�ݺ�ߣ

WEEDING WITH
ROBOTS:
Managing Collections in an
Automated Retrieval
System

Charleston Conference 2010
 Linda Masselink, Patricia Bravender
& Hazel McClure
Grand Valley State University
 Robert Kelly
Eastern Michigan University

Weeding with Robots: Managing Collections in an Automated Retrieval System

Why Automated Retrieval?
Space
Cost $$$$$

To Boldly Go Where No Man Has Gone
Before
 Challenges
 Content inventory
 Weeding

Random Access
 Sets
 Government documents
 Videos
 Microfilm

GVSU Loading Up
 3 sizes of bins 10” 12” 15”
 Barcodes
 Numbering by hand

WEEDING THE GVSU
BUSINESS
COLLECTION

Weeding
 Getting started…

Rick Lugg R2 Consulting
Legacy Print Collections
 From the Kent Study:
40% of books never circulated if they
did not circulate within the first 2
years of purchase
 No circulation in 6 years = potential
use is 1 in 50

Aggressive De-Selection
 No impact on users
 Content available at other places
 Susan Gibbons “user driven
collection”

Core Collection
 Core collection = books used
 Noncore collection = books not used
 WEED noncore

What holds us back from weeding?

Weeding Business Books
 Request for book information
 Target a small collection to start
 Pull each item from the retrieval
system

Discoveries
 Weeding can guide future purchase
decisions
 Discover “mistakes”

It is hard to weed good books
that are not used!

Considerations
 Key indicator - number of
circulations
 Books checked against:
 Bowker Book Analysis for relevance
 Web of Science for citations
 Choice Outstanding Academic Titles

Other Considerations
 World Cat listing for holdings at other
libraries
 Condition
 Cost of replacing

Next Group
 “No Circs” brought out first
 Slow process

Weeding
 Not a “someday” activity
 Schedule
 Limit your time

UNHAND ME YOU MECHANICAL
MORON:
WEEDING THE GVSU LAW
COLLECTION

The Steelcase Library in downtown Grand Rapids
was to become the home of GVSU’s 3,000 volume
law collection.

 GVSU accepted the Grand Rapids Bar Association’s
35,000 volume, 120 year old law collection.
 Bar collection accepted in
its entirety - was not weeded.
 Part of the collection put in
the reading room - majority
stored in the ARS where…

it became invisible to the human eye.

GVSU updated and added new material to the
collection for next six years
By 2007 apparent that:
 Use by the Bar was declining
 Use by GVSU students was low
 Many items duplicated (and paid for twice) in
on-line services such as Westlaw
 Cost of legal materials were skyrocketing
 A lot staff time required to keep updated

Why was use of law collection declining?
The following may have contributed:
 Relevance of materials to GVSU students- too
specialized
 Location of Steelcase library w/respect Bar
members
 Availability of materials on-line
 New barriers between users and the collection
 Difficulty of using law materials stored in an ARS
 New law school library built within ½ mile of

GVSU and GR Bar Association agreed that
GVSU could dispose of the collection.
 The Bar did not want the collection returned.
 Titles that GVSU did not want were to be
offered to Bar members and new law school
library in Grand Rapids.
 Any titles that remained after this procedure
would be shredded according to GVSU’s policy.

How the Robot Retrieved the
Bar
Collection from the ARS
 When collections were
merged, it was noted in the
catalog record of each item that
it had come from the Bar.
 This was done because the
merger was to be a three year
trial – if it didn’t work Bar had
the option of reclaiming its
collection.

First Weeding
 GVSU generated a list of every Bar donated
item from its catalog.
 GVSU determined which titles to keep and
which to discard according to its weeding
policies.
 The discard list was then circulated to the
local law library and members of Bar who
selected books.

 Shelving designated for holding selected
items.
 Monographs requested from the ARS,
reviewed and placed on a shelf for de-
accession from the collection by circ staff.
 If requested item was a multi-volume set,
only one volume was requested for review.

 Circ staff pulled multi-volume sets during
evening, low patron use times.
 Law books often consist of more than one
volume (often many) and each volume might
be located in a different bin.
 Minimum one minute per item, often longer
if many items are in queue.

Second Weeding
 Canceled and outdated materials next
priority –some from Bar list and some not.
 Worked from list of canceled material
generated by technical services by call no.
range

 Items processed in the same manner
as described earlier
 All items shredded
 6 months - 14,000 volumes

Third Weeding
 Remaining volumes from the original Bar list
being pulled and discarded using this process as
time and space permit.
 Majority out-of-date material collected over a
period of many years by Bar Association.

FULL STEAM AHEAD:
Weeding the GVSU Social Work Collection

Social Work Collection
 Had been weeded in past few years
 Still had a lot of chaff
 New librarian & turnover

Weeding 101
 Librarian was new to the profession & to
weeding
 Academic preparation
 Practical experience
 Jumped into weeding out of necessity

What I Did…
 Made lists based on
 Call number ranges
 Circulation stats
 Core lists
 Availability of quick/easy/inexpensive replacement
 Used this list to “pull” and withdraw books from
collection

& Why I Did It
 Provided a way to weed collection without
ready physical access
 Circ stats are a good predictor of future use
 Had to be cruel to be kind; weeding had to
happen
 I didn’t have many choices

Weed or Seedling?
 Encountered books from other collections
 Johnson Collection
 Criminal Justice
 Government Documents
 Multiple copies
 Books that were owned by many other schools
 Books that were in areas that weren’t very well
represented in the collection

Strengths of This Method
 Gets the job done
 Easily sectionable
 Lends itself to
 Automation
 Record keeping

Weaknesses of This Method
 Making lists was time consuming
 Dependent on circulation staff to pull items and
systems librarian’s team to get lists
 Making & manipulating lists was boring work
 Impossible to determine condition of items
 Maintains/encourages distance between
librarian & collection

Mistakes Made, Lessons
Learned
 Failed to consider some usage stats
 Didn’t have a nuanced enough LOC call
number list
 Some areas relevant to Social Work weren’t
covered
 List included many titles irrelevant to Social Work

Back to the Future: Ongoing
Plans
 Refining LC Call number list
 Solution to perpetual problem of lack of
weeding
 Periodic generation of lists & storage or
withdrawal
 Able to use some information gathered to
assist other librarians
 Method (with appropriate record keeping)
offers window into usage trends
 Using method with other collections

Possibility For Other Collections
 Public, Nonprofit and Health Administration
 Will need LC ranges
 Need to consider changing curriculum
 Need to be mindful of overlap of other collections
 Future possibility of automation/expanding
method across disciplines
 Relationship of weeding via this method to
Subject Collection policies

LOST IN SPACE
10 years later looking in all the wrong places

Why Inventory?
Robert G. Kelly 2010 Charleston Conference November 6, 2010
 Reconcile bin holdings with online catalog
Reduce ILL requests
Clean up catalog records (none,
incomplete, inaccurate)
 Accurate count of items
 Fill rate of bins
 Space
 Preparation for weeding of ARC

Material Types in ARC
 Type # of Items % of holdings
 Book 401,599 76.35
 Periodical 109,189 20.76
 Visual 9,229 1.75
 Nocirc 3,271 .62
 Audio 1,381 .26
 Video ref 1,223 .23
 Misc items 134 .03

Inventory time requirements &
processing
 Quick Visual
 To date inventoried X number of bins
 30 to 45 minutes depending on
size/format of the bin materials and
problem items
 Sequence of how bins are being called
(need to identify) and why this method

Arc Item Processing
 PC is set up to inventory mode and bin called
& delivered.
 Remove and scan each item within each
section and receive confirmation that the item
is in the correct section and bin.
 Will move items to different section within the
bin to ease space constraints.
 Once all items scanned, bin returned to rack
and another called.

Typical Issues
 Item not found
 Section displayed on screen as full but actually
has room.
 Requires high level of attention. System will
warn but easy to miss warnings. Needs to be
improved.

Prognosis/Results
 Most items are in correct bins
 Problems so far are items not linked to
bins (i.e. lost in space).
 Flexible…can start/stop as needed.
 Long term process: Will take a year to do.
 Hands on inventory provides opportunity
to also assess condition of materials.

Closing thoughts
 Hands-on real time inventory provides
opportunity to:
 Update linking to catalog so that correct item is
retrieved.
 Accurate determination of holdings and their
location.
 Assess condition of materials which may need
conservation.
 Inventory of the available space as well as
materials.

Recommendations
Completely and aggressively weed collections
before moving them into an ARS.
Ensure cataloging records contain information
necessary to isolate discrete collections if such
exist.
Develop procedures for weeding an ARS on a
regular basis, including periodic review of the
holdings in an ARS by call numbers and subject
headings.

Recommendations
Examine the feasibility of programming an
ARS to keep multi-volume sets in the same
bin.
Consider carefully whether an ARS is the
proper location for storage of multi-volume
sets and high-use items.

 “We are all robots when uncritically involved
with our technologies.” -- Marshall McLuhan

Resources
 Atkins, S., Weible, C. “Lost is Found”, Collection Management, 31:3
25-32, 2007.
 Bullard, R., Wrosch, J. “Eastern Michigan University’s Automated
Retrieval System, 10 years Later”, Journal of Access Services,
6:388-395, 2009.
 Gibbons, S. “Time Horizon 2020: Library Renaissance”,
http://hdl.handle.net/1802/10051 2010.
 Kent, A. “Uses of Materials: The University of Pittsburgh Study”,
Books in Library and Information Science (v. 26). New York: Marcel
Dekker, 1979.
 Schonfeld, R.C. & Housewright, R. Faculty Survey 2009: Key
Strategic Insights for Libraries, Publishers, and Societies. Ithaka S
+ R: http://www.ithaka.org/ithaka-s-r/research/faculty-surveys-2000-
2009/Faculty%2520Study%25202009.pdf
 Young, D.J. “Get to Effective Weeding”, Library Journal, 134:19 36,
2009.

�ݺ�ߣ

Weeding with Robots: Managing Collections in an Automated Retrieval System

More Related Content

Weeding with Robots: Managing Collections in an Automated Retrieval System

Editor's Notes