Emrah Soner Özdeş ☆ Turkey, 2015-01-16 15:52 (3582 d 09:48 ago) Posting: # 14299 Views: 7,633 |
|
Hi All; I normally conclude BE with WNL and R. Yet my manager is big fan of SAS and wants the BE results as SAS output as well. I have the code for ANOVA models as described in Guidances and able to obtain 90% Confidence Limits for differences with LSMEANS as you may see below. As a final step LSMean(i)-LSMean(j) values should be antilogged and multiplied with 100 to obtain upper and lower limits. What I wonder is how may I code this last step so that it can be included in the SAS Output. Any help will be appreciated. Regards |
d_labes ★★★ Berlin, Germany, 2015-01-19 09:20 (3579 d 16:20 ago) @ Emrah Soner Özdeş Posting: # 14301 Views: 6,076 |
|
Dear Emrah! ❝ As a final step LSMean(i)-LSMean(j) values should be antilogged and multiplied with 100 to obtain upper and lower limits. ❝ ❝ What I wonder is how may I code this last step so that it can be included in the SAS Output. Any help will be appreciated. The short answer: Ask your big SAS fan . To be serious: I suppose you are using Proc GLM .Place an ODS output statement before or within the Proc GLM code.Something like: ODS output LSMeanDiffCL=whatever_you_like_as_name; Take the dataset whatever_you_like_as_name and perform the back-transformation.The relevant varibles are diff, lowerCL, upperCL if I recall correctly. Take care of the order of treatm the diff and its CI will be calculated and change sign if necessary.A simple Proc Print call will show you the results.BTW: How did you arrive at T-R (i=2, j=1)? Mine is always R-T if treatm is coded that way due to lexical ordering.Hope this helps — Regards, Detlew |
Emrah Soner Özdeş ☆ Turkey, 2015-01-20 14:12 (3578 d 11:28 ago) @ d_labes Posting: # 14305 Views: 6,005 |
|
Hey Detlew; Thanks for your help. I finally made it by searching through some SAS tutorials. By ODS tracing I obtained the label of table, created a data set, and performed some math on the parameters and finally printed. ❝ BTW: How did you arrive at T-R (i=2, j=1)? Mine is always R-T if As I said SAS is not my favorite thing to work with. Yet, here is the PROC GLM of my code. Hope this helps: Regards Emrah |
jishnu ☆ India, 2022-08-06 21:02 (823 d 05:38 ago) @ Emrah Soner Özdeş Posting: # 23205 Views: 2,095 |
|
Hi Emrah, have some doubts. please message me on +91 9074206778. regards, Jishnu prasad Mumbai, India. Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post #5! [Helmut] |