Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
Linear Contrast
Message posted by Jeff P (via 208.37.105.21) on October 3, 2001 at 10:18 PM (ET)
Hi,
Can someone please explain to me why someone would use linear contrast? What is the value of using this procedure? What is the reasoning behind it? I know this is a complex question, and I can perform a linear contrast "by hand," but I really don't know why I'm performing the operation.
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: Linear Contrast
Message posted by Tomi (via 154.32.142.220) on October 5, 2001 at 4:40 PM (ET)
Are you referring to the use of linear contrast stretch in image processing or in some other context?I am only aware of its uses in image processing - as a method of adding depth.
Re: Linear Contrast
Message posted by Jack Tomsky (via 208.249.113.130) on October 5, 2001 at 6:27 PM (ET)
A linear contrast is a linear combination in which the sum of the weights is zero. As special cases, these include pairwise differences, the difference between the average of one subgroup from another, etc. The F test for a one-way ANOVA being statistically significant is algebraically equivalent to the existence of a statistically significant contrast. This property is used for multiple comparisons.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
