lASSIST Newsletter, Vci. 2, Nc. I (Spring 197d) THE NOHC LIEBAEY AS INFGR.IATICN CENTER AIJC DATA ARCHIVE PatricK 3ova National Cpinion Research Center INTRODUCTION Tn to pr igin searc its i munit probl With tnis Libra itsel many organ that which publi e ob j ovide ot t h Cen nf crm y It ems the paper ry , w t, s ways, izati Dart ' inf c. ective a desc he Nat ter (N ation b serves encount data ar descr e will ince th the en. T of th ormatio or rip ion ORC ase an ere Cfli ibe als e pub this t ion al Op ) Ll , the d so ve. s t he o ref Libra lie a NCRC agenc passe pa or t inio brar use me o th er t ry i rm Libr y \s t The scribed the se classif as a c book closely search researc inf orma all ma N03C, N search ety of chive f generat It IS service versity people researc about a Univers data fa of the houses estmgl text. General form of rect re and gro cial sc NCRC as nse y lib onven and alig inter h pro tion nner CEC s in ge publi or N ed wi def i data of C acqui h as ata 1 ity cilit Sec data y eno the U Soci the suit wth lence Libra a spe that rarie tiona ref e ned t est grams of f i of tuaie neral cs. ORC d tn N nitei arc h 1 c a g re a wel ocate maint y (wi lal from ugh i ni ver al Su I] ORC of t of NO ry c Cldl word s. 1 1 renc t of t ce w info s, , t It i ata ORC I- ° hive ata 1 a d el ains thin Sci ICPS n th sit y r vey Lib he RC a an best librae is us It is ibrary w e coll he cur re he intr It is th hich su rmation and surv c a wide £ the da and for particip ct the for th althoug from NO £ inior sewhere . a se the Di ences) B and, e presen copies The p rary is recent h nd of t per IS ne or- n Re- y and t com- f the cially thougn € NOBC o NORC £, m of the ary is hro jgh the te de- y, in ea to active ith a ection nt re- am ural e NORC pplies about ey re- vari- ta ar- data at ion - local e Uni- h many BC for nation The parate vision wnich inter- t Con- or the resent a di- latory he so- HISTORY CF NORC The history of NOEC includes changes in the size of NOSC and in the number, complexity and subject matter cf studies undertaken. NOHC was founded in 194 1 as a place where the then new technique of opinion sampling could be appxied in the neutral setting of a univer- sity. From 1941 to about 1960 the center nad a modest budget and a small staff in the Cnica,.j headquarters and in a New YorK Cit; office. NORC could be character- ized in that period as a researj.i center housed in an ola mansion or. campus; the staff was small enouga to "meet every day in the dinir.: room of that house ror coffee di.l cake. The during of naf of fo the U. tween these occupa Shirle t udes series which by the tratio versit no tha lona reig S . 194 late tion y St towa of some Ce n St y- table t per 1 sur n af f Depa 5 an l\' ar St r d me healt are nter udies St lod vev3 dlf s rtme d 19 the rest udy ntal h-r e stil for (CH uuies were on t cond nt of 57 (m 1947 iqe st of pop iilne latea 1 be in Healt AS) a conduct ci the serie,; he conduct acted for State be- ore about Nortn-Hatt udy, the ular atti- ss , and a surveys of g repeated h Adminis- t the Uni- A li serv icago nt a inten dex, fairs riod is in on te £, P e typ e tod dent ete. eful ow at out 1 e inc arch nds o brary ex e tiie p and New nd usefu ance of by subj studie from 19 dex cont xt with lus seme e and si ay, alth that it The i in acce Roper C 25, and iuded , tool in f anaiys isted 1 rot ess io York. 1 activ a quest ect, of s conduc 45 to A ains th percent mf orma ze. It ough we is any tern ind ssing th enter) , since t tne fil itself, is. n tnose nal staf One im ity was ion or the for ted in Pril, 1 e rull q age mar tion on is stil are not longer ex is g ese stu which nu he raargi e is a for cer days f in por- the item eign tne 957. ues- gin- sa n- 1 in con- com- uite aies mber nals re- tain Bes librar als u Dook a relate the s have a tant h done i rather survey then tained These used f have n taem u user i the f i are pr that t laes y ma sed nd 1 d f tudi con ealt n th ext res prod , i fil rom ot p to nter les. obao ime the i n t a i in 3 ourna o the es be cise h-rel e 19 ensiv ults ucing nclud es s time m ade date est. of ly a perio gues aed tudi 1 CO su ing COl 1 ated 50s) e ri fro th ing till tc t an due We cou uni d. tion file es llec b jec Qon ecti sur ies m ma em fore exi ime , atte to in rse, que inde s of and a tion t mat e (we on or vey r In ad of p ny a were ign s St a alth mpt t cost tend sin resou X, the mater i- small closely ter or still impor- esear ch dition, oil and gencies main- our ces. nd are ough we o keep and low to kee'J ce they rce for N t w j X >_' 1 1 o r 2 (3pril;.j 1978) r. a X 1 - t t .-. t 1 n were NORC mac h does li i o u r n a result ether tact m i ri i n m tact t f uncti Even in present NORC, tion ac The 01 publ tae per academe occur re t he Lib plexity creased well be the li and div sionai new sta tne lib tion ; request creasm .po. as t,.dt ti. ^tj idt: dl le ior and s J ert bee E i s -S u 1 t r. e I a as weli Cpjinion r r 3 m s t zatiga th li the tiv 196 ic lOd an yon tra €rs st f f rar as br It nd w as d th ear^ aria e vej. y wa n e t a s o r V t y r austi at no ^ re s :» 1 o n as p u News, w u^les OIiS. D r a r y w n a t e V e r t NCFC t; St ua y 19 to I. ijegan or pub quite i 1 ur d r e a a _: 1 i z r. e ts dlt s tnat is re that txlSLl .11 C.I c by NOK Publi as pr publi was o y air s, wh wor Ki lie m low. Os cna repr ese change At 'io h 1 c n ha y — t.-. studie Q the t d h e a 1 1 ry cox iried. aif gre mambers y f or h public and qu ref err nged NCRC's ion, j her aspe ever ai : nsequenc , umber an n jed ntd t: d ot; PC s< opic r. st iect Th w a cam isto CO lit esti ed t n progre s t roil I udies, s ions ex e NOSC p r. d c h a n g e to dep Lical in act i n c r ens wer c the li id I., a my jn st air ; riles time lea se'o Gallup ng t.ie cvared C and c coa- cbabiy c con- 1 1 e n a ector . en the rg for for raa- styie ust as cts or things e s for d com- ss i.n- erated c taat panaed rof es- ed end on forma- eased, e m- brary. THE CONTE.^EORAEY RECCED Or NOFC The brief preceding dis has largely to ao with the i research programs at K05C. NCEC founded its Survey R Service (SRS) to i:oriiiai.ly research services, chief! collection (but includm phases of survey research), social science community. T ter had previously contrac conduct studies zoi axt sponsors (a good example Stouffer communism study in conjunction with the Gallup zation) , but SRS was active moted as sucn. The major teristic or SRS tnat sho noted is that in most cases role was tecnnica^. — we nai xittle to ao with the planning of these studies a tie to do with anaj.y2ing d reLJOrtmg the results. Tni acleristic has implications library ir. its mrormation a aichive activities, as we see . Coincident with the founding oi S;iS were tiie start of the Great So- cietv poverty prograns which in- c^uata Luiit-m evaluation appro- priations. SKS was available to c ussion n ternal In 196J eseardh provide y data g ail to the he Cen- ted to r amural is t .1 e 1554 in o E g a n 1 - ly pro- cnar ac- ula be NORC s usually i nitia 1 n d iit- ata and s c .-; a r - for the no data scall carry out some of the massive data collection programs needed and be- came quite busy at it. The pub iences c one CO iences re atten e public Iped alo ANS-acti ere wer cm in hi a u a t e creased the CO ces -- BC Libra re resp rger aua lie nange uld s wante ti ve appa ng b on an e oth gher stude sophi mpute that r y a onsiv ience image or d m the ay that d the p to its r rentiy y such a Ps_ych er curr educatio nt popui sticatio r i h t h had eft nd force e t o a the social 19b0s, too the social ublic to be esults, and began to be, magazines as oio^y Today. en"Es -- "The n and in the ation, the n in the use e soft sel- ects on the d it to be larger and One a grea the da to 196 i n g da person pe rson ail, s as muc we see to be haps. At NOR dr amat chivin State Roper ter in symbol activi which deemph releas ve ys w ticipa devel t impa ta arc peop ta bu o^m hiv le manne impier h Qata now. more than C the ic ac g was Depar Public 1959. ic ot evolve asize ing t h as sig te oppo a tS or W str man fir t r th tme OK our on d d arc e S nil e r Our de- lve nnm opos uaie d c uld ivin ovid ded f el chiv e ar the uec emph nad g an It io s, m ondi not g p e tn to t t w e (w chiv sta ision asize to dJ arcn n. any q tion. get roper e run ry to ere u e hop es wh dies) ent in t on the N e movem were cer n a mor sea to a Times there si demand hat ther a i g h t f o y studie s t and p elated e deposi nt stud inion R his act policy ar chivi uring th hiving, tate Cep leant be evised p to try NORC as with t i ve is Our dat uite old The funds to ly, and ds itsel send t sable t ed to fi ich woul ue 1 9 CHC L ent. tainl e pe n ag were , mply for e was r w a r d d are robab to da ting ies esear is mo (if ng a e 196 The artme cause olicy 60s had ibr ar y: Prior y shar- rson to ency to after was not them as tended , per- today. ly most ta ar- of the in tne ch Cen- re than not our t NORC Os: to act of nt sur- it an- , at le a data ha fact an ex pan a are and man or ganiza do the could f, so we hose stu o a log nd appro d liouse ast, ar- that si ve NORC y m tion ar- not de- aies ical pri- many Aithougn data archiving nas been aeemphasized, we have not stopped archiving machine readable data. The old studies mentioned above — although not aix are worth equal attention -- still require much work beiore we would inflict them on any arcnive. We thereiore pro- vide sample survey data for second- ary use, but we do not coileet or keep data frcm other sources, ex- ^5 - lASSIST Newslettdr, Vci. 2, Nc. 2 (Spring 1978) Cfcpt those which may have been acquired tor a specitic project.. Just as our data archive activ- ity IS closely idtntirieJ with NORC production, so is our information activity. And just as data arc the result of a long line of activity, so too do our lof ormat ion services cover all ascects of these events, not only to flORC staff but also to the interested public. The mously and mor only in studies conduct sponds ation o get so spond . duction to scho tereste guestio cial s within search, met hods procedu mg) a matter. Our erated— acce of r e t r are wor cation subject start, tion t cases, request user as require we have are sup study d The which r teresti con trac of data 2-3 ye contact to seek serves least o the ori with s working passes sets ar usually will ge and all make th ers, i cases i best, since tape (e sur act i e pe th but ing to m n MO aeon We to cl g d cien the Q (e res nd vey ve t ople e re ais them est RC p e on als NCRC roup n VI e re ces, pur uest (eg: witn move hese ar suit o in re-ju ro je th and s th si ti eiv b view ions que in par ment day e in s cf th The es ts cts e 3 rovi £ur at ng ut o] hav stio terv ticu IS s, an terest thes e me t h Libra for i or tr taff t de an vey re might NOHC. ever t mainly surve e to d n wor iewer lar s f robl inf or ieval king of a or depe he u the ed, best d, we a CO plied irect em w mati at are on s stu met ndin ser resu ana we sup py- on or p ith u on is prese prim olut 1 d y on hod g on neea its o here can. ply t In ly wi ermis sing NOR the usu nt our m itive, ens. T a part is jus what in £. In f a stu we info If da hem, as any case th spon si on. enour- d more ed not e many cds ot ry re- n I o r m - ies to c re- intro- sear en be in- The he so- stay y re- c with ding) , tram- ub ject C gen- ai one ethods but we be lo- icular t the forma- most dy are rm the ta are suming , data £or or data fo eside a ng iim ts call after ars) ,with permis useful f whic ginal cmeone in the mere an e "wil ha ppen t tired cws us e data ncludin t turns origina we will xcept t NO bo. f o a our the sion pur h is anax who sam d mo led" sis of bian aval gar out 1 c no mayb xtram RC a Ai r pub et t feel origi to poses to ysis is e are re o to that givin ket p lable chi ve that opy o t ha etc ural re i thou lie ime ing na 1 use lear and pr a. f th NORC the g pe ermi to s. we f t ve u cop studies n an in- gh most release (usually is that sponsor the data not the n about tc talk es umably As time ese data What sponsor emission ssion to all ccm- In some have tne he data, sed the y a few tiiras) and since many original sponsors lack the institutional setting needed to care for tapes and documentation for any length of time. These aata sets are treatol in the same manner as NORC's own data.. THE i:iPACT OF USER DE.'IAND We of how of the sfaouiu fiae na sc lent set so concer of doc at NOR ersity which dence mentin change analys docurae turn user Libra af f ec ture o i f ic e me pr ned wi umen ti C: ho or compet over t g of s s in is ha ntat io now to aem and ry (an t tha f s urv nte r pr ior iti th tw ng or w t o c in for e with he ar urveys da ta ve cr n prob af f e d pe t wor ey r ise es . o has a r c h i ope w matio and chivi , an pro eated lems. onsi cts rhap k) esea may We ic VI ng ith n tak ng a d ho cess ad deration the work s how It and how rch as a help us will be problems sur ve vs the div- recuests e prece- nd docu- w recent ing and ditionai in serv quit r r cm help soph have quir of te pass The to g on da anot list ten pear day prag sa ti tent ne o the e a e di hig wit isti ser emen n fi ive care et a her of just s t to mati sf y for f the NORC L wide va verse i h schoo h a deb cated ious a ts. A nd ours way , t f ai wo data 3 use is thing things does n hat in day rou c rout those ser vie f undamen ibrary riety o nfor mati 1 stude ate topi profess nd compl s a res elves re o immedi rk that et in sh quite o to do to do , ot get a the hur tine, e only who are tai prob is that f users on need nts who c, to q ionals ex data ult we q acting, ate dema is regu ape for ften si in tne and most one. It ly-buriy we take and try most in j.ems we with s — need uite wno re- uite in a nds. ired sec- mply long of- ap-. ot the to sis- IM IQtl OF THE ARCHIVIST Now, everyone deserves atten- tion, one would presume the high schooler as much as the potential user of one of our surveys (which may lie "unarchi ved" , in unusable shape) . Wny should we pay special attention to the archive function when limited resources are stretched in the first place? We suggest that the scientific nature or survey research demands that special attention be paid data for secondary use. Tne overriding philosophy of the NORC Library in Its role as conservator of NORC surveys is that the entire history of a survey should be preserved so 46 - I A S 6 1 .S T K t w s * ci 1 1 .lab ;£ib . t. one is- t.ie ri- cse re- ie , rOCUMENTATIOt; AND ITS USES Idea ii e n t a t i data, b ixcat lo lished least , concius This id cai at stages uce wr our CO whicn r of sarv ing in field K s p e c 1 f i publica cation en OL Y t he n m findi to i ions eal 1 NOR or sj 1 tt en deboo ecogn t y do forma crk a cat lo tions and c then a s way) oraer ntexl draw s no C , s rveys doca ks CO izes camen tion nd t ns, and odmg £udy wou to cr , igen n f r t wh ince do men t ntai the tati on rain bibl tne ins ne rinai (not o Id p e r m 1 test th at tn tly jad cm tue ciiy imp tae V in fact ation. n infor broader en by 1 the s in g, au lograpfii usual da tr uction doc J- niy of t rep- e p JD- e ver y ce the data . racti- ari js pr od- Taus nation n a t a r e ncl ud- ampie , est ion es of ta io- Although the view of documenta- tion outlined above does not solve the daily crunch, it does give a basis for deciding wnat documenta- tion should include. It is tco bad that there is really very little pressure for good documentation even from the aost sophist icatea users. The rare exception is tne graduate student or professor work- ing in the area of survey methodol- ogy. In t largely documen vey . traditi which m ment mi one kne at abou widcspr y s i s " s transit data pr terns, " for mat i data a older s be kep jobs to now a p p Udl res ant do w o r K . ra o r e a o bv tne di V 1 .; ja le wo depe ;at 10 i n 1 1 1 :ns ide ssing . ab ; t r.e :ad u )f t wa I on f ;ce = s :here ) n d b ;alys 'Stem in cthe :ar s :arcn ;£ mo Cnt ;umen ;cr k L res rk of naent n pro cece c; pr It po inio oat o same se oi re p rom o ing has out d IS o s re g crae r pe to De er or St o canno ta tio organ tare Ti arc on Quce ntly oce s ssi b rma t tner 1 1 sta dcka pera to c been a ta dire r t rson tha re f hi t, h nth izat hiving, the guai d by eac , ther sing at le to s ion fro St udie me. Wl tistical aes an tot cont f e ra ting a loss f tccessi ur veys. d t R a t r c comma £. Tne t tne 1 n search a £ or he c w 6 V e r , an is re ion. T regjires ue are ity of h s u r - t were NOBC upple- m' what £ aone th the anal- d tne rolled s ys- cf in- r. g and The ecords r 1 c a t e trena divid- £SlSt- r own expect qui red he in- fewer to n a litt esea . A ed V as e u NOHC main docu ant may proauce bout his le need, rca, to s a con- ariacles used by n usable. Library tain the ment the der ived be re- orig inal records ana tends not intelligible informatio work, because there is at tnat point in the r communicate with others sequence tne construct anu final fern of data researchers tend to b For these reasons the emphasizes the need to original file and to derivation of import variables so that they constructed using th Ja ta.. Cider and more recant surveys pose very different sets of prob- lems for archiving at NOHC. An ex- ample of each will point out some of these. An example of an older study is the Hay, 196'4 occupational prestige study which is the basis ror the Hoag e-S iegel-Hossi prestige score (used in the 3eneral Social Survey). The most remarkable and revealing tning is that this study was not archived until February, 1976, Que to study director reluc- tance to allow secondary use. In tne not so distant future, studies such as these will be virtually ir- retrievable, for the simple reason that much of the data are saved only on aunched cards (which are warped now) with multiple punches, and the collective memories of the people involved with the studies tend to be more and more vague. Almost all NORC data prior to mid-1960s can be expected to con- tain some iaultiple punches which require special processing that in- creases cost (and tediousness, too) of archiving- This problem also makes it dixficult to create fresh copies of data. Time, in general, is a crucial factor in data so- trage, since even tape copies can be expected to become unreadable after a while. There is also the problem of reconstructing documen- tation for these older studies. In many cases we must rely on memories or oersonal fixes to recreate some crucial piece of information. Fi- nally, with the immanent loss of the last rew pieces of unit record equipment at NORC, we can expect that the solution for the problems of archiving older data sets will become even more difficult. The Continuous National Survey (CNS) was archived in December. 1975 and is an example, although extreme, of some problems with more recent studies (tais study was con- ducted rrom April, 197 3 to May, 1974). Although tne data are docu- mented and usable, they are only dVdilaole in the form or SPSS sys- tem files or character coded data derived from the SPSS files. The original data were not processed by - 47 - lASSIST Newsletter, Vci. 2, Nc. 2 (Spring 1978) the tr Rather, rield f ent irel ware. The or tireiy of the usable mentdti way to SPSS sy back to questio But agg piex do system ally mo associa aditi they ormat y wit which igina lest; chara becau on. veri stem the nnair revat cumen files re st tea w onal me were pu and pr h custom was ne 1 data an in cter cod se of a There i f y the files ( seven or es) . ing prob tation r as oppo raight f ith orig ans nched ocess desi ver u have terme ed d lack s vi guali short eigh A ies lem i eguir sed t orwar inal at in ed gnea sed bee diat ata of rt ua ty o of t th s s s th ed f o th d ma data Noac. free- almost sof t- again . n en- e form is un- aocu- lly no f the going cusand erious e com- or the e usu- te rial CONCiaSICNS Archival problems will ccntine at NORC -- of that there is no doubt,. If there was serious con- cern with archiving as a part of the methodology in the social sc ences, ana perhaps more of a reco nition that a scientific enterpri requires good documentation (a the means to get it) , we wouIcl more optimistic. As it is, we e pect to see more archival probie with recent studies and increasi difficulty with aata from old studies. S such that with of N fair sure must cont publ ad ju more job NOPC perh tion brar that a glum as a s the car ORC info ly well, s for cope. inue to ics as St ou r time w of Study studies aps a fa 's data ies. note peci e o rmat CO ser v For be a we c do can ster arc should , let ai libr f and c ion and nsiaer 1 ice wit the f u s respo an be, orities the al cumen t a fin a t rate, hives not end us point ary mand ommunica data , w ng the p a whicn ture we nsive to but we ana s 1 impor tion so heir way into tae nd data on out ated tion e do res- wa will our may pena tant that , at na- li- - 48 -