I’ve been thinking a lot lately about the possibility of running a multiple regression analysis in order to help stratify certain aspects of bigfoot behavior. This won’t prove or disprove anything but it could turn out to be a useful research tool.

First, here’s a little information concerning the definition of a multiple regression analysis;

regression analysis

n : the use of regression to make quantitative predictions of one variable from the values of another

also

A statistical technique used to find relationships between variables for the purpose of predicting future values.

The general purpose of multiple regression is to learn more about the relationship between several independent or predictor variables and a dependent or criterion variable. For example, a bigfoot researcher might record for each sighting the size of the bigfoot if witnessed, the size of any footprint found, the time of day or night that the sighting occurred, any limb twists/limb formations found, the season, the weather conditions, the geographic location, and a subjective rating of observed behavior. Once this information has been compiled for various sightings it would be interesting to see whether and how these measures relate. For example, it has been theorized that the bigfoot in the southern portion of the U.S. seem to be more aggressive than those in the Pacific Northwest. A multiple regression analysis, while not being able to prove or disprove this theory, can help correlate a geographic layout that will either shore up, or dilute the theory.

This is actually a very simplified example in that certain factors would have to be applied (observed behavior) in order to increase the reliability of the analysis. Also, it will be important to attempt to guarantee a good representative sample. Otherwise, the degree of error can become so great as to render the analysis inconclusive.

This will prove to be a daunting task. The compilation of data alone will be the hardest part of this project. There are literally thousands of reports that will have to be gone over to obtain the necessary information. Also, a template will have to be developed in order to correctly enter the data. Once that is all completed however, the analysis itself is really no more than a few keystrokes away.

I’m still trying to sort out the best way to approach this. I’ve got a friend who’s a statistics professor here at the local community college. I’m going to run these thoughts by him and see if he can get me started in the right direction.

If anyone else out there has any statistical background and you’d like to add to this (or better yet, point out any obvious errors) please fire it on in! As I said, I’m not an expert.

AND…if anyone is aware of a multiple regression analysis that’s already been done, please post a link here. This is one wheel I’d hate to reinvent.

sounds like quite a project fishy .what factors are you going to use to try to determine your findings?mostly agression vs. non-agression or are other factors going to be involved?(you've heard of the answer man,I'm the question man )is the data going to be used area specific?i.e. regional,state or otherwise.or are you going to use a broader template?and know the only thing I can remeber ever being noticed by any of the serious investigators is that miss patty's track was originally discovered elsewhere,five years earlierhows that for an oddity?good luck.

That's one of the main stumbling blocks I'm hitting right now, Jim. Aggressiveness would have to be a factored response. So, I'd have to assign a value to aggressive behavior as well as passive behavior. 0 would be the baseline value, meaning that no visible behavior at all was observed, although that would be rare. Then I have to determine how to array aggressive and non-aggressive behavior to define the factors.

Example:

minor vocalizations (grunts, growls) could be assigned a factor of 1

thrashing through undergrowth and rattling bushes could be assigned a factor of 2

screams could be assigned a factor of 3

thrown rocks and limbs could be assigned a factor of 4

and finally, bluff charges could be assigned a factor of 5

This is a very rough example in that the factors would need to be much better defined.

My hope is that once this data is compiled, it will be applicable to very small geographic areas as well as extremely broad geographic areas provided that there is enough representative sighting data. Like I said, "daunting"!

Personally Fishy I would swap your 2 and 3 around. Screams can be construed as a warning, but they can also just be from fear of the intruder. I think thrashing trees shows more aggressiveness than does screaming.

Just trying to help! :wink:
This sounds like a long project that is going to be a lot of hard work. If I can help in any way let me know. I'm sure you will do fine...

Thank you very much for your observation, Arkansan. And you're absolutely right. Just laying the foundation is going to be a long row to hoe and I'll need all the help I can get from researchers such as yourself.

Your insight is greatly appreciated. Also, before I would ever attempt to go ahead with a MRA, I'll post all the data here for critiquing first.

wow fishy,I knew you said it would be "daunting"but I think it may go beyond that.hopefully it will show some sort of behavior pattern that can be used to aid in searching for our hairy friends,likewise with arky let me know what you need and I'll do whatever I can to get it done.

Hi all

My main concern in a project like this one is to determine excatly what you are identifing ie. Learned behavior vs. Reactionary (primal) behavior.

Sasquatch sightings are GEOGRAPHIC they happen in the same area. That alone is a huge project to gather these. I plot all the ones i find on a map and then go out to investigate. But what is important is to kind out what its range is how far will it move in one day. But this technique is flawed if you think that if you are gonna just go up there and expect to see a Sasquatch. the forest up here in the Pacific northwest is really thick, you gotta get out and walk down the trail because it you stay at the road you have 5 -10 ft visablity.


Food is of the utmost importance to a sasquatch. You gotta have a lot of calories to move around a 800 lbs plus animal up a mountain side.



We need to try to identify what are the primal behaviors of this ape.

1. Movement (surveying trails )
Remote sensing maps: i.e if they NASA can make a map of the geographic features of the earth and a map of the reflectivity of the different types of vegetaion at different levels from the forest floor from the ground you would think that if the tree breaking is marking a territorial boundry; over time would generate a swath of secondary canopy in the forest (old growth) and around food sources giving its generalized territory.

2. Seasonal and regional food sources ( Fruits and vegetables, water sources, and tidal zone foraging)


These are some important factors to take into consideration.

thanks
ken

Ken, thanks for your input but please allow me to clarify a couple of things.



Behavior is only one small aspect. As I defined "regression analysis" in my OP, the use of regression to make quantitative predictions of one variable from the values of another, notice that what I'm proposing is a multiple regression analysis. If implied behavior (whether learned or reactionary) were the only question I was concerned with answering, then a straight linear regression analysis would suffice. (Even though that endeavour would prove to be a complete waste of time since behavior is one of the least known things about bigfoot.)



The same area in relation to what? Climate? Topography? Major water sources? Does one variable have an overriding effect, or is it a combination of several variables? Do these variables require a factor in order to better weigh their influence? Again, this is another reason for attempting the analysis. Not to gain a definite answer but to be better able to make more educated inferences in relation to the data.



I'm not sure if you're referring to your technique or the analysis I'm proposing here. But, if you're referring to my proposal, not only would it be flawed to think the analysis would put you on top of a sasquatch, it would be patently absurd. This is not something that will locate a sasquatch for you. My hopes are that it will help you in determining the where and the when in order for you to better plan a field investigation. The best it could ever be is still an "implied inference" with many unknowns thrown in for good measure. My main concern is being able to do what is possible to minimize the degree of error. That's going to be the bear!

Also, my plan is to use the BFRO sightings database for the purpose of this analysis. IMHO, their site has the most credible and scientific reports available. Eventually, once the initial data has been entered from the BFRO's site, I'll consider incorporating sighting reports from other websites that I deem credible. (Provided I'm not old, gray and senile by the time the initial analysis is complete. :wink:)



That's a good thought but I don't believe there's much of a chance of finding that info. First, there's been to much clear-cutting and reforestation that's gone on across the country over the years. Second, if you could find "troughs" through the forest growth with satellite imagery, how could you determine whether this swath had been created by a sasquatch or a tornado?

One more thing. Keep in mind that this is not a site specific study. There have been bigfoot sightings all over the continental United States as well as Canada. And it is my hope that once this is accomplished, you'll be able to make as much use of it in the PNW as I can make with it here in the southeastern U.S.

We'll see.

Fish

Oh! Welcome to the BFF! Good to have you on board.

Anything ever come of this?

I would be interested in knowing if there's a correlation between month of the year, the state, and # of sightings. Although I guess it would be skewed a bit since there are generally less hikers in the winter and some states have harsher winters than others.

But, it would be interesting to know if June is the most active month for Washington and November the most active for Texas.

I haven't had the opportunity to do anything other than a rather simple spreadsheet for Florida.

I haven't forgotten about this though. It's just on the backburner for the time being.

One problem is that I really need a good statistical software package in order to correlate the data across the board. And you are correct in that the data will have a certain degree of error and that will also have to be computed.

I've been looking into possibly purchasing SPSS as it's a fairly good package and should handle the job. I'm hoping that maybe Santa will keep that in mind for my stocking.

But I'm not holding my breath.

No you can't do that. You are a mammal after all. Nice ideas Fish.
Also: Good luck with Santa

Ok Fishbone, you got me....what's SPSS?

You can run a simple regression analysis on Excel or QuatroPro.

Lisa, SPSS is a statistical program that allows you to run in-depth, detailed analyisis' using multiple variables, such as all the different aspects that may have been noted during a sighting. It's mondo cool and fairly affordable. (Though I'd prolly rather buy some spiffy night vision first) . Here's a link to their site. SPSS

Chronic, you're right. Excel would work to run a simple linear regression analysis but it can't handle a multiple regression analysis, which is what you'd really want/need for the analysis to possibly be of any real use. Too bad there's no way to tweak Excel.

Thanks for asking that question Lisa!

I was too embarassed to do it.

The man definitely knows his stuff!



The only SPSS I could think of was the sound my date makes when she's feeling a little low...

Well I'm glad you ask to, because I was wanting to know also.

SPSS Statistical Package for the Social Sciences

Just out of curiosity has anyone done a statistical analysis of tracks collected and sightings of the animal?

This is whole purpose of my database thing. Only I would like to see if predictions as to where a sighting might occur could be generated by the stats.

I also thing the credibility factor of each report could clog up the works too.

Oh well!

Fish:

I realize this ia an old thread, but it popped up again...



I seem to remember John Green saying that Sasquatch seemed to favor areas with over a certain amount of annual rainfall.

They may be related to us Webfoots after all.

wow thats quite an undertaking fish. too bad there's nothing like seti@home for bigfoot, altho i once saw a yeti@home spoof.

i keep wondering about ... how do i word this?... do the females show themselves to men more often than to women, and do the males show themselves to women more often than to men.

anyway, best of luck with all that data entry. what a pain. i'd help if i could

gael

Hey , this guy Jim Keegan has been working on this mapping thing...check it out here:

http://www.geocities.com/Area51/Matrix/4867/us.html

very interesting stuff

-Janice

Gee, I guess Keegan just had a "gut feeling" that those
sightings in the great lakes were rabbits or "flash sightings"?

Or maybe they didn't support his hypothesis

Interesting, but I'd like to see more of the data that was used.

Not sure if Fishbone35 is still checking this link? I am curious to see if you have taken this regression study further? I am new to the forum and was considering doing a geographical relationship study in GIS. I'm presently searching for the best data available, and hoping you could give me some advice? I am specifically looking for any data(excell/Access or similar) showing X/Y coordinates, etc.. Thanks Nick

Hi Nick. Welcome to the BFF!

Unfortunately, I have yet to take up this task in earnest. I do however have an Excel spreadsheet with all the BFRO Florida sightings on the public database. It hasn't been updated lately so there are a few reports that have been added that aren't on the spreadsheet.

PM me your email addy if you like and I'll send it to you. As far as X,Y coordinates it's really hit or miss according to how much information was submitted with the reports.

Maybe it will prove to be of some help.

Tim

Now THIS is why this forum really shines. To heck with
the bickering and personal sniping. Kudos Nick and Fish
for getting on with something worthwhile, best of luck in
the endeavor.