topic Re: Journal quality regression output table 2. in SAS Procedures

Journal quality regression output table 2.

GKati — Mon, 19 Mar 2018 12:37:32 GMT

Hello Everyone,

Following on the post by zzecon

(https://communities.sas.com/t5/SAS-Procedures/Making-a-publication-quality-table-from-regression-results-using/td-p/396779)

I would also like to produce regression output like this one:

Using the following code, I get a slightly diffferent output.

proc format;

picture stderrf (round)

low-high=' 9.9999)' (prefix='(')

.=' ';

run;

proc report data=parmest nowd out=table;

column type dependent variable value1 value2;

define numord /noprint;

define variable /display ' ';

define type / display;

define dependent / across ' ';

define value1 /display;

define probt1 /display;

define value2 /display;

define probt2 /display;

compute value;

array cc _c8_ _c5_ ;

if type='stderr' then do;

call define(_col_,'format','stderrf.');

end;

else do;

call define(_col_,'format','8.4');

do i= 1 to 3;

if ~missing(cc(i)) then do;

if 0.05<cc(i) <= 0.1 then call define(_col_, "style", "style=[ posttext='*']" );

else if 0.01 <cc(i) <=0.05 then call define(_col_, "style", "style=[ posttext='**']" );

else if cc(i) <= 0.01 then call define(_col_, "style", "style=[ posttext='***']" );

leave; end;

end;

endcomp;

run;

This is what my output table looks like:

type	indavgcost		value1	value2
coefficient	1	Intercept	2154.033	2361.4119
stderr	1	Intercept	3.9858111	22.628486
coefficient	1	tag13	101.8107	73.756248
stderr	1	tag13	8.189885	8.1373222
coefficient	1	avgcontract	-736.7952	-578.251
stderr	1	avgcontract	10.062152	10.037246
coefficient	1	tag13*avgcontract	-80.84143	-100.5845
stderr	1	tag13*avgcontract	20.23273	20.126612
coefficient	1	ses	.	-74.14517
stderr	1	ses	.	2.3641456
coefficient	1	female	.	-54.42843
…..

I want the standard errors to be parenthesis.
I would like to see the stars next to the coefficient.

The mistake will be in the array of my proc report, but I don’t know what exactly.

Thanks for your help.

Re: Journal quality regression output table 2.

PaigeMiller — Mon, 19 Mar 2018 12:44:17 GMT

To get parenthesis around your standard errors, and asterisks next to the p-values, you need to create a PICTURE format, and assign that format to the proper variable in PROC REPORT.

http://documentation.sas.com/?cdcId=pgmmvacdc&cdcVersion=9.4&docsetId=proc&docsetTarget=p0n990vq8gxca6n1vnsracr6jp2c.htm&locale=en

Re: Journal quality regression output table 2.

GKati — Mon, 19 Mar 2018 12:46:24 GMT

Thanks for your reply. I thought I was doing that right before the proc report. See above. Did I do something wrong there?

Re: Journal quality regression output table 2.

GKati — Tue, 20 Mar 2018 12:49:14 GMT

I am trying to write a picture format for the asterics. This is what I have so for:

proc format;
picture stderrf (round)
low-high=' 9.999)' (prefix='(')
.=' ';

picture asterics (round)
low - 0.01 = 9.999 (posttext='***')
0.01 < - < 0.05 = 9.999 (posttext='**')
0.1 -high = 9.999 (posttext='*')
0= (posttext='')
.=(posttext='')
run;

Posttext doesn't seem to work.

I get the following error message:
187 picture asterics (round)
188 low - 0.01 = 9.999 (posttext='***')
________
22
76
ERROR 22-322: Syntax error, expecting one of the following: DATATYPE, DECSEP, DIG3SEP, FILL, LANGUAGE, MULT, MULTIPLIER, NOEDIT,
PREFIX.
Do you happen to have any idea what else I could use?

Re: Journal quality regression output table 2.

mkeintz — Wed, 21 Mar 2018 16:01:32 GMT

Instead of a posttext= parameter, use the same technique you used in the stderrf format:

proc format;
  picture asterics (round)
    low - 0.01            = '9.999***'
          0.01 < - < 0.05 = '9.999**'
          0.1 -high       = '9.999*';
run;

data _null_;
  do sig=.001 ,.005 ,.01 ,.03 ,.05 ,.1,.8;
    put sig=    @12 sig=asterics.;
  end;
run;

BTW, I think you left some holes in your value boundaries in the format (see the .05 formatted value).