topic Re: Macro Jointpair for longitudinal data in Statistical Procedures

Macro Jointpair for longitudinal data

Rosa_Sh — Wed, 15 Jun 2016 01:41:18 GMT

Hi,
I want to run the macro Jointpair to fix a Random-Effects Model for Multiple Characteristics. The input data has three variables,19 units and 14 moments of observation (longitudinal data) but the results show errors associated to matrix operations. Share with you the output log. May someone knows how to save the error.
Thanks!

Re: Macro Jointpair for longitudinal data

Rick_SAS — Wed, 15 Jun 2016 14:10:02 GMT

The error says that the ESTIMATE variable is not in a data set that you are reading. Theerror is most likely occuring on Line 1754:

1748 /*---------------------------------------*/
1749 /* Construction of pair specific D */
1750 /*---------------------------------------*/
1751 /*Dpair(k)*/
1752
1753 use _covariantieD where (pair=&k);
1754 read all var{estimate pair r c};
1755 close _covariantieD;

A few lines later you get another error:

1776 use _covariantieerror where (pair=&k);
1777 read all var{estimate pair r c};
1778 close _covariantieerror;

WARNING: Data set WORK._COVARIANTIEERROR is empty.

There is no way for me to determine why these errors are occuring, but you might try inserting a PROC PRINT or PROC CONTENTS on Line 1732 to see what these files contain before the program enters PROC IML.

Re: Macro Jointpair for longitudinal data

Rosa_Sh — Thu, 16 Jun 2016 01:40:32 GMT

Hi, I´ll try to insert those sentences. Thanks for you advice!