QUEBEC-RESEARCH-L Archives

Archiver > QUEBEC-RESEARCH > 2004-04 > 1081796053


From: "Bertrand Desjardins" <>
Subject: PRDH Problem
Date: Mon, 12 Apr 2004 14:55:00 -0400


Bonjour everybody,

A few weeks ago, I celebrated my 31st anniversary as a member of the
Programme de recherche en demographie historique (PRDH for short), of which
I am now in charge. The PRDH is a University research project aimed at
bettering our knowledge of the French-Canadian population of Quebec from the
initial peopling of the territory by French migrants, while training
students to become researchers in the field of demography. To accomplish
these goals, we created a data base made up of a transcription of the main
elements of the baptisms, marriages and burials from the catholic parish
registers and of the individual biographies and genealogies which can be
established from them.

Among other things, I bear full responsibility for the genealogical
decisions of the project, our data management, our displays, etc. I am very
proud of our work and I do not hesitate to write that I feel we have
produced the best information system of the kind on a specific population in
the world. So you will understand when a message like the one below is
brought to my attention, which includes the phrases "These are very serious
fielding problems with the PRDH, and should be
remedied as it is causing much misinformation and obvious confusion" and
"But, while we are at it, let me issue a warning! The PRDH is pretty solid
in compiling family associations in the pre-1750 period, but is frequently
(understandably) incomplete and down-right incorrect in the post-1750
period", even if I am at home enjoying the day off with my family, I hasten
to react...I will successively address the two issues brought up by Mr.
Barton.

[From: "Kevan Barton" <>
Subject: Correction: RE: PRDH Problem
Date: Sun, 11 Apr 2004 05:34:36 -0400

The PRDH has a tendency to mislabel a number of vital events depending on
how you choose to view the data in your pull. It is a database field
problem. I think the problem centers on the family view offered where it
gives familial data in four columns: Birth, Marriage, Death, and First Name
of the Child. This is what I see very frequently:
For a Birth in the family-view:
If there is no birth date in the baptism certificate, then the birth date in
the family view defaults to the baptism date. However, the data base in the
family-view will field the baptism date as the birth date. This is a
serious problem and should be remedied as this is causing much confusion.
Without looking at the specific certificates, you can't know for certain if
the fielded birth date in the family view is the actual birth date, the
baptism date, or both (born and baptized on the same day).
For a Death in the family-view:
If there is no death date in the burial certificate, then the death date in
the family view defaults to the burial date. However, the data base in the
family view will field the burial date as the death date. Without looking
at the specific certificates, you can't be certain if the death date in the
family-view is in fact the death date, the burial date, or both (died and
buried the same day).
These are very serious fielding problems with the PRDH, and should be
remedied as it is causing much misinformation and obvious confusion.]

All the above is true. There are two dates associated with both Baptisms
and Burials in the Parish registers, that is birth and baptism, and death
and burial. Catholics were obliged to baptize as soon after birth as
possible and for obvious reasons, burial was usually close to death, so
dates are usually close from one another.
In its repertory ovf vital events, the PRDH has always transcribed both
dates when available; so any case where there is only one means the priest
did not specify the other (ususally the birth or death date is then
missing).
As demographers, we had to have a birth date for each of the individuals so
their age at different events could be ascertained. So we defined a variable
for that which stored the birth date if it was available, and the baptism
date as a proxy if it wasn't (knowing it was cetainly very close to it).
Same for death/burial date.

This date is the one which is displayed in the Birth/Deat columns. Thus
"Birth" does not identify "Birth date", but "Information about Birth date,
which can be baptism date in some cases". One has indeed to go back to the
transcription of the corresponding act to know exactly. I readily admit this
can cause confusion.

The late Rene Jette was very adament this should be remedied and it was on
the CD version of the PRDH families 1621-1765, where both dates are
distinguished in the individual files (but not in the family files, so not
to overcharge the displays). The site will be updated in a few weeks and I
will have the displays of the site changed to remedy the situation.

[But, while we are at it, let me issue a warning! The PRDH is pretty solid
in compiling family associations in the pre-1750 period, but is frequently
(understandably) incomplete and down-right incorrect in the post-1750
period. An excellent example is the marriage association of Ignace Levesque,
husband of Marie Catherine Raymond and Marie Genevieve Laplante. The PRDH
ascribes another marriage to Ignace even though this other lady is still
producing babies from her first husband. This is really bad news and I've
seen this bad connection cropping up in data bases (i.e. the Ancestral
File). The PRDH is not performing well in some contexts and is being the
cause of bad data! ]

I have more difficulty with this affirmation. It is not acceptable to state
one mistake and write "is FREQUENTLY incomplete and down-right incorrect in
the post-1750 period". Imperfections in the data in the parish registers,
the multiplication of homonyms with time and the necessity to handle hundred
of thousands of informations necesseraly cause mistakes to happen. Everybody
benefits from the work of their predecessors. Tanguay was the first to do
systematic Quebec genealogy 1621-1750s and made a lot of mistakes. Godbout
came after him and made less. Jetté came after them and made even less for
the 1621-1730 period (but committed a few). We came after Jetté...But nobody
had done 1760-1799 completely and systematically before us, hence...
Look at the numbers : there are some 75 000 marriages before 1800, which
means some 150 000 filiations have to be established. If 150 mistakes are
found in the PRDH information on filiations, you have a 99.9% exact data
base as to that part of the information. Find 1500, and you have 99%. I can
vouch you will never finf 1500 bad filiations in our data, so I can't accept
our information would be equated with "FREQUENTLY incomplete and down-right
incorrect".
In the case stated by Mr. Barton, the priest wrote in the marriage of Ignace
Lefevre to Genevieve Laplante he was the widower of Marie Boucher. This
information was processed without benefiting from the linked Baptisms, so a
mistake was made where Ignace was confused with the only Lefevre-Boucher
couple around. Actually, Ignace's former wife was Marie Raymond, whose
mother was a Boucher, inducing the priest to write a faulty name in the
marriage act. I have corrected the data and it will show up correctly in the
updated version of the site in May.

I would argue the best thing to do when one finds a mistake is not to cry
wolf and issue a warning the PRDH is being the cause of bad data, but to
write it in to us, so we can make the correction, to the benefit of
everyone. This said, I recognize our mistake and...I do believe there are
not many of them around!

[Comments wanted.]

You got them! Please note I appreciate people working seriously at genealogy
and as the years go by, the information will keep getting nearer to
perfection...if such a level exists!!

The best to you all,


BERTRAND DESJARDINS Ph. D.
Prog. de recherche en démographie historique
Centre interuniversitaire d'études démographiques
Dép. de Démographie, Université de Montréal
C.P. 6128, succ. centre-ville
Montréal (Qc), Canada H3C 3J7
Tél.: (514) 343-7613 Fax: (514) 343-2309



This thread: