[Babase] Conversion Errors - Round 1
Catherine Markham
babase@www.eco.princeton.edu
Fri, 26 Aug 2005 10:08:36 -0400
Hi all,
I have finished correcting the conversion errors that Leah emailed me on
Tuesday. Leah's list of errors is cut/pasted into this email and my
notes are inserted.
Details of steps taken with corrections are fully described in the
Babase Changes document (babase_changes.doc), but I wanted to make note
of the errors/changes in the listserv archive as well.
Catherine
Leah Gerber wrote:
> CONVERSION ERRORS
> August 22, 2005
> Leah Gerber
>
> 1. CYCLES Sname = TT , CID’s 1827, 1828, 1829
>
> I queried cycles for those cid’s and the sname is TTX for all three. Not
> sure if this is still a relevant error. Could this have been fixed in
> Babase since Karl downloaded the data?
Yes, I changed the sname of TT to TTX in Biograph, Members, Census,
Cycles, Cycstats, and Repstats on 14 April 2005. Date of Karl's data
download discussed on Babase listerv (subject name of "Babase date for
conversion comparison").
> 2. DEMOG Cenid = 1673602, 1676301
>
> There are two entries for these cenid’s.
Corrected. Discussed on Babase listserv (subject name was "Dupe Cenids
in Demog").
> 3. UNO (F) Censused after her statdate of 09/05/94. Group 3.0
Corrected. Problem was that UNO's death was never entered in Babase.
> 4. OVA (F) Censused after statdate of 28/03/78. Group 1
Corrected. Appears to have simply been a data entry mistake.
> 5. SUR (F) Censused before birthdate. Group 1.1. It looks like
> the census for SUR on 20/10/93 (corresponds to birth) was entered as
> 20/10/92. The group was not actually censused on 20/10/92. So, I believe
> she is in the database for a year before her birth due to this error.
> Not sure if others were placed in grp 1.1 on this day by accident as well.
Corrected. SUR's birth date was 4 July 1993. I deleted all rows for
her in CENSUS and MEMBERS prior to this date.
> 6. There was a conversion error “no name when sname not NULL, not an
> abortion”. I checked all names that were blank when an sname existed.
> All of these had a sex of “U” so I think they are abortions. I am
> thinking this error has already been corrected. Have you done anything
> with abortions recently?
Corrected. None of these were actually abortions - this was a mixed bag
of issues mostly resulting from Lodge group births during periods of
spotty monitoring. Solution in most cases was to add info to the name
field.
> 7. WAD (F) Censused one day before her birth (02/09/93), group
> 2.0. She is also in members that day
Corrected. Appears to have simply been a data entry mistake.
> 8. WAV (M) Censused on the two days before his birth (20/09/93),
> group 2.0. He is also in members these days.
Corrected. Appears to have simply been a data entry mistake.
> 9. 12/20/95, group 3. Many individuals censused twice. One with a
> status of “A” and one with a status of “M”.
> 10. 12/31/95, group 3. Many individuals censused twice. One with a
> status of “A” and one with a status of “M”.
Both 9 and 10 corrected. These were both problems that resulted from
the data entry of incomplete censuses on 20 December 1995 and 31
December 1995 as (1) manual presences (status = "M") AND (2) complete
censuses with presences/absences. The Census data now reflects only
confirmed presences - it does not infer data for individuals not
observed and does not mark absences.
> 11. 9/18/00, group 2.2. ATO and LAL censused twice. One with a status
> of “A” and one with a status of “M”.
Corrected. Error had to do with the not marking these new infants as
present in the group on the first day they were observed (first
observation date for both was 18 September 2000).