iAiiSIST Newsletter, Vol. 2, No. 3 (Summer iy7tt) NUTF.S ON THE UISTRIBUTION UF LABUK IN A SOCIAL SCIENCKS DATA INFOHMATiON NETWOKK. Haul tvan Peters Social sciences Intormatin Utilization Laboratory University Center lor International Studies University of Pittsburgh FHAMEWUKK Tne point-oi-view wnicn i apply IMlHUUUCllON to MKUFs and networks can be expressed fairly simply. Within ine concern of these brief the last five to ten years we have remarks is to contribute to the witnessed a very important adjust- aeliberations having to do with ment in the aspirations of those networks and Machine-Readable Data involved in the "data library move- Files (MKUFs; by calling attention raent," to use a single descriptor to the fact that one way to proceed to refer to those activities navng IS to aistribute labor judiciously to do with secondary uses of social among network participants. To science data. We used to hear distribute labor is to distribute a rather frequent mention of national very large cost, and, therefore, and even international data librar- investment, associated with network les. indeed, with varying degrees operation. Distribution of invest- of success, some nations have ment often shows an added benefit embarked on such a course. More ot distributing commitment with as and more, though, we are focusing many people as possible having a our attention on perceived stake in network success national/ international efforts or failure. The distribution of directed at organizing mtorraation labor also invites specialization about data rather than data there- by network participants so that selves. 1 agree with this change those in the best position to of "target" and am optimistic that accomplish certain tasks are this dream can be realized, encouraged to do so without dis- traction irom other matters. This The sort of thing 1 have in mind modularity fosters incremental when 1 think of a "data information growth and is thus well attuned to handling system" (UlHS; can be smiting tinanciai arrangements and sketched as tollows: a changing population of partici- pants. Most of all, though, care- 1. The information flow ful distribution of labor is an begins with the primary organizing style uniquely appropri- researcherC s) . Through ate to grass-roots movements strug- the efforts of this agent gling to legitimize their inter- a set of social science ests. It IS style that can be data is created. if we found in the history and operation are lucky, a codebook is of lASSlST itself. For all of also prepared. The flow these reasons 1 would like to take ends with a secondary a moment to pursue this thought to researcher, who saves see what it has to offer. time and money by learn- ing of another - by - lAiiSioT Newsletter, Vol. i", No. i lijuinmer lyytS) researcher's data and using them. within "organizational reach" ol these research- ers things are consider- aDiy simpiitied il some data handling agent (UHA) IS available. 'Ihis agent could be a data archive or library, a statistical consulting service, a computer center, an aca- demic/research library, etc. Assuming the avail- abilty ol such an agent, either knowledge ol the researcher's data or actual responsibility lor them can lollow. 11' such an agent does not exist then knowledge ol' the existence ol research data IS more problem- matic; both the dissemen- ation and acquisition ol data and inlormation thereabout is more hapha- zard , resting upon an uncertain set ol proles- sional incentives. ol system design, operation, and user groups . j. UHAs ol' mation economy scope Kather all pos as need ually s tern ne with a tlhis simple simpler tage 1 justi f i as both ducer s wnateve unit d this e and qui ler a d handling in the ol than con sible r ing to b urveyed , ed only 11 possi IS by problem .; Anot s that ably be the pr and con r bib rives th stablish te poten ata sys ar cov ceiv esea e CO th ke ble o me UHA re imar sume iiog e s es a t Id inlor- tem an ea of erage . ing 01 rchers ntion- e sys- ep up UHAs. ans a merei y ad van- can garded y pro- rs ol' raphic ystem ; rare entity A tio est "sh pre By led kee 1m The sys kno My was tha ent Cth b) res all to que tag ice unn lar "ro tho to ing to minimal da n system ablished ared" inlo sent state word-ol-mo ic catalog ps abrea dings and problems tern are a wn to thi favorites te of UHA t all hav ries in m e tip of t a waste earcher el have to UHA askin stion ain , the berg; ; a ecessar y ger , bust" UH se with e under wri /promotion operations ta inlorma- would be If the UHAs rmation, the of affairs, uth and per- updates one St of new acqusitions. with such a 11 too well s audience, are : a ) a effort in e redundant ailing lists he iceberg) ; of secondary fort in that go from UHA g the same repeatedly tip of the nd , c J an bias toward financially As , I.e., nough money te market- in addition we have beg to think 1 national or data inform agent CDIHA tive record by some mea to the UIHA would be e central int (. computer-b course ; . 1 bank would issue prod provide se major produ publ ication un , therefore, n terms of a international ation handling ; . Uescrip- s of some type ns would flow Ihere they ntered into a ormation bank ased , of he inlormation be used to ucts and to r vices. The ct would be a containing YU - iAii^J.i>T Newsletter, vol. ^, No. J lijumraer lyytt) tne descriptive records searcn service vendor, and a set ol' suitable such as Lockheed, iiUU , indexes. ijucti a publica- BH3, etc., or both. tion could taKe the lorm of a directory, whereby each edition replaced the It may seem that the problems faced previous one, or a per- by LlHiJs are simpler than those lodical, whereby each faced by DIHSs. 1 submit that this issue reported new impression can be traced to the records and cumulative lact that LlHbs have become a fam- indexes are printed once iliar teature ol academic lite and per year. Two services that this fact can cloud our recol- would be provided, both lection of how things used to be in of them involving custom- the 1S)5Us and even, for many, the designed searches of the 19faOs through the early 1970s. 1 information bank: a; believe that the similarities are Desired data could be strong enough to invite coraparision described and the entire so that the "light" UK LIHS experi- bank could be searched ence can be shed on UIHS develop- Ithe "retrospective" raent. Perhaps even a co-develop- search mode); b) Desired ment of these two types of systems data could be described is in the otling; this is certainly and recent additions my hope and 1 don't believe 1 am could be searched on a unique in this respect, periodic basis tthe "cur- rent awareness" search mode) . ANALYSIS The DIHS schematic just outlined resembles the structure of litera- My analysis of the distribution ture iniormation handling systems of labor in DlHSs begins with the ILiHbsj as tne combined talents of observation that IHSs always have library, computer, and mtormation one or more "bibliographic units" scientists have brought them into CBUs) which are the life-blood of being: their operation. 1 believe that three types of BUs can be found in 1. A primary researcher can a DIHS: be viewed as an author while a secondary 1. The most complete type is researcher can be viewed that ot the actual data as a reader. and codebook themselves. These can be collectively '^. A DriA can be viewed as an referred to as a "study academic journal or a documentation." It may book publisher. both be that these have been collect materials from transferred by a primary authors and make them researcher to a DHA. The availabe lor use by read- DIHA snould not assume ers. that this is the case and Should provide for parti- i. Tne DlHA can be viewed as cipation by individual a secondary publisher, researchers who desire to such as Sociological maintain discretionary ADStracts, or a computer control over their data - Yl - lAo^iU'i' Newsletter, Voi.. c", No. i liiummer lyva; i. even tnough they are willing to release infor- mation about them. I'he second type can be called a "study descrip- tion" CiJUK by this 15 meant a somewtiat exhaus- tive char acter izt ion ol the data: the research problem lor which they were produced, notable aspects ol' the methodol- ogy, time-1'rame, popula- tion, etc., and comments about condition, accessi- bility and so I'orth. 1 propose that the prepara- tion 01 this be regarded as the responsiblity ol the individual primary researcher, UHA , or both, i wish that it was rea- sonable to think in terms of a "standard bU scheme" but i think that the most we can hope for in the near future is the pro- mulgation of a set of guidelines, much as aca- demic journals offer to their contributors. 1 regard the 5U an DIHS muc that of journal L1H5. T the bibl sent to t UIHA the lyzed and tion" (iiC pared . constitut entry t informati would be the ent LIHSs , vi article citation , keywords . this d its h as the artic he 5U lograp he UIH i)U wou a " s ) wou These e th o the on ban very r les z , au title , abst Ihe concep role 1 re indiv i le in would hic A. At id be tud y c Id be 3Cs e ac gro k. s i rn i 1 a bound thor n jou r act , uiriA w t of in a gard dual a be unit the ana- ita- pr e- wold tuai wing They r to in ame , r nal and ould have its LIHA jour var i cont expe for . ble crea file had 5U r This one- indi coll libr othe thes to be a receipt is in 1 nal art ance in ent cted an It wo for t tea mi of all received eproduct could at-a-tim viduals , ection ar les , r basis e two al s fl of ts r icie fo shou d uld he crol th and ion be d e ba a basi or in tern ex ibl SUs eceip s ; rmat id accou be po UIHA iche/ e iiUs to o servi one SIS , n en s , on bet ative e in as a t of some and be nted ssi- to film it ffer ces . on a for tire for some ween s . 1 have refe without nav in specific For the pur concern is believe an are forthc efforts in one, has areas. My ing what 1 with these red to ing sa they poses not a swers oming proces Actio inter thin BUS. the Id m wou of t pre to from s an n Gr est k se iiUs and uch about Id look 1 his paper ssing one. this ques a number d lAiiiilST, oups in IS in desc hould be iJCs what Ike . this 1 tion of for both rib- done This discus their processi to mind anothe sion of labor have presented arrive at a Ul filed alter a and entered i bank. The r that all aspec the direct c ratner than their own SCs that a valuab uniformity wo this means, the case insot tne :5Cs consis sion ng b r que impl thin Hi) w 5C h nto eason ts of ontro havin , is le i uld IThis ar as tent of y a stio icat gs, here ad b the wh the 1 g tha f no be IS the wit the BUs and UlHb brings n with divi- ions. AS 1 a i)U would It would be een composed information y 1 propose SC be under of d UIHA, UHAs submit t 1 believe t necessary achieved by particularly "coding" of h the provi- - Y^ iAiiiiiST Newsletter, Vol. t! , No. i tiiuraraer lyVbJ sions ol a vocabulary control simplil ication ol bu preparation protocol are concerned. iiucn pro- and transmission would hasten tocols are now regarded to De a development ol' SD standards. The necessary requisitie of reliable most advanced possibility along IHS operation.; 1 also believe, on this line would entail the SC the other hand, that the majority information bank being available in ot the inl'ormation provided by the much the same way that the Uhio bL should be imbedded in the bU. uollege Library tenter CUCLCJ has This would ease the i>C preparation made available Library ot Congress labor requirement at a UIHA. tven CLC) cataloging data. These ideas though this is the case it is are exciting and technologically important that a UIHA be responsi- feasible. Nevertheless, without a ble tor the accuracy, completeness, firm base of support 1 fear that and general quality of the SCs. setting our sights on these equip- Specific, identified accountability ment cont igurations as a tirst step in this area is extremely useful. of development is ill-advised. l believe, to say it once again, that 1 have two additional observa- we are involved in grass-roots tions on this subject. Part of my organizing and should use distribu- position vis-a-vis the division ot tion ot labor concepts appropriate labor attendent with SU and iiC pre- to this, paration is based on my experience with journal article abstracts in There are also division of labor the L1H3 represented by United concerns to consider at the output states Political Political science end ot the system. in my scheme a Documents lUi>P:Sl)}. This experience UIHA would broadcast the entries in shows that, in general, there is its information bank by means of quite a difference between publication, directory or periodi- abstracts written by authors and cal style, and search services, those written by document analysts retrospective and current aware- in a L1H3. My belief is that the ness. To every extent possible, abstracts ot analysts are better the preparation ot the publication than those ot authors in the senses should be done at a UIHA using that they are more accurately des- advanced information-base manage- criptive ot an article's actual ment concepts such as those we use contents and that they are more for USPSU. A sophisticated pro- adequate predictors of source docu- gramming system, for which many ment relevance. My second observa- models are available, can produce a tion is that even though l have SC listing and all desired indexes described a situation in which SUs from the same "linear" file as is arrive at a UIHA by mail, l believe used tor searching purposes. This that the UIHS should encourage sub- means that searching and publishing mission of SUs in machine-readable can conveniently interact so that torm. This may also have a bearing the intormation bank can be pack- on the standardization ot i>Us aged and repackaged in a very issue. It is possible to conceive dynamic fashion with little or no of UHAs presenting iiU data to a additionak keyooarding labor. Ihe UIHA through use ot an interactive most desirable output from such a program housed in the UIHA compu- system would oe "camera-ready copy" ter. It IS also possible to con- for the publication m question; ceive ot a uiHA making a program given this, the only "external" available to accomplish the same cost to be incurred by the publica- thing using the computers of the tion would be that of printing and UHAs. Whichever, the - 1i - irtSSiST Newsletter, Vol. d. No. i Uummer lyYttJ Dindi inter compu raent , would encod WHO a nave cond i an ar the such costl logic ng. nal ter- t be ed in oes ; pri tion rang puDl a th y an ally It the control based t hen the to prod agnetic most nters wti s should ement wh ication ing wou d it IS necessa UIHA or ypese best uce tapes metro do the ereby is Id be no 1 ry. did not have some sort ol' tting equip- al ternati ve appropr lately for someone politan areas Under no UIHA accept the text of rekeyboarded ; inord inately onger techno- Ihe reasoning applying to provi- sion of search services differs from that of publication. Here the co-operation ol a uiHA with other agencies is not only possible, it IS desireable. The minimal method ot service delivery would entail the contacting ol the UIHA by someone interested in having a search performed. A discussion ot the characteristics of desired data would be held with a "retrieval specialist" who would compose "search strategies," review results, and lorward what seemed to be the most relevant entries in the information bank to the inquirer. The experience of LlHijs leads me to believe that, depending upon volume, tne "turn-around" time for such a system would run irom lU to lY days and would occassional ly take even longer than this. Evi- dently, a more time-responsive approach would be preferrable. The manner in which this could be accomplished would entail the pro- vision of access to the bU file by way of an on-line, interactive search and retrieval systeiri. IWq UIHA could engineer such a system itsell or It could transmit its information bank to a vendor of searcn and retrieval services, someone like Lockheed, iiUU, or BHi> to name the current three major commercial actors. 1 favor the latter approach primarily because it woud reenlorce what 1 consider to be opmen secto It w which UlHb this opera tant ind Iv their resul form and r answe an ex tint r of t ould would in It sort tion . thing idual own ts as of on- etr lev r . trem he he i also hav e ot s He wou or sear soon line al ely promi 1 iteratur n formatio reduce e to be b tlorts to earch an gardless , Id be to a UHA ches and as possi , interac system wo sing devel- e searching n industry. the costs orne by the estaDlish d retrieval the impor- enable an to pertorm to receive ble. iiome tive search uld be the Two footnotes can be added to this discussion. birst, 1 believe, consistent with my reported oeliefs throughout this paper, that the place to start is with the provi- sion 01 search services through a retrieval specialist attached to a UIHA. It is a tried and true tech- nique. It concentrates investment in labor rather than equipment, it can be initiated quickly, and it provides a powerful evaluative framework. Nevertheless, the eye of the UlHb would always be on the on-line, interactive approach and as soon as adequate financing Decame available and oroad popular interest and support became evident the ulHi> would move in that direc- tion. It is also important to note that such a movement would not ren- der obsolete the retrieval special- ist role at the UIHA. My experi- ence with LlHSs, again, provokes this observation. Some individuals and UHAS , tor various reasons, will never want or be able to assume responsibility for their own searching and retrieving activi- ties. Others will perform searches so infrequently that they will desire to consult with someone with more experience. in short, retrieval specialists attached to any IHS always add value to compu- ter-based search and retrieval procedures in a way analagous to Y4 iAiiSiST Newsletter, Vol. ^, No. i liiummer I^YBJ tne value added by a carpenter to a 1 am lelt with tne impression tnat nammer. mucn more needs to be said. My intention has been to direct atten- My second footnote is tnat the tion to the question ot who will do UIHA should nave the capaoiiity to what in the UlHbs we are all envi- transmit the iiC intormation bank sioning. This seemed to be worth into any requesting agent. This doing because much of what I have means that suitably equipped and been reading and hearing recently motivated individuals and UHAs has been distionctly oriented could acquire the entire informs- toward equipment concerns. It tion bank and updates thereto so occured to me that it was once that tney could process their own again time for someone to raise the searches locally rather than issue of the people dimension ol remotely. For agents with a high the data information production, volume ol' searching activity this acquisition, organization, and dis- possibility will allow greater cost semination process under discus- control than the other arrangements sion. This is what 1 have discussed so far. It will also, attempted to do. l believe that need l say it again, distribute this concept of "distnou- labor in a useful way. tion/div ision of labor" which has occupied my attention in this paper IS extremely important and can pro- vide a very concrete principle CUNULUlalUN guiding how we choose among compet- ing courses of action. I hope that These "brief" remarks have wound its articulation and examination up as rather "extended" remarks but will lead to further discussion. YtJ -