topic Re: Equation from Proc transreg in Statistical Procedures

Equation from Proc transreg

Miracle — Sun, 17 May 2015 15:24:31 GMT

Dear all,

How are you?

I'd like to ask if it's possible to write out the equation from the proc transreg below?

Can the regression coefficients below be written as y=6.87157+3.65544male+...+0.47702score?

and be interpreted as one unit increase in x will increase y by 0.53355?

Thank you very much.

ods graphics on;

proc transreg data=rtm solve test nomiss plots=all;

model spline(y/ nknots=9)=opscore(male)

opscore(age18_19) opscore(age20_24) opscore(age25_34) opscore(age35_44) opscore(age45_54)

spline(x/ nknots=9) identity(score) / cl detail;

output out=trans predicted residuals mrc coefficients;

run;

ods graphics off;

proc print data=trans; where _TYPE_='M COEFFI'; run;

obs _TYPE_ _NAME_ TIntercept Tmale Tage18_19 Tage20_24 Tage25_34 Tage35_44 Tage45_54 Tx Tscore

1325 M COEFFI y 6.87157 3.65544 6.04942 5.01395 4.95410 5.90719 3.10483 0.53355 0.47702 .

Re: Equation from Proc transreg

plf515 — Tue, 19 May 2015 10:51:46 GMT

Yes. That is how to read them (I am assuming each of the tage variables are coded 0 and 1).

If you have actual age, I would recommend using that, instead of the categorical variables for age.

Re: Equation from Proc transreg

Miracle — Wed, 20 May 2015 07:21:21 GMT

Hi Peter.

Thanks for your reply.

I tried what you suggested by fitting spline to the continuous age variable instead.

I later examined the residuals of my spline regression but I'm not sure if the residual plot and the Q-Q plot indicate any severe heteroscedasticity and non-normality?

Perhaps you could shed some light me on this?

Thank you very much.

Re: Equation from Proc transreg

plf515 — Wed, 20 May 2015 10:13:42 GMT

It does seem to exhibit some of each; whether that will be problematic is another question - I don't know of good tools for figuring out how great a violation relates to how problematic a regresssion

Re: Equation from Proc transreg

Miracle — Fri, 22 May 2015 01:05:23 GMT

Thanks Peter for your response.

I'll go and explore a bit more.

Re: Equation from Proc transreg

Miracle — Tue, 09 Jun 2015 06:05:34 GMT

Hi Peter.

How are you?

Let's not worry about the non-normality and heteroscedasticity for now.

Can I please ask you about my interpretation of the spline regression coefficients in the first post whether it is correct?

Thank you very much and have a good day.

Re: Equation from Proc transreg

plf515 — Wed, 10 Jun 2015 11:30:17 GMT

Yes, you're right in your interpretation.