topic Re: prog reg in SAS Procedures

prog reg

a_d_a — Thu, 22 Jan 2015 01:45:37 GMT

i want to ask why my output of standard error is zero. here my command that i used

data MP02;

input N LAI SPAD HEIGHT DIAMETER BLUE GREEN RED NIR;

cards;

2.22 1.99 78.27 5.93 2.47 8.12 11.99 8.61 53.29
2.35 2.07 65.30 5.67 2.37 5.11 8.84 4.98 65.95
3.39 2.91 67.77 4.23 2.50 6.82 10.81 7.32 52.35
2.46 2.85 76.43 4.97 2.29 7.00 10.53 7.32 54.81
2.37 1.85 74.17 5.81 2.50 5.01 8.13 4.81 65.08
2.27 2.22 61.20 5.29 2.54 9.29 16.34 11.74 55.12
2.45 1.95 71.30 4.12 2.35 4.87 8.07 5.13 54.27
2.16 3.30 72.33 5.05 2.30 5.55 8.75 5.50 64.41
2.52 2.07 69.53 3.99 2.46 5.91 9.46 6.17 54.29

;

proc reg;

FORWARDSELECT:model N= LAI SPAD HEIGHT DIAMETER BLUE GREEN RED NIR/SELECTION=FORWARD;

proc reg;

MAXR:model N= LAI SPAD HEIGHT DIAMETER BLUE GREEN RED NIR/SELECTION=MAXR;

proc reg;

MINR:model N= LAI SPAD HEIGHT DIAMETER BLUE GREEN RED NIR/SELECTION=MINR;

proc reg;

RSQUARE:model N= LAI SPAD HEIGHT DIAMETER BLUE GREEN RED NIR/SELECTION=RSQUARE;

proc reg;

ADJRSQ:model N= LAI SPAD HEIGHT BLUE GREEN RED NIR/SELECTION=ADJRSQ;

PROC PLOT; PLOT N*LAI N*SPAD N*HEIGHT N*DIAMETER N*BLUE N*GREEN N*RED N*NIR;

run;

is there anyone that can suggest what i should do?

Re: prog reg

Reeza — Thu, 22 Jan 2015 02:01:15 GMT

You have 9 observations and 8 variables in your actual data?

Re: prog reg

a_d_a — Thu, 22 Jan 2015 02:07:36 GMT

i averaged according rep

i have three trt, three rep.. in each rep i take 8 plants and averaged for each data

Re: prog reg

Reeza — Thu, 22 Jan 2015 02:17:33 GMT

Ok...I think you should use a different analysis method if you have replicated measures, but that isn't the question you asked.

That being said, its your full model - where you have all 8 variables included- that is causing the issues. Given the DF=8 its not surprising. A intro stats text can probably explain it better, I'm rusty with the exact calculations.

proc reg data=mp02;

FULL:model N= LAI SPAD HEIGHT DIAMETER BLUE GREEN RED NIR;

run;

Re: prog reg

a_d_a — Thu, 22 Jan 2015 02:38:17 GMT

the standard error still 0. is the problems is due to my data? how i can do the regression in correct way?

Re: prog reg

Reeza — Thu, 22 Jan 2015 03:00:13 GMT

You don't have enough data to generate standard errors.

I don't know what your actual data looks like so hard to say. Start a new thread with sample data if you'd like. Best to post these types of question in the Statistical Forum as well

Re: prog reg

a_d_a — Thu, 22 Jan 2015 03:02:58 GMT

thanks reeza