Hello, I am trying to conduct a difference in difference (DID) analysis to examine the effect of an intervention on the prevalence of smoking using a national cross-sectional survey dataset (2010-2015). The outcome of interest is a binary variable (smoking: yes/no), and I am comparing states with and without the policy of interest.  I have been asked to run both a linear probability model (using proc surveyreg) and a logit model (using proc surveylogistic) to examine the effect of the policy in treated versus reference states. I have to get the following estimates:         1) The average difference in the probability of smoking (treated versus reference states) from a linear probability model (i.e., estimate [95% CI])         2) The pre-post changes in the odds of smoking (treated versus reference states) from a logit model (i.e., odds ratio [95% CI])   Question 1: Does the coefficient obtained from the "lsmestimate" and "estimate" statement represent the DID estimate (i.e., the average difference in the probability of smoking)? If not, how can I get the DID estimate as a probability with a 95% CI? Here is the code I was using: proc surveyreg data=survey_data; domain gender; stratum stratum_var; cluster cluster_var; class policy_time intervention covar1 covar2 covar3; weight weight_var; model smoking= policy_time intervention policy_time*intervention covar1 covar2 covar3 /CLPARM solution vadjust=none; estimate "Diff in Diff" policy_time*intervention 1 -1 -1 1; lsmeans policy_time*intervention; lsmestimate policy_time*intervention "Diff in Diff" 1 -1 -1 1; run; Question 2: How can I obtain the changes in the odds of smoking (treated versus reference states) as an odds ratio with a 95% CI?  Here is a sample code: proc surveylogistic data=survey_data; domain gender; stratum stratum_var; cluster cluster_var; class policy_time intervention covar1 covar2 covar3; weight weight_var; model smoking= policy_time intervention policy_time*intervention covar1 covar2 covar3; run;   I appreciate any insight you can offer.       ... View more

Hello,   I am currently working on a longitudinal dataset that contains a lot of 0/1 variables (~90) . The dataset looks something like this (please note there is no pattern to the variable names):   ID     initial           mid1         final              PCI 100     0                 .                 1                    . 100     .                  0                .                     . 100     .                  .                 .                     1 200    1                  .                 0                    . 200    .                   .                 .                     1 300    0                  .                 .                     . 300     .                  .                 1                    .    300     .                 0                  .                    . 300     .                 .                   .                    0       I need to create a new 'summary' data set where that contains only one observation per ID. The dataset should look something like this:  ID      initial           mid1         final              PCI 100     0                 0                1                    1 200     1                  .                0                    1 300     0                  0               1                    0   How can I do this?   The only way I can think of doing this would be to create new variables that sums the 0/1 values for each variable and outputs the last value (using retain and first. last.). Unfortunately, this will take too much time and will require creating a lot of new variables.  Is there a simpler way of doing this? I appreciate any help you can offer.   ... View more

Hello, I need to graph the mean and SE of the variable COST for 2 groups of people (Age_group="Younger", "Older") in one plot. The plot is basically graphing the trend of cost over time. The two groups of people need to be next to each other so that their SE do not overlap and I need a line connecting each point for both groups. I know my code has to look something like what I listed below.    What can I add to the code below so that the mean & SE for the two groups will appear next to one-another for each year, and how can I customize the marker color and lineattrs for each group?   Title "Average cost per year"; proc sgplot data=final; scatter x=year y=cost/ yerrorlower=lowercost yerrorupper=uppercost markerattrs=(color=black symbol=SquareFilled)errorbarattrs=(color=black); series x=year y=cost/group= age_group lineattrs=(color=blue pattern=2); XAXIS label="Year" VALUES= (2001 TO 2007 BY 1); YAXIS label="Average Cost" VALUES= (0 TO 50 BY 10); run;   I appreciate any help you can offer. ... View more

Hi,   I am working with multiple data sets. There are a total of 4 data sets per year (~250 variables total across all 4 datasets and I have to keep all of them),and I have 10 years of data.   I need help coming up with a macro to minimize the code I have to write.   Here is a list of the 4 data sets for year 2000 (I have 4 data sets each for years 2000-2015):   *List of data sets for year 2000; clin_2000_phys_exam clin_2000_lab clin_2000_group clin_2000_final   *List of data sets for year 2001; clin_2001_phys_exam clin_2001_lab clin_2001_group clin_2001_final   ......and this repeats until we reach year 2010     What I am trying to do is something like this for each year:   *Merging "clin_2000_phys_exam" data with "clin_2000_lab" data by form; proc sort data=clin_2000_phys_exam; by form; run; proc sort data=clin_2000_lab; by form; run;    data phys_lab_2000; merge clin_2000_phys_exam  clin_2000_lab; by form; run;   *Merging the newly created "phys_lab_2000" data with "clin_2000_group" and "clin_2000_final" data by ID;   proc sort data=phys_lab_2000; by ID; run; proc sort data=clin_2000_group; by ID; run; proc sort data=clin_2000_final; by ID; run;   data 2000_final data; merge phys_lab_2000 clin_2000_group clin_2000_final; by ID;  run;   *Repeat the code above for all 10 years (years 2000-2010) until each year has a merged final dataset; *Concatenate data together to create a long dataset that contains data for all 10 years;   data year200_year2011; set 2000_final data 2001_final data 2002_final data 2003_final data 2004_final data 2005_final data 2006_final data 2007_final data 2008_final data 2009_final data 20010_final data; run;   I need help coming up with a SAS macro that does this so that I don't have to rewrite the same line of code for each year since the only difference would be the name of the dataset.    I appreciate any help you can provide.  ... View more

Hello,   I have multiple SAS data sets that are measuring the same thing across multiple study visits. Unfortunately, the variable names are not the same across the data sets even though they are measuring the same thing (with the exception of ID). The data I have looks something like this:   Name of data sets:        Visit_1                                          Visit_2                                       Visit_3  Variable names:             ID       200          101                   ID        300      102                    ID         400          103   Variables 200, 300 and 400 are all measuring whether the visit was attended. Variables 101,102 and 103 are all measuring the results of a test.   I am trying to create one data set that looks something like this:   Variable names:      ID       Visit_number      Visit_attended        Results                                  001           1                          1                          0                                  001           2                          1                          0                                  001           3                          1                          1                                  002           1                          1                          0                                  002           2                          0                           .                                  002           3                          1                           0   How might I go about solving this issue?    -Thank you ... View more

Hello,   I am trying to create a graph where the x axis will be time CD4 count was measured (before diagnosis, at diagnosis, 2 months after diagnosis, 4 month after diagnosis, 8 months after diagnosis and 10 months after diagnosis) and y axis will be the CD4 count. I want to label the grid values so that I can see: "before diagnosis", "2 months", "4 months" and so on on the x-axis. The partial code I want to use is below. How do I achieve the formatting of the "Time of CD4 measurement" values in the graph?    proc sgplot data=one; panelby state; refline 0/ axis=x lineattrs=(color= red pattern=2); series x=time y=CD4_count lineattrs=(color=gray pattern=1 thickness=1); colaxis label= "Time of CD4 measurement" grid values= (0 to 10 by 1); rowaxis label= "CD4_count"; run;   I appreciate any help you can offer ... View more

Hello,   I need to use stabilized IP weights for my analysis because of the large weights I get when I use the regular IPW (due to large variance). I calculated the regular IPW using the following code:   *Calculating probability of participation/no drop out given covariates; proc logistic data=ipw ; class gender race income education smoking BMIcat ; model responder = gender race income education smoking BMIcat; output out=weights predicted=prob; run; *Computing inverse weights; data inverseweights; set weights; if responder=1 then ipw = 1/prob; if responder=0 then ipw=1/(1-prob);  run;   *Merging dataset containing inverse weights with full dataset; data weight.ipwanalysis; merge inverseweights fulldataset (in=a); if a;  by id;  run;   *Final Adjusted model weighted by IPW; proc glm data=ipwanalysis ; class bmicat race; model FinalScore=bmicat race age; weight ipw; lsmeans bmicat/ stderr pdiff cl; run;  quit;   How can I calculate and use Stabilized Inverse Probability Weights (instead of regular IPW) so that I can weight the mean differences I get from my final adjusted model (proc glm). I appreciate any help you can offer ... View more

Hello,   I am trying to see if non-proportionality exists for a cox proportional hazards model that is adjusted for multiple factors and is stratified by gender using a plot of -log(log survival) vs. log (time). I would like to include a plot in my final submission.    I know that I can't use "proc lifetest" because that can't adjust for other factors.  I am thinking that I might have to do something like the code below and then use proc gplot to graph log(-log survival) vs. log(time). How can I create the plot that I need?   proc phreg data=final; model Survival_time*sensor(0)= Age BMI Smoking/ties=exact; strata Gender; Title "Cox proportional hazards model stratified by Gender"; baseline out=graph loglogs=llog survival=surv logsurv=lsur;  run;   proc gplot data=graph; plot llog*(I DON'T KNOW HOW TO GET THE LOG TIME)=Gender   I appreciate any help you can give me ... View more

Hello,   I am currently working on a project where I want to see the OR of having a cardiac event for: underweight, overweight and obese bmi categories at each of the age values I listed in my code (ages 50 & 60). For some reason my code below only outputs the OR for age and bmicat individually and the OR for age at 50 and 60.  Is there something with my code? How can I get the output I want  for the odds ratio(listed below)?    Code used: proc logistic data=final; class bmicat(ref="normal")/param=ref; effect agespline=spline(age/naturalcubic);                                 *Wanted to allow for curvature by using spline; model cardiac(event="yes")= agespline bmicat/ rsq ; oddsratio age/at (age= 50 60); oddsratio bmicat/at (age= 50 60); run;   Output I am trying to get for the Odds Ratios: underweight vs. normal at age=50  overweight vs. normal at age= 50 obese vs. normal at age=50    underweight vs. normal at age=60  overweight vs. normal at age= 60 obese vs. normal at age=60  ... View more

  Hello,    I am a new SAS user and I need help with clarifying any of the following concepts:   1. When conducting multiple linear regression using mostly continuous predictors, and one categorical predictor that was coded as 0/1 (for instance Sex: 0= male and 1=female), can you use proc reg (since the indicator variable is already created) or do you have to use proc glm with a class statement?   Can I use a code like this: proc reg data=new; model BMI = Sex Age Height Weight; run;   2. Is there a difference between stepwise selection and lasso selection when trying to identify the best fit model?   I appreciate any insight you can offer for any of the questions.  ... View more

  I created a data set to help explain my question. In my example, researchers are trying to understand the impact of different types of  Diet plans on change in BMI over 3 study visits.     ID Visit  Sex  Age   Height   Weight (lbs)     Diet (Type)   1    1     0      23       1.75       200                  0 2    1     0      18        2.1        300                  1 3    1     1      .           1.5        195                  2   1    2     0      24        1.75      180                  0 2    2     0      19          .           321                 1 3    2     1      29        1.5           .                    2   1    3     0      25        1.75     170                   0 2    3     0       .          2.1        280                  1 3    3     1       30        1.5      135                   2   I want to remove all observations for an ID if it has any missing values (I want a new data set that ONLY contains ID with observations for all visits). So the new data set would only include ID1 (since ID2 and ID3 have missing values). My desired output is:    ID Visit  Sex  Age   Height   Weight (lbs)     Diet (Type)   1    1     0      23       1.75       200                  0   1    2     0      24        1.75      180                  0   1    3     0      25        1.75     170                   0   I am trying to apply this concept to a large data set with ~50 variables. So l would like help to get the desired output without listing the 50 variable names in a procedure. I think I can do this by using proc sql, but I am confused on how to do this.    ... View more