cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
Caetreviop543
‎09-22-2023
  • 93 Posts
  • 0 Likes Given
  • 0 Solutions
  • 6 Likes Received

Re: PROC LOGISTIC CONTRAST STATEMENTS

by Obsidian | Level 7 in Statistical Procedures
‎03-23-2018 05:23 PM
2 Likes
‎03-23-2018 05:23 PM
2 Likes
Hi Reeza,   I figured out what was wrong with my calculations...I was dividing the log odds and then exponentiating, rather than exponentiating the log odds to get the odds, and then dividing by the odds to get the odds ratios.    Yea, I think your right...the odds ratio statements are easier. I'm not really sure the utility or advantage of using the estimate statements.    Thanks for your help!   Emily  ... View more

PROC LOGISTIC CONTRAST STATEMENTS

by Obsidian | Level 7 in Statistical Procedures
‎03-21-2018 07:25 PM
‎03-21-2018 07:25 PM
Hi everyone,   I am trying to use contrast statements in proc logistic to show estimates at different levels of X, and that the linear relationship between my dependent and independent variable isn’t constant.   I was able to find documentation here:   http://support.sas.com/kb/35/189.html   I just have a few questions:   How does SAS calculate the estimates at different levels of X with control variables in the model? The example seems to use the formula B0+B1x+B2x2, but when I add control variables to my model the output and results based on the formula don’t match.   What does the ‘intercept 1’ statement refer to? I assumed it meant setting B0=1 when obtaining the estimates (based on B0+B1x+B2x2), but that doesn’t produce the correct answer.   Where exactly does the formula B1+B2(2x+1) tie in? I thought it corresponds to the unit change difference between two estimates, where a one unit difference would be B1+B2(2x+1), a two unit difference B1+B2(2x+2), etc. It also seems that way, because in the contrast 'logOdds@11-logOdds10' statement you plug in the final value for (2X+1), although I don’t understand why you then double that for a two-unit increase. The author explains that the difference between the two estimates is simply estimate two minus estimate one, so why then are the final values for (2x+1) incorporated in the code?   Help on any of the questions is greatly appreciated!   Emily  ... View more

Re: %if %then %else %str macro functions

by Obsidian | Level 7 in SAS Programming
‎02-14-2018 02:21 PM
‎02-14-2018 02:21 PM
Ok. thanks for your response. I was using %str function because it was less code. I did eventually resort to %if %then %do %end, but I think Astounding's response is the most succinct/straightforward. ... View more

Re: SAS predicted probabilities vs. hand calculation

by Obsidian | Level 7 in SAS Programming
‎02-14-2018 02:13 PM
2 Likes
‎02-14-2018 02:13 PM
2 Likes
Originally, I was struggling with how to multiply the class level to the correct class dependent coefficient (in my first post). There are three levels for MRT, so I needed to multiply each one with the corresponding class-dependent coefficient.   I ended up doing this:   %macro predictor;   data new; set old; %if MRT=1 %then %do; pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU)+(MRT*MRT_P1)))); %end; %else %if MRT=2 %then %do; pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU)+(MRT*MRT_P2)))); %end; %else %do; pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU)))); %end;   %mend predictor; %predictor;   Astounding showed me a much more straightforward, simple way:   if mrt=1 then temp=mrt_p1; else if mrt=2 then temp=mrt_p2; else temp=0; pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU) + (temp) )));   Once I did this, discrepancies between SAS predicted probabilities and those using the formula were minuscule. Thanks Reeza!   ... View more

Re: %if %then %else %str macro functions

by Obsidian | Level 7 in SAS Programming
‎02-13-2018 07:20 PM
‎02-13-2018 07:20 PM
Thanks!!   So you can't use macro functions when trying to modify variables in a data step? Is that correct? I've used %if %then %do statements and macro variables in data steps and not had any issues. I guess I've only used %if %then %str in a proc though.   Regardless, your response was super helpful! Creating the temp variable resolved my issue with differences between the predicted values produced by SAS and by hand.   Emily ... View more

Re: SAS predicted probabilities vs. hand calculation

by Obsidian | Level 7 in SAS Programming
‎02-13-2018 06:45 PM
‎02-13-2018 06:45 PM
Thanks Reeza!   I accounted for rounding errors by outputting the coefficients (outest=coefficients), and merging them with the predicted probabilities (pred=pred). I then applied the formula in SAS, so both the coefficients and predicted probabilities come from the same model and are unrounded.   I did, however, alter my model using param=ref:   proc logistic data=temp desc outest=coefficients; class EDU (ref='0') SEX (REF='0')  MRT(REF='0') / param=ref; model outcome=AGE SEX EDU MRT; output out=pp pred=pred; run;   Is this correct?   Despite still not identical, using param=ref greatly reduced discrepancies between SAS's predicted probabilities and those obtained by hand. The min. diff is now 0, the max, 0.0295, and the mean, 0.0097480.   There must be some reason for why they are still not exactly the same, but regardless, thanks for your help!   Emily ... View more

SAS predicted probabilities vs. hand calculation

by Obsidian | Level 7 in SAS Programming
‎02-13-2018 05:46 PM
‎02-13-2018 05:46 PM
  Hi,   I have a follow up question to the original posted earlier today (listed below in bold).   How come there are slight differences in the predicted probabilities between those calculated by SAS:   proc logistic data=temp desc; class EDU (ref='0') SEX (REF='0')  MRT(REF='0'); model outcome=AGE SEX EDU MRT; output out=pp pred=pred; run;   and using the standard formula:   pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU)+(MRT_P*MRT);   Assuming Intercept, AGE_P, SEX_P, EDU_P, AND MRT_P are the coefficients provided by SAS, multiplied by the actual values per person.   The min. difference between the predicted probabilities produced by SAS and by hand is .00003, the max, .0815, and the mean .0219.   Thanks!   Emily   Hello,   I am trying to use the %str function to modify my code. Here it is:   %macro predictor;   data new; set old; pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU) %if MRT=1 %then %str( + (MRT_P1)))); ); %else %if MRT=2 %then %str( + (MRT_P2)))); ); %else %str( ))); ); run;   %mend predictor; %predictor;   When I plug what's in the string functions individually, it works, so I know it doesn't have anything to do with how I coded the content (i.e. missing parentheses/semicolons). I've also used this function before in similar contexts, and never had a problem.    I'm getting the dreaded "There is no matching %IF statement for the %ELSE...A dummy macro will be compiled," error.   Any help would be appreciated.   Thanks in advance!   Emily ... View more

%if %then %else %str macro functions

by Obsidian | Level 7 in SAS Programming
‎02-13-2018 05:17 PM
‎02-13-2018 05:17 PM
Hello,   I am trying to use the %str function to modify my code. Here it is:   %macro predictor;   data new; set old; pred=1/(1+exp(-1*(Intercept+ (AGE_P*AGE)+(SEX_P*SEX)+(EDU_P*EDU) %if MRT=1 %then %str( + (MRT_P1)))); ); %else %if MRT=2 %then %str( + (MRT_P2)))); ); %else %str( ))); ); run;   %mend predictor; %predictor;   When I plug what's in the string functions individually, it works, so I know it doesn't have anything to do with how I coded the content (i.e. missing parentheses/semicolons). I've also used this function before in similar contexts, and never had a problem.    I'm getting the dreaded "There is no matching %IF statement for the %ELSE...A dummy macro will be compiled," error.   Any help would be appreciated.   Thanks in advance!   Emily ... View more

Re: Nested do loop tertiles

by Obsidian | Level 7 in SAS Programming
‎11-20-2017 02:46 PM
‎11-20-2017 02:46 PM
Thanks Tom! That's extremely helpful; I would have never thought to do it that way. I appreciate it! ... View more

Re: Nested do loop tertiles

by Obsidian | Level 7 in SAS Programming
‎11-20-2017 10:17 AM
‎11-20-2017 10:17 AM
Thank you for your reply, ballardw.   I did change the initial from i to K, however, I received an error saying "Mixing of implicit and explicit array subscripting is not allowed."   I think it's because I included the array reference for only one variable, then tried to incorporate two stand-alone do statements. Ideally, I would like to create the tert variables without an array reference. Not sure if that's possible...   Thanks   Emily ... View more

Re: Nested do loop tertiles

by Obsidian | Level 7 in SAS Programming
‎11-17-2017 05:43 PM
‎11-17-2017 05:43 PM
Ok, thanks that makes sense.   So, is this correct?     data temp; set rank_tert;   array tert(i) TERT1-TERT10;   do I=2.5 to 3.4 by 0.1; do J=3.5 to 4.4 by 0.1; do i=1 to 10;   if 0<PA_AVG<(I) then tert(i)=1; else if (I)<=PA_AVG<=(J) then tert(i)=2; else tert(i)=3;   end; end; end;   keep PID_PDE TERT1-TERT10;   run;     Thanks ... View more

Nested do loop tertiles

by Obsidian | Level 7 in SAS Programming
‎11-17-2017 04:59 PM
‎11-17-2017 04:59 PM
Hi,   I am trying to create 10 tertile variables (TERT1-TERT10), which based on a variable titled PA_AVG on a five point scale, increase by 0.1, and have lower bound second tertile cut offs starting at 2.5, and ending at 3.4. The second tertile upper bound cutoffs start at 3.5 and end at 4.4, increasing by .10.   So, category one in TERT1 would be 0 to 2.4, category two 2.5 to 3.5, and category three, 3.6 to five. Category one in TERT2 would be 0<CAT_ONE<2.6, category two, 2.6<=CAT_TWO<=3.6, and category three, 3.6<CAT_THREE<=5, and so on until TERT10 with a second tertile cut off of 3.4 and 4.4 has been completed.   I am using a nested do loop, but the code isn’t executing (it just keeps running):   data temp; set rank_tert;   do I=2.5 to 3.4 by 0.1; do J=3.5 to 4.4 by 0.1; do TERT=1 to 10 by 1;   if 0<PA_AVG<(I) then TERT=1; else if (I)<=PA_AVG<=(J) then TERT=2; else TERT=3;   end; end; end;   keep PID_PDE TERT1-TERT10;   run;   Not sure what I am doing wrong, but it may have something to do with how I am creating the 'TERT' variable? Thanks in advance for any help!   Emily   ... View more

Re: macro do loop

by Obsidian | Level 7 in SAS Programming
‎10-25-2017 03:27 PM
‎10-25-2017 03:27 PM
Hi Tom,   Great, thanks! That's super helpful. Is it always necessary to put a period after the macro variable and before the &i, if doing a do loop within a macro?   Emily ... View more

macro do loop

by Obsidian | Level 7 in SAS Programming
‎10-25-2017 02:34 PM
‎10-25-2017 02:34 PM
Hello,   I am trying to run six, stratified logistic regression models. Those which are sorted/stratified by PA_AC1, PA_AC2 and PA_AC3 have the primary predictor PA_AVG, whereas the predictor for those sorted by NA_AC1, NA_AC2 and NA_AC3 is NA_AVG. This is the (non-working) code I've come up with:   %macro mod(var1, var2);   %macro nest;   %DO i=1 %to 3;   proc sort data=work.last; by &&var1&i; run;   proc logistic data=work.last; model outcome=&&var2; by &&var1&i; run;   %end;   %mend nest;   %mend mod; %mod (PA_AC, PA_AVG); %mod (NA_AC, NA_AVG);   Thanks in advance for any help!             ... View more

Re: nested/more efficient macro for proc logistic

by Obsidian | Level 7 in Statistical Procedures
‎08-29-2017 12:58 AM
‎08-29-2017 12:58 AM
That's interesting! I wouldn't have thought to do it that way...   Thanks Reeza! ... View more
Contact Me
Online Status
Offline
Date Last Visited
‎09-22-2023 03:28 PM
Likes from
View All
Quick links