Left Menu Right Menu

ASA Data Project: Data Reports Cover Letter

To: Lynn Chmelir, Chair, Collection Development and Management Committee

From: Nancy Nathanson

April 6, 2005

Article Supplier Analysis Project: data reports

We have received 17,668 entries for the Article Supplier Analysis study. All but one member institution (COCC) submitted a report, ranging from 34 entries (MHCC) to 2,201 (OSU).

Each report was reviewed and corrected as necessary so that it could be used in a single file of combined data from all members. In most cases, we consulted with the submitting library to get corrected information or agree on how we would “fix” the data. The number of individual entries in the study is very large, and the data problems shouldn’t effect the outcome in a significant way. I have included some notes about the nature of the data in case it might influence how you interpret or factor a confidence level for the results

Data from each report was loaded to a single Access table. Depending on the information desired, results were prepared from Access queries, or from data exported back to a single Excel file for other filtering and calculations.

Notes about the data fields

Title

Due to the varying data entry practices, some titles will be non-matchable (abbreviations; periods vs. no periods at end of title; etc.) A report specific to analyzing titles is not prepared at this time.

There are only 857 records with a title and no ISSN. Of those, 710 are “unique” titles (in other words, some of the articles received were for the same title). There are probably a few more matches that we could find with additional review; this would reduce the number of unique titles by a little bit.

ISSN

A few entries might be monographs (suspicious ISBN’s; investigated and discarded a few, haven’t yet reviewed about a dozen more.). Some bad data was corrected (for example, first character omitted, bracket at end of field); 7 entries lacked both title and ISSN. Differing formats so that ISSN’s were not matching; many lack hyphen. We corrected some here, and asked for some to be re-submitted. We haven’t fixed them all.

16,804 entries with ISSN. 7,312 unique ISSN

articles received with same ISSN or title:

more than 1 Rec’d: 3,095 unique ISSN(18% of all received)

more than 4 Rec’d: 751 unique ISSN


 

ISSN

# of articles received

unique ISSN

 

1

4,217

 

2

1,384

 

3

637

 

4

323

 

5

201

 

6

127

 

7

88

 

8

80

 

9

47

 

10

32

 

>10

176

 

 

 

(note: 25 ISSN with more than 20 requests)

 

7,312

 

 

Format

Some entries in this field contained characters other than e, m, or p (commas, blanks, periods, other alpha and numeric characters). Most were corrected, only a few with extraneous data had to be deleted.

p e

Print and Electronic

7,037

45.9%

 

p

Print only

4,456

29.0%

 

p e m

Print, Electronic, and Microformat

2180

14.2%

 

e

Electronic only

1,058

6.9%

 

p m

Print and Microformat

574

3.7%

 

m

Microformat only

23

0.1%

 

e m

Electronic and Microformat

15

0.1%

 

subtotal

 

 

 

15,343

(no data)

 

2,325

 

 

total

 

 

 

17,668

Year

Some Year fields contained numeric data that had been reformatted and could not readily be interpreted as a “year”; sometimes we were able to guess at the year intended; also sometimes contained a span of years (e.g. 1973-74) or specific dates or months. No year field in 163 reported entries. Ultimately we had to eliminate the year field data in a couple dozen odd entries, and a few dozen others were changed by library staff and re-submitted.

Bib records

The data in this field is not uniformly reliable for several reasons: lack of understanding about concept of bibliographic records and about catalog display screens, especially difference between Index browse and Title browse screens where number of entries is displayed; uncertainty about counting records owned only by CRL, etc.

Borrowed From

Borrowed From institutions (Suppliers): instances of bad data were corrected where possible, such as “11” instead of “1”, “v” instead of “u”, etc.; and a small number (fewer than 10) were deleted in the last day.

The attached reports

I have prepared several reports in a single Excel workbook. The source of data may have been Excel or an Access query.

 

Recd_from_Alliance

Total number of articles received for each member, and how many received from an Alliance member as lender. For example,

Clark reported 59 articles; 35 (59%) supplied by an Alliance member

WWU reported 1,291; 54% supplied by an Alliance member

UO reported 2,171 articles; 13% supplied by an Alliance member

Avail_from_Alliance

Number of articles received from any source, where an Alliance supplier was available. [1]

Avail_from_Alliance-#ofowners

The number of Alliance members owning the title: 0,1,2,3,4,5,6 or more

Owners

Detail report showing, for each member (Borrower), the number of articles owned by each Alliance member. For example, of everything Clark received, 9 articles were owned by CWU, 11 by EOU, 13 by Evergreen, etc. Looking at the columns, Clark owns 241 of the articles that were received, CWU owns 1,386, etc.

Unique_hldgs

Unique holdings, or Sole Ownership: the number of times this member is indicated as the only one owning the article (presented two ways: based on no data for any other member, or only “n” or “u” or “blank” for every other member). For example: Where there is no ownership data recorded for other members (“blank”):

UW is sole owner of about 4.2% of all articles in the study,

WSU is sole owner of about 1.2% of all articles

Date

Year Article Published. About 18% of all articles were published before 1990, 32% published 1990-1999, and about 50% published in 2000 and later.

Suppliers-sums

Number of articles received from each Alliance lender