Invalid Reference Value?

[CraigSmith]

Hi all,

Have attached screenshots and have copied code / outputs below - struggling to include basic analysis on smoking variable. Any idea how to fix this? 

Many thanks in advance,

Craig

Data dictionary is as follows;

smok

1093

Created

141

number

Smoking never/former/current

0

Never

1

Former

2

Current

99

Missing

Had formatted as follows:

PROC FORMAT;
VALUE SMOK
0 = 'Never smoker'
1 = 'Former smoker'
2 = 'Current smoker'
99 = 'missing'
;

And was using the following for a logistic regression 

proc logistic data= ARCAGE;
FORMAT SMOK SMOK. DRINK DRINK. SEX SEX.;
CLASS SMOK ( REF = '0: Never smoker')
DRINK (REF = '0: Never')
SEX ( REF = '2: Female' )
BMIC ( REF = '2' );

model CASE = AGE10 SEX SMOK SMOKP PROLEV DRINK ALCOIC BMIC /rocci RL SELECTION = F;
run;

1007 PROC FORMAT;
1008 VALUE SMOK
1009 0 = 'Never smoker'
1010 1 = 'Former smoker'
1011 2 = 'Current smoker'
1012 99 = 'missing'
1013 ;
NOTE: Format SMOK is already on the library WORK.FORMATS.
NOTE: Format SMOK has been output.

NOTE: PROCEDURE FORMAT used (Total process time):
real time 11.71 seconds
cpu time 1.07 seconds

1014 proc logistic data= ARCAGE;
1015 FORMAT SMOK SMOK. DRINK DRINK. SEX SEX.;
1016 CLASS SMOK ( REF = '0: Never smoker')
1017 DRINK (REF = '0: Never')
1018 SEX ( REF = '2: Female' )
1019 BMIC ( REF = '2' );
1020
1021 model CASE = AGE10 SEX SMOK SMOKP PROLEV DRINK ALCOIC BMIC /rocci RL SELECTION = F;
1022 run;

ERROR: Invalid reference value for SMOK.
NOTE: The SAS System stopped processing this step because of errors.
NOTE: There were 3969 observations read from the data set WORK.ARCAGE.
NOTE: PROCEDURE LOGISTIC used (Total process time):
real time 0.01 seconds
cpu time 0.00 seconds

[PaigeMiller · Accepted solution]

The formatted value is 'Never smoker' so you want

CLASS SMOK ( REF = 'Never smoker')

[CraigSmith]

Thanks so much Paige - got it working!

Many thanks,

Craig