but painstakingly chosen to be reliable.

There is a proof that the only unbiased linear estimator is to divide the sampled values by the sample fraction to obtain an unbiased estimate of the bear population. Each sample is 1/4 of the population. Dividing by that fraction is the same as multiplying by 4.

Pick one of the 4 samples at random and multiply that population by 4.

The population estimates are: 20, 8, 16, and 44. The actual population is 22.

Now you are going to say "That's worthless." Fine, I'm not the one that did the bad job of sampling.

It is not worthless. I clearly understand what you are getting at. I am just trying to point out that there are flaws even in the best estimates. As you said, we can not physically count every single bear. It is cost and time prohibitive.

Clearly you don't understand sampling. They not only admitted they were low, but stated why they were low and took corrective action. They are not tracking bears. They are estimating populations. very different.

Maybe I read this wrong. Was the researcher working for the Wisconsin DNR? It was my belief, erroneous perhaps, that this was an independant university researcher who discovered this, not as a plan to count bears for the state, but as his own project. If I was wrong I appologize. I had been assuming the DNR just was being lazy. If they had this man working on their problem then they are actually correcting the problem, not ignoring it. Still, they must have been somewhat lazy in the past.

Back tot he main topic here. The fact that an estimation technique can be flawed is not new. The problems with sampling are well understood. The issue here is that sampling does not apply to big foot. To prove the existence of big foot you don't sample. You simply find one and you are done.

Edited December 9, 2009 by DieChecker

[stereologist]

I'm not a statistician by trade, but I have worked in an obscure field that uses statistics. I was working on the sampling of geometric quantities, one of which is estimating the numbers of things. A sample as you recall is used to represent a population. From the sample an inference is made about the population. Where things go astray is when some of this is turned around and some claim is made to find a representative sample.

In the case of counting, a fair sample is one in which every member of the population has an equal chance of being counted. For the bears you need to have every bear being given an equal chance of being counted. So if the baits are distributed in such a manner that only some bears can access the baits or only some bears would find the baits, or whatever, then the sampling breaks down. The next part of the sampling is the hunter. If the hunters preferentially shoot darker colored bears, or larger bears, then not all of the bears have a equal chance of being counted. So if smaller, younger bears tend to eat the bait and hunters shoot older, larger bears, then the sampling would not be a fair sample. The population is not sampled equally.

The problem with animals is that they roam. It's not like counting jelly beans in a jar. In the case of the bears the sampling being done here can be thought of this way. Take a bag of white beans. Reach into the bag and grab a bunch. Count how many beans were grabbed. Now use a marker to place a mark on the beans. Throw all of the beans back into the bag and shake them up. Reach in and grab another handful. The beans are either marked or not marked. Count the two groups. This can be used to estimate how many beans are in the bag. The estimate can be improved by marking the beans and returning them to the bag and repeating the process.

A poor sampling protocol is one that ends up with a high variance. That means that the samples are very different from each other. If a sampling protocol can produce similar results each time, then the sampling yields a precise result. When the samples vary quite a bit the estimates are not precise. It is possible to adjust the sampling to improve the precision. A new innovation in sampling is called the proportionator.

Wikipedia article on the proportionator

Random sampling is important. That is an important way to make sure that all bears have an equal chance of being counted. Random sampling is also an important means of making sure that an estimate is unbiased. Unbiased means

that the mean tends to the true answer, i.e. an accurate answer. A precise estimate means that the estimates are close to each other, but not necessarily close to the correct answer.

[stereologist]

I didn't see that the DNR were being lazy. Making mistakes in sampling and estimation is easy. My favorite is this 6-sigma crap. The stated failure rates are true only for normal distributions, i.e. bell curves. In some data you could have 30,000 failures per million and still be 6-sigma compliant because the distribution is not a normal distribution.

[Wookietim]

I've seen those before. They look like photos of a chimpanzee to me.

To be honest, a wild Chimp in the PA woods would be about as remarkable as Bigfoot...

[+DieChecker]

A poor sampling protocol is one that ends up with a high variance. That means that the samples are very different from each other. If a sampling protocol can produce similar results each time, then the sampling yields a precise result. When the samples vary quite a bit the estimates are not precise. It is possible to adjust the sampling to improve the precision. A new innovation in sampling is called the proportionator.

OK. I think I understand what you are talking about now.

I didn't see that the DNR were being lazy. Making mistakes in sampling and estimation is easy. My favorite is this 6-sigma crap. The stated failure rates are true only for normal distributions, i.e. bell curves. In some data you could have 30,000 failures per million and still be 6-sigma compliant because the distribution is not a normal distribution.

I know what you mean about the 6-sigma. It seems to be a fad that everyone is crazy about everyone knowing it.

[stereologist]

I knew you'd get it DieChecker. In sampling things like land use, or organism, or cell distributions it turns out that systematic random sampling yields a lower variance than independent random sampling. In many saves the variance is proportional to 1/n instead of the 1/sqrt(n) for independent random sampling.

PS I worked on 64K memories back when they were new.

[KennyB]

Oh sigh, me oh my. Come the day when I browse the forums and see a post about the sightings of an actual Bigfoot and discover it can't be disproved or ridiculed. Sigh.

It will have to be a dead one and at least a dozen biologists and other 'scientific' types will all have to agree that it's really a bigfoot and even that might not be enough. Dedicated skeptics won't believe anything unless they do it themselves. By the way, Bigfeet aren't the only homonoids roaming the North American forests. There's also the Almas. KennyB

[Agent. Mulder]

It will have to be a dead one and at least a dozen biologists and other 'scientific' types will all have to agree that it's really a bigfoot and even that might not be enough. Dedicated skeptics won't believe anything unless they do it themselves. By the way, Bigfeet aren't the only homonoids roaming the North American forests. There's also the Almas. KennyB

well, the Almas are in mongolia. and russia area.

[Neognosis]

will have to be a dead one and at least a dozen biologists and other 'scientific' types will all have to agree that it's really a bigfoot and even that might not be enough. Dedicated skeptics won't believe anything unless they do it themselves. By the way, Bigfeet aren't the only homonoids roaming the North American forests. There's also the Almas. KennyB

That's quite a cynical overstatement. Most of us "skeptics" (I prefer to call myself "reality based") would settle for any evidence. But it has to be evidence. Not the silliness that some of you claim is evidence.

[~Onyx~]

It will have to be a dead one and at least a dozen biologists and other 'scientific' types will all have to agree that it's really a bigfoot and even that might not be enough. Dedicated skeptics won't believe anything unless they do it themselves. By the way, Bigfeet aren't the only homonoids roaming the North American forests. There's also the Almas. KennyB

I would settle for a dead one that was cleverly disguising itself as The Mongolian Death Worm, which would then make discovering it only TWICE as unlikely as being hit by lighting while hitching a ride on the wing of a 747, but hey, since we're theorizing and hypothesizing.

[Cryptid_Control]

I would settle for a dead one that was cleverly disguising itself as The Mongolian Death Worm, which would then make discovering it only TWICE as unlikely as being hit by lighting while hitching a ride on the wing of a 747, but hey, since we're theorizing and hypothesizing.

Huh..?

[Johnder]

I think the Jacobs thing was a juvenile Sasquatch based on its proportions only especially after seeing this

http://www.youtube.com/watch?v=cUTh2VFG8Z4

[+HerNibs]

I think the Jacobs thing was a juvenile Sasquatch based on its proportions only especially after seeing this

http://www.youtube.com/watch?v=cUTh2VFG8Z4

Bear with mange.

Nibs

[Johnder]

No I don't think so, it's more like the chimpanzee especially when the proportions were measured and they were almost identical.

Edited February 10, 2010 by Johnder

[stereologist]

No I don't think so, it's more like the chimpanzee especially when the proportions were measured and they were almost identical.

The measurement of proportions of a 3d object from a 2d image is difficult due to foreshortening issues. The error in using these images is likely in some cases to be huge.

[Johnder]

The measurement of proportions of a 3d object from a 2d image is difficult due to foreshortening issues. The error in using these images is likely in some cases to be huge.

Not when they take the pictures again within the same year before the tree the camera was in grows or the ground is covered with years of leaves. This case they did it within the first year so the sizes of the proportions are very accurate and they weren't of a bear. The engineer and scientists that worked on this knew what they were doing cause I seen this done on Monster Quest the other night; they did the same thing. It still might be an escaped chimp or a human but I have hopes that it wasn't. These Sasquatch stories we hear all the time just might be real.

[+HerNibs]

Not when they take the pictures again within the same year before the tree the camera was in grows or the ground is covered with years of leaves. This case they did it within the first year so the sizes of the proportions are very accurate and they weren't of a bear. The engineer and scientists that worked on this knew what they were doing cause I seen this done on Monster Quest the other night; they did the same thing. It still might be an escaped chimp or a human but I have hopes that it wasn't. These Sasquatch stories we hear all the time just might be real.

Are there pictures of the tree during the day, from the same distance, with something to use as a guide?

Nibs

[stereologist]

Not when they take the pictures again within the same year before the tree the camera was in grows or the ground is covered with years of leaves.

That actually has nothing to do with it. The problem is what was measured and what ratios were taken. The ratios taken from a 2d view of a 3d object change as the object rotates in front of the camera.

I thought you might have taken the measurements. But you say you saw the measurements taken. Can you explain what was measured?

[+DieChecker]

There is a youtube video that explains all the ratios and whatnot. I think all this talk of measurements comes from that.

http://www.youtube.com/watch?v=k2Z1I7u-rzw

[psyche101]

No I don't think so, it's more like the chimpanzee especially when the proportions were measured and they were almost identical.

I do not agree, the head is far larger, and the line has clearly been extended to exaggerate the back arch.

The example in this report (JOURNAL OF WILDLIFE DISEASES) shows a Bear of similar proportions.

I think the BFRO are digging themselves in deeper on this one. In their depicted superimposed example

What the heck is the proposed beast supposed to be doing? Scratching his head without using his hands? Mooning his Brother? hey mum check this out.....

You guys have seen the pics of what approached the cam traps before (same night) this alleged Juvenile Sasquatch was captured?

Check this out.

After the Bears left, a Juvenile Sasquastch shows up...............

Things that make you go HHHHRMMMZZZZZZZZZZZZZZZZZZ............

Edited February 10, 2010 by psyche101