Managing the Electronic Collection with qualitative and quantitative data A case study: the Wiley-Blackwell collection at the University of Milan Tiziana.

Managing the Electronic Collection with qualitative and quantitative data

A case study: the Wiley-Blackwell collection at the University of Milan

Tiziana Morocutti and Federica Zanardini

Part 1. Context and methodology

SISTEMA BIBLIOTECARIO DI ATENEO

University of Milan (year 2009)

• 65,000 students• 2,500 professors and researchers• 2,000 staff

• 9 faculties• 139 programs of study• 20 doctoral schools and 73 specialization

schools

UniMi Library system (SBA*)

• 250 FTE staff

• 1.5 M books• 25,500 print journals (7,500 current subscriptions)• 120,000 loans/y

• 9,000 purchased e-journals• 170 databases (bibliographic and FT)• 1.6 M downloads FT (e-journals)/y• SFX / Metalib / Ezproxy

* Sistema Bibliotecario di Ateneo

SBA and Digital Library costs

2009 budget:– UniMi 710 M €– SBA 8.1 M €

(permanent staff costs

excluded)

Bibliographic materials:– ER: 1 M € (2005) 3.2 M € (2009)– Print (print&online included):

5.1 M € (2005) 3.7 M € (2009)

ER acquisitions:– Directly from publishers 40%– CILEA DL (consortium) 30%– CARE (national contracts) 30%

SBA budget

1,03,2

5,1

3,7

1,61,2

0,0

2,0

4,0

6,0

8,0

10,0

2005 2009

ER Print other

Managing the economic crisis

Analysis of the relationships among…

content usefulnes

s

prices

(gears or mines?)


The first case study: the Wiley-Blackwell collection

2005-2009 contract

print subscriptions + online

access to the Full collection

(1,188 titles in 2009)

2010-2012 renewal

- E-only deal

- consortium purchase

(mirroring and backfiles

ownership)

Data:• Usage statistics (2008 JR1)• Economic data (2009 price list, contract terms)• Bibliographic data (2009 titles, subject coverage)• Demographic data about users (2009)• Results of a qualitative survey among users (2009)

A web-based survey to assess the perceived value

• Population: entire faculty (2,440 units)• Respondents: 25% in 40 days

Limits:• Subjective evaluation• Ambiguity (importance referred to user’s actual activity or

to relevance in the research field?) a generic perceived value that could be used together with usage data to measure usefulness

• No information about impact


Users were asked to select Wiley titles they consider important, specifying if essential or simply useful

Part 2. The Wiley-Blackwell collection: data analysis


Subject coverage


Title price ranges


Usage distribution

• 2008 downloads = 157,606

• 30% of titles => 85% of usage

(subscribed and unsubscribed)

• 4% of titles never used

• Below the threshold value “100”: 23.000 downloads in a long tail-like distribution (unlike Anderson’s model)


jou

rnals


Perceived value distribution (qualitative survey)

• “Which Wiley-Blackwell journals are important for you?”• 94% of titles were selected by users• 35% of titles were selected at least by 10 users

Relationship between usage and perceived value

• Usage and perceived value are related?

• Data are displayed on a scatter plot as a collection of points corresponding to titles

• => higher density in the area where there are low-usage low-valued journals


Relationship between usage and perceived value

0

1000

2000

3000

4000

5000

6000

7000

8000

0 20 40 60 80 100 120

selections

dow

nlo

ad

s

Correlation between usage and perceived value

• Is there a linear correlation between the two variables?

• Pearson’s index of linear correlation (0< R < 1)– R = 0.55 titles with #downloads > 100– R = 0.35 titles with #downloads <= 100

• A higher linear correlation between usage (actual

usefulness) and perceived value (perceived usefulness) in case of high-usage journals could be interpreted in terms of “conscious” usage of the resources?



Ratio between downloads and selections

• Are there anomalies in the relationship between usage and perceived value?

• Extreme values in the ratio M=(downloads)/(selections) give interesting information:

• M=∞ => titles used but not selected = underestimation?

Niche journals the users of which did not answer the survey? • M=0 => titles not used but selected = overestimation?

• Anomaly index

Journal ranking 1/2

• Usefulness is defined through an algorithm combining data about usage and perceived value:

U = (e + 0.2u) * downloadsU = usefulnesse = number of selections as essentialu = number of selections as useful

• Titles were given a score and ranked• By adding prices (p+e) the ranking list can be used

to calculate savings in relation to cancellations



Journal ranking 2/2

Remarkable savings can only be obtained by giving-up a significant number of titles

The first 300 titles cost as much as the entire Big Deal

Part 3. Conclusions


Results and findings

• Performance indicators• Journal ranking

• Hypothesis: hit content usage nonhit content usage


The long tail-like usage of nonhit content

Are information needsatomizedor elastic(depending on perceptions of availability)?

After the building blocks… tasks for the future:

• Enhancing the assessment method measurement of content usefulness evaluation of content impact on research activity development of proper statistical methods to analize

together qualitative and quantitative data

• Supplying practical instruments for supporting collection-development decisions starting a benchmark activity of Big Deal packages



Thank you!

[email protected]@unimi.it

Managing the Electronic Collection with qualitative and quantitative data A case study: the Wiley-Blackwell collection at the University of Milan Tiziana.

Documents

impact slide

useful slide

usage data

perceived value usage

wiley titles

journal packages slide

usage distribution

useful titles