Hi, I am trying to perform logistic regression but am facing rare events (~0.07%) out of a total sample of 200,000+ observations. I understand that one method is to perform stratified sampling. But I also read that Firth method is possible too? (Logistic Regression for Rare Events | Statistical Horizons) Can I check if Firth method is appropriate for rare events? ... View more

Hi, I am trying to perform oversampling on my dataset (~200,000 observations), which consist of a flag variable of value 1 or 0. I want 100 samples with each sample to contain all the observations with flag=1 (~100 of them in total) and then randomly select ~750 of observations that have flag = 0. However, I seem to have some difficulty in getting what I want. I ended up with ~10,000 observations in total for each sample. And sometimes, my code takes forever to run. Can someone advice me on what is wrong? My code is as follows: data oversamples; do sample=1 to 100;         set fun;         do i = 1 to _N_;         if flag= 1 or (flag=0 and ranuni(sample+7320) < 0.003) then output;         end;     end; run; Thanks for any advice. ... View more

Hi all, I am trying to use proc import to create a SAS database. In particular, one of my columns contain CUSIP data, in which some are 9 digits numbers while some contain a mix of numbers and characters. Thus, I want to make SAS read in that column as character rather than numeric variables, however, the first few rows of my data are all numeric rather than a mix of numeric and character, thus SAS reads in that column as numeric. Is there a way to just change that column format into character while keeping the other columns as default? I have quite alot of columns so would hope not to have to redefine all columns. Thank you for any help! ... View more

Hi, I am relatively new to SAS and I have a question on what Proc should I use. I have a dataset that is categorized by two classes, e.g. Scenarios A,B,C and also Status Active Inactive. I hope to create a table where I can get the mean, median, s.d.,min,max for each category and also to test the differences in both mean and median between Status Active and Inactive for each of the Scenarios A,B,C I mean something below: Test for Difference between Active and Inactive samples                                    t-statistic                         Wilcoxon Z Scenario A                       2,2**                                  2.8** Scenario B                       3.1***                                 2.1** I know I can use the proc means to get the descriptive statistics that I want, but after reading the guide, Im not sure how to proceed to get the mean and median difference test. Must I separate my dataset into 3 dataset based on the 3 scenarios and then run proc ttest and proc npar1way for each of the 3 sub datasets? Or is there a better way that I can use for the combined dataset and generate the desired table format? Thank you in advance for any help offered! ... View more