A Model for Assessing Relative Interest in E-books Compared to Print

Steven A. Knowlton, M.L.I.S.

University of Memphis

Libraries

A Model for Assessing Relative

Interest in E-books Compared to

Print

Tenn-Share Fall Conference

Nashville Public Library

September 26, 2014

What library statistic

would you like to have –

That hasn’t been

invented yet?

How the heck can we

compare

e-book usage stats

to print circulation stats?

Why it would be useful We use circulation statistics (often in conjunction

with ILL statistics) for collection development and

assessment

Areas of strength/weakness

Weeding

Changing allocations by disciplines

Why it’s so hard to compare

themE-book Usage Print Circulation

Every time a reader opens a book, it counts

If downloading is allowed, checkout is often shorter than print

Casual interest (peeking at the first chapter) counts as much as deep reading

OVER-represents interest

User may open the book numerous times during a check-out, but it only counts as one circulation

Extended checkout periods (esp. for faculty) mean that some interested users may be kept away

Casual interest is not counted as a circulation (perhaps “in house” use)

UNDER-represents interest

Why I want to compare

them Many questions remain regarding whether our

patrons like e-books

Polarized reactions on the reference desk

Campus reactions vary widely

We have some stereotypical ideas (“science is best

in e-books”) but no firm ideas about what suits the

needs of our patrons

Nonetheless… I kept trying.

While poking around, I found this:NUMBER OF

CIRCULATIONSTITLES

NUMBER OF E-BOOK USES (EBSCO ONLY)

TITLES

0 87,199 0 61,968

1 8,134 1 4,071

2 1,219 2 1,418

3 235 3 726

4 71 4 408

5 15 5 248

6 6 6 177

7 1 7 122

8 0 8 80

9 1 9 62

My epiphany

It is much more common for a book – whether print or electronic – to have no recorded use in a given year than to have any level of use

That means…

The level of use is a “marginal

variable”; it is not crucial to

making informed collection

decisions. We can simply

look at whether a book was

used, not how much.

So…

We don’t have to try to make comparisons between print circulation and e-book usage statistics.

We just have to look at whether the book received any use.

I’m a genius!!!!

Or not… other people have

used the same test: Robert Slater, “E-books or print books, “big deals” or local selections—What

gets more use?” Library Collections, Acquisitions, & Technical Services 33

(2009): 31-41.

Marilyn Christian Son and Marsha Aucoin, “Electronic or Print Books: Which

Are Used?” Library Collections, Acquisitions, & Technical Services 29

(2005): 74-79.

Justin Littman and Lynn Silipigni Connaway, “A Circulation Analysis of Print

and E-books in an Academic Research Library,” Library Resources and

Technical Services 48 (2004): 256-62.

Terry Bucknell, “The ‘Big Deal’ Approach to Acquiring E-books: A Usage-

Based Study,” Serials 23, no. 2 (2010): 126-34.

But…They don’t measure local preferencesThey measure whether e-books get more use than print books, and in which disciplines e-books get most use

But they don’t control for the size and strength of e-book collection vs. print collection

New epiphany!

Combine the

binary test of

use with a test

of “strength of

interest” used in

assessing print

collections

Strength of interestAssume that the proportion of the collection represented by a discipline should also be the proportion of titles used.

IfHistory books = 10% of the titles heldThenHistory books should = 10% of titles usedVariation shows where our patrons like what we’re offering, and where they don’t find it helpful

Percentage of Expected

Use

Again, not a new idea (see

below).

But, if we combine the binary test

of use + percentage of expected

use,

we can see where we should buy

e-books and where we should buy

print

Terry R. Mills, “The University of Illinois Film Center Collection Use Study.” ERIC report ED 227 821 (1982).

William Aguilar, “The Application of Relative Use and Interlibrary Demand in Collection Development,” Collection Management 8

(1986): 15-24.

John N. Ochoa, “Use of Circulation Statistics and Interlibrary Loan Data in Collection Management,” Collection Management 7

(2003): 1-13.

DatasetOur e-book collections:

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

6000

19

61

19

62

19

63

19

64

19

65

19

66

19

67

19

68

19

69

19

70

19

71

19

72

19

73

19

74

19

75

19

76

19

77

19

78

19

79

19

80

19

81

19

82

19

83

19

84

19

85

19

86

19

87

19

88

19

89

19

90

19

91

19

92

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

# of titles

Dataset

1990 – first year more than 500

ebooks represented

Need to use records with LC

Classification numbers

Dataset

Decided upon:

All EBSCO and Springer e-books

published 1990-2014

All print books available for

circulation published 1990-2014

Dataset

E-books: n=99,077

Print books: n=92,101

How we generated the data

Export from ILS1. List of all circulating books published since 1990

With circulation figures for 2013-14 academic year

ILS has 2 years of circ figures, but – OOPS – 2012-

13 figures were accidentally wiped

2. List of all e-books published since 1990,

hosted by EBSCO or Springer

These records have LC Class numbers; records for

other hosts don’t; but EBSCO and Springer

constitute the largest sets by far

How we generated the data

Export from e-book platformsUsage figures for 2013-14 academic year

List only shows titles with some usage; have to

assume any title not on the list has zero uses

How I generated the dataManually merge the usage data into the ILS

file

How I generated the dataBig mistake:

Relying on title matching

(forgot to export ISBN from ILS)

- E-book data had only 245 |a

- Conference proceedings difficult to

distinguish

- Had to grab ISBN from e-book platform, then

look up in Amazon to match to subtitle in ILS

file

How I generated the dataConvert all numeric values of circulation or e-

book usage:

1 to YES

0 to NO

How I generated the dataMarked any titles that we held in both print

and electronic:

Used in print but not electronic

Used in electronic but not print

Used in both print and electric

Not used in either format

How I generated the data

Results of binary test

USED IN

2013-14 % USED

Complete

Dataset 18,294 9.5%

Electronic 8,940 9.0%

Print 9,354 10.2%

Results of binary test:

E-book -- high and low use by discipline

DISCIPLINE PERCENTAGE USED ABOVE/BELOW

AVERAGE

Military & Naval 4.2% -4.8%

U.S. history (local) 4.7% -4.3%

Physics 4.8% -4.2%

Agriculture 4.9% -4.1%

Asian History 5.1% -3.9%

Social Sciences (General) 13.2% 4.2%

Sociology 14.3% 5.3%Auxiliary Sciences of History 14.8% 5.8%

Psychology 15.1% 6.1%

Education 16.6% 7.6%


Print -- high and low use by discipline


AVERAGE

Astronomy 1.9% -8.3%

Geology 3.3% -6.9%

Geography 4.3% -5.9%

Biology 4.5% -5.7%


U.S. History 17.3% 7.1%

History (General) 17.6% 7.4%

U.S. History (Local) 18.7% 8.5%


African History 29.2% 19.0%


Print -- high and low use by discipline


AVERAGE

Astronomy 1.9% -8.3%

Geology 3.3% -6.9%

Geography 4.3% -5.9%

Biology 4.5% -5.7%


U.S. History 17.3% 7.1%

History (General) 17.6% 7.4%

U.S. History (Local) 18.7% 8.5%


African History 29.2% 19.0%


Gap – high use of print and low use of e-books

DISCIPLINE E-BOOK USE

COMPARED TO

AVERAGE

PRINT USE

COMPARED

AVERAGE

GAP

African History 3.3% 19.0% 15.7%

U.S. History (Local) -4.3% 8.5% 12.9%Social Sciences (General) 4.2% 15.8% 11.6%

Photography -3.4% 4.0% 7.4%

History (General) 0.1% 7.4% 7.3%


Gap – high use of e-books and low use of print

DISCIPLINE PRINT USE

COMPARED TO

AVERAGE

E-BOOK USE

COMPARED

AVERAGE

GAP

Education -1.6% 7.6% 9.2%

Astronomy -8.3% 0.6% 8.9%

Geography -5.9% 2.6% 8.5%

Psychology 2.1% 6.1% 4.0%

Anatomy & Physiology -3.0% 0.9% 3.9%

Books in both formats

ELECTRONIC

, BUT NOT

PRINT

PRINT, BUT

NOT ELEC

BOTH NEITHER

7.1% 8.3% 3.1% 81.5%

Having a title in both formats

increases its likelihood of use by

95%

Overall use is 9.5%; use of titles

available in both formats is 18.5%

Books in both formats

Probably indicates that patrons

have a strong format preference

If use is so much higher for titles

without a format restriction, then

they are likely bypassing books in

the format they don’t like

Books in both formats:

High percent of titles used in electronic but not in

print

DISCIPLINE PERCENT USED IN E-BOOK BUT

NOT IN PRINT

Education 13%

Non-English Literature 10%

Law 10%

Sociology 10%

Librarianship 10%


High percent of titles used in print but not in

electronic

DISCIPLINE PERCENT USED IN PRINT BUT NOT

IN E-BOOK

History (General) 17%

U.S. History (Local) 16%

Technology (General) 15%

African History 12%

Philosophy 12%


Biggest gaps

DISCIPLINE % P NOT E % E NOT P GAP

Technology (General) 15% 0% 15%

History (General) 17% 5% 12%

Social Sciences (General) 10% 0% 10%

African History 12% 2% 10%

U.S. History (Local) 16% 7% 8%

DISCIPLINE % E NOT P % P NOT E GAP

Anthropology 9% 5% 4%

Geography 5% 2% 4%

Education 13% 9% 3%

Non-English Literature 10% 7% 3%

Biology 6% 4% 3%

Percentage of Expected Use

General idea:

DISCIPLINE % OF BOOKS

HELD

% OF

BOOKS

USED

PEU

FACTOR

ECONOMICS AND

FINANCE

10.2% 6.6% -3.6%

POLITICAL SCIENCE 2.5% 2.1% -0.5%

ASIAN HISTORY 1.8% 1.8% 0%

MATHEMATICS 1.5% 2.4% 0.9%

SOCIOLOGY 7.7% 10.1% 2.4%

Percentage of Expected Use:

Worst and Best for E-booksDISCIPLINE % OF BOOKS

HELD

% OF

BOOKS

USED

PEU

FACTOR

Economics & Finance 13.1% 10.1% -2.9%

Engineering 8.9% 6.3% -2.6%

Physics 2.4% 1.3% -1.1%

Law 2.3% 1.5% -0.8%

Political Science 2.4% 1.7% -0.7%

Computer Science 7.0% 7.7% 0.7%

Psychology 1.5% 2.5% 1.0%

Medicine & Nursing 10.5% 12.9% 2.4%

Sociology 5.0% 7.9% 2.9%

Education 3.5% 6.5% 3.0%


Worst and Best for PrintDISCIPLINE % OF BOOKS

HELD

% OF

BOOKS

USED

PEU

FACTOR

Economics & Finance 10.2% 6.6% -3.6%

Engineering 5.4% 3.2% -2.2%

Biology 2.6% 1.2% -1.4%

Geology 1.5% 0.5% -1.0%

Education 5.6% 4.7% -0.8%

Philosophy 3.1% 4.0% 0.9%

African History 0.7% 2.1% 1.4%

Medicine & Nursing 6.9% 8.8% 1.9%

U.S. History 3.4% 5.7% 2.4%

Sociology 7.7% 10.1% 2.4%


Biggest GapsDISCIPLINE GAP IN FAVOR OF E

Education 3.8%

Sociology 0.5%

Medicine & Nursing 0.5%

Psychology 0.5%

Computer Science 0.5%

DISCIPLINE GAP IN FAVOR OF

PRINT

U.S. History 1.7%

Religion 1.4%

African History 1.3%

U.S. History (Local) 1.2%

American & English Literature 0.9%

Age of the book

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

40.0%

ALL

E-BOOK

PRINT

Linear (E-BOOK)

Linear (PRINT)

Why do we know now (about U of Memphis)?

• Some disciplines keep emerging with a

strong preference regardless of the

measure

(Education & Psych for ebooks;

History & Philosophy for print)

Not always what “conventional wisdom”

dictates!

• When there’s a preference for print, it’s

much stronger than when there’s a

preference for e-book

Why do we know now (about U of Memphis)?

• Newness of a book has more effect on use

of print books than e-books

• (And, by the way)

Those disciplines whose faculty tell us we

shouldn’t buy books are right! (Economics &

Finance, Engineering)

A Model for Assessing Relative Interest in E-books Compared to Print

Government & Nonprofit

print books

circulation analysis

disciplines ebooks

recorded use

library collections

print collectionnew

binary test of use

strength of ebook collection