lASSIST Newsletter, Vol. 2, No. 1 (Winter 1978)

THE IMPACT OF COMPUTER NETW0BKINf5 ON THE SOCIAL SCIENCE DATA LIBRARY

Alice Robbin

University of Hiscor.sin - Madison

ABSTRACT

This paper explores the factors which have constrained the social sci-
ence data library's participation in the use of computer networks as a
vehicle for accessing information. It also suggests why changes in the
situation can be expected and further suggests some of the ways that com-
puter network resource sharing will affect the social science data li-
brary structure and services.

data and computations, sharing re-
sources, and providing information
services" (EDUCOM, 1973). At the

INTRODUCTION same time, there continue to be a
number of factors which constrain

Traditional libraries are expe- the development of network resource
rienced in meeting information sharing and information servicing,
needs and in coordinating activi- These factors include:
ties among users. Computer tech-
nology has made it possible for the 1. technical considerations,
traditional library to service the which revolve around
user more quickly and (many enthu- processor configuration,
siasts of on-line data bases would software and communica-
add) more complicated than when the tions (Davis, 1972);
reference librarian relied on man-
ual methods for searching and re- 2. financial considerations,
trieving information. Computer which involve the nature,
technology has made it possible for size and distribution of
traditional libraries to avoid some monetary support;
costly duplication of human and
technical resources. 3. organizational and polit-

ical considerations.
While the traditional library is which include the struc-

a relative novice to computer tech- turing of the network,
nology the social science data provision and nature of
service organization (data library services, monitoring of
or data archive) has been linked to performance, source and
the computer and modern technologi- distribution oi authority
cal developments by the very nature and responsibility (Da-
of the medium of its collection, vis. 1972), degree of
machine readable data. The social control at the network
science data library, while tied to level and local level,
advanced technology, has continued and integration of the
to use traditional means for locat- local efrort into na-
ing, transferring and accessing ma- tional network efforts
chine readable data files fMRDFl (EDUCOM, 1973);
and for communicating its needs ana
coordinating activities related to U. legal considerations,
MRDF. Moreover, the social science which involve federal and
data library has neither benefited state legislative re-
from the set of experiences of the strictions, which at the
traditional libraries in resource federal level prohibit
sharing within a networking envi- monopolies and restraint
ronment nor utilized network com- of trade (Clavton Act of
puters to share resources and ex- 1914) and protect the use
pertise and to cooperate for more of communications serv-
efficient allocation of resources. ices as a public utility

(Federal Communications
Act of 1934, and Neumann,
1973) , and which at the

CONSTRAINTS IM THE USE OF state level are designed
CDH7IJTEt^N~T¥glKTNG to protect the outflow of
B7~7HE lOCIir state dollars for the
IklM^E D5T^ LIBRAII buying of non-state serv-

ices; and.
For several years now, as a

growing number of articles, mono- 5. user considerations,
graphs and books attests, computer which include knowledge
networking has become an "important of user neeas and
mode for remotely gaining access to



lASSIST Newsletter, Vol. 2, No. 1 (Winter 1973)

characteristics, ease of These small centers remain
system access, use and invisible to policy makers and thus
operation, a variety of when decisions are made about com-
services to assist in ef- puter use and about activities
ficient and productive which will involve interaction with
operations, education and centers outside the home institu-
training, and documenta- tions, these centers are never in-
tion. formed.

Even with these constraints, the One of the results of the lack
use of networking by libraries, re- of institutionalization is the per-
searchers and students, particu- ception of these services as non-
larly in the natural sciences, has legitimate and its staff as unpro-
increased during the last several fessional. The staff members are
years. This has not been the case not viewed as professionals by the
for most professional social scien- user community which employs their
tists, who have had little or no services, nor do the staff members
experience in network use, or for perceive themselves as professional
the social science data library, data specialists, although many
whose clientele are social scien- data center personnel are indeed
tists. experts in data processing and han-

dling. Although a staff member may
What accounts for the low level in fact be performing the work of a

of use of computer networks and why reference librarian or information
has there been no network resource specailist who searches and re-
sharing by the social science data trieves selected information upon
library? The reasons are struc- request, as indeed most staff at
tural (the result of political re- local data centers do, that staff
alities and historical accidents)

,

member usually does not recognize
economic, sociological, and exper- the role he perf orms--t ha t is, can-
ientiai. Structural factors in- not assign a name to the function
elude the lack of data services, of being performed. He usually lacks
institutionalization of the inrorm- a methodology for the tasks he is
ation service at the local level, performing. Professional training
and of professionalization. provides tools and products (re-

sources) , an explanation for the
Efforts at coordination and re- activities performed by certain in-

source sharing have been made at dividuals, and a Metholodogy for
national and international levels, task performance. But. in most
but only minimal efforts have been cases, the data center staff member
made to encourage the development has not been trained as a profes-
of infrastructures at the local sional.
level. Major archives have main-
tained their dominance and have^ Why should the lack of profes-
contrary to public expressions or sionalization affect the use of
support, done little to encourage computer networking and resource
the development of the local data sharing? Reference work implies
center. But, networking depends on knowledge of and understanding of
the creation of local "nodes" and the nature and potential of availa-
without the local effort, national ble information resources. If peo-
networking will not be successful. pie are unaware of resources, they

cannot utilize them. The informa-
Most organizations which provide tion specialist today is made aware

data services are structurally of on-line data resources at the
weak, existing as appendages to one introductory course level and re-
unit of a larger parent organiza- ceives training in informational
tion, rather tnan as an independent (bibliographic) data base creation
unit within the parent organiza- and manipulation. The housekeeping
tion. Hith only tenuous funding and maintenance functions performed
support, the staff must dedicate by libraries are facilitated by a
its efforts to maintaining services networking environment. In other
with a continually eroding funding words, our data services personnel
base. The staff therefore has few are uninformed of the potential use
or no incentives to develop or em- of networking facilities because
ploy networks to communicate with they have not been trained as pro-
information services outside its fessional information specialists,
local environment. In generalj
these data services are small ana Economic reasons also explain
operate on the periphery of comput- why the social science data library
ing activities of their parent or- has made little use or computer



IAS5IST Newsletter, Vol. 2, No. 1 (Winter 1978)

structu
sof twar
ineffic
sources
i n e X p e n
the use
tor) h
revise
ter pro
assista
an ext
data,
or crea
anaiyti
ceptabl
sis pro
society
support
patabil

res,
e CO
lent

si ve
ran
ave
docu
gra
nee
enae
proc
ting
c ca
e as
cess

ha
was

ity.

SO
uld

us
Staf
com

d St
had
ment
sup
De

d p
essi

so
pabi
pect

s h
te:
and

that
be wr
e of
f ti
modit
af f (

litt
ation
port
lays,
roces
n g t h
f twar
litie
s of
In ot
ad en
dupl
dela

special
itten,
comput

me has
y , so t
and adm
le ince
/ prov
and bet
as a r

s of
em and
e with
s have
the d a t
her wor
ough m
ication
ys.

purpose
and for

ing re-
been an

hat both
inistra-
ntive to
ide bet-
ter user
esult of
locating
locating
certain

been ac-
a analy-
Qs, tne
oney to
, incom-

Th
commi
techn
purpo
servi
bly b
velop
tangi
proce
mg i
the
user.
which
whose
short
been
gener
port
of p
solve
at a
lecte
forma
vant
diffi
-they
It is
suppl
the t
for e
prese
commu
cient
commu

MR
from
to th
a f e
varia
probl
prima
use.
reali
of tr
his h
are p
the d
guter
le t

unive
compu
side
physi
ganiz

e society has made
tment (investment)
ology for direct
ses, but not for
cing purposes. Tn
ecaase it is far e
measurement tools

bie byproduct than
ss, which informa
s. Products can b
funding agent an

But, informati
rarely offer a "

"product" is rare
time after the

supplied, have be
ate sufficient fu
its intended goal,
articular data
particular proble

particular moment
d from a very larg
tion, most of whic
to the problem,
cult to justify i
must be taken on
difficult to demo

ying information
otal social cost
xample. Althoug
nts a useful te
nicating informa
ly, the resources
nications are not

DF CO
seve

e pop
w to
bles.
ems h
ry c
the

stica
ansmi
ome s
rohib
ata' s
fund

o tha
rsity
ting
the

caliy
ation

llections ra
rai hundred
ulation of th
literally t

Although t
ave not prove
onstraints in
social scient
lly have the
ttmg the in
ite. Transm
itive. Utili
home site r

s which may
t researcher

prohibits
dollar expen
campus. Wh
transfered f
to another.

an
in
appl
inr

is i
asie

to
to

tion
e of
d p
on s
prod
ly v
serv
en u
nds

th
requ
ms,
in t
e se
h is
It

ntan
fait
nstr
will
of r
h ne
chni
tion

to
aval

nge
obse
e 'J.

hous
ecnn
n t

ne
ist

form
issi
zing
equi
be u
bee

the
ditu
en
rom
cos

enormous
computer
icat ions
or mation
s proba-
r to de-
judge a
judge a
ser vic-

fered to
otentiai
er vices,
uct" and
isibie a
ice has
nable to
to sup-
e supply
ired to
required
ime^ se-
t or in-
irrele-
is verv

gibles--
h alone.
ate that

reduce
esearch,
tworking
que for

ef f i-
provide

lable.

in size
rvations
S. , from
ands of
ological
o be the
tworking
does not
ssiblity
ation to
on costs
data at

res com-
na vaiia-
ause his
use of

res out-
ata are
one or-

ts which

are
calc
for
ment
and
data
tion
data
pear
tne
dist
if
made
prod
ment

T
face
to
for
whic
sele
user
the
acce
remo
or ob
ther
quir
acco
er s

.

sues
data
more
remo

incu
ulate
proce
ation
over h
is b
in V

prod
s to
cost
ribut
this
, ho
ucer '

rred
(sta

ssing

eadf!
ilied
estme
ucer,
have
of a
ed in
physi
w do
s ca

are
f f an
the

rcell
Th

for
nt i

and
no pr
data
this

cal
we p
iital

fair
d c
r eq

ing
us,
the

ncur
th

oble
fil
fas

tran
rote
izat

ly
ompu
uest
and
the
cap

red
e se
m ju
e wh
hion
sf er
ct
ion

easy to
ter time
, docu-
postage,
buyer of
italiza-
by the
Her ap-
st if ying
en it is

But,
is not

the data
invest-

heor
of

writ
the

h a
cted

data
ssin
te u
lems
e ar
e re
unti

T

an
ev

te a

et ic
it,
e an

CO
d1 us
f ll

uch
wo

g th
ser

ne
e ac
prog
ng s
here
bout
d ho
iden
cces

ally
it
ac

mput
ts c
es
that
uld
e d
wou

ed
coun
ramm
yste
are
who

m
t wh
s ar

, at 1
appear

countin
er bil
osts o
by the
the 1

not be
ata fil
Id. A
to be

ting on
ing of
m of m

philo
should

uch, w
en net
e invol

s qui
g al
ling
f ac
statu
ocal
bill

e,
var
re

es wh
the

ost
sophi
. P^Yuich
worki
ved.

on the
te easy
gorithm
system

cessing
s of a
user of
ed for
but the
iety of
solved,
i c h re-
current
comput-
cal is-
for the
become

ng and

There are a variety of
ical reasons which expla
social sciences data lib
community has not made u
computer networking and
sharing. In general, the
user community is not p
use electronic means to
formation and communicate
other. Computer networ
requires a moderate f
with computers. Como
while becoming more comm
social sciences, is usual
to one or at most two cou
mester in each discipline
is largely for a class p
an exercise in data hand
information or data m
Although the profession
scientist may from time
frustrated in his inabil
cate MRDF sources, in
relies on his colleague
"invisible college" for i
on sources of data and
themselves. This actio
ates the lack of support
library, since it perpe
fact that the best data a
chived (but if one knows
people, one gets access
rormation) and reside i
hands, and presumes tha
brary staff can provide
sistance in information s
the social scientist.

sociolog-
in why the
rary user
se of the

resource
library 's
repared to
access in-
with each

king still
amiliarity
uter use,
on in the
ly limited
rses a se-
, and use
roject and
ling, not
anagement.
al social
to time be
ity to lo-
general he
s and the
nf ormation
the data

n perpetu-
for a data
tuates the
re not ar-
t h e right

to the in-
n private
t data li-
little as-
eeking for

- 5 -



lASSIST Newsletter, Vol. 2, No. 1 (Winter 1978)

Re
resou
sense
the
makes
whom
on k
which
the
who a
ratri
EDUCO
bert
tor o
tium
hinise
mariz
with
have
about
thems
lytic
worki
needs
cial
conce
bert
tists
ing,
socio
commu
expla
netwo
and t
what
cial
its u

source sharing
rces are known,
, deemphasizes
invisible colie3
individuals les

they know and
nowii-g informat
are machine bas

assistance of
re specialists
eval. Soni.e yea
M conference, R
(19 73, p. 14 9),
f the Inter-Univ
for Political R
If a political s
ed social scie
regard to r^lRDF;
three needs: (

data, (2) acce
elvjs, and (3)
capaoilities.

ng meet these s
? The discussio
scientists was c
p t of networking
concluaed that
were not ready
Thus, it has be
logical aspects
nity which prov

'or th
which

nation
rking use: lacK
he lack of under
networking is an
scientist could
se.

implies
and

the imp
e, sin
s depenu
more dep
ion res
ed and r
i n te r m e dm infor
rs ago,
ichard H
former

ersity C
esearch,
cientist
ntists

'

He sal
1) mfor
s s to t

h

a vaiiabl
Kow coul
ocial s
n group
onfusea
, a n d H
social

for ne
en perha
of the

ide the
low- lev
of expe
standing
d why t
beneiit

that
n this
act of
ce it
ent on
endent
ources
eguire
iaries
mation
at an

of f er-
direc-
onsor-

and
, sum-
needs

d they
mation
e data
e ana-
d n e t -

cience
of so-
by the
or fer-
scien-
t w o r k -
ps the

user
best

el of
rience
about

he so-
from

new methods of organizing and

CHANGES IN ORIENTATION

For
of comp
in netw
must be
izing i
creatin
product
product
must b
standar
order t
and to
ity. A
nomic,
tial fa
will CO
use,
some op
ture s
network
velopme
tion t
require
ment

.

velopme
social
ices, (

for des
informa
profess
age the
1 1 o n t h
transfe

data
uter
ork
inf

nfor
g an
s.
3 de
e cr
as I
o r
enha
Itho
see

ctor
ntin
rece
timi
ocia
ing
nts
hat

or
Thu

nt o
scie

crib
tion
iona
inf

at
r of

11
net

reso
rast
mati
d u

To
scri
eate
or
etri
nee
ugh
iolo
s de
ue t
nt d
sm i
1 sc
acti
stem
info
gani
s, w
f a
nee
deve
ing

Is a
orma
cost
the

brarie
works
urce 3
r u c t u r
on ser
tilizi

sha
bing t
d.
their
eve th
the p

the St
gical
scribe
o con
eveiop
s war
ience
vities

from
rmatio
zation
e are
n i n f r
data

lopmen
and c
3) re
re nee
tion,
s of
infor

s to
and t

ha ring
es f o
vices
ng inf
re re
he se r
There
produc
e mf
roduct
ructur
and e

d in S
strain
me nts
ranted
data

T
the

n a bo
and

seeinq
astruc
librar
t of s
ontroi
cognit
essary

pr oauc
mation

make use
o engage
, there
r organ-
and for

ormation
sources

,

esources
must be
tion 1 n
ormation
's util-
al, eco-
xperien-
ection 1

network
suggest
for fu-
iibrary

hese de-
recogni-
ut MRDF
manage-
(1) ae-

ture or
y serv-
tandards
iing the
ion that
to m a n

-

recogni-
tion and
warrant

managing access to inrormation, and
(5) understanding of various tech-
nological advances which the social
scientist can utilize to access in-
formation and transfer and retrieve
data more efficiently and effec-
tively.

A
that
esse
tent
tion
grow
data
need
ophy
nati
ble,
orde
tial
in vo
comi
pubi
and
cial
Chan
uced
acti
spec
age
Rapi
mote
file
tran
SIS
anal
guir
tent
the
his
thes
ble.
are
the
will
chin
sear
need
suff
retr
and
ess.

cceptance among researchers
secondary analysis of data is

ntial to realizing the full po-
ial of expensive data collec-

has been increasing. The
ing cost of creatinq complex

riles to meet a variety of
s is resulting in a new philos-

that these data represent a
onal resource, publicly availa-

to be widely' disseminated in
r to realize their full potan-

Inter-disciplinary research,
Iving a variety of data, is be-
na of increasing importance for
ic oolicy planning and making
for" sustained analysis of so-
, political and economic
ge. Complex data files prod-
by large-scale data gathering

vities nave induced a need for
ialists who can organize, raan-
and document the information,
d reductions in the cost of re-
ly accessing tnese complex
s is making it unnecessary to
sfer to a local site for analy-
purposed. More sophisticated
ytic techniques are being re-
ed in order to realize tae po-
iai of these complex files, and
social scientist is readjusting
perspective on requiring that
e techniques be locally availa-

The researcner and analyst
finding it necessary to support
creation of structures which
organize a collection of ma-

e readable data, assist the re-
cher in locating data for his
s, and provide the analyst with
icient documentation to make
ieval of statistics possible
to eliminate error in tne proc-

Fail
trie val
recogni
data ri
the man
informa
recogni
tity o
very di
possibl
evant
evaluat
ation c
process
etv mus
gies fo
mforma
society
tion of

ures
ha v

tion
les
ner
tion
tion
f in
f f ic
e, t
bits
e th
olle
its

t Q
r CO
tion
' s r
its

th
need
of

tha
form
ult,
o lo
of

e ju
cted
elf,
evel
llec
whi

ecor
elf.

to
o t h e
Ther
t th
atio
ind

cate
inf

alit
and
an

op r
tmg
ch 1
dkee

nr or
dere
col
be

r CO
e i
e en
n 1
eed
and

orma
V of
* th
d th
at io

an
3 a
ping

matio
d a
lecti
organ
ilect
s a
ormou
s mak
somti
seie
tion
the

e col
at th
nal
d ret
bypro
ana

n re-
growing
ons of
ized in
ions of
growing
s auan-
ing it
mes im-
ct rel-
and to
inf or m-
lection
e soci-
strate-
r ieving
duct of
evalua-

- 6 -



lASSIST Newsletter, Vol. 2, No. 1 (Winter 1978)

This has led to a reevaluation experienced data handler who has
of the importance of libraries and never paid much attention to the
information services and the grow- quality of documentation and for
ing need for individuals who have whom the invisible college has op-
expertise in information selection erated effectively to maRe it pos-
and retrieval, organization, man- sible to obtain tne data he needed,
agement , dissemination, and docu- This support comes at an appropri-
mentation. The reference function ate time: During the last several
is becoming recognized as a crit- years, data information specialists
icai activity for the supply of in- have been working on guidelines for
formation to the user comraunitv. documenting MRDF, ranging from var-
The individual who serves as a ret- ious types of bibliographic de-
erence librarian for MEDF will in scriptions, products such as cata-
the near future be viewed as a pro- log records, indexes and
fessional. classification schemes and stand-

ardized study descriptions, to file
Support for the establishment of and variable level descriptions,

data libraries is the result of Software, such as the Interchange
recognition of the cost of social file, developed by Roistacher and
research: Data libraries represent Noble (1976), will obviate the ne-
savings in scarce resources with cessity of rewriting codebooks for-
their potential for collecting m matted for different statistical
one location studies which are of software packages. Thus, there is
value to a variety of inaividuals; support in the user community for
providing a systematic description the data professional's concern
of these studies to facilitate lo- about documentation and an apparent
eating and utilizing data re- willingness to accept ti;e recommen-
sources; reducing duplication of dations of these specialists so
purchases; providing centralized that better descriptions of riRDF
expertise in rile creation, proc- will be available,
essing, and description; and, pro-
viding a basis for a data services These developments encourage an
infra structure to facilitate ac- atmosohere in which computer net-
cess to data at a reduced cost worKihg can be accepted as a viable
through memberships in consortia and desirable means to (1) access
and through exchanges, enhance com- information processing services
munications about data resources, outside the local environment, (2)
and lay tne foundation for the de- share resources which will provide
velopment of data information orod- economy of scale or operation to a
ucts to benefit the user community. number of participants of the so-

cial research process, and (3)
Part of the failure in informa- share intellectual resources and

tion transfer and retrieval is due cooperate in joint programs (Rois-
to the lack of standards for docu- tacher, n.d.). NetworKS will pro-
mentation. In general, MRDf docu- vide a mecnanism for more effective
mentation has been poor and biblio- communication, cooperation and co-
graphic control unaeveloped. Sue ordination of information through
Dodd, of the Social Science Data services such as a social science
Library at the University of North data librarv.
Carolina, has commented extensively
on this through the the Classifica-
tion Action Group of lASSIST (1977
a- b, c) as have others involved in THE IMPACT OF NETWORKING
the development of documentation OK THE SOCISI
standards (Nielsen, 1977; Mochmann, SriEUCE CITIJ LIBRARY
1977; and Robbin, 1977). The so-
ciai science data librarian has of- The social science data library
ten been unable to locate files be- is a soecial purpose library or in-
cause no title or producer formation service which has been
statement was provided. Analysts created to respond most directly
do not know how to acknowledge tne and immediately to its special
use of secondary sources of data in clientele, a diverse community of
their publications. Study descrip- users composed of social science
tion providing brief histories of a researchers and students, policy
data rile have never contained ade- planners and analysts. It is lo-
guate descriptions of the data. cated in a variety of settings,

within government, commercial or-
In the past year, however, we ganizations, foundations, and aca-

have seen growing support for docu- demic institutions. The data li-
mentation standards. It is signif- brary may be part of a computer
leant tnat support for documenta- center, a larger library (facii-
tion standards comes not only from ity) , a (general) information cen-
the socia.. scientist who does not ter, a data archive (in the Euro-
have easy or regular access to data pean sense) , a social science
or IS not iinkea to a major iniorm- department, a research institute,
ation network, but also from the



lASSIST Newsletter, Vol. 2, No. 1 (Winter 1978)

market
may be
ganizat
which
(in mos
tunding
stances
may be
ing arr
of spec
may not
brary • s
provisi
(which
expert!

The
closely
activit
vide (e
tents a
ements
library
technic
clien te
els of
and tec
priat e
of the
techn ic
the inf
compreh
materia
collect
ble els
materia
other

,

analysi
the da
agement
quired
tion.

search organi
independent

The orga
library is

ases) the lib
urce, except
en part of it
ported by ex
ements for t

products (w
e a byproduct
gular activit

of specia
be the resu

re
an

ion
the
t c
so
wh

sup
ang
ial

b
re

on
may
se)

coll
as po

ies an
ff icie
nd ret
within
staf

al t
le, w
suDsta
hnical
inform
libra

al as
ormati
ensive
1 desc
ion an
ewhere
is fro
the p

s, an
ta to
and

for th

zati
ser

niza
imbe
rary
in t
s ac
tern
he
hich
of

ies)

It

on , or
vice or-
tion in
dded is
' s major
hose m-
tivities
al fund-
creation

may or
the li-
or the

services
of staff

ectio
ssibl
d is
nt) a
rieva
the

f 's
raini
ho ha
ntive
skil

ation
ry ' s
sista
on.
know

ribed
d

Fa
m one
repar
d the
the
analy
e dat

reflec
e the die
organized
ccess to i
1 of selec
collection
substanti

n g pr o V i
ve differi
, methodo
Is, with
on the c
collecti

nee in ut
The staf

ledye of r
in its re

tentialiy
e transfer

location
ation of d

relation
c o m p u te r a
sis softwa
a in the

t s as
ntele* s
to pro-
ts con-
ted el-

The
ve and
des a
ng lev-
logical
appro-

ontents
on and
ilizing
f has a
esource
ference
a vaila-
of the
to an-

ata for
ship of
nd man-
re re-
col lee

-

The
mber
dium
issio
blish
on 1
d/or
rvice
tiona
y be
velop
llect
ta m
ta a
d rec
nat io
es.
ta 1
ecial
r ized
spons
on o
cessi
d di
on re

data lib
of activi
of its

n." A da
er and pr
n machi

may pr
s much li
1 library
responsic
ment, da
ion, da
anagement
rchiving
ords mana
n, and d
In gener

ibrary wh
purpose
by three

ibie for
n behalf
oning mat
sseminati
guest.

rary
ties
coll
ta 1
oduc
ne
ovid
ke t

T
le f
ta
ta

(or
geme
ata

icn
libr
fun
loc
of

eria
ng t

engag
relate

ection
ibrary
er of
r eadabl
e inr
hat of
he data
or stud
pre para
process
data a

acces
nt, dat
ref eren

howev
respon

ary is
ctions:
ating
its cl
Is it a
hese m

es 1
d to
and
may
info
fe

or ma
the
lib

y de
tion
ing
naly
sion
a di
ce s
er

,

ds a
cha

I
info
lent
cgui
ater

n a
the
its

be a
rma-
f or m
tion
tra-
rary
sigh
and
and

sis

,

ing)
sse-
erv-
the

s a
rac-
t is
rma-
ele ,

res,
ials

The three basic functions of
reference, accessioning and disse-
mination require access to differ-
ent types of inrormation. The ref-
erence function requires access to
information about the existence,
availability and logical structure

of M
requ
abou

iir
soft
file
tena
stru
rela
ture
The
acce
need
rele
tion

RDF.
ires
t t
ical
data
ware
sto

nee,
ctur
tion

to
diss
ss t
s f
vant
of

The a
acce

he his
struct
to man
and

rage,
qual

e and
ship of

the
eminati
o infor
or effi
data f

inforraa

In
is inv
and ma
lar ba
as an
tion a
organi
cializ
for d
used t
ucts

.

brary

'

ers
potent
It ela
tion
upon r
inform
w h ic h
collec
variet
scribi
resent
about
data
cessf

u

manage
tions
which
and ca
format
produc
Stat f
retrie

othe
olve
nage
sis.
inte
nd t
ze i
ed p
ata

ge
T

S CO
appr
lall
ssif
to f
eque
atio
it
tion
y of
ng t
ace
dat

reso
1 i
ment
net
per

n su
ion,
e in
tra

ving

r wor
Q in
ment

Th
rmedi
he us
nr orm
resen
resou
nerat
o de
llect
opria
y use
ies t
acili
St.

on
then

Th
info

he CO
ess t
a res
urces
nf orm

req
work
form
pply

o
forma
ined
the

ccession
ss to
tory of
ure, rel
agement
compute

processi
ity of
document
the lo

physical
on funct
mation
cient r
rom a la
tion.

ds, the
inf orma

processe
e dat a
ary betw
er, to
ation t
t or pot
rees wh
e stati
velop th
ion, the
te info
ful dat
he gat he
tate it
The lib
acquire

integr at
e librar
rmation
llection
ools for
ources
t hems el

ation ga
uires a
among o
a simil
each ot
rganizat
tion pr
in elas
informat

ing function
information
the data,

ationship of
and analysis
r hardware,
ng and main-
the logical
ation, and
gical struc-

strueture.
ion requires
about users'
etrieval of
rger collec-

data library
tion seeking
s on a regu-
library acts
een inrorma-
retrieve and
o meet spe-
ential needs
ich will be
stical prod-
a data li-
staff gath-
rmation on
a resourses.
red informa-
s retrieval
rary records
d materials
es into the
y produces a
products de-
, which rep-
information

and for the
ves. Suc-
thering and

communica-
rganizations
ar function
her with in-
ions which
oducts, and
sifying and
ion.

W
pute
mg
br ar
coll
mana
enga
orga
pote
tion
othe
serv
zati

hat
r ne
on t
v 's
ecti
geme
ges,
niza
ntia
ship
r 1
ices
on.

then i
tworki
he soc

st r
on , in
nt

ind
t ions?
1 for

of
nf orma
w i t h i

s th
ng
lal
uctu
form
oces
rela

N
af f

the
tion
n th

ef
nd re
scie

re,
ation
ses i
tions
etwor
ect in
data
and

pa

feet o
source
nee da

ser
seeki

n wai
hip to
king h
g tne
libra
comnu

rent'

o

f com-
shar-

ta li-
vices,
ng and
ch it
other

as the
rela-

ry to
tation
rgani-

Networking appears ro indicate a
trend toward 'centralization of
services. The experiences of the
1950 's have shown that centraliza-
tion of services was not cost ef-
fective and has not provided more
effective and better user services;
however, in the last two years, we
have seen a move to centralize or-
ganizations which perform similar



IAS3IST Newsletter, Vol. 2, No. 1 (Winter 1978)

services (e.g., the Carter
Administration's efforts to reduce
"inefficiencies" in government by
reorganizing its bureaucracy). The
central computing center appears to
be making an effort to exert con-
trol, both directly and indirectly,
over other organizations which per-
form functions related to comput-
ing. After several years of reas-
sessing its role in the parent
organization, the computing center
appears to be moving toward'efforts
to centralize computing services
and computation activities within
the Darent organization. It ap-
Eears that the computing center may
e successful in its efforts be-

cause of its size (and budget) and
the mythology of expertise which
the computing center perpetuates.
In an era of continuous inflation,
there are increasing pressures on
the administrators of the parent
organization to maintain existing
facilities which service the larg-
est number of users and require the
largest budgets.

The social science data library,
specializing in assistance to so-
cial scientist who have tradition-
ally not been big users of comput-
ing services, is not able to
generate large-scale support.
Thus, I would predict that as net-
working becomes an integral Dart of
the activities of a computing fa-
cility or large information serv-
ice, data libraries whicr. do not
have strong. independent consti-
tuencies within the parent organi-
zation, will De subtly and not so
subtly pressurea by the parent or-
ganization either to be absorbed by
a computing center or to have its
functions provided by another or-
ganization which is responsible for
information services. It will be
hard to counter the trend toward
centralization of services, espe-
cially because the computing center
provides experience and know how
which are strong arguments for ex-
tending services to a wider market,
one which is covered by a social
science facility like the data li-
brary. But, the trend should be
resisted because it will mean that
special needs of a specialized user
community will most likely go un-
met .

In
the po
comput
ties p
format
frees
in a
data 1
which
and s
Networ
to a c
tion

prin
tent
ing
rovi
ion
the
free
ibra
pro

ervi
king
ompu
whic

ciple, n
ial for
centers
ding com
service
data 11
market

ry can 1
vides th
ces at
, howeve
ting cen
h has

etwor
inde

and o
puter
s.
brary
situ

ook
e be
the 1
r, po
ter ,

opera

king
pena
ther
-rel

Ne
to

atio
for
St
owes
ses
an o
ted

creates
ence for
f acili-

atea in-
tworking
operate

n: The
a seller
products
t rate,
a threat
rganiza-
like a

monopoly, because it affects its
utilization and revenue. Thus, the
computing center will make every
effort to control the outflow or
computing dollars elsewhere.
Ratner than entering into a free
market environment and upgrading
the quality of its services, the
computing center retrenches and be-
gins to exert pressure on the par-
ent organization's policy makers to
stem the flow of money elsewhere.
While it may be impossible to pre-
vent the user commuaity from taking
its money elsewhere, the computing
center may convince adminstrators
to make it difficult or at least
inconvenient through a variety of
bureaucratic measures to buy serv-
ices outside. I would predict that
this is a short term response and
in the long run it will be desira-
ble to develop competitive capabil-
ities which will favor the user by
helping to reduce prices for compa-
rable services and to improve serv-
ice quality and service availabil-
ity ^Neumann, 1973, p. 23). Thus,
I think that in the long run net-
working will free the data librarv
from overwhelming dependence on its
local computing center and at the
same time provide the data library
with better services from its local
computing facility. Protectionism
has never worked to the long term
benefit of the protected— as eco-
nomic history has shown in the last
century

.

N
rela
user
work
face
staf
time
ices
tion
suit
comm
syst
rath
for
1973
meth
been
leve
ing
guir
tica
have
prov
lar
Thes
carr
fund
mean
use
for
libr
more
suit
toma
pabi
expe

etw
tio

ed
ba

ati
uni
em
er
fu

a

,

ods
n

Is
is
ing
' I
ide
wit
e
led
ing
tn

or
use
ary
in

/ =i

ted
lit
rti

orking will also affect t
nship between staff a

Most data libraries ha
on a one-to-one, face-t
sis with their users. T
has provided extensive a
>nsuming user support ser
training, tutorial inform
documentation and human co
on, because its us
ty has preferred to use t
on an "as needed basi

than to spend time prepari
tare system use TNeuman
p. 3) and because differe

of user assistance ha
ecessary to meet differe
of expertise. When networ
introduced as a means of a
new information and stati

roducts, the library wi
o extend its resources
services for users unfami

h other operating system
services will probably

out without addition
support. Networking wi

e development of and great
automated interactive too
r support services. Da
staff will probably beco
volved in computers as a r
nd in the development of a
and interactive support c

les (and ironically, t
se required for the

ne
nd
ve
o-
he
nd
V-
a-
n-
er
he
s"
ng
n,
nt
ve
nt
k-
c-
s-
11
to
1-
s.
be
al
11
er
Is
ta
me

- 9 -



lASSIST Newsletter, Vol. 2, No. 1 {Winter 1978)

development will propel the data duplicatinq a copy of a data file
library toward the computing center each time there has been a request
which has more experience in sys- from outside its local environment,
tems development)

.

(This is in contrast to a book
loan, where one copy of a book cir-

Networking will make the user culates and there is no need to d u-
community even more aware of the plicate a copy of the book each
necessity for good documentation time a request is made for its
for data, software and operating use.) Other reasons which explain
systems. While documentation has why data have not been distributed
been primarily hard copy, network- on an interlibrary loan basis in-
ing will probably increase the elude the extensive capitalization
trend toward automating instruc- investment that a data library
tion, updating through the termi- makes in developing its collection,
nal. and development of systems di- difficulties in the physical trans-
rectories and catalogs of services fer of data and repeated use of the
and products. medium on which the data are stored

(typically magnetic tape) ; need to
Probably the greatest opportu- prepare a data file in a physical

nity that networking presents for structure compatible with the host
the data library is in the area of environment to avoid additional
retrieval of information about the processing; and, far more human
existence and availability of MRDF, time required to prepare data for
As documentation becomes automated, an external environment. Further-
networking will make it possible to more, it is not yet economically
search for information contained in feasible to transmit data remotely
"on-line data bases" of directories from one site to another because
of data holdings, contents of data transmission speeds are too slow
files, and codebooks and other doc- for the quantity of data typically
umentation for data files located analyzed by the social scientist,
at institutions far away from the and because transmission costs are
local data library. still too high.

Data libraries will have an op- Obtaining a copy of a data file
portunity to participate directly has been the only way that a datam cooperative ventures to create library increases its collection
data bases containing information (and thereby justifies the number
on the contents of data files which of personnel required to maintain
are located at local data centers. the collection, since quantity is
There will of course be non-trivial always a more tangible measure of
administrative and financial prob- service than quality) „ It has al-
lems to be resolved. But, network- ways been assumed that when a data
ing presents an unprecedented op- file is needed that it must be ac-
portunity to create resources of quired, even when, as in most
utility to a wider user community cases, what the user community does
tnan now served by small local data is prepare a statistical overview
centers and to enhance the quantity of the population in the data file
and quality of information now and prepare some inexpensive, pre-
available. liminary statistical results.

(This is called, "getting a feel
As networking becomes a more ac- ror the data" and most researchers

cepted activity in the generation begin their projects in just this
of statistical products, we can ex- way.) In many cases, the data
pect to see an increasing amount of which are acquired by the data li-
statistical analysis done remotely, brary on behalf of a user and are
although it is doubtful that there reviewed in this manner, are re-
will ever be more remote than local jected as not meeting the user's
access and analysis of data. Re- needs; and the researcher never
mote access to data files will af- completes a detailed analysis of
feet the data library: its collec- the file,
tion, how much time it allocates to
the accessioning process, and how For every data file which is ac-
it (and the user community) pay(s) quired, scarce resources of time
for data which are paysically not and money must be allocated to in-
in the data library's collection. tegrate the file into the collec-

tion. This accessioning process is
At present, a data library's expensive because the data must be

collection grows by acquisition of checked to verify that what is de-
a copy of data archived and main- scribed as its physical structure
tained by other data libraries and actually is and that the descrip-
repositories. It has never really tive materials (documentation) ac-
been feasible for the data library companying the data allow the user
to engage in an interlibrary loan to understand the logical structure
type of data exchange because the of the data and to carry out sta-
library has had to bear the cost of tistical analysis of the file, and

- 10 -



lASSIST Newsletter, Vol. 2, No. 1 (Winter 1978)

because duplicate copies must be
made to protect the data for future
use.

Thu
plicat
costs
tion
data f

and th
ting d
anothe
discou
sharin
tional
cooper
tensiv
brary
a data
at uios

s, because
e copies
associate

investment
rom one 1
e infeasib
ata remote
r, data 1
raged from
g methods

librarie
ative netw
e investme
in acquiri
file wnic

t a few ti

or
d wi

an
ocat
ilit

ibra
uti

used
s th
orks
nt b

h w
mes.

the need
a data

th capit
d witn
ion to an
y of tra
rom one s
ries Lav
lizing re
by more

rough re
There

y the da
and maint
ill be ac

to du-
f ile

,

aliza-
moving
other,
nsmit-
ite to
e been
source
tradi-
gional
is ex-
ta li-
aining
cessed

In
wouxd
limin
taken
acces
cause
sis w
acces
ferre
attac
for a
distr
could
acces
user
is ac
were
cover
data
distr
costs
charg
avail
cess-

Re
one
data
anoth
cal p
inevi
unf am
data
feren
ducin
vol ve
reque
inevi
indiv
elimi
view
acces
the
resou
tion
acces

a ne
be

ary a
plac

sioni
exte

as in
s of
d to
hed t
ny CO
ibuto
be p

s of
or f

o

cesse
to p
ed th
and c
ibuto
to

es i

able

tworki
acquir
nalysi
e or w

ng cou
n a e d a
tenaea
the da
the us
o acce
nsulta
r woul
aid ei
the f

i

r each
d.
ay a
e cos
onsuit
r coui
all it
cross
for n

ng envi
e Q only
s oft
hen acq
Id be 1

nd det
Cos

ta cou
er . A
ssing
tion t
J provi
ther fo
le by a
time t

If the
one tin
t of a
ing ass
d ca Ic
s file

all
etwork

ronm
af

he
uisi
usti
aile
t fo
Id b
fee
the
hat
de.
r fi
n in
he d
rem

e f
cces
ista
ulat
s an
data
rem

ent data
ter pre-
data had
tion and
fied be-
d analy-
r remote
e trans-
would be
data and
the data

A fee
r s t time
dividual
ata fiie
ote user
ee wnich
sing the
nee, the
e access
d apply

files
ote ac-

way o
trans
er

:

robie
tably
iliar
to b
t com
g sta
d in
St-
tably
iduai
natin
which
sioni
data
rces
whicr.
sed i

access v
f reduc
f e r fro
by elimi
ms of da
occur b
with pr

e access
gut ing
if and

proce
reducing
occur d
copies

g the bu
takes

ng proce
library

'

for acce
may be

n the f

la
ing
m on
nat i
ta t
eca u
oduc
ed w
envi
comp
s s i n
tl Dl

ue
of
rden
plac
ss

;

s a
ssio
very
ture

netwo
the

e ii
ng th
ransf
se pe
i ng c
ithin
r onme
uter
g th
e la
to pr
a da

of
e du
and,
lloca
ning
intr

rking is
cost of
brary to
e physi-
er which
ople are
opies of

a dif-
nt ; re-
time iu-
e data
g s which
ocessing
ta file;
data re-
ring the
reducing
tion or
informa-
equently

The data library has justified
its information gathering, acguisi-
ton of new materials, ana retrieval
and disseaiination of selectea in-
formation on the basis of the in-
formation's potential use for a va-
riety of individuals in its user

community, a
small part o
utiiizea on a
working makes
the cost of a
from the infor
individual wh
cialized infor
a reassessment
resources for
ment tasks wh
performs on a
lewer resourc
records ma nag
sioning basica
lection, reso
cated to ga
retrieving s
upon request,
tomated' user
which create
which could p
wide variety o
erse user comra
tinuing to
services orie
sc ientist

.

Ithough o
f its CO
regular b
it possi
cquiring
mation se
o require
mation an
of the a
inf or mat

ich tne
regular

es are a
ement (w
lly is)
urces can
thering
electea
and to de
support
intor mati
otentiall
f Individ
unities

,

provide
nted to

nly a very
llection is
asis. Net-
ble to shift
information

rvice to the
d the spe-
d to justify
llocation of
ion manaqe-
data library
basis. If
llocated to
hich accBS-
of the col-
be reallo-

inf ormation,
inf or mat ion

veloping au-
capaailities
on products
y benefit a
uals in div-

while con-
specialized
tne social

S
and
rela
form
Comp
tech
will
coram
muni
idlv
sear
mess
will
time
shou

I
fore
tion
tor
not
"cos
over
prod
libr
in t
two
inai
and
bill
coram
data
stjon
is,
a ne
serv
an D
than
der "

ices
seem
tion
plis
on t
get
IS ,

more

ucce
man

tion
atio
uter
nolo

as
unic
ty's

ch f
age

re

id t

ssf u
agem
ship
n
net

gica
sist
at in
n ee
It

or i
in t
ach
Requ
here

1 in
ent
s t
anion
work
1

th

3
^^

as e
sho

nf or
ne n
eve

est
fore

f ormati
require
o coram
g or
mg is
deveioD
e data"
s and i
f f icien
uld be
mation
etwork
rvone a
and re
be sho

on gathe
s a set
unicate
ganizati
an excel
ment w
library

ts user
tiy and

easier
by placi
system w
t the
sponse
rtened.

ring
or

in-
ons.
lent
hich

in
com-

"?o
hich
same
time

think
e a rea
ship bet
and loca

j ust 1
ting out
head and
uction
ary (and
ne socia
or ganiza
vidual u
have qui
ties to
unities.

librar
d more 1
i n f o r u t
ed-only
ices may
per at ion
in ddva
for an

On t
that m

al plann
hed if r
he basis
evened
some org
than o

that netw
ssessaent o
ween tne da
1 service d
n the tec
" data and
direct CO

and process
t tl u s to t

h

1 relation
tions whicn
sers quite
te diifere
their dif

It may
ies will b
ike indivi
ion will be
basis. Thi
be paid f
is perfor

nee with a
ticipated f
he other ha
ore effect
ing can on
esource sha

that "eve
out in the
anizations
thers, but

or Kin
f th
ta di
ata 1
h n i q u

pas

ing
e use
ships

res
diff

nt re
feren
well
eg in
duals
acqu

s mea
or on
med,
"blan
ut ure
nd, i
ive a
ly be
ring
ry thi
end"

will
ove

g will
e rela-
stribu-
ibrar v,
es for
sing on
or data
to the
r) , but
of the

pond to
e r e n 1 1

V

sponsi-
t user
be that
to re-
; that
ired on
n s that
ly when
rather

ket or-
ser V-

t would
nd ra-
a c c o m -

is done
ng will
; that
benefit
r time.

1 1
-



lASSIST Newsletter, Vol. No. 1 (Winter 1978)

all organizations will benefit
about the same amount. Information
sharing would argue for its cost
being borne by the largest number
of potential users possible. Cer-
tainly, computer networlcing offer
this possibility.

SUMMARY

This paper presents a cursory
view or the factors which have
nstrained the use of computer
tworking and resource sharing by
e social science data library and
s user community. These factors
e structural (the result of po-
tical realities and nistorical
cidents), economic, sociological
d experiential. These factors
11 continue to constrain network
e. but recent developments sug-
st that we will see greater use
networking by the data library

d social scientists. There is
owing acceptance of secondary
ta, growing costs of creating
mpiex data files, need for spe-
alists who can organize, manage
d document these data, rapid re-
ctions in the cost of remotely
cessing these complex files, rec-
nitioD that collections of data
les need to be organized and well
cumented, reevaluation of the im-
rtance of libraries and informa-
on services, and support for the
tablishment of data libraries,
ese developments would seem to
courage an atmosphere in which
tworking would be accepted and
en as a viable and aesirable
ans to access information proc-
sing services outside the local

environment, share intellectual re-
sources, share resources to provide
economy of operation, and provide a
nechansim for more effective com-
munications, cooperation and coor-
dination of information.

He should not expect to see net-
working affect radical changes in
the structure, services, staffing,
collection and relationships with
other data centers. 3ut, we can
expect that networking will produce
a push toward centralization of
computer-related services within
the parent organization in which
the social science data library is
located, in the future a freer mar-
ket in which to buy computer serv-
ices, more user support services
which are computer-based, greater
demands for library staff expertise
about computing by the user commu-
nity, better documentation of in-
formation materials, interactive
information products whicn describe
the contents of HRDF, remote access
for preliminary statistical analy-
sis of data, the shifting of re-
source support from the library to
the user, fees for services
rendered rather than for services
anticipatea (as memberships in con-
sortia are) , and a reevaluation of
the relationship of data distribu-
tor-archive to the local data li-
brary.

The conclusion is that in gen-
eral computer network and resource
sharing should provide better serv-
ices, although the social science
data library may be integrated into
a larger information services or-
ganization.

12 -



IASSI3T Newsletter, Vol. 2, No. 1 (Winter 1978)

REFERENCES

aiities of
ks tor
Princeton

,

eraniversity
il, inc.. ,

Dodd, Sua. Report of the Joint
United States-Canadian Action
Groups on Classification.
lASSIST Newsletter. 1,2 (1977),
5-TT7

Report of the Joint
Canadian-Uniteds States Action
Groups on Classification.
IASSIST Newsletter. 1,3 (1977),
5-107

The Emerging Priority in
Bringing Bibliographic Control
to Social Science .Machine
Readable Data Files (MRDF)

.

IASSIST Newsletter. 1,4 (1977),
tt^th:

EDUCOM. Planning for National
Networking. Proceedings of the
EDUCO;! Sprinq Conference, 1913.
TPrrnce^on7~NJ:~ E'DWCUIJ TEe
Interuniversity Coaiaunications
Council, Inc.

,

' 1973) .

Hofferbert, Richard. Networks and
Disciplines. CoEmou Themes and
Concensus; Report ang
Ciscussion oT worlcsIToFs.
TPrince^on, UJ: 'E^TS'COn, The
Interur. iversity Communications
Council, Inc. , 1973)

.

Mochmann, Ekkehard. Information
Access at the Data Item Level.
SIGSOC Bulletin. 6 (2,3).

Neumann, A. J. Review of Network
Management Problems and Issues.
NB3 Technical Note 795.
T^aslTingron , DC: U. H.
Department of Commerce, October
1973) , 22-23.

Network User Information
Support. U., S. Department of
Commerce, December 1973.

Nielsen, Per. Information Access
at the Data File Level:
Documentation Prerequisites on
the File-Level Data Base Inauiry
Process. 3IGS0C Bulletin. '6

(2,3).

Robbin, Alice. Managing
Inforniation Access through
Documentation of the Data Base.
SiGSCC Bulletin. 6 (2,3).

Roistacher, Richard C. The Data
Intercnange File: A First
Report. CAC Document No. 207.
Urbana, IL: The University of
Illinois, 21 June 1976.

and Noble, Barabara B.
Computer Network Support of
Social Research Communities.
Unpublished paper. Urbana, IL;
University or Illinois, n,d.

13