Confirmatory Factor Analysis Error, Proc Calis, message file not found - how to fix error

[kjwainwright]

The SAS version I'm using is 2016, 9.4 (TS1M6), SAS/STAT 15.1

I am doing an exploratory factor analysis and confirmatory factor analysis with 16 binary variables and keep getting this error for the confirmatory factor analysis (proc calis):

NOTE: Convergence criterion (GCONV=1E-8) satisfied.
ERROR: Message file "t0l8en" is not found.
NOTE: The data set MUR.OSTAT has 79 observations and 23 variables.
NOTE: PROCEDURE CALIS used (Total process time):
real time 2.61 seconds
cpu time 0.15 seconds

Any ideas how to address this error? I just can't find anything on it. 

Just in case it helps, here's my code. This is one way I ran it, I ran it a second way and had the same error with some additional warnings (below):

Proc calis data=mur.ms_medical_history3 outstat=mur.ostat;
lineqs
A1ATTENTIONMEMORYrc = p1 fONE + e1,
A1DIZZINESsrc=p2 fONE + e2,
A1DOUBLEVISIONrc = p3 fONE + e3,
A1EYEDISCOMFORTrc = p4 fONE + e4,
A1FATIGUErc = p5 fONE + e5,
A1SLURREDSPEECHrc = p6 fONE + e6,
A1WALKINGrc = p7 fTWO + e7,
A1MOVINGLIMBsrc=p8 fTWO + e8,
A1LOSSBALANCErc = p9 fTWO + e9,
A1COORDINATIONrc = p10 fTWO + e10,
A1WEAKNESSLIMBsrc=p11 fTWO + e11,
A1MUSCLESPASMsrc=p12 fTHR + e12,
A1PAINMUSCLESPASMsrc=p13 fTHR + e13,
A1NUMBNESsrc=p14 fFOU + e14,
A1TINGLINGLIMBsrc=p15 fFOU + e15,
A1VISIONLOSsrc=p16 fFOU + e16;
STD
e1-e16 = vare1-vare16,
fONE fTWO fTHR fFOU = 1 1 1 1;
COV
fONE fTWO = COV12,
fONE fTHR = COV13,
fONE fFOU = COV14,
fTWO fTHR = COV23,
fTWO fFOU = COV24,
fTHR fFOU = COV34;
pathdiagram diagram=standard useerr;
Run;

Here is another way I ran it where I get the same error, with some additional warnings:

From the log: 

NOTE: A utility disk data set is being used.
NOTE: Convergence criterion (GCONV=1E-8) satisfied.
WARNING: The estimated error variance for variable A1MUSCLESPASMSrc is negative.
WARNING: Although all predicted variances for the observed variables and latent factors are
positive, the corresponding predicted covariance matrix is not positive definite.
It has one negative eigenvalue.
ERROR: Message file "t0l8en" is not found.
NOTE: The data set MUR.OSTAT has 97 observations and 23 variables.
NOTE: PROCEDURE CALIS used (Total process time):
real time 2.50 seconds
cpu time 0.25 seconds

The other way I coded it: 

Proc calis data=mur.ms_medical_history3 method=wls modification outstat=mur.ostat;
Factor
ONE ===> A1ATTENTIONMEMORYrc A1DIZZINESSrc A1DOUBLEVISIONrc A1EYEDISCOMFORTrc A1FATIGUErc A1SLURREDSPEECHrc,
TWO ===> A1WALKINGrc A1MOVINGLIMBSrc A1LOSSBALANCErc A1COORDINATIONrc A1WEAKNESSLIMBSrc,
THR ===> A1MUSCLESPASMSrc A1PAINMUSCLESPASMSrc,
FOU ===> A1NUMBNESSrc A1TINGLINGLIMBSrc A1VISIONLOSSrc;
Pvar
ONE = 1.,
TWO = 1.,
THR = 1.,
FOU = 1.;
pathdiagram diagram=standard useerr;
Run;

Any ideas? 

[Rick_SAS]

I alerted a SAS developer who works on PROC CALIS. He says that the missing message is from an optimization routine that CALIS calls. He believes that if CALIS runs, you can probably ignore the error about not finding the message file. It does not seem to be an error about the optimization or modeling fitting itself. The routine might be trying to log a NOTE but for some reason cannot resolve the message.

If you are not comfortable with this answer, then submit your program and data to Technical Support. If the problem is data dependent, that will enable the developer to discover which message is not available.

It is also possible that you have a corrupted message file, which would be a system/installation problem. Technical Support can also advise you on how to resolve that.