Title: Tables of Contents in the Catalog and Their Impact on Usage
1Tables of Contents in the Catalog and Their
Impact on Usage Karen D. Miller Monographic/Digi
tal Projects Cataloger Michael S.
Babinec Assistant Head, Bibliographic Services
Dept. Northwestern University Library Evanston,
Ill. 60208
2- About Northwestern University Library
- NUL is a research library with approximately 4.8
million volumes held on campuses in Evanston and
Chicago, Illinois. - TOC fields have been added to bibliographic
records since 2002 using Blackwell Book Services
TOC service. - Over 61,000 Blackwell TOCs have been added to
MARC 505 fields at NUL. - At 1.05 per record, we wanted to determine if
they made a statistically significant difference
in circulation.
3TOC in the catalog
4- Study Objectives
- Sought to replicate a 1997 study done at the
University of New Mexico Health Sciences Center
Library by Ruth C.T. Morris - Hypothesis that the inclusion of Table of
Contents (TOC) fields in bibliographic records
influences circulation
5- Study Objectives
- Do other factors interact with the TOC field to
influence circulation? - We examined the effect of
- Library location
- Subject matter, indicated by Dewey class
- Year of publication
6- Previous studies
- Focused on smaller collections, homogeneous
collections. One study used a health sciences
library, one a collection of IEEE conference
proceedings. - Used smaller sample sizes than NULs study (some
as small as 180 books). - One looked at any kind of enhancement, including
TOC fields, book covers, author biographies, and
reviews. - All determined that TOC enhancement increased
circulation and several concluded that prior
circulation was also a factor.
7- About the study
- Study sample randomly chosen from over 5,000
bibliographic records with TOC enhancements
available from Blackwell - 14 month study period June 2006 through July
2007 - 7-month study period was long enough for patrons
to check out and return most books - Both periods encompassed roughly similar
academic cycles - In-house browse data not as robust as
check-out data
8- Defining the sample
- We included only bib. records
- For monographs
- From circulating locations
- With circulating status
- With Dewey numbers
9- Refining the sample
- We excluded
- Course reserve material, as it needs no help to
circulate! - Books assigned to the New Book Alcove and to
study carrels - Books held at a Circulation desk, which cant
easily be browsed - Any bibliographic record with more than one
associated item
10- Study Methodology
- 4,633 bibliographic records split into two
groups, stratified by Library Location and Broad
Dewey Class (at the hundreds level). - Study Period 1 June - December 2006 no TOC
enhancements were added so that we could collect
circulation (check-out) and browse (in-house
usage) statistics. - Study Period 2 January July 2007 TOC
enhancements were added to Group 2 (the test
group) but not to Group 1 (the control group). - At the end of the study, we compared circulation
increases for the two groups.
11- About the Study
- We used SPSS for Windows to perform the
statistical analysis and used outside statistical
review of the sample stratification, study
design, analyses, and interpretation. - Gary Strawn, our systems librarian, added a MARC
035 subfield that indicated a bibliographic
record was in the study and to which study group
it belonged. - Library IT assisted in pulling data from the
Voyager back-end databases.
12Library Location distribution
13Distribution of Dewey Classes
14Distribution of Year of Publication
15- Study Methodology
- We looked at several ways of measuring results
- A binary variable indicating whether a book had
any use (circulation plus browsing) in Period 2.
This is the variable that the University of New
Mexico used. - A binary variable that measured whether there
was an increase in the number of times a book was
checked out of the library from Period 1 to
Period 2.
16- Study Methodology
- We examined and dismissed several approaches
- using in-house browse data. While nearly 40
of books were checked out, fewer than 18 had
browse charges - measuring how much circulation increased between
the study periods
17- Statistical analysis
- After examining the Year of Publication
distribution, we narrowed our study down to 3
variables - Study Group (whether a bibliographic record had
TOC enhancement or not) - Library Location, with 5 values
- Broad Dewey Class (hundreds level)
18- Statistical analysis
- Relationships between each of the independent
variables and the dependent variable were
examined and each had merit on its own. - Chi-square tests indicated that all but the
Dewey Class were independent predictors of
circulation increase.
19(No Transcript)
20(No Transcript)
21(No Transcript)
22- Statistical analysis
- Logistic regression was used to analyze the data
- well suited to categorical independent variables
and a binary outcome variable (increase or no
increase) - measures the likelihood of a certain event
occurring, in this case, an increase in
circulation. - likelihoods are expressed with respect to
different categories of the independent variables
in the model.
23(No Transcript)
24(No Transcript)
25(No Transcript)
26- Statistical analysis
- Dewey classes not evenly distributed and too
closely associated with Library Location
dropped from final model. - Residual analysis was performed and showed
- No patterns of poor fit
- No evidence of individual cases that influenced
the model, but the Main Library influenced the
model due to the number of cases it represents in
the study data.
27- Study results
- Enhancement of catalog records with TOC fields
the most significant factor in determining
whether a book had an increase in circulation. - A book with a TOC-enhanced Voyager record was
24 more likely to have a circulation increase
than a book without a TOC-enhanced record. - Location is also a significant predictor of
circulation increase-- Math the most likely
Location to see an increase, then Africana, with
Schaffner the least.
28- Study results
- Replicating the University of New Mexico study
We looked at only whether a book was used or not
during the second period of the study - TOC enhancement was still significant, making a
book 22 more likely to be used, but not as
significant a factor as whether the book had been
used previously or not. - A book that had previously been used in-house or
checked-out was 57 more likely to be used again
than a book that had not been removed from the
shelf before.
29(No Transcript)
30- Outstanding Questions
- Isnt the TOC enhancement just as likely to
convince a patron that a book is not what she or
he wanted as it is to get the book checked out of
the library? We dont know and maybe well never
know. We cant read minds, only databases. - Our speculation is that the TOC enhancements
result in an increase in recall, which turns up
books that might not have been otherwise. Whether
this results in a corresponding increase in
precision is still unknown.
31- Additional Research
- Our next study will examine keyword searching.
We hope to answer - What percent of keywords were found only in the
TOC field? - What percent were found only in the subject
heading fields? - What percent were found in transcribed fields?
- We hope to see you next year in Chicago with the
exciting conclusions of our keyword search study!
32- Many thanks to.
- Gary Strawn, Authorities Librarian/Library
Systems Analyst, for help interpreting Voyager
coding and SQL. - Stu Baker, Assistant Director of Library
Technology, for giving us the go-ahead to pester
the Voyager DBA and Gary Strawn. - Suzette Radford, Head of Circulation Services,
for guidance on where to get data and how to
interpret it. - David D. Morris, Director of Statistics at
WebbWrites Clinical Writing Statistics - David P. Nichols, Master Statistician, SPSS Inc.
33 Please direct questions and comments to Karen
D. Miller Monographic/Digital Projects
Cataloger k-miller3_at_northwestern.edu Michael S.
Babinec Assistant Head, Bibliographic Services
Dept. m-babinec_at_northwestern.edu Slides are
online at http//presentations.ala.org Papers
will be published as part of the ALA conference
proceedings.