[WikiEN-l] Age fabrication and original research

[Rob]

The reason I balk at using the SSDI or the census is I don't think we
should be using primary sources in this manner.  There are numerous
pitfalls, including many errors of spelling and fact, to using these
sources. Historians and journalists should be evaluating these
sources, not us.  In this particular case, editors are using a primary
source to disprove reliable secondary sources, which are plentiful and
unanimous (until now, see below) when it comes to the birthdate.
Isn't this the kind of primary source research that we always
discourage Wikipedians from doing?

The fact that original secondary sources were wrong in this case is
immaterial.  Errors in secondary sources should be a reason to dig up
more secondary sources, not to make a point using primary ones.  In my
initial message I expressed doubt that a historian or journalist would
ever write about such a minor point.  I was wrong; the relative posted
a link to a Fortean Times article which cited as sources for the real
birthdate a newspaper article from the 70s and a biography of Dixon.
When I get my hands on those through the magic of interlibrary loan,
I'll update the article accordingly.  This has convinced me that we
shouldn't be using those primary sources.  It's tempting, certainly.
An editor can just type a name into the SSDI webpage or a census
database and say "I've done research!".  That's easy, but tracking
down decades-old articles and biographies requires going to the
library or having a familiarity with databases you can only find in a
library.  But that's the kind of hard work we should be doing.