Finding the p value

[bok11]

I am taking a statistics class, and i have some problems Understanding a formula.

The formula in question is like this:

p-value=2P(T>|t_obs,beta_i|)

I have calculated the t_obs,beta_i, but i do not understand what i am supposed to do in this next step. specifically, i have no idea what P is, and i suspect (T>|t_obs,beta_i|) should be a binary value? like 1 or 0?

 

[hlsmith]

Can you add any more content, such as the statistical test you are performing or a screenshot of the actual problem?

 

[bok11]

This is the method i should use

 

[bok11]

This is the task i am asked

 

[hlsmith]

Well it seems like you may be doing a two-sided t-test based on a regression coefficient. The 2P, is the multiplication of the p-value you had looked up on the t-distribution, times 2, since it is a two-sided hypothesis, meaning it could be different by being significantly more or less, no directionality in the difference just that it is different from the null.

 

[bok11]

Thanks for clarifying, it helps yet i do not understand how i do the computation at step 3.

 

[bok11]

What i dont understand:

What is T ?
how do i do the computation of (T> |t_obs,bi|)?

 

[hlsmith]

Did you calculate a test statistic in step 2? If so, what is it?

 

[bok11]

The calculation i did for step 2:


β_2=-0.0024044
β_0=4.073
σ_g=0.004414



t_(obs,β_i )=(β_2-β_0)/σ_g =-9557.107

 

[hlsmith]

Well that value can get converted into a p-value if you know your degrees of freedom. I believe it is usually n-1.

 

[bok11]

I know my degree of freedom to be 573.

 

[hlsmith]

Well you are lucking out, I am waiting to see if someone replies to my post, so I am in the cyber-neighborhood.


Look at this wonderful page I found just for you bok11:


http://www.cyclismo.org/tutorial/R/pValues.html


I didn't punch all of your numbers in but I quickly got the following based on your t-statistic:


5.565753e-11

Are you following all of your above steps and understanding what all of the values mean?

 

[hlsmith]

Sorry I got sloppy above and didn't insert all of the numbers. I believe it should be this:


2*pt(-abs(-9557.107),df=573)


Which you may need to force to show you more digits if desired or otherwise it is, p<0.0001.


With experience you will recognize that 9557 is a big number.