Dunnet test - Programming issue

[Maicon]

I'm having an issue when programming SAS to use Dunnet test: instead of use the treatment "-1" as reference for dunnet test, it is using the treatment "0". So it is comparing treatment "0" with each other treatment. Please, see the attachment with the SAS code and result. What can I do to make SAS use the treatment "-1" as reference?

In addition, the same are occurring in the contrast analysis: instead consider treatments 0, 1000, 2000, and 3000 excluding the treatment "-1", it is performing the contrast using "-1, 1000, 2000, and 3000" and excluding the treatment "0". 

The treatment -1 means a control treatment and did not macth with the other treatment levels. Such code programming usually work, but now I'm using SAS online and it did not work in the same way.

Thanks in advance.

[Reeza · Accepted solution]

Don't post your code as an image, use the icons in the forum to insert your code.

You need to modify your CLASS staement. You can set the reference level to first or last. If the level you need is not one of these, then you need to recode your data such that it is. 

http://support.sas.com/documentation/cdl/en/statug/68162/HTML/default/viewer.htm#statug_mixed_syntax...

---

@Maicon wrote:

I'm having an issue when programming SAS to use Dunnet test: instead of use the treatment "-1" as reference for dunnet test, it is using the treatment "0". So it is comparing treatment "0" with each other treatment. Please, see the attachment with the SAS code and result. What can I do to make SAS use the treatment "-1" as reference?

In addition, the same are occurring in the contrast analysis: instead consider treatments 0, 1000, 2000, and 3000 excluding the treatment "-1", it is performing the contrast using "-1, 1000, 2000, and 3000" and excluding the treatment "0". 

The treatment -1 means a control treatment and did not macth with the other treatment levels. Such code programming usually work, but now I'm using SAS online and it did not work in the same way.

Thanks in advance.

---