|
|
SPSS Folks,
I wonder if anyone might have suggestions for a conditional data merge . I
have a unit-record data file with student level information from two sources
(student information systems (SIS) and FAFSA data). These two data sources
contain two types of campus identifiers: The Office of Postsecondary
Education ID and the Integrated Postsecondary Education Data System (IPEDS)
ID. SIS data are for all students who enrolled in public institutions in a
given region. FAFSA data are for all students from the same region who
applied for aid, though data is available for students who enrolled in
anywhere. These SIS and FAFSA data are merged together into a single flat
file for a given year (let's call it the annual file).
I have institutional level data from IPEDS that I want to merge on to the
annual file for a particular analysis. I want to match cases first with the
OPE IDs, then using the IPEDS ID (for all those students enrolled in public
institutions, but who did not apply for aid). Any suggestions about ways to
do a conditional merge without splitting the data file?
Here is what I have done/ thought of:
. I am using MATCH FILES /TABLE function
. I can merge with the annual file using OPE ID or IPEDS without any
problem
. I could create two IPEDS files with different variable names,
merge with the OPE ID then merge the other file with the IPEDS ID, finally
creating a single variable (for example, 'sector_a' and 'sector_b' would
become 'sector'') that would be used in later analysis . However, this seems
somewhat inelegant and I am hopeful that some syntax wizards out there will
have a 'cleaner' suggestion.
. I have used the two key variables with the 'BY' command, the OPE
ID and the IPEDS ID, making sure the files were sorted ascending. If I
understand what this procedure does, however, it appears to match cases by
both variables concurrently, rather than sequentially.
Best,
Jake
|
Locked
