Similar Function to Excel's T.Dist.2T()

Specific Domains Statistics
1

Seems straight forward but how would I get a p value if I have calculated the test statistic and have the degrees of freedom.

This example uses the method in Zar(1996) for comparing regression equations (Chapter 17)

the formula for the t-statistic is

t = (b1-b2) / (sb1-sb2)

where b1 and b2 is the calculated coefficient of two different lines here the values are:
b1=1.353712741
b2 = 1.249425897

and the errors for those coefficients are:
sb1 = 0.32542874
sb2 = 0.197671413

Using the formula for the t-statistic I get an answer of:
t = 8.480670589

alpha = 0.05
d.f.= 17

Essentially what I am looking for is a similar function to T.Dist.2T(test statistic, degrees of freedom) in Excel.

2

The Distributions package has a T distribution that you can use:

using Distributions

2 * ccdf(TDist(df), abs(t_value))
3

Thank you but what is ccdf and where does that come from?

4

From an Excel page:

image
image2306×961 90.2 KB

With Julia:

julia> using Distributions
julia> 2*ccdf(TDist(60),1.959999998)
0.054644929975920895

The ccdf is the complementary cumulative distribution function defined in the Distributions.jl package for the given distribution, which is TDist(60) or the T-distribution with 60 degrees of freedom.

5

I see that in the univariate distributions page now. Never would have thought to look for it there.

Thank you so much