I’ve been working with SAS Enterprise Guide 6.1 with SAS On-Demand on Virtual Box and I can’t find much to complain about. It is slow, but since I’m usually doing 14 things at once, while I’m writing for a task to complete I can read an email, download a file or answer a phone call. Why bother? Why not just use something quicker? Well, there is the free part. I expect to be teaching at a different university this fall with a higher proportion of low-income students (stay tuned), so “free” would be good. Virtual Box is also free.
Then there is the ease of explanation of some topics with graphics, a big soapbox issue for me. Also, the SASHELP library has some ready to use data sets. Here is just one example, from the Heart data set.
Then I selected from TASKS > GRAPH and clicked on the picture of a stacked bar chart.
I dragged sex under Column to Chart and Status under Stack.
Under the ADVANCED tab I selected Percentage as the Statistic used to calculate bar and clicked the box next to Specify one statistical value to show for bars and picked percentage from the drop down menu.
With a little pointing and clicking I have a good first look at our data. It is obvious that there were more women than men in the sample – 55% to 45%. It’s also obvious that the odds of a man dying were much greater. While men were about equally likely to be alive as dead, women were twice as likely as being alive as dead.
So, from this little chart we can start talking about odds, then move to odds ratios. We can also raise objections to looking at just a single variable related to death. Were the women in this sample younger? Certainly age would be related to death. What other variables should we include?
So, here you have a sneak peek at the beginning of my class on categorical data analysis at WUSS in November. You should come. It will be fun. There will be pictures. And cookies.
Even potentially pictures of cookies. I have to go now because I am hungry.