Survey methodology

I've recently been asked to assess a large collection of
water-damaged animation production backgrounds (original watercolour
paintings and related cels) and have some questions concerning the
survey methodology I am using. To begin, the collection is
consistent throughout in terms of the support materials and media;
based on this information I have decided to use a random sample
methodology. It was suggested to me by the owners of the collection
that I look at a higher percentage of the more seriously affected
portions of the collection than those that were less affected. To my
way of thinking this would essentially create two separate surveys.

Since I am not a statistician I can't say for certain whether or not
there would be any appreciable difference in my results if I went
ahead and looked at the collection in this manner (or if the results
would be skewed) or even if it is worth making a distinction between
the two differently affected portions of the collection (though I am
inclined to do as my client wishes and look at the backgrounds as
essentially two distinct collections in order to waylay any concerns
they have).

I've read various articles on this subject in the past but am having
trouble tracking down said articles through regular channels (i.e.
BCIN). Any advice regarding my specific questions or links to
insightful articles would be most appreciated.

Andrea Reichert
Paper Conservator in Private Practice
2712, rue Springland
Montreal, Quebec
Canada


                                  ***
                  Conservation DistList Instance 20:35
                 Distributed: Monday, January 15, 2007
                       Message Id: cdl-20-35-015
                                  ***