stats homework

[Flipx99 5,000+ posts]

I have homework in my stats class but I was too cheap to buy the book. Does anyone know how to determine the mean of errors?

Click to read more...

 

[tRiGgEr 5,000+ posts]

http://www.rit.edu/~uphysics/uncertainties/Uncertaintiespart1.html

//content.invisioncic.com/y282845/emoticons/confused.gif.e820e0216602db4765798ac39d28caa9.gif

 

[Flipx99 5,000+ posts]

hopefully that is correct. Thanks

 

[swimfreak26 10+ year member]

The mean of any set of numbers is just the sum of that set, divided by the quantity of numbers you have.

You're most likely going to have multiple errors in a system, and you basically just add them up and divide by the number of different errors you had. You'll notice in general, that some of the errors will essentially disappear because their size is insignificant in comparison to others.

Now, there is a slight issue that you may not be aware of though. The question MAY be asking about the uncertainty, which is different all together. The uncertainty in an addition/subtraction equation (used to find your final variable) is the uncertainty (error) in each, added together...unless the measurements are independent of one another (most cases), where now the uncertainty is equal to the sqrt(u1^2 + u2^2 + ...).

If you are asked to do this on examples where your variable is calculated by multiplication/division (i.e. p = mv) then you have to add the fractional uncertainties in quadrature and multiply by the calculated value of the calculated quantity. So (I will use "d" for the ucertainty symbol here)....

dp = p * sqrt( (dm/m)^2 + (dv/v)^2 )

I think I gave you way too much information, and probably the first line is all you need, however this stuff will come in later I guarantee it, very important in measurements that you make in a lab.

 

[sumone 10+ year member]

Are you looking for variance or standard deviation?

if variance, take the mean of all trials; call it M.

add the total of (M-nT)^2 where M is the mean you just found & nT is the value for trial "n"

divide this total by number of trials minus 1 (n-1, since I'm guessing this is an unbiased calculation)

that gives you the mean(average) variance

take the square root if you want standard deviation.

ex:

n1 = 1, n2=2, n2=3

M = (1+2+3)/3 = 6/3 = 2

(M-n1)^2 + (M-n2)^2 + (M-n3)^2 = (2-1)^2 + (2-2)^2 + (2-3)^2 = 1 + 0 + 1 = 2

divided by n-1 = 2 / (3 - 1) = 2/2 = 1

 

[chinny 5,000+ posts]

i hate numbers

 

[ToyTundra 10+ year member]

i hate numbers

I

I hate stats though

 

[Acidburn 5,000+ posts]

I
I hate stats though

x2+1000000

stats is the biggest load of bullshit evar

 

[AllStar1500bd 5,000+ posts]

x2+1000000
stats is the biggest load of bullshit evar

Stats 201 is the most pathetic college course I have ever been in. Class average on our last midterm was 68% //content.invisioncic.com/y282845/emoticons/laugh.gif.48439b2acf2cfca21620f01e7f77d1e4.gif My score was 91% //content.invisioncic.com/y282845/emoticons/biggrin.gif.d71a5d36fcbab170f2364c9f2e3946cb.gif

 

[ighettoboyi 10+ year member]

...or are you looking for STANDARD error?

Sd / sqrt 'n'

statistics is ok. the only math/science class i've gotten a solid A in

 

[swimfreak26 10+ year member]

...or are you looking for STANDARD error?
Sd / sqrt 'n'

Incase you didn't know what this meant, Sd is the standard deviation, and standard error is also known as the standard deviation of the mean, or rather, your uncertainty in your mean value.

Worth while equation, though, you should probably buy the book...never a bad choice IMHO.

Along with everyone, I hate stats as well, but these kind of things really aren't in depth statistics in my opinion...these are things you need if you're in any science field, to report data and to understand it.

P.S. I love numbers too //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif ........but, I hate differential calculus.