Prev: Converting a SPSS file to SAS with Value and Variable labels
Next: SCAN function
From: TSA on 26 Jul 2006 19:06
After running my program, I received the messages that "74
observations are not included because of missing values" in the log
file.

Based on the file that I imported, there should only be 43 missing
values (denoted with a ".").

Is there anyway to get SAS to show me what values are being considered
"missing"?

Thanks! I look forward to your responses.

Tonya

From: Eash on 27 Jul 2006 23:45
Do a PROC FREQ on the column with a /missing option - will give you the
missing value frequencies.

for eg:

data test;
infile cards truncover;
input a ;
cards;
1
2
..
..

1
2
;
run;

proc freq data=test;
tables a/missing;
run;

Output:

a Frequency Percent
----------------------------------------------
. 3 42.86
1 2 28.57
2 2 28.57

Note :

when you read numeric variables, the Blanks as well as .'s will be
considered as Missing values, as shown in the example..that may be a
reason for you count mismatch on the missing values.

Thanks
EasH

TSA wrote:

> After running my program, I received the messages that "74
> observations are not included because of missing values" in the log
> file.
>
> Based on the file that I imported, there should only be 43 missing
> values (denoted with a ".").
>
> Is there anyway to get SAS to show me what values are being considered
> "missing"?
>
> Thanks! I look forward to your responses.
>
> Tonya

From: TSA on 28 Jul 2006 12:54
Thanks for your suggestion. The proc freq shows only 44 missing values
as well. Could missing values for other variables be causing other
observation to excluded from the analysis?


Eash wrote:
> Do a PROC FREQ on the column with a /missing option - will give you the
> missing value frequencies.
>
> for eg:
>
> data test;
> infile cards truncover;
> input a ;
> cards;
> 1
> 2
> .
> .
>
> 1
> 2
> ;
> run;
>
> proc freq data=test;
> tables a/missing;
> run;
>
> Output:
>
> a Frequency Percent
> ----------------------------------------------
> . 3 42.86
> 1 2 28.57
> 2 2 28.57
>
> Note :
>
> when you read numeric variables, the Blanks as well as .'s will be
> considered as Missing values, as shown in the example..that may be a
> reason for you count mismatch on the missing values.
>
> Thanks
> EasH
>
> TSA wrote:
>
> > After running my program, I received the messages that "74
> > observations are not included because of missing values" in the log
> > file.
> >
> > Based on the file that I imported, there should only be 43 missing
> > values (denoted with a ".").
> >
> > Is there anyway to get SAS to show me what values are being considered
> > "missing"?
> >
> > Thanks! I look forward to your responses.
> >
> > Tonya

From: Bjoern on 28 Jul 2006 14:19
Depending on what you are doing that could also be the case. E.g. if
factors/covariates were missing in an analysis.

Additionally there are other missing values than the plain ".", so if
those could potentially occur, then I believe a safer check is "<=.z"
instead of "=.".

TSA wrote:
> Thanks for your suggestion. The proc freq shows only 44 missing values
> as well. Could missing values for other variables be causing other
> observation to excluded from the analysis?
>
>
> > TSA wrote:
> >
> > > After running my program, I received the messages that "74
> > > observations are not included because of missing values" in the log
> > > file.
> > >
> > > Based on the file that I imported, there should only be 43 missing
> > > values (denoted with a ".").
> > >
> > > Is there anyway to get SAS to show me what values are being considered
> > > "missing"?
> > >
> > > Thanks! I look forward to your responses.
> > >
> > > Tonya

From: TSA on 30 Jul 2006 19:16
Thanks for your help, all. When I factored it was indeed the missing
values for my covariate that were causing the discrepency.

Tonya

 | 
Pages: 1
Prev: Converting a SPSS file to SAS with Value and Variable labels
Next: SCAN function