topic Re: Identity line for regression in New SAS User

Identity line for regression

bkq32 — Thu, 05 Mar 2020 04:18:34 GMT

I'm running a simple linear regression of WEIGHT_1 vs WEIGHT_2.

proc reg data=have;
 model weight_1 = weight_2;
run;

How can I add an identity line to the fit plot and make the axis ranges the same?

Re: Identity line for regression

Reeza — Thu, 05 Mar 2020 04:24:15 GMT

Consider using SGPLOT instead?
You can use the REG statement, then use a LINEPARM statement for the identity line and you have full control of the axis using the XAXIS/YAXIS statements.

Re: Identity line for regression

bkq32 — Thu, 05 Mar 2020 05:26:54 GMT

Thank you, @Reeza. In order to keep the ANOVA and parameter estimates that I'd get from PROC REG, I'm thinking I'll just piece the output from PROC REG and PROC SGPLOT together.

data have;
 input id weight_1 weight_2;
 cards;
1 80 81
2 150 155
3 190 170
4 201 203
5 204 211
6 180 188
7 170 171
8 177 169
9 90 82
10 99 101
;
run;


ods trace on;
proc reg data=have;
 model weight_1 = weight_2;
 
 ods exclude diagnosticspanel;
 ods exclude residualplot;
 ods exclude fitplot;
run;


proc sgplot data=have;
 title "Fit Plot for weight_1";
 reg y=weight_1 x=weight_2 / cli clm;
 lineparm x=0 y=0 slope=1 / lineattrs=(color=red);
 xaxis values=(50 to 250 by 50);
 yaxis values=(50 to 250 by 50);
run;