topic Re: Not receiving RELRISK plot in frequency table in SAS Programming

Not receiving RELRISK plot in frequency table

Vicente95 — Sun, 11 Jul 2021 18:35:07 GMT

Currently, I am not receiving the appropriate Odds Ratio and Relative Risks table in my SAS below when using the ameshousing3.sas7bdat dataset. The expectation is that SAS produces the table in the Results Viewer. However, that is not happening in my use case. Does anyone know what I am doing wrong? I have attached the SAS dataset to this question along with a screenshot of the output that I am receiving.

PROC FREQ DATA = WORK.ameshousing3;
	TABLES fireplaces * bonus / CHISQ RELRISK NOROW NOCOL NOPERCENT;
	FORMAT bonus bonusfmt.;
	TITLE "Association Analysis - Chi-Sq and Odds Ratio";
RUN;

Re: Not receiving RELRISK plot in frequency table

Watts — Sun, 11 Jul 2021 19:22:19 GMT

The odds ratio and relative risks are provided for 2x2 tables (tables that contain two rows and two columns). It looks like your table is 3x2.

Re: Not receiving RELRISK plot in frequency table

Vicente95 — Sun, 11 Jul 2021 19:45:42 GMT

Ok, thank you for the response. How can I fix it?

Re: Not receiving RELRISK plot in frequency table

Watts — Sun, 11 Jul 2021 20:06:08 GMT

You can decide what are the objectives, what are the appropriate analyses, what do you need to compute, and so on.

https://sciencing.com/calculate-odds-ratio-contingency-table-8782587.html

Re: Not receiving RELRISK plot in frequency table

tarheel13 — Sun, 11 Jul 2021 20:15:25 GMT

You can also get odds ratio from proc logistic.

ods graphics on ;
proc logistic data=stat1.ameshousing3 plots(only)=(effect oddsratio (type=horizontalstat));
	model bonus(event="1") = fireplaces / clodds=both;
run;
ods graphics off;

Re: Not receiving RELRISK plot in frequency table

Vicente95 — Sun, 11 Jul 2021 20:24:06 GMT

Oh wow! Thank you for letting me know. On the same topic of PROC LOGISTIC, how can I fit a binary logistic regression model with bonus as the outcome variable and basement_area, fireplaces, and lot_shape2 as the predictor variables? In this model, I want to specify "Regular" as the reference group for the lot_shape_2 variable and "0" as the reference level for the fireplaces variable.

Would it look something like this?

PROC LOGISTIC DATA = WORK.ameshousing3
	PLOTS(ONLY) = (EFFECT (CLBAND X = (FIREPLACES)) ODDSRATIO (TYPE = HORIZONTALSTAT));
	CLASS lot_shape_2 (PARAM = REF REF = 'Regular') fireplaces (PARAM = REF REF = '0');
	MLogit1: MODEL bonus(event='1') = basement_area lot_shape_2 fireplaces / clodds=pl;
	UNITS fireplaces=1;
	ODDSRATIO 'Comparisons of Type' lot_shape_2 / DIFF=ALL CL=PL;
	TITLE 'Bonus Eligibility Status Model';
RUN;

Re: Not receiving RELRISK plot in frequency table

tarheel13 — Sun, 11 Jul 2021 20:42:17 GMT

Looks right.