Usage metrics: tools for evaluating science collections by Michelle Foss Leonard, Stephanie Haas, Donna Wrublewski, and Vernon Kisling Marston Science Library, University of Florida, Gainesville. ACS 2010, Boston, MA
Why Metrics? Why Now? Higher education will be increasingly viewed as a business, and calls for accountability and for quantitative measures of library contributions to the research, teaching and service missions of the institution will shape library assessment programs and approaches to the allocation of institutional resources. ⁿ UF is moving toward a business based budget model. Important for the UF Libraries to justify collection spending AND how the collections are being used by the patrons. (new patron driven models) Lead to new collection policies & procedures. ⁿACRL Research Committee: Environmental Scan 2007
The current studies provide an initial assessment of use data associated with the monograph & journal collections at the Marston Science Library (MSL), University of Florida. The overall strategy of this project is to analyze audience-based circulation patterns, purchases, and e-book/journal usage statistics from the academic years July 1, 2008 to June 30, 2010. Such analyses provide an evidence-based framework for future collection decisions.
Potential Audiences
Potential Audiences
Science Monographs Purchased $127,141
Purchased monograph circulation
Circulation by User Group for subclasses of Q (Science)
Circulation by User Group for subclasses of S (Agriculture)
Circulation by User Group for subclasses of T (Technology)
The text file of subclasses was uploaded into TagCrowd http://tagcrowd.com and filters chosen.
2009 Circulation of ALL Monographs QD Subclasses (n=845) Highest use areas: QD1 Serials QD96 Spectra QD381 General
QD Chemistry monographs borrowed by graduate students (n= 651)
2009 Circulation of Monographs Purchased in 2009 QD Subclasses Highest Use Areas: QD462 Quantum Chemistry QD116 Electrochemical Analysis QD96 General Works
LC subclass analysis using TagCrowd 1. The TagCrowd is extremely effective in helping to analyze data through visually clustering. 2. Finer granularity can be obtained by extending the call number extensions; however, decimal points must be converted to other characters. Example QD272.2 must be converted to QD272d2, where d=decimal
Findings from circulation 1. Graduate student borrowing far exceeds borrowing of either faculty or undergraduates, except in fisheries, zoology, and chemical engineering. 2. Based on total potential users, borrowing average is less than 1 book per user per year. 3. LC subclasses with the highest circulation activity were identified.
900 In-house Usage, Jan-June 2010 Q, S, T n=4237 800 700 600 Usage 500 400 Books Journals 300 200 100 0 Q QA QB QC QD QE QH QK QL QM QP QR S SB SD SF SH SK T TA TC TD TE TH TJ TK TL TN TP TS TX LC Range
Print vs e-use for titles [40 titles circulated 43 times; online usage was 431, or a 100% increase. ]
E-book Usage 12000 Science E- books (Springer 2009) n=6652 10000 8000 # of Unique Titles 6000 # of Uses 4000 2000 0 Q QA QB QC QD QE QH QK QL QM QP QR QS QU QV S SB SD SF SH T TA TC TD TE TF TH TJ TK TL TN TP TS LC Subclasses
LC Class Expenditure by LC Class LC Class E- Books Purchased by LC Class Average Cost per e- book Uses in LC Class Average Cost Per Use B $430.44 B 4 $107.61 11 $39.13 D- F $984.68 D- F 11 $89.52 92 $10.70 G $712.14 G 4 $178.04 15 $47.48 H - HF $2,738.43 H - HF 27 $101.42 105 $26.08 HG - HV $2,022.45 HG - HV 23 $87.93 128 $15.80 J $573.12 J 6 $95.52 20 $28.66 L $399.29 L 4 $99.82 18 $22.18 M $249.71 M 2 $124.86 17 $14.69 N $130.73 N 2 $65.37 8 $16.34 P $1,792.85 P 15 $119.52 76 $23.59 Q $3,926.74 Q 33 $118.99 121 $32.45 R $2,280.38 R 22 $103.65 73 $31.24 S $893.48 S 5 $178.70 15 $59.57 T - TP $3,173.15 T - TP 30 $105.77 199 $15.95 TR - TX $287.87 TR - TX 4 $71.97 12 $23.99 U $28.80 U 1 $28.80 2 $14.40 Grand Total $20,624.26 193 $106.86 912 $22.61 Total e- book uses Pilot; e- books used once not purchased Pilot; uses of e- books purchased (May - Oct 09) Post Pilot; purchased e- book uses (Nov 09 - April 10) No. of e- books purchased - reused (Nov 09- April 10) % of books reused (150/193) Avg. cost per use of e- books purchased (May 09 - April 10) 1,974 418 912 644 150 78% $8.06
Study conclusions concerning monographic use Monographic use by patron status is now evidence-based with graduate students accounting for 58% of the circulations. LC classes of greatest interest have been identified: QA, QH, QC, QD, S, SB TK, TA, TP, Current purchasing decisions accurately reflect usage 40 print titles were also available in electronic copy and online usage was 100% greater. This finding has major implications for future collection development decisions.
2009 Print and Electronic Usage - JACS 4000 Print Archive Electronic Archive Print Current Subscription Electronic Current Subscription 3500 3000 2500 2000 1500 1000 500 0 Dec 08 Jan 09 Feb 09 Mar 09 Apr 09 May 09 Jun 09 Jul 09 Aug 09 Sept 09 Oct 09 Nov 09 Dec 09 Number of Usages Usage Month
2009 Print and Electronic Usage Journal of Organic Chemistry (ACS) 2500 Print Archive Electronic Archive Print Current Subscription Electronic Current Subscription 2000 1500 1000 500 0 Dec 08 Jan 09 Feb 09 Mar 09 Apr 09 May 09 Jun 09 Jul 09 Aug 09 Sept 09 Oct 09 Nov 09 Dec 09 Number of Usages Usage Month
Study conclusions concerning data mining capability Theoretically, library functions supported by computers and databases should be quantifiable; however, acquiring meaningful data requires multiple extraction runs, data striping, and recombination. Currently, extracted data that can assist in collection development strategies is limited to LC classes and/ or subject headings. Specificity of LC class appears stronger than subject headings, but can future text analysis programs provide accurate subject extraction? Visualization software and functionality can effectively be applied to library data.
Study conclusions concerning data mining capability Data archiving will become an issue in future metricsbased library studies. Most E-book services do not provide useful statistics to assist in collection development decisions. Title level usage statistics are needed for package subscriptions. E-book metadata is incomplete, either because of service provider decisions or local modifications. LC Call numbers and/or subject headings should be mandatory to permit data mining and analysis.
What does all this mean? What is the impact? Based on metrics and usage statistics: recommend CM policies & procedures that encourages the purchase of e-books; initiate discussion with e-book vendors to provide more convenient ways to gather and analyze usage stats. Continue to use patron-driven acquisition models for books: MyiLibrary; course reserves, ILL, and distance learning Adapt collection weeding policy for hard copy based on usage statistics: moving books into off-campus storage recommend policy for weeding hard copy with e-book equivalent.
Questions? Michelle Leonard Science & Technology Librarian University of Florida mleonard@uflib.ufl.edu Donna Wrublewski Chemistry Librarian University of Florida dtwrublewski@ufl.edu This work is licensed under the Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States license: http://creativecommons.org/licenses/by-nc-sa/3.0