SAS Customer Recognition — Steven Howe
Professor and Head A&S Psychology
University of Cincinnati
How long have you been using SAS®?
What SAS products have you used in the past? What products and solutions are you currently using?
Almost exclusively Base SAS and SAS/STAT®.
What is a problem you have solved using SAS?
Around 1982, I helped Public Demographics Inc. complete a project for New York State in which we produced demographic analyses of each census tract in the state. We aggregated the tract-level data up to as large of a geographic area as possible, limited to the size of the county, and then used report-writing features of SAS (pre-ODS) to prepare the documentary materials to submit to the federal government for the certification of Areas of Chronic Economic Distress.
What is the most innovative way you have used SAS?
Around 2002, I was working with a colleague on a problem of Appalachian migration. The US Census Bureau had released all of the migration stream data on DVD. There were approximately 12 million potential county-to-county streams (not that all of them had migrants), and there were hundreds of versions of each actual stream (e.g., by age group, educational group, etc.).
The only way the Bureau had been able to get all the data onto a limited number of DVDs was to use packed binary (PB) format. But they had provided no means to access the data other than a user interface that allowed only a very limited amount of data to be retrieved at a time. I had never worked with PB data before, but with formatting assistance from a colleague at the University of Missouri, I wrote a program to access a file, macroized it to work across tables and across states of origin, and pulled all the data off the DVDs and stored it in SAS data sets. The job took 48 hours to run, in total. Later, when I told a content specialist at the Bureau what I had done, she was very excited. I offered her the program and she, in turn, offered me the chance to specify a table for the 2010 migration files. Sadly, she left the bureau before I was able to get my reward.
What is your most memorable SAS moment?
I gave a workshop on using 1980 Census Bureau products via SAS at the 1982 annual meeting of the American Statistical Association with Jack Brown, who was then head of the Ohio Data Users Center in the Department of Development. When I taught the technique for aggregating tract segments up into whole tract records (using SET, BY and RETAIN techniques), I encountered a huge rush of excitement from the audience, at least a good many of whom had not been aware of SAS' ability to program across observations. That's when I knew I was good.
How has SAS changed in the time you have been using it?
What I tell my students is that there was a point around 1982 when I knew maybe 75 percent of SAS. Since then, I get deeper and deeper into parts of it, and certainly have forgotten very little (although talking to a colleague the other day did throw me because he used the && sequence in a macro, and I had once known that and forgotten it), but now I wonder if I am expert with even 10 percent of SAS. SAS was actually my fourth home package, after OSIRIS, BMD/BMDP, and, briefly, SPSS. About a year ago, I finally got around to using ODS table definitions to conduct a macro-based bootstrapping analysis in which I reached into the statistical output tables to grab a few key results, appended them to the end of an expanding data set with one observation/sample, and then analyzed the results. It was fun, but caused me to reflect how much harder it is to teach SAS now to new users. I still need most of the time to cover the basics and then use the last couple of weeks to bring in colleagues to show very sophisticated applications. Each year, the students' eyes get wider and wider and it's harder and harder to convince them that what they know is, in fact, the basis for what they just saw.
Have you ever attended a SAS Users Group meeting or SAS Global Forum? If yes, please list them.
Has your work with SAS been influenced by any other members of the SAS Community?
Oh my, yes. I explained above how all I needed from a colleague was a formatting example to build a complex program around. And I have colleagues and former students with whom I exchange ideas regularly. More often than not, though, I get my ideas from my academic partners who are not SAS users, but who need problems solved. I've always been able to get the job done in SAS (except for some forays into R and Mplus for specialized analyses).
If you could point a new SAS user to one resource, what would it be?