What is the "estimate" in the SolutionR output of the proc mixed. Is this the random slope?

[lgutm18]

Hi folks,

I am running this linear mixed-effect model, where I am outputting random effects using SolutionR:

ods  output SolutionR=BLUPs_idno;

proc mixed data=trial;

  class idno  sex;

  model cognition = time brainvol WMH FA icv AGE sex / s;

  random int time / subject=idno type=un solution;

  run;

When I look at the contents of SolutionR, I observe the following variables: 

Alphabetic List of Variables and Attributes

DF	Num	8	BEST4.
Effect	Char	9
Estimate	Num	8	D8.4
Probt	Num	8	PVALUE6.4	Pr > |t|
StdErrPred	Num	8	D8.4	Std Err Pred
idno	Num	8	BEST.	PARTICIPANT ID NUMBER
tValue	Num	8	7.2	t Value

I am interested in obtaining the Random Slope for time per individual (as specified in the model).

So, I am wondering if the "Estimate" value from this SolutionR, IS indeed the random slope per person?

OR

if the "Estimate" corresponds to the random errors, and in order to obtain the random slopes I should sum this value to the fixed effect I obtain in the Solution of Fixed effects. 

The mean i obtained for the estimate of the slope is 0 and the SD= 0.00631836. Which made me think I must add the value of the estimate for TIME as a fixed effect to get the random slope per person. Is this correct? OR is the "Estimate" directly the random slope per person?

Thanks!

[Ksharp]

You want this ?
https://support.sas.com/kb/37/109.html

[jiltao]

SOLUTIONR table shows the intercept and slope "deviation" from the overall intercept and slope, respectively. So you are correct -- you would need to add the fixed effect solutions (the overall intercept and slope) to the SOLUTIONR values to obtain the subject-specific intercept and slope. The usage note pointed out by @Ksharp provides some sample code for this.

https://support.sas.com/kb/37/109.html

Hope this helps,

Jill

[lgutm18]

Thank you very much for your answer. So, I am trying to emulate a measurement published in another article.  Per the methods section, this measurement is defined as "the random slope parameter estimated in the model" OR "the residual slope of Y; that is the person-specific deviation from the rate of decline expected at the marginal level". 

My intuition is that this refers to the "Estimate" measure provided by solutionR (BLUPs), per this other document as well: https://support.sas.com/documentation/cdl/en/statug/63962/HTML/default/viewer.htm#statug_mixed_sect0....

Or would the correct approach be to sum the fix effect as well, to obtain the total person-specific slope? If I add the fixed effect, will the interpretation still be "the person specific deviation from the expected rate of the decline at the marginal level". 

Thank you!!