One of the continued challenges for preservation resources is the demand for objective data to make informed retention and withdrawal decisions. Discussions within the shared print and print repository communities have circled around the integral question pertaining to the selection process of books to be incorporated into a national shared print system, namely the minimum number of copies such a system must maintain. The challenge has been knowing the condition of those volumes that are withdrawn or retained, since all decisions have been based solely upon a shared catalog where partners do not have data to know the condition of others’ volumes. This conundrum led to a national research initiative funded by the Mellon Foundation “Assessing the Physical Condition of the National Collection” to create a baseline of understanding of the actual condition of the national collection in research libraries and collections. The project undertook an extensive assessment capturing data from 500 “identical” volumes each from 5 different research libraries and analyzing the dataset to answer the following questions: What is the general condition of library collections in the 1840-1940 period? Can the condition of collections be predicted by catalog or physical parameters? What collection assessment tools help determine a book’s life expectancy? Filling the gaps in knowledge for understanding the physicality of our collections is helping us identify at-risk collections and explain the high percentage of dis-similar “same” volumes due to the impact of paper composition. Predictive modelling and simple assessment tools allow more accurate prediction of good and poor-quality copies of books, as well as what is typical and atypical for specific decades.