Condition of the University of Maryland Libraries Collections
Contents
Introduction
The following report highlights findings about general collections condition from the 1989 Report called "Preservation Survey of General Collections" by Paul Koda, former Preservation Officer of UM Libraries. This condition survey analyzed and described the depth and scope of deterioration in the collections.Only the general book and periodical collections were surveyed. Special collections, which require special treatment, and reference collections, which have extraordinary use, require different survey techniques and were not included in this survey.
Methodology
The methodology for the Survey was adopted in part from the studies conducted at the University of North Carolina at Chapel Hill, the Smithsonian Institution Libraries, and Syracuse University. The major differences were in the larger sample size (3183 items for UM) and the shorter amount of time used to conduct the Survey (only two-and-one-half days for UM
- Sample: As of June 30, 1988 the holdings of the UM Libraries identified for the survey comprised 1,711,051 cataloged volumes, excluding the volumes in Reference, Reserve, Special Collections, Special Locations (eg. off-campus storage areas), Non-Print and Government Documents. The initial step in the Preservation Survey Project was to determine the sample size that would be drawn from each of the following libraries whose holdings are shown in parentheses:
- Architecture (35,411)
- Art (59,461)
- Chemistry (52,132)
- Engineering and Physical Sciences (EPSL, 200,403)
- McKelding (1,096,965)
- Music (39,546)
- Undergraduate Library (227,133)
- Random Numbers: The Libraries consulted with the UM Computer Science Center's statistical services unit to generate random numbers that could be assigned to the total shelves in each of the libraries involved in the study and to individual volumes on shelves based on 25 volumes per shelf. In the absence of a volume to which a random number could be assigned, the next immediately available volume was assigned for retrieval. The elements included in the stratified random sampling were:
- location and floor (one digit)
- range (one to three digits)
- section (two to three digits)
- volume (one to two digits)
- Labeling: In preparation for the generation of random numbers, every floor, range, section and shelf in each library were counted and their configuration recorded. The information was used by statistical services to generate numbers, so random volumes could be retrieved for examination and data collection.
- Questionnaire: The questionnaire was an adaptation of the form used first at the University of North Carolina at Chapel Hill and then at the Smithsonian Institution Libraries. It was designed to meet the following requirements:
- simple to fill out, using a check system that was easy to key into a computer
- on one side of standard 8-1/2" by 11" piece of paper
- gather yes or no information for computer compilation, cover and container materials and condition, and textblock materials and construction
- use a three-part design that captured storage information, cover and container materials and condition, and textblock materials and construction
- arrange the elements in a pattern that reflected a "natural" order of examining a book, from taking a book off the shelf to determining its textblock consolidation
- use standard preservation and conservation nomenclature
- Sampling and Statistical Methodology
Sampling
The Fall 1987 ad-hoc study, while supporting the general hypothesis that the Library's collections were deteriorating, did not reveal "p", the proportion of easements in the population (i.e., McKeldin's Stack collection)possessing the characteristics contained on the survey form. Since "p", the parameter to be estimated in the course of testing the entire population of the UMCP Libraries remained unknown, it was decided that the minimum size for the sample "n" would be determined by using the formula for determining sample size for estimating proportions:
n=(z2pq)/d2
Where:
q = p
n = sample size
p = proportion of elements in the population possessing the survey characteristics
z = reliability factor
d = desired interval
The formula was derived from standard statistical measures designed to estimate "p" within 0.05 probability with a 95 percent confidence level.
As of June 30, 1988, the holdings of the UMCP Libraries to be surveyed comprised of 1,711,051 cataloged volumes. The initial step in the Preservation Survey Project was to determine the size of samples that would be drawn from the following libraries whose holdings are shown in parenthesis:
- Architecture (35,411)
- Art (59,461)
- Chemistry (52,132)
- Engineering and Physical Sciences (ESPL 200,403)
- McKeldin (1,096,965)
- Music (39,546)
- Undergraduate Library (227,133)
Application of the formula revealed a minimum size of 384 volumes required to estimate the proportions in the total collection possessing the characteristics of interest listed in the survey form. It was decided that the following sample sizes would be drawn from each library as follows:
- Architecture (N; holdings surveyed = 35,411; n = 400)
- Art (N; holdings surveyed = 59,461; n = 400)
- Chemistry (N; holdings surveyed = 52,132; n = 400)
- ESPL (N; holdings surveyed = 200,403; n = 400)
- McKeldin (N; holdings surveyed = 1,096,965; n = 800)
- Music (N; holdings surveyed = 39,546; n = 400)
- UGL (N; holdings surveyed = 227,133; n = 400)
- Staff Utilization: Seventy-six persons participated in the survey, ranging from undergraduate student assistants to associate directors of the Libraries. The retrieved sample books and periodicals from each library were placed on book trucks and placed at the ends of tables in pre-designated areas. Surveyors selected volumes from the book trucks. Information was gathered and recorded; completed Survey forms were placed in the volumes; and the volumes were put back on the book trucks.
The sample size for McKeldin was 800 since McKeldin's holdings are heterogeneous in that they contain, more than the other libraries, materials representing all subjects covered under the Library of Congress (LC) Classification Schedule. Moreover, the actual physical distribution of the collections are spread out over seven floors, with one floor containing materials covering the entire range of LC subject classes, and with the remaining six floors containing materials in call number order. Thus, a stratified random sample size of 800 was deemed advisable for the survey.
- Brittle Materials: The extent of brittleness in paper-based library materials is determined by folding a paper corner back and forth until it separates from the leaf, but not to exceed five folds. (This is done only for test purposes on sample volumes and never on special collections materials). A double-fold (once forward and backward) is the standard and is called a complete fold. Brittleness is usually defined as two or fewer complete folds. In using this standard, the Survey found that 16.4% (280,612 items) in the collections is in jeopardy when anyone simply turns the leaves of a book or periodical. If the Libraries decided, for whatever reason, to call brittle those items which are Zero Folds (i.e. it was impossible to fold the paper even once without it breaking apart), then there are 111,218 items literally too brittle to use.
- Acidic Materials: The survey showed that 28.2% of the paper in the collections are highly acidic and 46.8% of the paper in the collections have some acid. These figures are high but not outside the range reported by other academic research libraries.
