proc glimmix multinomial syntax error

[mlien89]

I am relatively new to performing glimmix procedures, and I am wanting to carry out a multinomial analysis for the following dataset, and perform a power analysis. I am receiving a syntax error shown below. I've attached my code for reference. any debugging would be appreciated!

data idc;
input line $ seeding @@;
fields=15;
n=40;
do idc='none','mild','mod','sevr';
	input p @@;
	do field_id=1 to fields;
        ptst=((n*p)/100);
		mu=round(ptst,1.0);
		output;
	end;
end;
datalines;
S 25 5 17 25 53 
S 50 8 13 32 48
S 150 14 22 19 45
R 25 8 60 21 11 
R 50 15 64 14 7 
R 150 18 70 6 6 
;
proc print;
run;
proc glimmix data=idc;
class line seeding field_id;
freq mu;
model idc (order=data)=line|seeding/link=cumlogit htype=1,3 dist=multinomial s;
random intercept/field_id;
ods output tests3=a;
parms (.12)/hold=1;
  estimate 'No IDC line=S seeding=25' intercept 1 line 0 1 seeding 1 0 0 line*seeding 0 0 0 1 0 0,
           'No IDC line=S seeding=50' intercept 1 line 0 1 seeding 0 1 0 line*seeding 0 0 0 0 1 0, 
           'No IDC line=S seeding=150' intercept 1 line 0 1 seeding 0 0 1 line*seeding 0 0 0 0 0 1, 
           'No IDC line=R seeding=25' intercept 1 line 1 0 seeding 1 0 0 line*seeding 1 0 0 0 0 0,
           'No IDC line=R seeding=50' intercept 1 line 1 0 seeding 0 1 0 line*seeding 0 1 0 0 0 0, 
           'No IDC line=R seeding=150' intercept 1 line 1 0 seeding 0 0 1 line*seeding 0 0 1 0 0 0, 
           / ilink;
  estimate 'No+Mild IDC line=S seeding=25' intercept 0 1 line 0 1 seeding 1 0 0 line*seeding 0 0 0 1 0 0,
           'No+Mild IDC line=S seeding=50' intercept 0 1 line 0 1 seeding 0 1 0 line*seeding 0 0 0 0 1 0, 
           'No+Mild IDC line=S seeding=150' intercept 0 1 line 0 1 seeding 0 0 1 line*seeding 0 0 0 0 0 1, 
           'No+Mild IDC line=R seeding=25' intercept 0 1 line 1 0 seeding 1 0 0 line*seeding 1 0 0 0 0 0,
           'No+Mild IDC line=R seeding=50' intercept 0 1 line 1 0 seeding 0 1 0 line*seeding 0 1 0 0 0 0, 
           'No+Mild IDC line=R seeding=150' intercept 0 1 line 1 0 seeding 0 0 1 line*seeding 0 0 1 0 0 0, 
           / ilink;
  estimate 'yg1+2+4 t=40' intercept 0 0 1 trt 1 0,
           'yg1+2+4 t=80' intercept 0 0 1 trt 0 1 0, 
           'yg1+2+4 t=100' intercept 0 0 1 trt 0 0 1 0,
           'yg1+2+4 t=120' intercept 0 0 1 trt 0 0 0 1 0,
           'yg1+2+4 t=160' intercept 0 0 1 trt 0 0 0 0 1 / ilink;
  estimate 'yg1+2+4+5 t=40' intercept 0 0 0 1 trt 1 0,
           'yg1+2+4+5 t=80' intercept 0 0 0 1 trt 0 1 0, 
           'yg1+2+4+5 t=100' intercept 0 0 0 1 trt 0 0 1 0,
           'yg1+2+4+5 t=120' intercept 0 0 0 1 trt 0 0 0 1 0,
           'yg1+2+4+5 t=160' intercept 0 0 0 1 trt 0 0 0 0 1 / ilink;
run;
data power;
set a;
alpha=0.05;
fcrit=finv(1-alpha, numdf, dendf, 0);
power=1-probf(fcrit, numdf, dendf, numdf*fvalue);
run;
proc print;
run;  

[sbxkoenk]

Hello,

[ I moved this post to Statistical Procedures board ]

Without even studying your model, I can already say you cannot specify a variable-name as an option behind the forward slash (/) in the RANDOM statement.

If this variable of yours is identifying the subjects in the model, then specify :

SUBJECT= (Identifies the subjects in the model)

followed by your variable name.

Koen

[SteveDenham]

I recommend looking into using LSMESTIMATE statements rather than ESTIMATE statements - the syntax is much more compact and intuitive.

SteveDenham