Interpreting the FINAL LINK REPORT and UNLINKED HEADING REPORTS
At the completion of each authority control job, LTI creates a Final Link Report summarizing the results of the authority processing. Libraries also receive a list of unlinked headings as an ASCII text file. Explanations and examples of these reports are provided below.
Final Link Report
In the Field Link Summary there are three primary link statuses. Library headings can link to an LC name or subject authority record, to an LTI name or subject authority record, or not link to any authority record. In the first two statuses, the heading can link fully (the primary heading plus any subordinate parts), or only partially with the authority record (the primary heading but not all of its subordinate parts).
LTI defines the word "link" to mean that the text of the primary heading($a) taken from the library's catalog record, but not necessarily all its subordinate parts, corresponds to and is replaced by the text of the authority record heading. A link can be considered "full" even though the library's heading contains data not present in the authority record heading. For example, floating tables are used to validate language of work ($l) in title and author/title name headings, while general ($x), chronological ($y), and form ($v) subdivisions appear in many subject headings. LTI's validation checks that any additional data complies with AACR2 and current LC cataloging practice and these subfields are properly coded.
The Field Link Summary divides fields into three categories: 1) library headings that linked to an LC authority record, 2) library headings that linked to an LTI authority record, and 3) library headings that could not be linked to either an LC or an LTI authority record. Tags are grouped into name name, series, and subject fields. While series headings are considered part of name authority control, for reporting purposes LTI distinguishes between name and series headings.
Within each category are listed the number of fields that were changed or modified during the final link routine, as well as the number that were not changed (designated "SAME") during the final link. It is important to note that changes made to records during pre-authority control clean-up (e.g., setting of title field filing indicators, MARC Update processing, deletion of subfields $e, $w, and $4) are not counted in the Final Link Report. In other words, headings designated as "SAME" may have been changed at an earlier point in LTI's processing.
Statistics given in the Dahlgren College sample report are from an academic library that received limited review authority control. The report is largely self-explanatory. It shows that there are 454,062 topical subject headings (650 fields). Of these 444,309 (99.6%) were linked to an LC subject authority record. This is a "derived number" found by adding the CHANGED and SAME columns and then dividing that number by the total headings-i.e., 7,962 + 444,309 = 452,271 / 454,062 = 99.6%. One sees that another 1,613 (0.0036%) topical subject headings were linked to LTI authority records. Only 178 topical subject headings (0.0004%) could not be linked to any authority record. Totals and percentages are given at the foot of each column.
Four cumulative statistics appear following the field-by-field breakdown: "Headings linked to LC," "Headings linked to LTI," "Headings linked to LC or LTI" and "Headings not linked." For U.S. library databases created in accord with national cataloging standards, LTI guarantees that 95% or more of the library's controlled headings will be linked to either an LC or LTI authority record. For this reason the combined statistic line is bolded. If the library has opted to restore certain 830 fields back to 440, the number of 830 fields converted to 440 is listed on the following line.
At the end of the report is the number of LC name and subject authority records that have linked to the library's headings, the total number of linked national level authority records, and the ratio of LC authority records to catalog records.
Following authority control processing, LTI generally returns to the library slightly more than one LC name and/or subject authority record per bibliographic record for databases up to 250,000 catalog records. This ratio will be smaller for public libraries because they tend to have multiple titles for the same author. As the number of catalog records increases, the ratio of linked LC authority records to bibliographic records decreases. For example, approximately 675,000 LC authority records might be extracted for an academic library database of one million catalog records. A large academic library of 4.7 million records authorized in early 2007 resulted in the extraction of 2.2 million LC name and subject authority records. Linked LC name and subject authority records are written to separate files and deduped so that the same record appears only once, except when the local system requirements differ.
Final Link Report statistics reflect field counts going into the final link processing, rather than field counts after the final linking. For example, some 650 fields may be retagged as 651 fields after the final link, or some 100 fields retagged as 110 fields. These tag level changes account for occasional slight variations between Field Link Summary counts and post-processed field counts given in LTI's RecStat report.
DAHLGREN COLLEGE
FINAL LINK REPORT
F I E L D L I N K S U M M A R Y June 22, 2007
**LINKED TO LC** **LINKED TO LTI** **NOT LINKED**
FLD FREQ CHANGED SAME CHANGED SAME CHANGED SAME
100 217,644 15,393 195,469 256 4,496 76 1,954
110 3,405 142 3,078 8 127 3 47
111 960 46 683 18 211 1 1
130 1,900 114 929 62 532 32 231
240 17,362 3,350 13,917 28 59 2 6
243 1 1 0 0 0 0 0
700 197,126 15,817 169,665 278 4,183 298 6,885
710 45,092 1,713 40,550 132 1,817 33 847
711 766 54 584 8 85 2 33
730 4,374 134 1,968 41 1,230 42 959
800 2,294 457 1,815 16 5 0 1
810 83 22 57 3 0 0 1
811 48 1 23 24 0 0 0
830 72,631 7,934 57,691 673 4,623 153 1,557
600 55,834 5,376 49,951 26 349 9 123
610 12,202 879 11,081 16 167 7 52
611 412 16 381 0 14 0 1
630 12,174 715 10,968 6 183 14 288
650 454,062 7,962 444,309 53 1,560 36 142
651 61,028 1,749 59,005 23 229 2 20
Genre Subject Headings
lc 10,292 0 0 7,303 179 2,806 4
gsd 245 0 0 46 199 0 0
msh 2 0 0 0 2 0 0
oth 1 0 0 0 0 0 1
TOT 1,169,938 61,875 1,062,124 9,020 20,250 3,516 13,153
(%) (100%) ( 5%) ( 91%) ( 1%) ( 2%) ( 0%) ( 1%)
Total records: 254,576 Total headings: 1,169,937
Headings linked to LC: 1,123,999 ( 96.1%)
Headings linked to LTI: 29,270 ( 2.5%)
Headings linked to LC or LTI: 1,153,269 ( 98.6%)
Headings not linked: 16,669 ( 1.4%)
830 -> 440: 47,621
New name/title/series related auth recs: 255,237
New subject related auth recs: 48,855
Total Linked LC Authority Records: 304,092
Ratio of Linked LC Auth. Recs to Bib. Records: 304,092 / 254,576 = 1.2
Unlinked Heading Reports
A byproduct of every authority control project is an ASCII text file of unlinked headings. In the Unlinked Headings Report (ULH) only headings with a subfield $a that could not be validated against either an LC or an LTI authorized heading are included. Separate lists of unlinked name/title (1XX,7XX) fields, series (4XX, 8XX), and subject (6XX) headings are provided. Unlinked headings are arranged alphabetically within these three broad heading categories. Because a single heading may have been used in hundreds of records, the control number of the catalog record in which the unlinked heading appears is not included in the printout.
Unlinked headings can be loaded into a word processor for review purposes. Keep in mind that the vast majority of cases, the only thing "wrong" with these headings is that no LC or LTI authority record exists for the heading. Most libraries do not systematically search unlinked headings. To enable libraries to examine only those headings that will result in the greatest number of potential fixes, the frequency of the heading is also indicated off to the left of the heading. Part of a sample page for each category is shown below.
Upon request LTI also makes available a separate unlinked headings report that includes all partially validated Headings (PVH). A heading is listed as "partially validated" if $a linked to an authority record but a subsequent subfield did not (i) link to an authority record or (ii) or validate as a proper use of a valid floating subfield.
Unlinked Name & Title Fields -- 7/21/2007 p. 3
. . .
X00-1-$aAllard, James Anthony.
X10-2-$aAllegri Quartet.
X00-1-$aAllen, Belinda.
X00-1-$aAllenby, Marlon.
X00-1-$aAllen, C. L.
X00-1-$aAllen, Drane.
2 X00-1-$aAllen, Hugh.
X00-1-$aAllen, Lewis.
X00-1-$aAllen, Sylvan.
2 X00-1-$aAllen, W. Peter.
2 X00-1-$aAllison, Beth.
X30-0-$aAll men are created equal.
X30-0-$aAll music guide to jazz.
X10-2-$aAll Souls Gathered Choir.
X10-2-$aAll Souls Orchestra.
X00-1-$aAlmani, Maryam Parvin.
X00-1-$aAlman, Stephanie Bernice.
X10-2-$aAlmaz Film Production.
X00-1-$aAlmeida, Acâacio de.
2 X00-1-$aAlmeida, Morgan.
X00-1-$aAlmgren, Susan.
X30-0-$aAlmighty God in the Lord Jesus Christ. $f1985.
. . .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Unlinked Series Fields -- 7/21/2007 p. 1
. . .
Y30-0$a1000 years of Russian liturgy.
Y30-0$a1987-: Peterson's annual guides/jobs.
Y30-0$a1st Midland book ed.
Y30-0$a20/20 Friday.
Y30-0$a20th century organ masterpieces.
Y30-0$a20th century voices on cassette.
Y30-0$a21st century renewable energy series.
Y30-0$a2nd opinion health debate series.
Y30-0$a"30 cross curricular units."
Y30-0$a50 ways series.
Y30-0$a90s.
Y30-0$aAACTE Accreditation resource series.
Y30-0$aABC News Time forum.
Y30-0$aABC Sports home video library.
Y30-0$aAccess unlimited.
Y30-0$Acclaimed video lectures (Biblical Archaeology Society). $nSeries
2, Set IV. $pHonor and shame ; $vd.
Y30-0$aAccountancy forum ; $vprogram 2.
Y30-0$aAcropolis Americana/Bicentennial.
Y30-0$aActor's Notes.
Y30-0$aAeolian band and orchestra library ; $vno. 127.
Y30-0$aAfrica.
Y30-0$aAfrican voice.
. . .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Unlinked Subject Fields -- 7/21/2007 p. 4
. . .
630-0$aClarel.
600-1$aClark, Peggy, $d. 1959.
610-2$aCleveland American Indian Center.
630-0$aClockwork orange.
651- $aCoelhos (Recife, Brazil)
630-0$aCollection of Psalms, hymns and spiritual songs.
630-0$aCommerce of the prairies.
630-0$aCommonplace book of John Milton.
630-0$aComus.
630-0$aConnecticut Yankee in King Arthur's court.
610-2$aConservative Amish Mennonite Conference$xHistory.
650- $aConsumer education (Middle school)
650- $aConsumer mathematics$Study and teaching.
3 650- $aConsumer mathematics$Study and teaching (Secondary)
630-0$aContra gentes.
610-2$aCoral Way Elementary School (Miami, Fla.)
630-0$aCorâonica de Adramâon.
610-2$aCouncil on Accreditation (National Recreation and Park
Association/American Association for Leisure and ecreation)$xHistory.
630-0$aCrâonica de veinte reyes.
610-2$aCumberland County Court House (Carlisle, Pa.)
610-2$aCumberland County Lyceum.
600-0$aCustaloga,cDelaware chief, $d. 1776.
. . .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Provisional Authority Records
Because most local systems with an authority control module can create authority records for unlinked headings as new records are loaded into the system, most libraries do not need vendor-supplied provisional authority records. For libraries requesting them, separate provisional authority records are created for every library heading that did not link fully to an LC authority record.
For example, the LCSH heading Automobile racing, which may be subdivided geographically, may occur in the library's database in the following headings:
650 0 $aAutomobile racing$zOhio. 650 0 $aAutomobile racing$zGeorgia. 650 0 $aAutomobile racing$zCalifornia 650 0 $aAutomobile racing$zMichigan.
The LC subject authority record for "Automobile racing" will of course be pulled and included in the file of LC subject records. In the above example, if the library also wanted to receive provisional authority records, separate authority records would be created for each of the full headings, including the place subdivision.
As a guideline, the number of provisional authority records created will approximate the number of LC name and subject authority records returned to the library. In other words, for a database of 100,000 catalog records, the library might expect to receive about 120,000 LC name and subject authority records and another 120,000 provisional name and subject authority records.
Any heading appearing in the unlinked and partially linked heading reports will be extracted for creation of a provisional authority record. However, there will be many more provisional authority records than unlinked headings.