lect6 - IE 410 Lecture 6: More on Multiple Comparisons...

Info iconThis preview shows pages 1–14. Sign up to view the full content.

View Full Document Right Arrow Icon
IE 410 Lecture 6: More on Multiple Comparisons Pairwise Comparisons. So far we have learned: FSD LSD Scheffe's Method As ways to test all pairs of means
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Next are methods based on the "Studentized Range" First, what is the Studentized range? Suppose we have " a " observations Y i ~ NID( μ i , σ 2 ) Let W = Max[Y i ] - Min[Y i ]
Background image of page 2
That is, W is the range of the sample Suppose we have an estimate of σ 2 (call it S 2 ), with v degrees of freedom, which is independent of the Y i 's. Let S be the square root of this estimate.
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Finally let q = W/S. q is the "Studentized Range". We will denote it as q( a ,v) to remind ourselves of the degrees of freedom
Background image of page 4
q(a,v) is a random variable. The larger " a " is, the larger we expect q( a ,v) to be. (Why)? The distribution of q(a,v) has been derived and tabulated in table VIII in the appendix
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Now in the case of pairwise comparisons we have y i. ~ NID( μ i , σ 2 /n) And MS E /n as an estimate of 2 /n We can thus find a Confidence Interval on: W = Max[ y i. ] - Min[ y i. ]
Background image of page 6
Under H 0 : all means are equal The (one-sided upper) CI is : T( α ) = q( α ,a,N-a) n MS E That is: Pr{Max( y i. ) - Min( y i. ) T( α )} = 1- α
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
TUKEY"S TEST Tukey's test says that any pair of treatment averages further apart than T( α ) are declared significantly different. Can think of it via the plot of treatment averages. Slide an interval of width T( α ) underneath to compare means.
Background image of page 8
y 3. y 4. y 1. y 2. * * * * T( α )
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
NEUMAN-KEULS TEST Example: Suppose we have 5 treatment averages STEP 1. We use q( α ,5,v) n MS E to look at the whole range.
Background image of page 10
If all averages in range, stop: * * * * * In this example, all means not different
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Otherwise , if not all averages in range, * * * * * Then STEP 2. Use q( α , 4 ,v) n MS E to compare all (both in this case) ranges of 4 points
Background image of page 12
* * * * If all points in range, stop . In this example, 1 and 5 are different,
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 14
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 08/06/2008 for the course IE 410 taught by Professor Storer during the Fall '04 term at Lehigh University .

Page1 / 48

lect6 - IE 410 Lecture 6: More on Multiple Comparisons...

This preview shows document pages 1 - 14. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online