Fwd: Strange behaviour in LMM with Repeated measures

[Mark Webb-5]

I have a variable [Country - Nominal Code 1 or 2] that, for some reason, won't be accepted by the model.
It seems to clash with School [another variable] . If school is removed Country works.
I get a message in the Estimates table "This modified population mean is not estimatable".
What does this mean ?
In the Type III Fixed Effect Table I get Zero degrees of Freedom for Country - I would expect this to be 1.
Wherever Country is used - degrees of freedom drop to zero.

Selected output & my syntax pasted below.

Any suggestions ?
Regards

Model Dimension
		Number of Levels	Covariance Structure	Number of Parameters	Subject Variables	Number of Subjects
Fixed Effects	Intercept	1		1
	Time	2		1
	Country	2		1
	School	8		6
	AgeShort	5		4
	IncShort	5		4
	Time * Country	4		1
	Time * AgeShort	10		4
	Time * IncShort	10		4
	Time * School	16		6
Repeated Effects	Time	2	Heterogeneous Compound Symmetry	3	Id	255
Total		65		35
Type III Tests of Fixed Effectsa
Source	Numerator df	Denominator df	F	Sig.
Intercept	1	237.148	1548.222	.000
Time	1	236.401	.049	.825
Country	0	.	.	.
School	6	239.946	1.300	.258
AgeShort	4	239.369	1.626	.168
IncShort	4	238.379	.643	.632
Time * Country	0	.	.	.
Time * AgeShort	4	238.601	.978	.420
Time * IncShort	4	237.622	1.996	.096
Time * School	6	239.165	1.652	.134
Estimatesb
Country	Mean	Std. Error	df	95% Confidence Interval
				Lower Bound	Upper Bound
1 Swaziland	.a	.	.	.	.
2 Zambia	.a	.	.	.	.
.a This modified population mean is not estimable.

DATASET ACTIVATE DataSet1.
MIXED GenScale BY Time Country School AgeShort IncShort
  /CRITERIA=CIN(95) MXITER(100) MXSTEP(10) SCORING(1) SINGULAR(0.000000000001) HCONVERGE(0,
    ABSOLUTE) LCONVERGE(0, ABSOLUTE) PCONVERG(0.000001, ABSOLUTE)
  /FIXED=Time Country School AgeShort IncShort Time*Country Time*AgeShort Time*IncShort Time*School
    | SSTYPE(3)
  /METHOD=REML
  /REPEATED=Time  | SUBJECT(Id) COVTYPE(CSH)
  /EMMEANS=TABLES(Time) COMPARE ADJ(LSD)
  /EMMEANS=TABLES(Country) COMPARE ADJ(LSD)
  /EMMEANS=TABLES(AgeShort) COMPARE ADJ(LSD)
  /EMMEANS=TABLES(IncShort) COMPARE ADJ(LSD)
  /EMMEANS=TABLES(Time*Country)  COMPARE (time) ADJ(LSD)
  /EMMEANS=TABLES(Time*AgeShort)  COMPARE (time) ADJ(LSD)
  /EMMEANS=TABLES(Time*IncShort)  COMPARE (time) ADJ(LSD)
/EMMEANS=TABLES(Time*School)  COMPARE (time) ADJ(LSD).
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD