NEWSLETTER vol. 4 no.2

USER SERVICES IN A DATA LIBRARY

Laine G.M. Ruus
Data Library, University of British Columbia

There seems to be some confusion as to just what is

meant by user services. Although the term is relatively

new to the library literature, the concept dates back to

at least 1876 when, amongst others, Samuel Swett

Green was beginning to argue "the disireableness

of ..personal intercourse between librarians and

readers " Dictionaries of library science or librarianship

do not yet define the term, nor do the library

administration texts that 1 have consulted

David Nasatir (1973) has outlined the components of

user services in a data library as consisting of

dissemination of data files; analyses on demand for

users; training of users; consultation on such subjects

as mathematics, statistics, methodology, data analysis,

etc; the conducting of training programs; a "current

awareness" function acquainting local users with

parallel research being done elsewhere as reflected in

the catalogues of holdings of other archives; and the

creation of machine-readable codebooks. Alice Robbin

(1977) has defined user services as consisting of

reference (or finding the right data for the right user)

training; data and documentation reproduction and
dissemination; data preparation, processing, and

analysis; project planning; and instruction and

orientation. These then are my terms of reference

when I speak of user services in the data

library/archive context.

Samuel Rothstein (1%1) defines reference service as

"the personal assistance given by the librarian to

individual readers in pursuit of information..." Given

this definition, his "reference" = our "user services."

His three levels of reference service are minimum,
middling, and maximum.

His "minimum" or conservative level of service is that

in which the librarian is merely a guide to the use of the

collection, and the user is encouraged to be as self-

sufficient as possible through user instruction and
orientation techniques—a policy of "laissez-faire."

"Middling" service offers a great deal more service to

the individual user (especially, in an academic library,

to faculty and graduate students), including in-depth

searching of the literature, compiling bibliographies,

and generally doing a fair amount of the user's work

for him. On the other hand, "conservative" service is

given to the undergraduate student, on the

assumption that learning to use the library and its

resources is part of the educational process, and the

student should therefore not be spoon-fed

"Maximum" or liberal service relates primarily to the

situation of a special or corporate library, where the

librarian's raison d'etre is to do that part of the research

that involves the literature, leaving the part that

mvolves the laboratory (or whatever) to others

qualified in other fields. That is, the emphasis is on the

delivery of information, rather than on delivery merely

of books, journals, etc., which might contain the

information. The information delivered is authentic,

relevant, and founded "on the impeccable scholarship

of the Librarian" (Rothstein, 1%1)

Ti.e application of this scheme to the data library/data

archive spectrum is not difficult.

On the conservative side there is, for example, the

new, local-service data library, quietly growing in the

bowels of some university structure--in an academic

department or computing centre. Because at first it

must concentrate on developing internally (you must

have a collection before you can provide services based

on it), this data library offers only basic services to

users: acquisition of machine-readable data files (mrdf)

on request, epedally if the researcher knows already

where a file is to be had; maintenance of data files as

they are received from outside sources without much,

if any, cleaning or checking for coding, wild punches,

inconsistencies, etc.; access to codebooks as they are

received from the supplier; and some basic

consultation on problems involving data processing

and statistical procedures. Difficult problems are

referred to experts: computing problems to the

programmers in the computing centre, statistical

problems to the statisticians in the statistical centre,

data identification problems perhaps to the librarians

in the library.

The middling level of service is exemplified by the

same local service data library several years later, when
it has grown both internally and in its services. Because

the collection is now large, with many massive and



NEWSLETTER vol. t, no.2

intricate mrdf, the data library has developed an on-

line inventory describing in great detail the collection,

to assist not only users but the data library staff as well

in identifying and locating specific files. Data files

acquired in a "dirty" state are cleaned, because it is

easier to clean them at once, while one still has contact

with the principal investigator, than several years later

when a researcher needs the data and it is found to be

unuseable. Codebooks are routinely converted to

machine-readable form, because this is the most
satisfactory way of ensuring that any user can get

access to a copy at any time. (And how else does one
economically supply a copy of a Gallup survey

codebook to a political science class of 60 on one day's

notice?) Special programs are developed, and special-

purpose subfiles of data files are prepared, to make
things easier (and cut down on hand-holding) for

novice researchers, and more espeaally, as the only

way to provide adequate service for 200 freshman
commerce students who every year descend on the

data library for their annual excerrise with CRSP stock

price data. (Although the staff has grown, it still cannot
provide consulhng to 200 students, and the only way
to give them sahsfactory service is to make the

procedure as "idiot-proof" as possible.) Regular

orientations, and some impromptu ones tailored to

special courses, are given to introduce novices to the

facility and its services, and there may even be a

manual describing how to search the inventory, how to

mount data files, special programs developed for

often-used files, etc. There is now a variety of staff

experhse that can be called on for consultaHon But the

"toughies" are still referred to the experts in other

departments, and there is shll stnct adherence to the

basic principle that the user should do the actual work
himself, because this is after all part of the educational

process, and the "true" researcher learns to be self-

reliant.

And at the 'liberal" far end of the spectrum, I envisage

a special purpose data library embedded in a research

institution, or a corporation, or a government
department. Much of the work of this data library is

involved in the actual creation of new data files, the

maintenance of on-going data bases, and the

secondary analysis of existing files. The data library

saff here are experts in computer programming,
statistics, sampling, etc.; these experts are part of the

research team of any project, handling such details as

the technical aspects of research design, the research

instrument, actual data gathering, and, later, data

analysis and interpretation. Here there is no question
of the user doing his own work. The level of service

given depends on the expertise of the staff; there is no
need for such ancillary services as orientation tours

and courses, of "idiot-proof" data files and

documentation for novices, for there are no novices.

The data base management system describing the

collection is designed for maximum efficiency, and
data files and documentation are machine-readable
and very clean, because this is the most efficient means
of maintaining and updating them.

Each of these levels of service has its own immediacy of

purpose, which level of service a data library or archive
approaches is dependent on its user community and
the constraints placed on it by available funding and
staff Certainly the level of sophistication possible in

the maximum service archive is far beyond the

developmental capabilities of the small, local-service

data library operating on two and a half people But the

techniques can be transported, and techmques which
result in greater efficiency for the end user will also

generally result in greater efficiency for the data

archive staff as well. My dream, from the point of view
of our small, local-service data library serving an
academic institution, is eventually to make user

services so efficient that no user need darken our door
again, except to persuade us to acquire a file that we do
not already have. There are, 1 think, several

components to this. The first is to make information
retrieval as efficient as possible, then to make
documentatior access as efficient as possible, and
finally, to maxe data access and software access as

efficient as possible These efficiencies need not, of

course, be implemented in quite this order; one
normally does the easiest things first. But 1 am going to

treat them in the order in which most users approach
them

Efficiency of information retrieval t)ecomes cntical

when the data library's collection grows beyond the

point where any staff member can remember what
every file in the collection contains. It then becomes
necessary, not only for the sake of users, but also for

the sake of staff, to be able to quickly find all particulars

about any given data file—not just the pnndpal
investigator, title, or date of collection, but the

individual variables contained in it. The consensus of

practice seems to point to some form of data base

management system (dbms). It is a little difficult to

ascertain who has such systems operating. We know
that Roper Center has had an in-house information

retrieval system in operation for some years. We also

know that ICPSR has offered (in October of last year)

on-line access to its inventory through TELENET.
SPIRES seems to be favourite among MTS
installations, being in use at the University of Alberta

Computing Centre, Stanford University Libraries, and

U.B.C. Library. Other systems are being used by the

University of Wisconsin-Madison and the University



NEWSLETTER vol. k no.2

of Washington The Data Clearinghouse for the Social

Sciences in Canada also had developed a data base

management system, although its present fate is

unclear. Among these systems, the amount of

information offered on any given file varies

immensely. None that I have seen, however, includes

quite the detail envisioned for the data documentahon
system that IFDO is supporting, which contains 16

pages of (mostly optional) fields per record. The
characteristics most desirable in a dbms for this

purpose would seem to be that it;

• be flexible enough to handle a variety of file formats

and interrelationships.

• be easy and cheap to maintain.

• have a powerful and flexible searching capability

• be EASY to teach users to search.

• be available at all times that the computer is up.

In other words, as a user service, it should be capable

of informing a user at a remote location of what files a

.
collection contains, and should provide sufficient

information to access the data without his visiting the

data library

There is another aspect to the question of efficient

information retrieval, the retrieval of information vis-a-

vis data files not held in the local collection--that is, the

identification and location of data files held
elscwhere-for eventual acquisition This is not a

problem that can be solved at the level of the local

service unit. Present services are dependent on local

collections of the inventories of other data libraries or

archives, government departments, etc; literature

searching; and the intuition of the local data library

staff. What is needed is a union catalogue of the

holdings of all known dissenvinators of mrdf, and some
efficient means of access to information on what new
data files are being created. The movement by ICPSR
and the Roper Center towards on-line remote access to

their inventories is a mojor step towards information

retrieval. The data base developed by the Data

Clearing House for the Social Sciences in Canada,
which included the holdings not only of Canadian data

libraries/archives, but also the holdings of government
departments, the commercial sector, and private

individuals, did for a time provide a union catalogue of

Canadian holdings of Canadian data files, albeit in a

restricted subject area. The data base was
unfortunately not made remotely available, and the

DCH is now defunct (no cause and effect relationship

is implied). What is needed is a similar product at the

international level.

Efficiency of documentation retrieval is more involved.

Documentation comes in a variety of forms and
formats Some data files have no documentation at all,

others have as their documentation only the second
paragraph of a personal letter addressed to person X,

and yet others have exceedingly complex
documentation that is much longer than the data file

that it describes. Documentation can be in hard-copy
or machine-readable. The utility of machine-readable

documentation for user services is obvious. It is the

only way that documentation can be made maximally

accessible, since in the optimum system, all machine-
readable documentation is accessible at all times that

the computer is up. Placing extra copies of

documentation in a local library does not serve quite

the same purpose, since I know of no library that is

open as long hours as the computing centre. And a

local computer system with a network of remote

terminals scattered throughout an institution gives

more immediate access than the library, which is

invariably at least a twenty minute walk from where
the user is (a very important consideration on the West
Coast where it rains a great deal). However, to make
the documentation machine-readable is not quite good
enough. No one wants to read through the whole of

the Six Nation Project codebook (Inkeles, n.d), which

occupies 9 linear inches of sh ,'lfspace, to find the

twenty variables wanted In order to maximize the

efficiency of documentation retrieval, what is needed is

some computer-assisted means of searching the

codebook, so that one can retrieve just those portions

describing variables containing specified key words.

The Roper Center's proposed on-line system is

designed to retrieve at this level of information. But

there are many codebooks which are not amenable to

this type of treatment, such as that describing the

CANSIM system, which contains over 300,000 time

series and as many variable labels. Maintaining a data

inventory is one thing, but maintaining this level of

information in an on-line data base is, for a small local-

service operation, at present quite unfeasable.

Maximizing data access efficiency is dependent on the

efficiency of tape mounting procedures in the local

operating environment. However, it should be
possible to develop a system such that, based on the

information given in the information retrieval system,

it is possible to gain access to all data files at all times

that the host computer is functioning, not just at those

times that the data library is open. Desirable features of

data access would include; a minimum number of

commands to access a file (whether on disc or tape);

and maximum simplicity of commands--i.e.,
mechanical details such as mode, blocking factor, label,

reel number, and position should be transparent to the



I$T NEWSLETTER vol. 4 no.2

user. In addition, a data access system can be designed

to maintain data security (e.g., check the permit status

ot computer IDs), collect statistics on tape mounts,

perform an SDI funcbon (referral to updated editions

of files), and reinforce the moral conditions of data

access (such as by cautioning against unauthorized

copying for dissemination of files by users).

Of course, efficient access to data and documentation
is of little value if one does not also have access to

appropriate software. Software acquisition, creation,

and support is very much dependent on the

institutional structure within which the data

archive/library exists. The smaller a local-service data

archive/library, the more likely it is to be heavily

dependant on the software resources of the host

computing centre, with little possibility of influenang

the computing centre's decisions as to hardware and
software procurements, software to be supported, and
future software (or hardware) developments. If the

computing centre in question is user-oriented (which is

the exception rather than the rule), if it provides

comprehensive documentation, user orientation,

introductory courses on commonly used statistical

packages and programming languages, and extensive

programming consultation, then the data
archive/library need not provide these services. It, on
the other hand, the computing centre is not user-

oriented, the problem of providing effiaent access to

appropriate software becomes serious, and the

solutions are neither easy nor cheap As the amount of

computing expertise dedicated to the data

library/archive (or employed in it) is increased,

however, an individualized level of service can be

provided. This may extend to special-purpose

housekeeping programs, to improve the efficiency of

such invisible services as data file cleaning,

rationalizing inventory, tape mounting and security. It

may also extend to visible services such as writing a

formula for the retrieval of data from a complex data-

base, or creating a program to provide uniform access

to and manipulation of aggregate data across several

censuses. As the amount of computing expertise

dedicated to user services is increased yet further, and
basic software for common problems has been

provided, yet more elaborate services become possible,

such as the development of special purpose software

Once one has all these systems operating, of course,

one must teach one's users to use them. Seminars,

special class presentations, and short courses seem to

be the normal approach. A good addition is a user's

manual, wntten not for the computer programmer (as

most computing documentation is), but for the novice

user. A history professor, with almost no computer

experience, should be able, with the help of your
manual, to sit down at a terminal, search your
inventory for any files containing, eg , illegitimacy

rates in Ontario in the 19th century, have printed a

copy of the (short) codebook on the local high-speed
printer, and mount your data file, all on a rainy

Sunday afternoon.

And then, of course, there arc the special purpose
subfiles, teaching packages, and special purpose
documentation for the 200 Commerce students.

Designing this kind of service requires more effort to

be put into instructor-education than into student-
education It is essential that the instructor have
designed the assignments so that they are appropriate

to the data files he wants to use The data library staff

should be given ample warning of the assignment It

must know the expectations of the instructor, and the

level of consulting he or she will give the students as

part of the course With adequate preparation by the

data library staff, the onslaught can be relatively

painless, and will have a marvelous effect on tape

mount statistics

Throughout, I have been considering only "minimum"
and "middling" service levels of the data
library/archive. At both levels, the user's loing the job

is considered part of the educational experience, and
self-reliance is something to be fostered 1 shall not
consider here the extent and nature of special services

offered in the maximum level data archive

Each of the above components is composed of

transparent services and apparent services. Apparent
services I define as those which the user can see and
appreciate. In this category I include the basic

provision of information services; acquisition on
demand of mrdf; consultation services of all kinds;

provision of some kinds of dissemination services

(e.g., relieving a principal investigator of the

responsibility of maintaining and disseminating his

data file himself); insuring efficient access to

documentation, data, and software; orientation
activities; provision of a user's manual, and so on.

Transparent services are those which the user seldom
notices or consciously appreciates, including the

extensive level of reference services involved in the

identification and location of fugitive mrdf, "on spec"

acquisition of mrdf, rationalization of tape library and
inventory procedures, routine data file cleaning and
machine-readable codebook creation, and creation of

special-purpose files to support special-purpose

software, such as a map file to support graphic

retneval of census data.



NEWSLETTER vol. 4 no.Z

Why this distinction? User services are usually the

raisoii d'etre of the facility, and the reason for its

conhnued funding. Apparent services are directly

user-oriented Transparent services, although often

refinements of apparent services, are more often staff-

oriented, and only indirectly contribute to that

desirable phenomenon--the satisfied user My
conclusion is that, at the outset, a data archive/library

should concentrate on apparent user services, in order

to culhvate an active and supportive user community,
and only later in its development, when this has been
accomplished, develop transparent user services.

REFERENCE TOOLS FOR MACHINE-READABLE DATA FILES

John G. Kelp
Laboratory for Political Research

University of Iowa

It would appear to be a rather awesome task to try to

summarize the current state-of-the-art in data

reference; for surely after 16 years of steady growth in

archives, reference matenals giving access to the data

in such archives should be exhaustive Yet this is

clearly not the case, although some attempts to provide

useful reference tools in this area have appeared over

the past decade. In the brief report which follows, I

have singled out for discussion three categories of

reference tools for machine-readable socal science data

which seem well-established

Those reference tools which appear to play the most
prominent role in directing users to appropriate

machine-readable data are: (1) data catalogues

describing the contents of individual social science data

archives and data libraries; (2) directories describing

the contents of more than one archive, or directories

within special topical areas; and (3) penodicals, like .; s

data, which have attempted to report at regular

intervals informahon on the holdings of social science

data archives

An attempt will be made to examine each category of

reference tool from two very personal perspectives: (1)

the user consultant who is continually asked to locate

data files which must meet a number of very special

condihons; and (2) the editor and compiler who must
try to IcKate all known data files relating to certain

topical areas and acquire informabon on the most
recent acquisitions of data archives In the former role,

one is frustrated by the inadequacies in reference tools;

in the latter role, one is amazed that we have come as

far as we have.

Data Catalogues

Lists of holdings, guides to resources, archive

directories, or data catalogues are available from most
individual data repositories Probably the most well-

known of those documents which describe individual

archives would be ICPSR's Guide Ic Resources and

Scn'ices, issued on a nearly-annual basis since

sometime in the I960's. Others of this j(enre would
include the recently-issued SSRC Survey Archive Data

Catalogue, the Sleiiimctz Archwes: Catalogue and Guide

((1978), the B-A. S.S. Invenlairc des Archives Disponsibies

(1975), Catalcf^ue of Machine-Readable Records in the

National Archives of the United States (1975), and the

Lokahseringsoversi^l of the Danish Data Archives (1978),

to mention just a few.

Physically, documents of this type are soft-cover,

book-length descripHons of the holdings of a major
archive. These are real publications, meant for broad

distribuhon to a national or international clientele.

Entries are often arranged according to some broad

subject classification scheme which might include such
major headings as community and urban studies, elites

and leadership, mass political behavior, social welfare,

religion, the international system, legislative and
deliberative bodies, etc. The individual entries usually

include title, author or data collection agency,

population covered and/or sampling scheme, time

period of study, number of cases and variables,

distribution restrictions, and a brief abstract

summarizing the purpose of the study and the focus of

major categories of variables.