Quality Control of Chinese Monographic Records: A Case Study

Journal of East Asian Libraries Volume 1998 Number 116 Article 6 10-1-1998 Quality Control of Chinese Monographic Records: A Case Study Fung-yin K. Simpson Follow this and additional works at: https://scholarsarchive.byu.edu/jeal BYU ScholarsArchive Citation Simpson, Fung-yin K. (1998) "Quality Control of Chinese Monographic Records: A Case Study," Journal of East Asian Libraries: Vol. 1998 : No. 116, Article 6. Available at: https://scholarsarchive.byu.edu/jeal/vol1998/iss116/6 This Article is brought to you for free and open access by the All Journals at BYU ScholarsArchive. It has been accepted for inclusion in Journal of East Asian Libraries by an authorized editor of BYU ScholarsArchive. For more information, please contact scholarsarchive@byu.edu, ellen_amatangelo@byu.edu.

QUALITY CONTROL CHINESE monographic RECORDS CASE STUDY fung yin K simpson university illinois purpose study east asian cataloging productivity U benefitted RLIN RUN CJK OCLC CJK systems since 1980s recently developed information many developments technology windows online environment concept resource sharing cataloging between RLIN RUN CJK OCLC CJK resulted great efficiency CJK cataloging these developments what current state chinese cataloging records order answer question study first reviews recent developments cataloging chinese materials I1 then go explore quality chinese monographic records checking critical errors affect retrieval comparing error rates among cataloging records various sources library congress LC LC adapted records done institutions LC copy OCLC member institutions member OCLC tape loaded RLIN RUN member institutions member RLIN recent developments chinese cataloging developments information technology 1980s RLG later OCLC developed CJK chinese japanese korean systems allowed catalogers display input CJK scripts bibliographic databases 1990s resource sharing established between RLG OCLC allow CJK records exchanged moreover both RLG OCLC seeking international cooperation augment bibliographic databases expand functionalities searching inputting displaying CJK records two examples accomplishments suffice RLG incorporated five libraries china arid arld thirteen RLG member libraries create international union catalog chinese rare books imprint years range 1080 1795 1 OCLC tape loaded nearly 300000 japanese records waseda baseda university information network WINE database december 1995 2 both RLG OCLC developed continuing technical improvements systems OCLC CJK plus system introduced may 1993 windows features development made much easier easler edit CJK bibliographic records using copy paste functions another convenience availability multiple screens rip flip through several bibliographic records november 1994 CJK plus system migrated PRISM service searching capability expanded include phrase search boolean combined keyword searches useful searches program pinyin pincin wade giles romanization conversion well record validation became available stage 1995 RLG released 31

software compatible windows following year inputting zn methods CJK characters expanded include pinyin pincin wade giles romanizations alongz 7 character components inputting method 3 above functionalities advanced CJK cataloging user friendly technologically advanced cataloging environment resource sharing CJK cataloging RLG OCLC tape loaded each others member librarys CJK monographic records since november 1989 however editing needed clean tape loaded records order meet inputting standards each systems example involves use word division aggregation RLIN CJK special symbol called aggregator used tojoin individual chinese characters together form semantic unit tape loaded chinese records OCLC do aggregators aggregatory required RLG hand vemacular vernacular punctuation spaces RLIN acceptable OCLC furthermore tape loaded records both databases usually lack call numbers 4 OCLC member libraries enhancing tape loaded records appropriate call numbers since december 1993 OCLC tried retain call number RLINCJK records tape loaded OCLC database part selected RLG membersrecords records implemented initially number records lacking call number among tape loaded records decreasing constant editing extra space proper punctuation still required LC copy cataloging fall 1993 library congress started treat copy cataloging standard activity order increase cataloging output reduce arrearage external source records bibliographic utilities OCLC RLIN sources adapted purpose records done copy cataloging identified lc Iccopycat 042 field 5 these records examined reexamined approved LC development should increase accuracy records databases methodology although small case study done local institution statistical analysis should provide important observations concerning quality control chinese bibliographic records data consisted 380 chinese monographic records selected OCLCs worldcat WorldCat OCLC online union catalog processed between october 1995 february 1996 asian library university illinois urbana champaign monographic records processed during period considered although random sample cataloging record OCLC database reasonable assume records processed during period typical rather exceptional advantage approach each item physically available verification along matching bibliographic record 32

order limit scope current publications avoid retrospectively converted records reprinted titles titles published before 1987 excluded bibliographic records entered before 1990 selected titles included 31 titles published late 80s 349 titles published 1990 1995 each bibliographic record following fields examined date entered worldcat WorldCat imprint date presence LC classification number presence LC subject headings cataloging institutions these fields well numbers errors entered microsoft EXCEL spread sheet table I1 shows 10 records illustrate structure data data spread sheet exported SASJMP statistical analysis TABLE I1 LIST ELEMENTS bibliographic RECORDS record E yr pyr P yr LCCN LCSH institution error 1 94 93 y y member 0 2 95 93 y y lc copy 0 3 95 93 y y lc 0 4 95 94 y y member 0 5 95 93 y y member 0 6 95 94 y y member 0 7 95 94 y y lc copy 1 8 94 94 y fiction mriln rlin alin 0 9 93 92 y fiction member 1 10 90 87 y y mriln rlin alin 1 completeness bibliographic records addition checking existence LC classification number LC subject headings error counts constructed checking crucial errors likely affect retrieval record examples include mis romanization tag code errors incomplete omitted fields correlation between error rates cataloging institutions determined compared results analysis among 380 sampled records 89 records cataloged LC 23.4 234 59 records cataloged through LC copy 15.5 155 152 records cataloged OCLC member institutions 40 80 records cataloged RLG member institutions tape loaded OCLC 21.1 211 see table 2 percentage distribution study suggests LC contributed 211 roughly quarter chinese records remaining three quarters contributed member institutions either OCLC RLG 33

percentage TABLE 2 cataloging institution institutions number LC 89 23.4 234 LC copy 59 15.5 155 member OCLC 152 40.0 400 mem RLIN 80 21.1 211 tape loaded total 380 100 completeness chinese records provision oflc ofle LC classification numbers LC subject headings adding parallel chinese scripts core fields three basic criteria evaluate completeness chinese records these 380 records 43 records lacked LC classification numbers among them 16 records without LC classification numbers result inappropriate tape loading see table 3 six records did assign LC subject headings excluding fiction literary works six romanization records lacked parallel chinese scripts required standard CJK cataloging study approximately 14.5 145 chinese bibliographic records incomplete respect basic requirements CJK bibliographic records almost these incomplete records provided member institutions except LC minimal level levei level 7 record without LC classification number TABLE 3 LC classification NUMBER institution yes total LC 1 85 86 LC copy 0 62 62 member OCLC 26 126 152 mem RLIN 16 64 80 tape loaded total 43 337 380 34

error counts error rates addition completeness bibliographic records error rate another vital measurement quality control bibliographic record each record carefully investigated fixed fields variable fields checking elements based requirements program cooperative cataloging PCC non roman core records 6 common errors categorized fivemajor groups code errors rule errors misspelling ISBN errors additions additional entries see table 4 TABLE 4 LIST ERRORS FIELDS field name code err rule err misspell ISBD err additions TOTAL fixenc 4 4 fixcont 4 4 hyill hxill fixill 9 9 fixdates fixothers 8 8 020isbn 2 7 9 04x 1 5 6 fix dates 10 10 09x 2 2 lxxrom 2 8 3 2 15 lxxver 2 8 2 2 14 245 Rom 5 14 23 13 55 245rom 245 Ver 5 12 4 3 24 245ver 246fi 250rom 246ti 2 3 1 8 1 15 250 Rom 3 2 5 250 Ver 3 2 5 250ver 260 Rom 12 21 8 0 41 260rom 260ver Ver 1 12 5 0 18 3xx axx 1 4 32 7 14 58 4xx&8xxr 2 1 4 4 4 15 4xx&8xxv 2 2 2 4 10 axx 5xx 6 4 6 5 21 6xx axx 6 17 6 3 6 38 7xxrom 1 10 5 7 7 30 7xxver 7 2 5 14 TOTAL 62 108 114 75 71 430 14.4 144 25.1 251 26.5 265 17.4 174 16.5 165 100 165 35

code errors incorrect omitted MARC coding tags indicators subfield codes failure encode fixed field required bibliographic input standards counted additions tape loaded records LC classification number retained enhanced encoded level remained L less full then did count code error contraventions ofaacr2 LC rule interpretations LCRIs categorized rule errors incompletemissing Incomplete missing transcription title publication information treated rule errors contradictions LC authority file name entries counted category addition regular english typographical misspelling errors several types misspelling errors chinese records mis romanization incorrect wade giles romanization letters diacritical symbols improper use omission hyphens personal geographic names wrong chinese characters vernacular fields extra spaces between romanized womanized words vernacular words counted ISBN errors due different cataloging standards ofoclc OCLC RLG tape loaded recordsfrom RLG RLO accordance OCLC requirements CJK chinese japanese korean cataloging unnecessary spaces word separations frequently appear records situation occurred many LC records cataloged through RLIN frequent error occurrences table 4 seen field 245 title statement field 260 imprint information high error rates finding similar lei zengs bengs research both OCLC RLIN samples 7 however zeng indicated inadequate ISBN punctuation spacing missing entry inadequate space editing three frequent errors followed misspelling romanized womanized words 7 according study misspelling womanized romanized fields frequent error followed rule errors then ISBN errors both fields since 1994 OCLC CJK system implemented PRISM environment problem improper spacing drastically reduced full record basis editing tagging errors detected corrected through record validation therefore inadequate spacing problems mentioned zengs bengs research reduced significantly years current study misspelling romanized womanized fields frequent error justified controversy unpopularity chinese romanization system even native speaker take time learn leam how romanize accurately enough search database 8 inaccurate chinese characters rarely occurred regularly used characters cannot found OCLC CJK character input codes rule errors fields 245 260 consist mostly improper missing transcription title information inconsistency between romanized womanized vernacular fields zeng pointed same error occurrences study common complex description chinese title page various forms title headings authorship publication information colophon page often states publisher well distributor long official title names furthermore required parallel fields roman vernacular apparently exhausted patience chinese catalogers missing entries inconsistency 245 260 common apparently due human error while inputting record rather unfamiliarity aacr2 LCRIs 36

another common rule error name entry conflict LC authority entry fields ixx lxx lax 6xx axx 7xx axx many CJK personal names share same romanizations CJK authors often use various forms headings publications however LCs authority file provides roman form CJK personal names cannot effectively distinguish name another 9 consequently often confused catalogers particular chinese personal names name entries records added types ofromanization adjust need patrons necessary catal ogers choices name entries romanization conflict LCs authority entries these should added locally kinds errors critical affect nonetheless affect retrieval particular additional name entries subject entries important I1 mentioned earlier six 380 records lacked subject headings excluding fiction literary works however catalogers should take care note subject headings provided order convey complex subject matter provide added name entries indicate authorship editorship important record ISBN known accurately ISBN important key elements searching foreign publications databases case searcher familiar language system romanization used database error rates among cataloging institutions 430 errors identified descriptive parts 374 records six records romanization lack parallel chinese fields counted maximum 10 errors each yere 490 errors 380 records descriptive cataloging mean error rate per record 1.29 129 compared among cataloging institutions LC smallest error rate 0.80 080 0120.12 followed LCs copy cataloging 0.864 0864 olg0.16 016 errors OCLC 1.71 171 019 olg0.19 019 errors tape loaded records RLIN 1.35 135 0210.21 errors see table 5 addition above errors descriptive cataloging subject headings 43 records lacking LC classification numbers 26 OCLC records 16 tape loaded records LC minimal level level 7 record consequently mean error per record 1.40 140 LC still maintained least error rate 0.81 081 OCLC 1.89 189 errors tape loaded records 1.55 155 errors per record order test significance difference between means tukey kramer honestly significant difference HSD method used overall significance level 5 10 tukey kramer analysis shows member library error rates OCLC tape loaded significantly higher LC rates original copy cataloging 37

ERROR RATE TABLE 5 cataloging institutions institution number mean std itd error LC 89 0.79775 079775 0.12293 012293 LC copy 59 0.86441 086441 0.15937 015937 member OCLC 152 1.71503 171503 0.19461 019461 mem RLIN 80 1.35000 135000 0.20905 020905 tape loaded mean error rate per record 1.289474 1289474 member institutions result does indicate RLIN records fewer errors OCLC required input standards composed members these two cataloging utilities different six romanized womanized records included study cataloged OCLC member institutions may CJK workstations available cataloging needs tape loaded records RLIN OCLC preselected RLIN member libraries least case instance counting records paired chinese scripts excluding six romanized womanized record OCLC 1.37 137 mean errors descriptive cataloging very close error rate 1.35 135 tape loaded records conclusions baseline study done leazer rohdy 1993 showed LC produced 91.2 912 bibliographic records domestic publications 61.8 618 records great britain 382 records foreign publications11 study 23.4 234 publications 38.2 chinese bibliographic records cataloged LC maintains high accuracy 0.81 081 errors per record member institutions cataloged 75 chinese monographs database resource sharing CJK cataloging exchanging bibliographic records between OCLC RLIN expanded both databases increased availability CJK bibliographic records timely manner member institutions foreseeable future OCLC RLIN likely convert foreign MARC records cataloging databases conversions doubt provide timely access foreign bibliographic resources north america benefit collection development reference catalogers may spend time enhance these records order maintain high quality control databases complexity chinese records difficult maintain quality control above study error rates shows frequent type error 26.5 265 errors see table 4 mis romanization various fields widely recognized romanization cannot properly differentiate thousands unique chinese characters effectively however nonetheless remains common method integrate chinese records unified online catalog environment north america currently online public access catalogs academic libraries do CJK displays regardless flaws romanization system access 38

currently use should therefore standardized should clear guidelines catalogers library patrons follow second common category errors rule errors made 25. 1 errors 251 calculated study see table 4 incomplete omitted transcriptions entries frequent errors type compared regular english bibliographic records chinese bibliographic record require lengthy complex process create parallel fields romanization chinese scripts important apply concept simplification cataloging chinese materials especially field 245 field 260 even crucial comprehensive CJK cataloging manual incorporates rules special guidance CJK catalogers follow present aacr2 workbook east asian publications published 1983 12 needs revised adhere current rule changes current cataloging envirom environment nent least moreresearch research should conducted bibliographic control CJK materials complicated cataloging issues addressed standardization cataloging achieved NOTES 1 five year RLG chinese rare books project successfully completed dec 1996 RLIN 1 237 focus 2 japanese records waseda baseda university added 21910 OLUC janfeb Jan Feb 1996 OCLC newsletter 3 RLIN terminal windows version 3 june 1996 rl17vfocus RLIN 202 asya 1994 quality timeliness chinese japanese monographic 4 tsao jai hsya reocrds RLIN database library resources & technical services 386063 63 5 LC expands copy cataloging Z program 1993 cataloging service bulletin 62 292 9 6 cooperative cataloging council non roman core record task group final report 1994 7 zeng lei 1993 quality control chinese language records using rule based data validation system part 1 evaluation quality chinese language records OCLC OLUC database cataloging 0 classification quarterly 162566 66 8 mccloy william brokaw 1993 cataloging chinese legal materials cataloging & classification quarterly 17181195 195 9 yu abraham J 1995 enhancement linkage online name authority file bibliographic database committee east asian libraries bulletin 106117 17 east4sian 10 sail sall sali john ann lehman JMP start statistics guide statistics data analysis 39

using JMP JMP insoftware software belmont duxbury press 1996 11 leazer gregory H margaret rohdy 1994 bibliographic control foreign monographs review baseline study library resources & technical services 39 2942 12 association asian studies committee east asian libraries subcommittee technical processing AACR 2 mork Work workbookfor east asian publications madison university wisconsin madison libraries 1983 40