The Health Sciences Center Library at Saint Louis University holds
approximately 120,000 volumes devoted to supporting curricula for
schools of Allied Health, Medicine, Nursing, Orthodontics, Public
Health and the Center for Health Care Ethics. In addition, it
supports the patient care activities of a tertiary care hospital and the
active research of a diverse faculty. It is part of a university system
including law, divinity, general academic, and aeronautics libraries.
The Library has been free from budget cuts since the 1970s.
However, it has not enjoyed adequate budget increases to meet
current needs, let alone meet additional demands generated by new
programs. Over a short period of time, the Library was asked to
support new programs in occupational therapy, dermatology and an
emergency medicine residency program, as well as expansions in
nursing and public health. The university libraries are routinely
asked to submit approximate budget figures for what it will cost to
support these new students and faculty. In the past faculty were
queried for acquisition suggestions and figures were quoted which
did not include media, computer assisted instruction (CAI),
Interlibrary loan (ILL) or other library activities necessary to support
new users.
Before developing a new acquisition's formula, a search was
conducted to find one that was predictive. Clapp - Jordan[1] and
Kohut[2] are familiar formulas in the academic community. These
formulas are used to take a known amount of money and divide it
across disciplines in an equitable fashion. The following data are
usually included: total literature output of the discipline and its
average price, university department size, credit hours taught,
library use, and total department majors. The formula is applied to
the existing materials budget and the money is allocated to each
department. Fund accounting is the process of monitoring the
allocations.
While exploring existing formulas,[1 -5] it became apparent none
were adequate. The formulas provided some valuable guidelines,
however they help only if the amount of money available is known
in advance. These formulas do not predict money needed to support
a new area of study. The following formula is predictive and
captures the impact of various kinds of users and disciplines on the
use of the library's resources.
Unique institutional elements gathered from local data sources,
include: Numbers of faculty, undergraduates (including medical
students), master's degree and Ph.D. students by discipline; Impact
value on the use of the book and journal collections Impact value on
the use of the audio-visual collection, by format Average cost of each
media format by NLM classification number
All acquisition formulas acknowledge a difference in demand by
different classes of users by weighting users based on faculty status
or level of degree program. For the purposes of this formula, Budd
and Adams'[3] relative use factors are modified. The number of
faculty is multiplied by a relative use factor of 2.5, the medical
undergraduate number by 0.2, other undergraduates multiplied by
0.5, the master's number by 0.7 and the doctoral number by 3.0. The
number of faculty and graduate students by level and discipline
came from an internal university planning guide.
Because medical students utilize a broad cross section of the
collection, their total number is divided equally across disciplines.
The sum of the relative use factors times the number of users by level
represents the total user equivalent by discipline.
The formula balances a small department, dominated by faculty and
doctoral students, with a large department, comprised
predominantly of undergraduate students.
Use by subject area varies. In this Library there is little activity in
parasitology (QX) and heavy utilization of neurology (WL).
Therefore, it is inappropriate to apply the same value to each
discipline when they are used to vastly different degrees. < P>
A library that circulates or has access to reshelving data for its
journals could employ those statistics to determine usage in each
discipline. Statistics from book circulation may also be employed.
Because this Library neither circulates its journals, nor maintains a
reshelving count, nor has an automated circulation system to
account for book usage, those data are not available. Employing an
ILL statistic assumes that users who have discovered the ILL service
are the same people who make demands on the library's collections.
In 1984 this author performed a study at another university which
demonstrated the greater the circulation the greater the interlibrary
loan demand by discipline[10]. As part of copyright compliance,
documentation is available for journal articles ordered over
PHILNET, the Washington University School of Medicine
maintained interlibrary loan network and NLM's Docline. Call
letters were assigned manually to each borrowed journal based on
the first significant subject in the title. For example: Obstetrics and
gynecology was assigned WQ, obstetrics, not WP, gynecology, and
the Journal of anatomy and physiology was assigned QS, anatomy,
not QT, physiology. The inequities are not significant when the
thousands of interlibrary loans for one year are taken into account.
Using the ILL statistic inserts the importance of the ILL costs into the
formula. The percentage of total interlibrary loan for one year by
discipline, provided a use impact factor which is multiplied against
the average book and serial value.
Audio-visuals provide another challenge because their use is not
recorded. An alternative procedure needed to be devised. The
Library's Educational Media Department published catalogs,
arranged by format, of its current collection. Call letters were
assigned to slides, software and videotapes in the same manner as
they were for the interlibrary loan requests. Impact factors were
developed. The rationale for employing an impact factor based on
the current media holdings, rather than on use, is dependent on the
fact that virtually nothing in these formats is added on speculation.
Every purchase is made with faculty input. Media materials mirror
the curriculum for which they are intended or they would not be
owned. It is a collection finely tuned to its user population.
Developing an impact factor for the user population and the average
price by material type does not adequately reflect the importance of
the materials to the collection or to the clientele. Nearly 80% of the
budget is devoted to the purchase of serials. Books account for about
18% and audiovisuals the remaining 2%. Further, health sciences are
driven by journal literature. Therefore, it is essential to weight the
impact values to indicate their relative importance in the library.
Each is weighted in accordance with its share of the materials
budget. Therefore, the weighting factor for journals is .80; for books
.18, and for audio-visuals .02.
For example QS books cost 1.05 of the average cost of health related
books. Column D is the User Impact Factor derived from
Interlibrary Loan demand. It indicates the demand on the collection
by discipline. Of the examples in Table I, QU (biochemistry) is most
heavily used. The Audio Visual Average Impact Factor reflects the
contents of that collection. The three media impacts were added
together and divided by three to determine the average use (impact)
of educational media on the discipline (Table 1, Column E). In this
example QS is heavily represented in the collection and presumed to
at by multiplying the Journal Value Factor by .8 and adding that to
the Book Value Factor which is multiplied by .18 indicating their
relative value to the health sciences. That total is multiplied by the
User Impact Factor which is finally added to the Audio Visual
Average Impact Factor multiplied by .02. In this example QU has
the highest Total Relative Cost : (.8*1.78 + .18*1.21)*1.3 + (.02*0.53) =
2.15. It was necessary to incorporate the impact factors and value
factors into one statistic. Column F of Table 1 unites the two Impact
Factors: (1) the ILL data (Column D) and (2) audiovisual holdings
(Column E), with the two Value Factors: (1) the amounts over or
under the average price of the books and journals (Columns C and E)
, and (2) the value of the type of materials to the library's clientele.
Journals were factored highest at .8, books at .18 and audio-visuals
at .02. The formula reads: (.8B + .18C)D + (.02)E = F. The name for
this factor is Total Relative Costs. Cost information for media is
unavailable.
In Table 2 Columns G, H, I, J and K are the actual number of faculty
and students by grade level derived from either the Planning guide
or by dividing the medical students by all medical disciplines. The
relative use factors presume that the use is higher for faculty and
increases by degree level for students. Students in a doctoral
program have the highest use factor (3.0), which exceeds the faculty
use factor (2.5). Total user equivalents by discipline is the sum of the
number of faculty and students (by type) each weighted by the
respective relative use factor (boxed under the column labels in
Table 2). For example, 37 faculty members in WY (Nursing) are
multiplied by 2.5; 400 undergraduate nursing students multiplied by
0.5; 219 master's students multiplied by 0.7 and 0 Ph.D. students
multiplied by 3.0. These are added together to equal 445.8, Column
L: (G*2.5) + (H*0.2) + (I*0.5) + (J*0.7) + (K*3.0) = L (Table 2).
Column M = F/26.96. Column N = M*$500,000 (Table 3).
The Allocation per User Equivalent is derived so that the number of new users expected can be multiplied by that dollar figure. That figure is arrived at by dividing the Materials Allocation, Column N, by discipline, by the Total User Equivalents, Column L (Table 4) : N/L = O.
Nine underlying assumptions were made during the creation of the
formula. 1. Data from outside vendors can be standardized, and
new users and new programs can be categorized using the NLM
classification system. 2. Medical students use the collection evenly
across disciplines as they proceed through four years of medical
school. 3. Each discipline is used differently. 4. ILL demand
mirrors in-house library use by discipline. 5. The quantitative
content of the audio-visual collection reflects the demands made
upon it. 6. The user weighting statistic is valid, e.g. use varies by
user class. 7. Journals in any format are the most important
information sources currently accessible in the health sciences
library. 8. The budgeted percentages for materials is valid as a
relative weight for the materials. 9. Past use of the collections and
interlibrary loan demand by NLM derived subject will predict future
use. These assumptions will be tested over time as trend data from
management reports of the Integrated Library System are collected.
To date the formula has been employed in three ways. The university's graduate board requested financial information as it evaluated the cost of adding several programs in public health. After ascertaining the number of new faculty and students predicted by grade level, reasonable financial projections were made. A clinical agency of the University supports the Library with an annual contribution. The cost of the clinical use of the collection was determined to justify a request for increased support. By adding together the total dollar amounts for the typically clinical subjects, it was demonstrated that about half of the materials budget is devoted to clinical purchases. This, with other documentation, resulted in a significant increase in the agency's contribution. Recently, an accrediting body inquired as to what was spent on that discipline in a year and, because of the formula, a prompt response was possible. Further refinements of the formula are planned as additional information becomes available. Data are being collected so that an audio-visual value factor by format can be included. Further, circulation and in-house reshelving data should become available in the next 12 - 24 months and those statistics can be used to validate and enhance the user impact factor. Hopefully, other libraries will test the formula and report on its validity.
A copy of the spreadsheet with the formula embedded will be provided to those who send a floppy. The spreadsheet portion of Microsoft Works 3.0 was used on a MACINTOSH.
Acknowledgments The author is grateful for the statistical,
mathematical, moral and editorial support of A. Kent Rissman, Ph.D.
and the editorial comments of Kathy Gallagher, M.S.L.S. and Suzy
Conway, M.L.S. EBSCO graciously provided a forum for discussion
and refinement of the assumptions presented in the formula.