Back in June I wrote a paper/poster for the Association for Institutional Research Forum. As the title suggests, the papers is about how to create institutional research maps using SAS. You can use the map code for areas other than IR, but the focus for this paper was to show enrollment, study abroad, and alumni data visually. The full list of paper and code can be found at: http://www.cmu.edu/ira/infox/index.html
I am hoping to have a video of the process completed soon.
You can download the paper and poster by clicking the link below.
In the past, I've often encountered the question “What if we added one more measurement (question) to the end of this survey?”It seems that there is almost always someone who wants to add another question, typically at the end of the survey, to measure a question that had just popped into their head. It is important to avoid doing this at all costs. Asking a question just because it would be nice to know is considered poor survey design. If you simply create questions because they were interesting the survey would become too long, lack focus, and hurt response rates.
One of the better books I’ve read on survey design is called Asking Questions by Bradburn, Sudman, and Wansink. Here are their suggestions for designing a questionnaire/survey:
1. Resist the impulse to write specific questions until you have thought through your research questions
2. Write down your research questions and have a hard copy available when you are working on the questionnaire
3. Every time you write a question ask yourself “Why do I want to know this?” Answer it in terms of the way it will help you to answer your research question. “It will be interesting to know” is not an acceptable answer
The best surveys tend to have very clear and concise research questions that are formed before even one survey question is created. A well formed research question will help focus your survey creation so that you don’t end up with extra questions that don’t tell you anything.
Whenever I start working with someone now my first goal for them is to create a set of research questions. When it comes time to create survey questions I always ask: “Does this answer one of the research questions?”
Research questions tend to guide projects - well formed questions lead to a clear and concise project, while poorly formed questions can cause a project to lose focus.
One of the better presentations I attended at the SAS global forum was using SAS in social media. This topic is very interesting to me because I'm constantly made aware of new technology by my CMU colleagues. Students today use twitter and facebook to share information, but it's often hard extract what they are saying and make it useful. The paper can be found by clicking anywhere on this link, which will take you the SAS Dummy Blog
Naturally, I had to try this out for myself. I decided not to use the facebook component because I didn't see how it would be useful from a working perspective. I did however run the SAS code looking for specific tweets. Recently, CMU had their annual Carnival and University promoted using hashtag cmucarnival (#cmucarnival). I downloaded 89 tweets from twitter to my computer using SAS. I think the next logical step would be to use a text miner, but I do not have access to such a program.
The interesting idea about this type of programming is that it could produce some really interesting data on what students are saying about Carnegie Mellon. For instance, if students used specific hashtags (that's the way SAS finds the information) then it could be possible to find themes about what students are saying/thinking. "College food is gross #cmu" or "That computer final was tough #cmu" are just two examples where we might be able to find out real-time information instead of relying on survey data that only comes in once or twice a year. Cool stuff.
Oh, and in fairness, college food is typically bad anywhere you go. I've eaten a number of times on campus and the food has actually been quite good -- my example was just a typical tweet a student might send out.
Using SAS to analyze facebook status updates or twitter messages has some very interesting implications for Institutional Research.
Insert Arrested Development quote referencing the $80k spent on Buster's M.A. in Cartography:
Michael: "Hasn't everything already sorta been discovered, by like, Magellan and Cortez?"
Buster: "Oh yeah, yeah...."
Michael: "All those folks."
Buster: :Those guys did a pretty good job. But there's still ... you know …"
One of my main goals at the SAS conference in Vegas was to get better at making graphs and maps. I think I'll give you a tour of how my mapping abilities have progressed. Warning: the first map I show you could hurt your eyes and cause uncontrollable laughter. Honestly, I had no clue what I was doing.
MY EYES! THE GOGGLES DO NOTHING!