On Aug 20, 2008, at 2:44 PM, Jim Betz wrote:
If I'm way off base here please enlighten me as to why ...
There are just shy of 1500 members on this list. It would see to me that
if those of us who are interested took just a few hours each to inspect the
photos in our books and compile
RR, Car#, location, RR photo taken of/on,
date, book title, and author (page?)
info into some common spreadsheet type that it would not take too long
before we would have a very credible sample of freight car distribution
that would stand the various tests of variables such as year, season,
branch or mainline, industry or yard, etc. I guess it makes sense to
only do 'non home road cars' but we could easily do several types
of freight car. But one of the great parts about it is that we
introduce a fair amount of randomness - from just using so many different
photos - that will eliminate biases of stuff like the photographer's
preference for a particular road, regional biases, etc. And I don't
think you have to struggle to document each and every car - if you
can't see the number relatively easily just go to the next one/pic.
We might even end up with enough information and entries on several
cars that somehow get repeated over the years that we can get an
idea of how widely cars ranged.
About all we'd have to do would be to coordinate so that we don't
duplicate each other's work. Someone would have to combine the work
into one large aggregate and someone would have to do the analysis
of the data once we have it.
I think that there might be several problems with this approach but the biggest problem is that you need to evaluate the numbers for one place in one time period. You can't combine different places. If you do, you are creating an "average" and will end up, by default with the national average. Data from individual railroads could be analyzed for that road, but that gains nothing from combining the analyses into a common file.
Your database would absolutely tell us how far certain cars roamed, but the fact that cars roamed nationwide should not be a surprise (see the archives for lots of examples) <G>.