|
Size: 1602
Comment: Notes
|
← Revision 7 as of 2025-11-06 21:34:51 ⇥
Size: 1834
Comment: Relation between F and T-squared dist
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 16: | Line 16: |
The [[Statistics/HotellingsTSquaredDistribution|T-squared distribution]] is a scaled version of the F distribution. The following are both valid formulations: {{attachment:ftotsquared1.svg}} {{attachment:ftotsquared2.svg}} |
F Distribution
The F distribution is a continuous probability distribution function that represents the ratio of variances between two chi-squared distributed random variables.
Description
The distribution is characterized by the degrees of freedom for each of the chi-squared distributed random variables. They are ordered/referred to as numerator and denominator degrees of freedom/notated as d1 and d2 respectively.
The square of a t distributed random variable has an F distribution with 1 numerator degree of freedom and the same denominator degrees of freedom as the original t distributed variable.
The T-squared distribution is a scaled version of the F distribution. The following are both valid formulations:
Moments
The first moment of the distribution is d2/(d2 - 2) for d2>2.
Usage
Probability Tests
The F distribution is almost exclusively used for test statistics: see the F test.
Furthermore, the test is usually formulated such that the denominator degrees of freedom are very large. As an example, for a very large denominator degrees of freedom and a significance level of 5%, the critical F statistics are:
- 3.84 for 1 numerator degree of freedom
- 3.00 for 2
- 2.6 for 3
- 2.37 for 4
- 2.21 for 5
- and so on
- 1.00 for very large numerator degrees of freedom