A couple of weeks ago, my colleague Don Gray discussed how to get the most value out of your data. He stressed the importance of preserving both enrolled and non-enrolled data to facilitate future data analysis. But there is more to it than just storing your data to perform meaningful analysis. Data needs to be captured, transferred, and recorded accurately; otherwise you will quickly be looking at a "garbage in, garbage out" situation. In order to avoid putting yourself in this "trashy" state, I have listed several common data errors that I have found in my experiences as a researcher:
- Deleting non-enrolled aid: The only way to do proper analysis on how financial aid impacts enrollment is to look at the money offered to all admitted students, regardless of their enrollment status. Looking at offers for enrolled students only will give you a false sense of how your dollars actually influenced enrollment decisions.
- Not isolating different applicant types: Freshmen and transfers are two totally different populations. They could respond differently to grants and scholarships and they typically yield at different rates. Therefore, these two populations should be tracked separately.
- Keeping admissions information for enrolled students only: The further up the admissions funnel you keep this data, the better analysis you will have. For example, if you are looking at how a campus visit affects enrollment when you are only looking at enrolled students, you may be falsely correlating the two ideas. However, if you were looking at this for all admitted students, you will be painting yourself a better picture.
- Having too little data to do proper analysis: Say you just started to capture the legacy status of your incoming freshmen last year, and you see that it was lower than you expected. Does this one year give you an idea of your overall legacy population on campus? Absolutely not. You cannot have a true sense of it until you have multiple years of data.
- Too much data in too many places: Is the same financial aid or admissions data stored in more than place on campus? If so, does reporting from these two different sources create different results? You might want to look into streamlining your data storing process. The fewer places the data is stored, the fewer errors you’ll have, and the more relevant your analysis will be.
What is your institution doing with its data? Are you getting the information you need to meet your enrollment goals?
Image © iStockphoto.
About the author: Statistical Programmer & Analyst Tim Schuldt came to S&K in July 2009. Before joining the team, he was an Equity Research Intern with Credit Suisse in London, a Client Representative at Canandaigua National Bank, and a Financial Services Representative with First Investors Corporation.
Tim graduated from St. Lawrence University in 2008 with a B.S. in Mathematics, a B.A. in Economics, and a minor in Statistics.