[Babase] PREGS conversion error - no conceive date
Catherine Markham
amarkham at princeton.edu
Wed Mar 1 13:22:23 EST 2006
Hi all,
I'm writing to follow-up on a recent conversation with Karl about some
conversion errors popping up in PREGS.
The PREGS table has 4 columns: pid, parity, conceive, and resume. In
the FoxPro version of Babase, the conceive and resume columns in PREGS
list cids that link to the CYCLES table so that the user can get the
ddate/conception date for each conceive and the tdate/resumption date
for each resume.
The conversion error at hand is that there are 40 pids in PREGS that are
linked by the conceive field to CYCLES rows without ddates. In other
words, these pids have a cid listed in their conceive field of PREGS,
but that cid has a blank ddate in the CYCLES table. In most cases,
tdate and mdate is also blank.
After some quick querying (will check more thoroughly before making
changes), most of these 40 pids are associated with live births - only 2
are for fetal losses. Also, nearly all are for pids in the 101, 102,
103, etc. series indicating that the pregnancies probably occurred near
the start of monitoring.
Regarding birth dates and matgrps, the following was a clear pattern for
each of the errors. Again, seems to indicate to me that these were
births near the start of the project or - with Nyayo's - during gaps in
monitoring. Does that sound right?
Born into Alto's group: births from 1968-1971
Born into Hook's group: births from 1975-1981
Born into Lodge group: births from 1982-1985
Born into Nyayo's group: births from 1992-1994
Assuming that we do not have detailed sexskin data to determine exact
ddates for each pregnancy, one solution might be to enter ddates for the
live births based on birth minus average pregnancy duration (Karl and I
though this was possibly an option we had discussed before). Obviously
we'd have to do something different for the two fetal loss cases. Also,
would the average gestation be calculated separately for the Lodge group
pids?
Finally, Karl mentioned that we could code these calculated ddates in
Babase as "estimated". Karl, is the column marking estimated dates
something that will be a lot of extra work to program?
Sorry for the lengthy email,
Catherine
More information about the Babase
mailing list