topic Re: Fitting a quadratic regression for a continuous outcome and categorical predictor in Statistical Procedures

Fitting a quadratic regression for a continuous outcome and categorical predictor

palolix — Thu, 12 Feb 2026 22:31:09 GMT

Dear SAS Communities,

I would like to fit a regression model in order to predict the continuous variable "dry matter" (DM) when the fruit has an hedonic score of 5, 5.5, and 6 (categorical predictor).

According to Shapiro Wilk p-value for normality, the residuals do not follow a normal distribution.

In this case, is it correct to include a quadratic term in the model like this? I did not include a class statement since the regression line is being fitted.

proc glm data=one;
model DM= hedonic hedonic*hedonic;
run;

I would greatly appreciate your feedback!

Re: Fitting a quadratic regression for a continuous outcome and categorical predictor

Ksharp — Fri, 13 Feb 2026 06:54:46 GMT

https://support.sas.com/kb/22/912.html

Re: Fitting a quadratic regression for a continuous outcome and categorical predictor

palolix — Fri, 13 Feb 2026 21:49:12 GMT

Thank you Ksharp!

So, if the quadratic and cubic terms are significant then I should include the higher order in the model I guess

Parameter	Estimate	Standard Error	t Value	Pr > \|t\|
Intercept	14.85272543	2.22070211	6.69	0.0011
hedonic	6.71613239	1.82095677	3.69	0.0142
hedonic*hedonic	-1.19830813	0.41222284	-2.91	0.0335
hedonihedonihedoni	0.07633635	0.02720314	2.81	0.0377

Re: Fitting a quadratic regression for a continuous outcome and categorical predictor

Ksharp — Sat, 14 Feb 2026 02:43:45 GMT

Yes. According to the Doc, I think it should be .

Re: Fitting a quadratic regression for a continuous outcome and categorical predictor

palolix — Fri, 20 Feb 2026 02:13:13 GMT

Great, thanks!