Decisions, Decisions

I’ve been working with data from the National Student Clearinghouse (NSC) for a while now. A lot of wonderful information can be found in the NSC data, but the detailed return file can sometimes be a bit difficult. There are so many ways the data can be sliced, and it can sometimes be challenging to determine how best to work with the data to present meaningful information to stakeholders.

For those unfamiliar with NSC data, the detail return file for a Subsequent Enrollment (SE) search, lists all enrollment records for a person that occurs during or after a specific search date. This means that, depending on the search date, each person has potentially many, many rows of data in the return file, each reflecting a different enrollment. Students will have a record for each enrollment term at an institution, but they will also have multiple enrollment records in a specific term if they took classes at more than one institution that term (this can be true even if the student dropped a class at one of the institutions), or if they were taking a class the same term at a school that reports parts of the school differently (for example, “School A” and “School A Graduate”). There are also separate rows for graduation records (and sometimes multiple rows for the same graduation).

One of the major decisions that need to be made when working with NSC data is the trumping order for which record or records should be selected for each student. Trumping schemes for which record or records to keep are based on the particular project; the same return file could be utilized in different ways for different projects, each with a goal of highlighting different information.

Example of what some of the pieces of an individual detail return file might look like:

For example, one project may focus on graduates of a particular graduating class or transfer-outs who had a similar final term (no matter when they first enrolled) to examine subsequent enrollment and graduation at other institutions. While it would be important to keep the graduation record at the other school since we also want to examine success in this project, for those who didn’t graduate, do we want to use their first chronological record, or the last? Do we want to select 4 year schools over 2 years (but what if they graduated from a 2 year)?   Do we want to include all records, or cut those who weren’t enrolled for x number of days (and does that number need to vary based on a summer vs. a fall/spring term)?

Another project may focus on the progression & completion of a cohort, which would utilize multiple terms of data. However, it would still need to be decided if records with fewer than x number of days of enrollment should be cut, and how to handle multiple records in the same term. Additionally, as start and end dates at schools differ (and a few have winter terms that can overlap with the spring terms of other institutions), clear term definitions need to be set while knowing that there is no perfect answer.

Over the next few months, I’d like to share some of the SPSS syntax that I’ve developed in working with NSC data. For my first project, I’ll be posting a short video explanation & syntax to using NSC data to examine the enrolling institution for students who were admitted to your institution, but who did not enroll.