LINES display is not produced

[wbba222]

Hello,

I'm trying to check affects between dependent variables before running tests on my data set but I keep getting an error message when trying to check with LSMEANS.

data v10r2timing;
input plot rep trt fung$ timing$ moisture tw yield;
cards;
205 2 1 none NTC 19.2 52.7 177.8531246
205 2 1 none NTC 19.2 52.7 177.8531246
403 4 1 none NTC 19.2 52.3 169.5848331
104 1 1 none NTC 19.2 53.3 156.6275595
312 3 1 none NTC 20.9 52.4 163.9054244
110 1 2 Delaro V10 21.2 51.4 188.2760082
402 4 2 Delaro V10 21.1 51.9 183.5223473
207 2 2 Delaro V10 19.1 53.5 163.3169461
308 3 2 Delaro V10 20.6 52.4 161.3609817
309 3 3 Triva V10 20.4 52.7 203.6156326
406 4 3 Triva V10 21.8 51 178.5306603
112 1 3 Triva V10 21.1 51.4 165.3998759
208 2 3 Triva V10 20.8 51.6 171.6888006
109 1 4 Head V10 20.5 52.4 177.9064594
311 3 4 Head V10 19.4 53.2 179.8935096
411 4 4 Head V10 18.6 53.2 174.0723452
204 2 4 Head V10 19.8 52.5 170.6010268
102 1 5 none NTC 19.8 52.1 164.6462049
303 3 5 none NTC 18.1 53.2 162.3858654
407 4 5 none NTC 19.3 53 165.5471783
212 2 5 none NTC 18.9 53.7 165.5048254
101 1 6 Delaro R2 19.6 51.6 171.6519377
401 4 6 Delaro R2 19.5 51.6 209.7455212
202 2 6 Delaro R2 19.7 51.7 178.8284676
304 3 6 Delaro R2 19.9 52.8 163.7364178
105 1 7 Triva R2 19.8 52.1 182.225473
305 3 7 Triva R2 20.1 52.4 193.138672
210 2 7 Triva R2 20.8 51.8 176.9195559
412 4 7 Triva R2 20.8 51.6 165.8828991
209 2 8 Head R2 18.9 52.7 182.394198
302 3 8 Head R2 20.5 52.3 177.0418856
111 1 8 Head R2 19.4 52.4 173.3306563
408 4 8 Head R2 19.4 52.6 165.329749
405 4 9 none NTC 19.7 52.7 197.8860052
106 1 9 none NTC 20 51.5 181.034741
211 2 9 none NTC 18.7 52.8 159.1446537
307 3 9 none NTC 19.4 53 161.5618535
301 3 10 Delaro v10r2 20.3 51.7 200.4840663
206 2 10 Delaro v10r2 20.8 51 180.3101577
103 1 10 Delaro v10r2 20.1 52.3 158.5048153
404 4 10 Delaro v10r2 22.4 51 171.4601391
409 4 11 Triva v10r2 21.2 51.4 192.3629617
306 3 11 Triva v10r2 21.8 50.3 191.9365811
203 2 11 Triva v10r2 20.5 51.7 176.9360711
108 1 11 Triva v10r2 21.7 51.4 170.1238565
201 2 12 Head v10r2 20.3 51.5 181.6381435
310 3 12 Head v10r2 20.8 50.7 199.8918381
410 4 12 Head v10r2 22.6 51.3 169.4745887
107 1 12 Head v10r2 20.4 52.3 169.8077339

;

proc glimmix data = v10r2timing;
class fung timing;
model tw = fung timing fung*timing;
random rep;
lsmeans fung/lines;
lsmeans timing/lines;
lsmeans fung*timing/lines;
run;

proc glimmix data = v10r2timing;
class fung timing;
model moisture = fung timing fung*timing;
random rep;
lsmeans fung/lines;
lsmeans timing/lines;
lsmeans fung*timing/lines;
run;

proc glimmix data = v10r2timing;
class fung timing;
model yield = fung timing fung*timing;
random rep;
lsmeans fung/lines;
lsmeans timing/lines;
lsmeans fung*timing/lines;
run;

NOTE: Convergence criterion (ABSGCONV=0.00001) satisfied.
NOTE: Estimated G matrix is not positive definite.
ERROR: LINES display is not produced for the fung effect because p-values are not available for all
least-squares means differences.
ERROR: LINES display is not produced for the timing effect because p-values are not available for
all least-squares means differences.
NOTE: PROCEDURE GLIMMIX used (Total process time):
real time 0.36 seconds
cpu time 0.03 seconds

172
173 proc glimmix data = v10r2timing;
174 class fung timing;
175 model moisture = fung timing fung*timing;
176 random rep;
177 lsmeans fung/lines;
178 lsmeans timing/lines;
179 lsmeans fung*timing/lines;
180 run;

NOTE: Convergence criterion (ABSGCONV=0.00001) satisfied.
ERROR: LINES display is not produced for the fung effect because p-values are not available for all
least-squares means differences.
ERROR: LINES display is not produced for the timing effect because p-values are not available for
all least-squares means differences.
NOTE: PROCEDURE GLIMMIX used (Total process time):
real time 0.11 seconds
cpu time 0.03 seconds

181
182 proc glimmix data = v10r2timing;
183 class fung timing;
184 model yield = fung timing fung*timing;
185 random rep;
186 lsmeans fung/lines;
187 lsmeans timing/lines;
188 lsmeans fung*timing/lines;
189 run;

NOTE: Convergence criterion (ABSGCONV=0.00001) satisfied.
ERROR: LINES display is not produced for the fung effect because p-values are not available for all
least-squares means differences.
ERROR: LINES display is not produced for the timing effect because p-values are not available for
all least-squares means differences.
NOTE: PROCEDURE GLIMMIX used (Total process time):
real time 0.10 seconds
cpu time 0.07 seconds

[PaigeMiller]

You cannot estimate all the terms in your model. This is not a SAS restriction, the underlying math does not allow Least Squares Means to be estimated from your data. You can see this here:

This "Non-est" usually happens when you have cells in your design with zero data points in a cell. PROC FREQ will identify those cells with zero data points.

Based on the above, you could remove timing='NTC' and fung='none' from your data, and then you will get results. Whether or not that is what you want is up to you.

proc glimmix data = v10r2timing(where=(timing ^= 'NTC' and fung ^='none'));

[SteveDenham]

Alternatively, you could fit a MODEL statement with only the interaction term, and use LSMESTIMATE statements to derive the main effect marginal means and standard errors, and with a JOINT option, an F test. That would eliminate the issues seen, but carries the penalty of you having to do the LINES equivalent in a post-processing step.

SteveDenham