BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
Kristine14
Calcite | Level 5
I have run a correlation using the Spearman rank procedure, with the code below. I would like to use Bonferroni to correct the resulting p-values but am not sure how to integrate this into the code. Everything I have read applies Bonferroni to ANOVA.


proc corr data=three spearman;
var unanfl invas soprfl unprfl soprgr unprgr;
by trt;
run;
1 ACCEPTED SOLUTION

Accepted Solutions
lvm
Rhodochrosite | Level 12 lvm
Rhodochrosite | Level 12
You may want to seriously think of alternatives to Bonferroni if you have many correlations. This is because the Bonferroni method will result in very low power to detect nonzero correlations if you have many correlations. This approach actually sets the bound for the type I error rate, and is based on k independent tests. But the correlations are likely (highly) correlated with each other.

With a collection of k p-values (including from t tests of correlations), there are multiplicity adjustments that have a much higher power. These are not automatically done, but these are easily done using a data step or using PROC MULT with output from the CORR procedure. I suggest you read chapter 2 in the book Westfall, Tobias, Rom, Wolfinger, and Hochberg. 1999. Multiple Comparisons and Multiple Tests. (SAS Publishing). This shows how to do the analysis. You first need to decide on the adjustment and then follow the directions.

View solution in original post

2 REPLIES 2
Doc_Duke
Rhodochrosite | Level 12
Kristine,

The Bonferonni adjustment for multiple comparisons is easily done by the user (rather than the program). Just divide the alpha level by the number of tests under consideration and use that as the new alpha.

Since CORR does not put any *'s to indicate significance, that is really all there is to do.

Doc Muhlbaier
Duke
lvm
Rhodochrosite | Level 12 lvm
Rhodochrosite | Level 12
You may want to seriously think of alternatives to Bonferroni if you have many correlations. This is because the Bonferroni method will result in very low power to detect nonzero correlations if you have many correlations. This approach actually sets the bound for the type I error rate, and is based on k independent tests. But the correlations are likely (highly) correlated with each other.

With a collection of k p-values (including from t tests of correlations), there are multiplicity adjustments that have a much higher power. These are not automatically done, but these are easily done using a data step or using PROC MULT with output from the CORR procedure. I suggest you read chapter 2 in the book Westfall, Tobias, Rom, Wolfinger, and Hochberg. 1999. Multiple Comparisons and Multiple Tests. (SAS Publishing). This shows how to do the analysis. You first need to decide on the adjustment and then follow the directions.

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 2 replies
  • 9809 views
  • 1 like
  • 3 in conversation