topic Re: SAS Proc GLM test for difference intercept and slope in SAS Programming

SAS Proc GLM test for difference intercept and slope

BDORR — Thu, 27 Jun 2019 20:59:08 GMT

The Code below produces the attached graph and I would like to test whether the intercepts and slopes of the 4 polynomial regression lines differ. I have 4 treatment levels control and stage 1-3.

ods graphics on;

proc glm data=MeanTotTRlat PLOTS=ANCOVAPLOT(CLM);
class treatment ;
model TotTR = latR LatR*LatR treatment ;

run;

ods graphics off;quit;

Re: SAS Proc GLM test for difference intercept and slope

PaigeMiller — Thu, 27 Jun 2019 21:43:33 GMT

I have 4 treatment levels control and stage 1-3.

Stage does not appear in the code anywhere. Is it important to the problem?

I would like to test whether the intercepts and slopes of the 4 polynomial regression lines differ

proc glm data=MeanTotTRlat PLOTS=ANCOVAPLOT(CLM);
class treatment ;
model TotTR = treatment treatment*latR treatment*LatR*LatR;
run;

The interaction between treatment*latr is the test to see if the slopes are the same, or not. If the p-value is <0.05 then the slopes are not statistically the same. The interaction treatment*latr*latr tests to see if the quadratic effects are the same across treatments.

Re: SAS Proc GLM test for difference intercept and slope

BDORR — Fri, 28 Jun 2019 12:49:19 GMT

Hi PaigeMiller (Esteemed Advisor),

The answer you gave below is great and addresses the issue of testing for differences in slopes. However I would also like to test for differences in intercepts for each regression line. Any Advice?

(reply in SAS Programming<> on 06-27-2019 05:43 PM )

________________________________

Re: SAS Proc GLM test for difference intercept and slope<>

I have 4 treatment levels control and stage 1-3.

Stage does not appear in the code anywhere. Is it important to the problem?

I would like to test whether the intercepts and slopes of the 4 polynomial regression lines differ

proc glm data=MeanTotTRlat PLOTS=ANCOVAPLOT(CLM);

class treatment ;

model TotTR = treatment treatment*latR treatment*LatR*LatR;

run;

The interaction between treatment*latr is the test to see if the slopes are the same, or not. If the p-value is <0.05 then the slopes are not statistically the same. The interaction treatment*latr*latr tests to see if the quadratic effects are the same across treatments.

Re: SAS Proc GLM test for difference intercept and slope

PaigeMiller — Fri, 28 Jun 2019 12:53:57 GMT

@BDORR wrote:
Hi PaigeMiller (Esteemed Advisor),

The answer you gave below is great and addresses the issue of testing for differences in slopes. However I would also like to test for differences in intercepts for each regression line. Any Advice?

Testing intercepts is the exact same concept as testing slope or quadratic effect. So I leave it as a homework assignment for you to figure out.

Re: SAS Proc GLM test for difference intercept and slope

BDORR — Mon, 01 Jul 2019 12:56:43 GMT

I don't need homework I need an answer. Can anyone else in this community provide an answer?