What inspired this Wish List Request?
This is just my personal suggestion, not from users.
Please describe the new or improved feature or solution you would like to see.
In JMP19, we can now calculate comparison within each sliced factor by "Multiple Comparison" command in Fit Model. It's called "Multiple Comparison", but it actually does not adjust all comparisons.
By using the glht function in R's multcomp libray, you can calculate adjusted p-value for general contrasts. In this method, the Monte-Carlo simulation explained in the following papers is used for calculating p-values and confidence intervals.
Alan Genz. Numerical computation of multivariate normal probabilities. Journal of Computational and Graphical Statistics, 1:141–149, 1992.
Alan Genz and Frank Bretz. Numerical computation of multivariate t-probabilities with application to power calculation of multiple contrasts. Journal of Statistical Computation and Simulation, 63:361–378, 1999.
Alan Genz and Frank Bretz. Methods for the computation of multivariate t-probabilities. Journal of Computational and Graphical Statistics, 11:950–971, 2002.
The following is an Example for the current JMP19 and R's multcomp.
<JMP19>
dt = New Table( "Test",
New Column( "x1", "Nominal", Values( [1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4] )),
New Column( "x2", "Nominal", Values( [1, 1, 2, 2, 1, 1, 2, 2, 1, 1, 2, 2, 1, 1, 2, 2] )),
New Column( "y", Values( [5, 8, 7, 4, 1, 4, 5, 1, 2, 3, 4, 5, 1, 2, 3, 1] ))
);
plat = dt << Fit Model(
Y( :y ), Effects( :x1, :x2, :x1 * :x2 ),
Personality( "Standard Least Squares" ), Emphasis( "Minimal Report" ),
Run(
:y << {
Multiple Comparisons(
Sliced Effect Estimates( Sliced Effect( :x1 * :x2 ), Slice Term List( :x1( "1" ) ) ),
Tukey HSD( 1 )
), Multiple Comparisons(
Sliced Effect Estimates( Sliced Effect( :x1 * :x2 ), Slice Term List( :x1( "2" ) )),
Tukey HSD( 1 )
), Multiple Comparisons(
Sliced Effect Estimates( Sliced Effect( :x1 * :x2 ), Slice Term List( :x1( "3" ) )),
Tukey HSD( 1 )
), Multiple Comparisons(
Sliced Effect Estimates( Sliced Effect( :x1 * :x2 ), Slice Term List( :x1( "4" ) )),
Tukey HSD( 1 )
)}
)
);
report(plat)[Outline Box("All Pairwise Differences")][Table Box(1)]<<Make Combined Data Table;
<R>
library(multcomp)
x1 <- c(1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4)
x2 <- c(1,1,2,2,1,1,2,2,1,1,2,2,1,1,2,2)
y <- c(5,8,7,4,1,4,5,1,2,3,4,5,1,2,3,1)
df <- data.frame( x1 = factor(x1), x2 = factor(x2), y = y)
rst <- lm(y ~ x1 + x2 + x1*x2, data = df)
summary(rst)
K <- matrix(0, 4,8)
K[1,5] = -1;
K[2,5] = -1;
K[3,5] = -1;
K[4,5] = -1;
K[2,6] = -1;
K[3,7] = -1;
K[4,8] = -1;
print(K)
rownames(K) <- c("Time=1", "Time=2", "Time=3", "Time=4")
summary(glht(rst, linfct = K))
<R's Result>
Linear Hypotheses:
Estimate Std. Error t value Pr(>|t|)
Time=1 == 0 1.000 1.768 0.566 0.962
Time=2 == 0 -0.500 1.768 -0.283 0.997
Time=3 == 0 -2.000 1.768 -1.131 0.705
Time=4 == 0 -0.500 1.768 -0.283 0.997
(Adjusted p values reported -- single-step method)
What would the value be to you and/or other users if this new feature is provided? In pharmaceutical industries, some analysts sometimes might want this kind of multiple comparisons.
Is this an external customer request?
No.
Is there a competitive software package that has this feature or solution that we should be aware of?
At least, the multcomp library in R can calculate adjusted p-values and confidence intervals for generel contrasts.
Will this impact a renewal, expansion, or adoption at a customer, and does this put an account at risk if we do not address this?
No.
