ORGANIZATION OF OFFICIAL REPRESENTATIVES SESSION

THEME: Building a house of bricks

(Lessons we learned from the "Three Little Pigs")

Presented by:
Fay Booker, University of Chicago
Laura Guy, University of Wisconsin-Madison

Table of Contents:

I. Introduction to the Session

II. Overview of ICPSR Membership

III. Local Data Services

IV. Interacting with the ICPSR

V. Building Local Partnerships

VI. Building Partnerships with the ICPSR

You can also view the slide show presentation associated with this talk. The slides include the fiscal charts discussed in Section II, Section D. Note that the organizational charts discussed in Section II, Part A. are currently not available in electronic format.

I. Introduction to the Session

A. Overview

The purpose of this session is to inform new ICPSR Official Representatives (O.R.s) and remind veterans of the service base that ICPSR offers as a means of making the most of everyone’s membership. We hope to add to your understanding of what to expect from ICPSR, how to use your membership to improve local research and instructional services and to improve the quality and quantity of services that ICPSR provides to you.

The subtitle of this session, "Lessons we learned from the ‘Three Little Pigs‘" may require explanation. To quickly summarize the fairy tale of youth, three little pigs set about the task of building new homes with a known villain in the neighborhood—the big bad wolf. One piglet chose to build a house of straw, another a house of sticks, and the third a house made of brick. Each piglet had reasons for selecting construction material—straw was most available and easiest to use, sticks were sturdier than straw and easy to use, bricks, while sturdy, required considerable resources. The piglet that built its house of straw completed construction first, but that house could not withstand the power of the wolf. Found homeless, this piglet joined its sibling who was building a house of sticks. While the house of sticks could withstand the "huffing and puffing" of the wolf, it could not withstand fire. With the house of sticks destroyed, the two homeless piglets joined up with the builder of the house of bricks to help complete construction. This house withstood the wolf’s challenges. There are two morals to this story. First, expediency is not a solution. Second, and more importantly, by working together, a strong edifice was built that withstood the challenges placed upon it.

Like the piglets in this story, we are building "houses"— services that are essential to the research and instructional mission of our institutions. At the local level, the actors in this story are faculty, researchers and students, data service providers, librarians, support staff, local and non-local "experts", computing resource providers, statistical support providers, and the like. These individuals and groups of individuals are our "partners". The villain might be any or all of the following: limited budget, lack of infrastructure, limited staff, time, indifference, etc. The local challenge is to build a service profile that can withstand the challenges placed upon it by the "villain". We contend that the ICPSR membership is one cornerstone of building a house of bricks at your local institution. Or, within the terms of our story—we are building a house (a data library) in a community (our institution) within which the ICPSR plays a major role. In fact, as well will see, it is with the ICPSR that we will form one of a series of partnerships within our community, the success of which will in large part determine how effectively we can do our jobs.

Viewed more broadly, we are also building a foundation with, AND FOR, the ICPSR. Our dues support the work that this organization does. As Official Representatives (O.R.s) we play an important part in determining the direction and orientation of the ICPSR's policies and programs. We must all be involved in building a house that can withstand the challenges that exist and will emerge as we enter into the 21st Century.

B. Introduction of Richard Rockwell (5 minute presentation)

 

II. Overview of ICPSR Membership

A. Organizational Structure of ICPSR

[Insert -- 3 Charts here: ICPSR Organization, Council Organization, and ISR/CPS/ICPSR Organization -- thanks to Pamela Schwarzmann. Please note: these charts are not available electronically at this time]

The primary organizational bodies are ICPSR itself, the ICPSR Council, and the Official Representatives. As we will see, both the ICPSR and its Council are subdivided into sections (in the case of the Council they are called subcommittees, in the case of ICPSR they are functional units). Official Representatives meet once every two years at Ann Arbor to discuss issues and share information.

The Preamble of the Constitution states that the ICPSR is governed "as a partnership between member institutions and the Center for Political Research of the Institute for Social Research [at the University of Michigan]". The basic organizational structure is as follows:

Issues:

B. Essential Information about Your Membership

  1. Constitution, Bylaws, Memorandum of Agreement, Council Minutes

Please Note:

The ICPSR was originally founded as a partnership between the Survey Research Center at the University of Michigan and 21 universities in the U.S. It now has over 330 member institutions in the U.S. and Canada and several hundred outside of North America.

It is an "organization of member institutions working together to acquire and preserve social science data, to provide open and equitable access to these data, and to promote effective data use" (mission statement).

Of the three documents the Bylaws are probably the most important from an O.R.'s viewpoint.

The Constitution consists of a Preamble and eleven Articles that spell out the nature of the organization, its purpose, its membership, governance, voting rights, and other items. These Articles are quite brief.

The Bylaws consist of ten very detailed Articles. For example:

(1) Conditions of Membership

Many responsibilities fall to the OR including insuring access eligibility, continuation of funding, etc. There are stiff penalties for leaving and leaving/returning. These include discontinuing use of material when leaving and financial penalties if rejoining after termination.

(2) Benefits and Obligations

Services for non-members (versus members). Now non-members have access to documentation and some data via web. The responsibility falls to O.R.s to ensure proper use of data. Types of improper use vary from institution to institution, but there are certain behaviors to watch out for (e.g., consultants doing "work", people redistributing data to friends and so forth). Members are to not redistribute data to other members without agreement of ICPSR.

(3) Executive Director

The Executive Director, and the Senior Staff through the Executive Director, is responsible to the Council for the conduct of ICPSR affairs and to the host institution for adherence to its rules and regulations.

2. Council Matters

Included in the materials we're giving you in this session are copies of recent Council minutes. These are usually posted to the ICPSR Web site, and are important reading.

One point to keep in mind is that there are numerous ways you can maintain the partnership between you and ICPSR and keep apprised of what is going on at ICPSR, if you are so inclined. Some people may enjoy following Council while others are more oriented towards the discussion lists, or talking to other O.R.s.

The Council depends heavily on the activity of standing committees, the membership of which is indicated in the minutes. The committees as of the most recent Annual Report are:

1. Archival Development

2. Computing and User Services

3. Long Range Planning and New Initiatives

4. Membership Relations

Council members are elected by the O.R.s to serve as the Executive Committee of the membership, with authority to act on behalf of the member institutions. The composition, election, and organization of the Council are set forth in the Bylaws.

For many years Council was made up of faculty from member institutions described as "leading scholars from many areas of social inquiry". More recently attempts have been made to "integrate" Council with members who are Official Representatives -- typically data librarians.

In the Bylaws the duties of the Council are clearly spelled out including the definition of organizational objectives; the determination of policies to achieve those objectives; establishing policies and procedures concerning financial and acquisition activities, etc.

C. Overview of ICPSR Services [Janet Vavra]

DRAFT--OVERVIEW OF RESOURCES AND SERVICES--DRAFT

Background

ICPSR Services

How the O.R. can HELP

D. Financial Profile

[Insert: Five Charts Here: ICPSR Expenses Paid from all Funding Sources, 1997; ICPSR Revenue from all Funding Sources, FY97; ICPSR Expenses Paid from Operating/Membership - 1997; ICPSR Net Income, 1993 - 1997; and Sponsored Research, 1993-1997 -- thanks again to Pamela Schwarzmann]

Of interest to O.R.s should be where ICPSR gets its money and how they spend it. Some interesting primary points include the fact that ICPSR is, despite all of the technological trappings, and probably like most of our own organizations, a "people shop"—meaning that salaries are a major expense item. Both Sponsored Research and Membership income sources (the two largest sources of income) contribute roughly equally to staff costs. Currently there are roughly 64 permanent full time employees of ICPSR. The breakdown is as follows:

Membership dues and miscellaneous income from the sale of codebooks, the ANES CD-ROM, individual memberships, and interest from ICPSR Equity account during FY97 accounted for 43 percent ($2,256,168) of total ICPSR funding ($5,187,625) as reported in the 1997 draft annual report. Revenue from Member dues only has increased 32 percent over the FY93-97 period (FY97: $2,181,661; FY93: $1,657,031). Sponsored Research includes Federal funding (40% of total ICPSR funding [$2,055,722] in FY97) and funding from private foundations and other nonfederal sponsors (5% of total ICPSR funding [$259,598] in FY97).

The five-year trend shows an increase in revenue from all sources of 26 percent and an increase in expenses from all activities of 17 percent since FY93 (FY97: $5,187,625; FY93: $4,109,964).

In FY 1997, which ended on June 30th, 1997, revenues were up 5.5% from FY 1996 (FY97: $5,187,625; FY96: $4,916,909). One of the cost savings recently realized is the move away from MTS. This, along with various efficiency measures has helped to make FY 1997 a "banner" year financially for the ICPSR.

E. Acquisition Statistics

We all know that even in our own data libraries that keeping acquisition statistics is difficult. Part of the difficulty rests in the nature of the data themselves -- they are comprised of files, studies, parts, records, and so forth -- which do you count?

For the first decade and a half of ICPSR’s existence acquisition statistics were not assembled (this according to Erik Austin).

Erik Austin goes on to add: "As a result, it was quite laborious to reconstruct information that would be comparable with what we began to record later. I’ve done that for three fiscal years before the mid-80s, to give some comparative purchase to the time series. Secondly, I’ve included BOTH data collections and data FILES acquired for each of the years shown. They are demonstrably not the same thing except in the very early years, when each collection typically included only one data file that had to be acquired, processed and archived, and distributed. Note the growth over time of multi-file data collections, as reflected in these statistics. The years referred to in the table below are fiscal years (e.g., 1964=1963-64) with the fiscal year running from July 1 to June 30 of the following year. Finally, the oddest of the numbers reported below requires one point of explanation. In 1996, the 4,175 files acquired included over 2,000 files from one collection, the 1990 Census STF 4 series (the oddly-organized detailed statistics by state by racial/ethnic group from the 1990 Census long form questionnaire)."

 

ICPSR DATA ACQUISITIONS, F.Y. 1964-1997

Year Data Collections Data Files
1964 34 39
1974 73 136
1984 172 673
1989 338 898
1990 265 800
1991 256 926
1992 259 966
1993 266 2094
1994 295 1169
1995 290 2211
1996 171 4175
1997 212 1975

 

We thank both Erik Austin and Pamela Schwarzmann for their significant contributions to this session.

 

III. Local Data Services

To this point, we’ve described an important part of the community in which you are building your house -- the ICPSR and its major components -- and we’ve provided a brief overview of the amenities that are available. We’re now going to shift gears and discuss the groundwork associated with building your house in your community. You might think of this as preparing the lot you’ve purchased for the impending construction of your home. In other words, you cannot effectively and efficiently make use of your membership without adequately preparing the environment for your institution’s utilization of this membership. Very few of us would build a quarter-million dollar house and not do landscaping!

How data services are implemented at your local institution is largely a function of local demand, history, finances, and existing human and computing resources. Surveys from recent years indicate that member institutions are incredibly diverse in terms of the services each provides, the computing systems used, and the strategies adopted to serve the local needs. Some institutions have long traditions of data services while others are creating these services "on the fly". Some attending this meeting are building upon the foundations created by predecessors, others are "recreating" services, and still others may be building for the very first time on new ground.

Jim Jacobs (UCSD) has a nicely laid out description Providing Data Services for Machine-Readable Information in an Academic Library: Some Levels of Service, The Public-Access Computer Systems Review vol. 2, no. 1 (1991): 144-160) of levels of services that any or all of us might consider offering. It is clear that most libraries will not be able to supply all levels of service for all categories, but based on resources, staff, and other variables, choices will have to be made concerning your own service policies. Its useful to have such a "mandate" or "mission statement" in writing that you can provide to users to help them put your library and its services into context. According to Jacobs, minimally, you need to decide what you will provide to clients in at the following categories:

1. Reference Services: Will you do data searches for the client? Will you advise clients about strengths/weaknesses of specific data collections?

2. Acquisition Services: Do you acquire EVERYTHING and ANYTHING that ANY CLIENT (faculty and/or student) requests?

3. Data Provision Services: How will clients typically obtain data? How will you deal with paper materials?

4. Data Support Services: Will you provide technical/statistical assistance to users for collections that you receive? What role, if any, will you play in helping users resolve problems with data? How will the types of applications and computing platforms available on your campus affect the types of data that you acquire as well as the format of those files?

To put these into a more practical context consider the following types of reference services you may want consider implementing (taken from: Geraci, Diane, Chuck Humphrey and Jim Jacobs. "Data: Services and Collections (Part 2)," in Management of Machine-Readable Social Science Information (ICPSR Summer School, August, 1996):

Obviously you could come up with similar lists for the other major categories that Jacobs outlines. Choosing what levels of service you will provide should not be viewed as intimidating or discouraging. In any library situation decisions must be made to delineate the services available. Thinking about these issues in a practical manner can be very useful -- they can help you define what you will do and what you some day hope to do. Having them written down will help both you and your users. However, even after the decisions are made, don’t be surprised if you find your users pushing the envelope (so to speak), and don’t be surprised if you find yourself complying!

A Comparison of Services at UW and UofC

We are going to provide a bit of information about how we have chosen to deal with these issues and others. In many ways it seems we are very different places doing very different things. However, we are bound together by the types and levels of services we provide. As to whether we are more similar than dissimilar, we think, there is no doubt. However, even between two "major" institutions like the UofC and UW, there are startling differences.

Here are "our" stories. A few caveats are in order.

 

A. Organizational History

UofC

The University of Chicago has been a member of the ICPSR for nearly 30 years. For almost all of this time, the OR has been a faculty member and the responsibilities of managing the membership have fallen to computing center staff as a SMALL part of their jobs.

In 1990, the Social Science Research Community broke away from Central computing and established a facility dedicated to the needs of Social Science Research. A cornerstone of the new center was the "data mandate"—the requirement that this operation assume the responsibility for acquiring, maintaining, and disseminating data to the research community. In 1993, the data library position was funded and I was brought on the scene.

My background to "data services" comes from the perspective of a "data user". As such, I've spent the last five years building and developing services that I hope have made quantitative research possible. Building services is an evolutionary process. My particular approach was to determine the services that key faculty felt were essential to complete their research. At the core of my operation is the mandate to integrate "social science language" and "computational realities" to enable research to go forth.

Staffing: 2.5 FTE (or at least should be by October)

UW

In the mid-1960s, at Wisconsin, the Data and Program Library Service (DPLS), and its administrative parent, the Data and Computation Center, were established as a result of an articulated demand for assistance in social science research problems related to large-scale data collection and computation. The DPLS was created because faculty members of the social sciences, especially in economics, political science and sociology, became convinced that the University needed a facility for managing the increasing quantity of available machine-readable social science data being produced on the campus and elsewhere. They recognized the importance of preserving data, collected often at considerable cost, which had significant subsequent value for other researchers and students. Thus, the DPLS was designated as the local campus repository for quantitative social science machine-readable data. Its major functions were defined as acquisition, storage, maintenance, and dissemination of data files.

Although the DPLS was considered a campus library, it was not incorporated into the campus library system in part because of the system’s inability to accept computer data as a legitimate information resource. This perception that data files were either too difficult or too obscure for traditional libraries persevered well into the 1980s. In the 1990s, the perception of data as a marginal resource has changed; the major functions of the DPLS have not. The DPLS continues to be both a data library, acquiring and facilitating access to data as they become available from individual researchers, other archives, national data repositories and profit and nonprofit agencies, and a data archive, acquiring, preserving and disseminating studies created by faculty, students and researchers on its campus.

In short, although we are mandated to serve the Social Sciences, in actuality we have become a campus-wide and worldwide resource. We are still connected administratively with a Micro-computing Lab; our major users are Economics, Sociologists, Demographers, Political Scientists, Health Sciences, Journalism, and Business.

Staffing consists of 2.2 Professional FTE Librarians, 1.0 Student Hourly, and 1.0 Administrative Support. We do all of our own computer support internally.

B. Computing Situation

UofC

A large capacity data storage server (UNIX-based) that serves data via NFS network to 19 workstations within the Social Science Computing domain. Current storage capacity = 27 GB disk.

Small (~100 MB) anonymous ftp site for non-SSRC client fetching space.

Coming soon: Win-NT data server for CD collections and PC-based data and applications. (The wave of the future?)

UW

For most of our history we were dependent upon the centralized campus computing services. Seven years ago or so we decided to become our own computer site (costs were going up and we were concerned about tape drive support). The decision was both good and bad, as we then became responsible for our own system maintenance (not a trivial task to be taken lightly!).

This last year we converted from UNIX to WinNT. We total over 20 gigabytes of disk space, and although we are moving some of our archival data "online", we do not make a habit of storing data online, per se. Our offline storage policy has undergone a definite shift since ICPSR made their holdings available online; we no longer archive ICPSR data locally.

In our WinNT system we support our own mail server, web server, ftp server, database, office software, and desktop publishing software.

Currently we have about roughly 25 gig of space spread out among several WinNT servers. The main WinNT file server houses the FTP and Web server holds about 15 gigabytes of disk space (the web site itself is in excess of 2 gigabytes). Two WinNT Workstations are dedicated for "Public Use" of standalone CD-ROM products (of which we now have several hundred) and web-based data sources such as Datastream, Ferret, and the like.

C. Users and Uses of Data

UofC

Overall, we have ~400 faculty, 1000 graduate students in the "core" constituency

85% of usage by active research projects (faculty and graduate students (Masters and Dissertation Level).

15% usage in curriculum (particularly statistics sequences) and undergraduates (who often need TONS of attention).

UW

The University has a population of 56,000 students and staff, all of which are potential "customers"; in addition, we serve a growing outside clientele via our online archive.

See the use statistics in the DPLS Annual Report. The report shows that about 75% of our usage is consistently within our college (the College of Letters and Sciences), and the majority of it is research-oriented, although the instructional-based portion is growing steadily. The actual breakdown of research versus instructional use is surprisingly close to UofC’s -- although class size may differ here (we see courses of 60 to 70 graduate students who are required to use the library for secondary analysis). We are also seeing a steady increase of Undergraduate use.

D. Data Services

1. Reference Services (We'll deal more with this subject later)

UofC

One of the most challenging features of my job! We provide extensive services up to the level of our/my (?) incompetence.

UW

Ditto to what Fay says. Doing effective reference work is a subject of special interest to me—See also my workshop on doing reference service in the data library:

  • http://dpls.dacc.wisc.edu/types/data_reference.htm
    •

    It is commonly acknowledged in the library profession that reference work is one of our most difficult tasks—in the academic research library or the specialized academic library, where we frequently develop long-term and intense relationships with our users, this is especially true. Our users often place extreme demands on our capabilities and our resources. This type of reference can be phenomenally rewarding, challenging, and enjoyable.

    2. Acquisition Policies and Services

    UofC

    In the old days when data were distributed on 9-track tape, we had much more leverage over acquisition rates than we do today. It was relatively easy to determine "preferences" as a direct function of tape costs. In the era of ftp, it’s quite easy to satisfy almost all client requests for ICPSR collections instantaneously; including those requests that are for multiple, large sized collections. I do not view this as a "negative attribute" of ICPSR's distribution policy. If data exist, they should be used. Since we have the capacity to store collections online perpectually, we do attempt to acquire key collections proactively.

    UW

    In the past when ICPSR data were ordered on tape, DPLS had a rather formal collection policy detailing which data sets we would acquire as soon as they were made available. In library parlance these might be thought of as "standing orders" -- the data were generally recognized to be of interest to our researchers. Other data sets were acquired upon request with the understanding that it would take several weeks to obtain them. The additional costs sometimes incurred by ordering particularly large data sets (like STF data) were usually shared with users. Currently there are only a very few data sets that we automatically acquire from ICPSR. We have never imposed limitations on the number of data sets users request (although we do monitor novices who seem to want "too much").

    As far as non-ICPSR data, we continue to work with users, their departments, and other units on campus to purchase data that we feel will be of interest to other researchers (we rarely purchase data that have limited usefulness). We generally offer to pay for half the cost, assuming the data will reside at the DPLS.

    3. Data dissemination

    UofC

    We’re "storage hogs". Once data is here, it’s kept on line "forever". [Question to raise: version concerns]. Data stored in "generic" space referenced...

    Most collections stored in ASCII, compressed format. We try to keep codebooks and supporting files (SAS/SPSS setups uncompressed).

    Users typically access electronic files on any one of 19 UNIX workstations in our domain. We encourage users to extract the variables they need and implicitly promise to keep the data online for the duration of their project.

    For users who don’t use SRC facilities, we have three strategies: a small cache (~500 MB) of anonymous FTP space or customized creation of diskette or CD-ROM (the latter is a new service to be instituted in 1998).

    UW

    DPLS locally stores non-ICPSR data (and a few select ICPSR holdings). Storage is primarily offline with data residing on both 4mm DAT and CD-ROM. The format of files varies according to platform: some are UNIX-based tar.gz, some are PKZIP. In the past we put great effort into storing data as uncompressed raw ASCII -- the most vanilla format possible. Now we put less emphasis on this, the feeling being that we are fairly safe with widely used and recognized compression algorithms.

    When a user requests ICPSR data we generally place their files (either compressed or uncompressed—its up to them) on our anonymous FTP site and give them plenty of time to download the files to their local disk space.

    Resources on this campus are as such that this paradigm rarely causes difficulties. Disk space is plentiful here, and many users have access to phenomenal computing resources.

    In the last few years we’ve had a stupendous growth of (non-ICPSR) CD-ROM-based products. These are accessed via standalone workstations in the library (allowing people to check them out them inevitably caused angst due to wrecked media).

    2. Documentation

    UofC

    Boring stuff about shelving at our site. Hardcopy Doc can be browsed or borrowed for 1 week (or thereabouts). As a rule, we don’t print electronic codebook files unless we consider it to be of widespread importance. Users printing documentation on shared printing equipment is a problem that we hope to address by "charging" for printing services. We, unlike the library, have no "teeth" to back up borrowing times and often resort to moral suasion to get things back.

    UW

    Ditto on the "boring but important" aspect of this subject. We catalog with our own totally unique cataloging system. This system has been in place since the mid-seventies and while it would be good for us to catalog in L.C. or Dewey, the amount of retrospective work would be prohibitive. Holdings are not in the campus OPAC (or whatever they’re called these days). DPLS has room with shelves. People pull off what they want and checkout is "overnight" (frequently abused and material sometimes disappears—we have no teeth either). One final comment is that many people still very much like to have printed documentation.

    4. "Data" Support.

    UofC

    We actually do statistical support, at least at the data step level. We have detailed knowledge about SAS, SPSS, and Stata.

    Training of users is key to our operation. We do a great deal of one-on-one consultation. As a rule, we don’t pull extractions for users. Notable exceptions are undergraduates who simply don’t have the tools to do it themselves (in most cases) or research projects that need something quickly that they PAY for.

    We also are a first point of contact when a user finds an inconsistency in the data. We verify the problem (98% of these problems go away due to either a misunderstanding of documentation or of skip patterns). In those cases where real problems exist, we make a judgement call as to whether we contact ICPSR or the researcher contacts ICPSR directly. [I have examples of both].

    UW

    Statistical consulting must go hand-in-hand with data dissemination. If you don’t do it in-house like Fay, you must have some other mechanism, be it formal (like paid consultants or a micro-computing lab) or informal (other friendly students, faculty, or researchers).

    We are administratively connected to a Social Science Computing lab. They have a very capable part time statistical consultant who works closely with us and fortunately is also intimately familiar with the ICPSR. DPLS staff has limited statistical abilities, and probably do more harm than good offering advice. Occasionally, under duress we will produce extracts or subsamples for someone.

    Besides the Microlab there are other possible resources, the use of which depends on the institutional affiliation (or, sadly, financial resources) of the person in need. The Microlab is limited to helping those in the "Social Sciences". The central campus computing center has statistical consulting but you must pay for it. Many departments have their own paid consultants (of varying quality). It is essential to develop good working relationships with these people (if for no other reason so that they know from where the data are coming and who to contact if there are questions they can not answer). We’ll talk more about this in the "Partnerships" segment.

    We do help people a great deal with data-related trouble-shooting. Due to the growth of the DPLS CD-ROM collection, we do consulting for the various front-ends. Insofar as ICPSR data go, we find that because the statistics consultants are often extremely unfamiliar with the data sets (and also of are varying capabilities), we out of necessity must be able to do some basic level of troubleshooting.

    IV. Interacting with the ICPSR

    [If only I knew then what I think know now]

    This section is based in part on a number of the basic concepts we are covering in this workshop: if you are building a house (a data library) in a community (your institution) then ICPSR might be considered your foundation (particularly if it is the primary source of your data). Among the many partnerships you will be building, your partnership with the ICPSR will be key. You can choose to look at this partnership in one of a number of different ways, for example, clearly you are part of a membership organization. But in a real sense it is also a vendor/customer relationship. You might also look at it in terms of a buying club. However you do decide to view it, it is essential that you (to put it crassly) make use of it and get your money’s worth.

    Later in the workshop we will be specifically talking about partnerships, in particular your partnership with the ICPSR, but first we feel it useful to discuss more directly interacting with the ICPSR.

    The efficient and effective utilization of your memberhip.

    How do you and your users get the most bangs for your buck? For some, if not many of us (like DPLS), the cost of the ICPSR membership may be your biggest non-salary expenditure. Many of us have to worry not only about how to make sure we get the most for our money, but we may even have to worry about justifying the cost of the membership. However, the ICPSR membership is an amazing resource (just ask one of your graduates who graduates to an institution that does not have one!).

    How can you use your membership as a foundation for creating/sustaining a high quality of service?

    It should go without saying (but we will anyway), that you need a basic set of applications readily available to you in order to even use ICPSR services. These are:

    The way ICPSR currently provides access to data is either via the Internet (thus you need tools that allow you to access information on the Internet) and via Periodic CD-ROM releases and special-order CD-ROMs (thesellatter cost extra). We envision that it would be difficult indeed to be an O.R. without Internet access and its associated tools.

    A. Identifying What’s Available from the ICPSR

    One of the "tragedies" of ICPSR’s success is that it no longer publishes the annual Guide to Resources and Services. For decades this was the main source of data availability from ICPSR. It was an outstanding resource for clients as well as O.R.s who needed information about the collections that are housed at the ICPSR. Its easily browse-able format (by broad subject category) helped users get past the lack of subject-classification that makes finding data sets so difficult.

    While you can find [almost] all of the information contained in the Guide it is packaged differently on the web site, and I would suggest that it isn’t quite as user friendly for us as ORs or for our clients. Two quick examples come to mind. First, overviews of ICPSR Series collections are now stored in each component study’s abstract, thus, it can be much more difficult to identify what a given collection is all about. Second, web-based information retrieval requires a considerable amount of "back and forth" navigation. The client who is "generally" interested in studying a specific topic can expect to spend a considerable amount of time at the ICPSR Website, whereas it would probably take half as long to simply scan the 10-15 pages of the Guide.

    www.icpsr.umich.edu

    New O.R.s would be well served to spend a couple of days browsing the copy of the old Guide (circa 1995) as well as the web site’s "Archival Holdings" to get a flavor of the types of collections that are available. Another very good way to get your feet wet (as we recommend again later in the session) is to read through the documentation of some of the more prominent data sets. Examples might be the General Social Survey, or the Current Population Survey codebooks.

    The ICPSR Archive

    1. Breadth and scope of the collection

    As a general guiding principle, the ICPSR is a very good source for national, survey samples. As Janmentioned earlier, ICPSR is home to several topical archives: criminal justice, aging, health. The ICPSR collection includes an incredible amount of data collected by Federal Statistics programs, including decennial census data, various months of the Current Population Surveys, SIPP, Annual Survey of Governments, mortality and natality data.

    2. Limitations of the collection

    This implies, however, that there are types of collections that you probably won’t be able to obtain from ICPSR. If you are hoping to get every data set that anyone could possibly need from the ICPSR, then you most assuredly will be disappointed. Data you won't find can include:

    3. Caveats about the timely availability of data

    It also bears noting that ICPSR doesn’t always make data available the moment it is received. This is one of the key issues raised in NAPA regarding the use of "FastTrack". Every data set undergoes some degree of processing by ICPSR staff to verify data consistency and accuracy. The featured article in the May 1997 ICPSR Bulletin goes into great detail about the steps taken to fully process archival data. ICPSR takes its archival role very seriously expending a great deal of effort ensuring that data files are fully documented and that respondent confidentiality is preserved.

    4. Making use of ICPSR's information sources

    With the preliminaries out of the way, how does ICPSR manage to keep it’s clients informed about activities, including the availability of data? The ICPSR hosts numerous electronic services aimed at assisting and informing Official Representatives. There are at least three major sources of information about the ICPSR. These are:

    For further information on these you can go to: ICPSR - Electronic Services List

    The most "salient" information for OR’s is the information contained about the "Archive". This is broken down into several sections. Including

    Unfortunately finding out whether ICPSR has or is about to release a particular data set, is not always as straightforward as one might hope. If you do not find it by searching the contents of the Archive, you will want to check FastTrack (FastTrack permits anonymous FTP retrieval of some of ICPSR's newly-acquired data collections, usually series data) and the Hotline (where newly released data collections and updates to existing collections are announced). If those fail you, you can always contact the Archive staff directly for further assistance.

    B. Determining What Your Client Needs

    1. Reference Interviews

    Reference work is without a doubt one of the most difficult things we do—it can be both very frustrating as well as incredibly rewarding! In fact, active Data Librarians will quickly tell you that the growth we’ve seen in the uses of data have led to corresponding growth in the amount of time we expend being counselors, advisors, and "data detectives".

    The article by Jim Jacobs we mentioned earlier poses a "continuum" of reference services ranging from passive referral in which you simply point a client to a resource to highly interactive reference in which you suggest specific data collections for the client to use.

    One of the key elements of your mission and service profile will need to clearly define what you will and will not do in this regard.

    As also mentioned previously, at last year’s IASSIST conference (do you all know what IASSIST is?) Laura did a workshop on conducting reference interviews in the data library. The entire workshop can be accessed at:

  • http://dpls.dacc.wisc.edu/types/data_reference.htm
    •

    In it there is lots of information not only about reference in a data library setting but more general tradition library reference theory, semiotics, information seeking models (is this boring stuff or what?!), and a detailed breakdown of the reference interview process. And while many of you may never wish to think about the reference interview in any great detail, basic ideas like learning to put the search into context, topic refinement, and so forth can be very useful.

    Although it is not uncommon to encounter questions, such as "Do you have the PSID?" it is equally (if not more) likely that you will encounter questions such as:

    Like many aspects of what we do, you can do it reactively or proactively by accumulating information and learning whatever you can to help prepare for your users’ questions. For those who choose to perform active "data reference", the reference interview should include at least the following elements:

    Furthermore, you should gather information about your client in this process. How sophisticated is the client relative to the task and time-frame at hand? For example, is your client a student trying to write a course paper? Is he or she a faculty member starting out on a new research project? Does your user have experience using data, computers, statistical software? All of these questions will shape the type of collections that you will refer the client to.

    DATA ARE LIMITED. There aren’t always data sets that meet the users research criteria, even when the question is incredibly fascinating!

     

    2. Searching the Archive for Data

    Once you have identified the basic parameters of the client's research needs, the next phase is to determine whether ICPSR has a collection or many collections that might be germane. Minimally, you should be familiar with the basic techniques used to navigate the Archive. The strategy you suggest or actually use will be conditioned by the specific research question. For example, consider a client who is interested in data on "appeals trials". There are at least three strategies to pursue:

    Searching the Archive is a fairly straight forward activity. The pull-down boxes beneath the search screen enable you select "Title, P.I. or Abstract" depending upon the type of search you're conducting. You can also opt to search for "words", loosely defined as text separated by spaces; or "strings", defined as a sequence of characters in a word. It bears noting the search engine does not allow "OR" searches. In the best of all worlds, your search will produce a reasonable number of collections satisfying the search criteria. For example, our search on appellate produces a list of eight. If no studies are returned, you might want to make your search more general, or try a synonym for your search term. If you fetch too many studies, try making the search criteria more specific.

    In many cases, you may leave it to your client to search the Archive for germane collections; however, there is a tremendous amount of information in the Abstract entry that is salient to you to do your job effectively.

    Summary: Collection Abstract

    Study Level Information

    In addition to "study" level information, the abstract contains information about the component data files called "parts" in ICPSR parlance. The part information includes entries about the file format (rectangular, hierarchical, relational, etc.) as well as data about the number of variables, number of observations, records per case, and the like. As a rule of thumb, you can calculate the size of files by multiplying the logical record length by number of observations by the number of records per case. This produces the byte count, so dividing by 1 million will generate a size in megabytes

    3. Who to call when you can’t figure it out.

    Unfortunately, Study Abstracts are just that--abstract. In many cases, it's not obvious to you or to your client whether there are any collections that meet your needs. In the case where electronic documentation exists, you might want to consider downloading the relevant files for browsing. The Web site does allow you to search for the presence or absence of specific strings in text based files; however, this method only returns the "lines" meeting your search criteria. If all fails, there are at least three types of persons who should be able to help you out:

    Ideally, the goal of the reference process is help a user identify a collection that is germane to the research question. At the core of the process, this typically means locating documentation for the user to read and understand.

    While most traditional academic books have a common format (table of contents, index, etc.), codebooks are not standardized, and they vary greatly in quality. New O.R.s would be well served by spending a bit of time reviewing codebooks to get a flavor of the various formats so that you can assist clients to interpret them.

    C. Considerations When Ordering Data

    Once you have figured out what the client needs, the next step will be to obtain the data collection(s). While data are "free" once you’ve paid the admission price, it isn’t necessarily the case that you need or want to obtain every single file that is available. The basic issues to consider are:

    Providing a user access to a data file from ICPSR may sound simple, but in reality it may be a complicated procedure. In yesterday's workshop given by Jim Jacobs and Chuck Humphrey a useful "decision tree" was used to help O.R.s through the data ordering procedure. This chart gives a very good representation of the various questions that must be dealt with when you supply ICPSR data to other individuals. These include ordering procedure, data formatting, data storage, and documentation formatting issues. The major focus of the tree is on the various choices that have to made made -- often in tandem -- by the data librarian and the user. These choices include how to download the data, in what format the data and the documentation should be obtained, and in what format they should be passed on to the end user. Answering these questions entails knowledge of both your user's resources and capabilities as well as your own.

    We reproduce their very useful decision tree here:

        Where to place the order 

  (removable media require use of cdnet to order / is there a netscape ver.?)

               |

       +-----------------+

       |                 |

    cdnet ftp        www.icpsr   <-- which to pick?  does it matter?

       |                 |

       +-----------------+

               |

   Is there a choice in documentation?

               |

       +-----------------+        

       |                 |

      No                Yes

       |                 |

       |        Which format to pick?

       |                 |

       +-----------------+

               | 

   What am i going to have to do to print it?

               |

               |

      Is there a choice in data (format & media)

               |

       +-----------------+

       |                 |

      No                Yes

       |                 |

       |        Which format to pick?

       |         Which media to pick?

       |                 |

       +-----------------+

               |

      Retrieve in compressed or uncompressed?  tar?

           Space needed to retrieve?

               |

       uncompress / untar

               |

       order verification

               |

       print documentation

               |

       prepare data file for local use

           What decisions need to be made to deliver locally?    

               | client's hardware/software

         - delivery format options

              ascii fix field text

              dif

              spss port

              system formats

              etc.

         - may entail zipping for local distribution, e.g.

     

    Related complications include how much checking you will do before handing off the data and documentation to your user. For example: will you confirm the file(s) match the specifications you received from the ICPSR? Does the codebook match the data? Are there any special notes associated with the data?

    In general, most of us use CDNet/FTP to order and retrieve data to our local sites. This has as much to do with tradition as with expediency. However, many of accummulate data requests and place orders for multiple data sets at the same time. While CDNet is clunky, it simplifies such orders. It's also the case that you can use the message facility in CDNet to order additional codebooks or to request a copy of a collection on CDROM or diskette. FTP'ing files can also be done in the background, thus enabling you to do other things while you wait for the collection(s) to be downloaded. There is one other feature of CDNet that is not yet implemented with WEB downloads. After placing an order for data, you receive a MANIFEST containing the titles of the collections you've ordered, format information about each dataset, and "special collection notes". This information is essential for data verification and for troubleshooting problems locally. For example, most statistical applications have a default record length (which translates into the default buffers allocated for retrieving data) which are often much smaller than the record length of the file. This information is typically found in the codebook, however, you can often troubleshoot problems with the information you have at hand from the manifest.

    EXAMPLE CDNET ORDER

    Note that there are screen dumps associated with this section that can be viewed by looking at the online slide show presentation.

    After you have placed an order, you are required to fetch the data within seven days from the ICPSR ftp server, ftp.icpsr.umich.edu . Once logged into this server, you will need to change directories where your new directory will correspond to the ICPSR Order Number that you obtained from your CDNet session. If you are using command-driven ftp (as opposed to a point-and-click version), you should verify BINARY transfer mode. Also, if your order contains many datafiles, you might want to consider turning off "interactive" prompting which requires you to confirm retreival of each file. This can be done by issuing the ftp command, prompt . If you do this, then you can issue mget * to retrieve all the files.

    When the data and supporting documentation files have been retrieved, most of us do a quick and dirty file verification. Typically, this is a primitive line count. For those working in an UNIX environment, the command syntax is: gzcat file.gz | wc -l . This verification simple confirms that the files we hold locally were transferred successfully. This does not tell you anything about the content of the files themselves (e.g. are there "binary" or other non- standard characters). As noted earlier, many of us DO NOT print copies of electronic documentation.

    D. Data Troubleshooting:

    Whether you want to or not, chances are you will be asked to help a user "solve" a problem with a data set. While you may not be a "SAS expert" or "SPSS Guru", you should have a basic understanding of how these programs work at the "data input" phase. You might want to check out:

  • http://www.spc.uchicago.edu/DATALIB/DLguides/Gdathand.html
    •

    This document lays out the most common strategies for helping users to setup SAS and/or SPSS runs for: rectangular, multi-record, hierarchical runs. We also have a very preliminary section on transferring data between applications.

    In our experience, approximately 90% of the data handling problems we deal with are "user/client" problems. Typically, these problems are most often a misinterpretation of the codebook or a "easily identifiable coding problem". (For example, misunderstood skip patterns, incorrect column locations, unfamiliarity with the particular format of the data -- like card-image, and so forth). As we state above, this is where it will really pay off if you can read a minimal amount of SAS or SPSS code. When dealing with these problems, it is useful to request that the user provide you with a copy of their statistical program, error log, and a description of the problem. Often it helps to sit down with the user and review the codebook. Always remember that in some respects the user may be more familiar with the data than you, and that this is a good thing. After some experience doing this, you will easily identify those common errors that seem to make up the majority of the problems.

    The remaining 10% of the problems that we encounter are truly data problems:

    As a rule, these "types" of problems should be referred to ICPSR, with the caveat that some problems are outside the "traditional" mission of ICPSR. We have discovered over time that ICPSR doesn’t "debug" problems on data sets that they "do not support"—those collections that they have not thoroughly processed [These are often referred to as "class IV" collections where the data and documentation are shipped out in same format as received.].

    As an OR, you need to make a judgement call as to whether you will take responsibility for pursuing the matter with ICPSR or whether you and the local researcher will pursue it jointly. At the U.W. we tend to encourage advanced users to contact the ICPSR directly as they often have a more intimate understanding of the data and the problem.

    E. Administrating the ICPSR membership

    1. Record keeping/Cataloguing.

    There is a difference between these two. What IS cataloging? Can you catalog your holdings in an OPAC? Do you want to have a local catalog of holdings or is simple record-keeping enough? How do you plan to reach your constituency, if at all? Do you need to worry about accountability issues?

    UofC

    We keep records, we do not "catalog". We massage the ICPSR manifests using a PERL script for incorporation into our local databases. We use this information for year-end reporting about the number of collections we have acquired as well as forecasting future storage needs. We also use this as a basis for providing information about our local installation of data, and this will be on our WEB site soon.

    UW

    DPLS does have a local catalog of holdings that we export to a web-based access system. Our holdings are not in the campus OPAC and we view our online catalog as being very important for outreach.

    We no longer "accession" (e.g., "catalog") the ICPSR data we get via FTP. Studies obtained via periodic CD-ROM releases are cataloged into our online database (title, author, abstract). We do acknowledge that this is a lot of work, but since our holdings are not in the campus online catalog, our own online catalog is extremely important for advertising our services and resources. We feel it is worth the effort to "beef up" our holdings this way.

    Paper documentation received for ICPSR data is put on the library's shelf even though (as stated) the data are not stored locally. (If this appears confusing, it is, we are still making our way through policy changes).

    There are lots of issues here having to do with the cost of storage, processing and so forth, as well as issues concerning making sure your data and doc are current.

    In our annual report we record not only "cataloged" studies but also ICPSR download statistics as well.

    2. Archiving data.

    UofC

    Every center has its own operating procedures. We "selectively" archive ICPSR data collections. Our criteria for archival candidacy is vague, but can be quickly summarized as follows:

    It’s quicker and easier for us to restore data from backup copies than to order from ICPSR given the online structure.

    UW

    DPLS is both an archive and a library. We pretty much no longer ‘archive’ ICPSR data, although, as stated, this policy is under consideration. I think our tendency is to download what is needed when it's needed and then distribute and delete. We have found the online access to data at ICPSR extremely satisfactory (our users are very impressed). The one problem is, of course, that not all data have machine-readable documentation.

    Our local archival mandate is extremely important and has been essential part of the increased support (e.g., increased funding) we have received. Today the amount of data coming in to DPLS from local Principle Investigators for archiving is greater than it ever has been. We attribute this in part to the visibility of our web-based online archive, and also staff enthusiasm and willingness to work very hard with Principle Investigators. Developing and maintaining the technology to ensure the continued preservation of our archival holdings has been an essential part of our work.

    There have been concerns expressed by faculty that we are not archiving ICPSR data. The concern revolves around the notion that things stored locally are a safety measure against the whims of technology. Again, as stated, our policies in this regard are still under consideration.

     

    V. Building Local Partnerships

    When you build a house in a community, you’ll find that you are inevitably building partnerships in the community. These partnerships can be with neighbors, various groups, commercial concerns, political entities, and so forth. While some of us may become very active in our community through building many such associations and partnerships, others may lead more private or solitary lives. Regardless, to enhance the quality of your life, or to facilitate solving problems, you will doubtlessly form relationships with others in your community.

    Of course our "building a house" metaphor can only go so far. As a data service provider, you do have the flexibility to form what partnerships you want, when you want. However, you may find that these partnerships are extremely important for the success of your enterprise.

    One very good example of local partnerships focuses on the fact that those who provide data services are often very well equipped to identify bottlenecks in the execution of research. The ability to identify these bottlenecks and to identify those who can help alleviate them is an important part of facilitating your researchers' work. Identifying bottlenecks is only one way that we as data professionals can work as an intermediary at a profound level and become an important part of our client’s research team. From reference services to computing services to data services, we can participate in a variety of ways to fulfill our users’ needs. Doing this involves forming a variety of partnerships with a wide range of individuals, groups, and institutions.

    The types of local partnerships you will need to or want to build are highly dependent upon your organizational set up and affiliations. They will depend upon the environment within which you work and that which surrounds you. They will depend on the expectations of your users and the resources available to you.

    Very few of us, if any, work in a vacuum. We work closely with other professionals including reference librarians, statistical consultants, purchasing departments and the like. Putting together a mission statement is of utmost importance in defining your position within your local community. A mission statement will define what services you will and will not do in terms of "Levels of Service" and in context of your membership. Questions to be answered include will you or will you not store data for people? Who will you or won't you help? Will you offer statistical consulting? Will you pay for miscellaneous items like copies of codebooks?

    Obviously a primary local partnership will be with your clientele. You will need to identify who they are and their needs. This statement seems simplistic, but it is deceptively so. Your user base may change, grow, shrink, or evaporate. Many of us have seen an increase in the number of undergraduates using our services, or the number of classes requiring secondary research. Another issue is keeping in touch with your clientele and their interests. Once you have identified your users its very useful to know what they are doing. One way of accomplishing this is by monitoring publications, working papers, departmental seminars and so forth.

    You will need to identify those local experts who can assist you in fulfilling the needs of your clientele. These may be faculty, librarians, computer technologists, departmental chairs and administrators, or college deans. Again, the environment in which you work will in part determine the specifics. Historically, the earliest manifestations of data libraries and archives tended to be associated with academic departments, research institutes, or computer centers. In the early and mid-1990s trends affecting traditional libraries, particularly in the United States, suggested that collaboration between reference library staff and data library staff was becoming critical to providing responsible reference and research support. In recent years there have been numerous instances of data services being absorbed by campus library systems.

    The organization in which the membership resides also has "needs". These needs may include securing funding to pay for the membership, promote the membership, or facilitate use of the membership. Other needs may focus on procuring staff support of all types—professional, technical, student hourly, or administrative.

    Local partnerships may be critical for the continued well being of your organization and the membership. Developing a base of faculty who appreciate and value the services you provide can assist you in numerous ways—from fending off budget cuts to obtaining additional resources.

    Local partnerships will be important for the development and expansion of your services. Data not obtainable from the ICPSR but in demand may need to be purchased using local "consortia". In these circumstances the ICPSR membership can sometimes be used as leverage. At DPLS we often justify other departments purchases of non-ICPSR data by pointing out to them that they benefit greatly from our funding of the membership cost.

    Local partnerships can be built by attending departmental brownbag lunches, meetings of local professionals and experts (library groups, technology groups and the like), and other similar sorts of events. Your involvement in any of these groups will be dependent upon your own skills and interests, but any involvement outside the strict purview of your unit will reflect positively upon it. It is a way not only to secure support and assistance when needed, but to spread the religion, so to speak, of what you do and the services you provide.

    Much of the type of assistance we offer our clientele is detailed and long-term. The relationships we form often last for years. Research projects, whether a dissertation or an analysis for publication, entail developing long-term relationships between users and data service providers simply because the projects can go on for extended periods of time. Again, these become partnerships and offer us the chance to become a valuable member of a research team. These partnerships enable us to work as an intermediary in many aspects of our users’ research. They also provide the perfect opportunity to obtain feedback from those who use our services. Obtaining feedback from users should be an on-going process to measure the effectiveness of your operation.

    As we have seen, it is important to evaluate who your users are and what they need, as well as the effectiveness of the services you offer. Because much of our work is done one-on-one, we have ample opportunities to measure our effectiveness by obtaining feedback directly or indirectly from our users. Often it is as simple as asking about the value of the resources you offer, learning why certain resources were selected or rejected, and acquiring knowledge of the content of resources (such as data files) in the context of their use.

     

    VI. Building Partnerships with the ICPSR

    As mentioned earlier, by becoming members in the ICPSR, we invest considerable resources; and at least for U.S. institutions, it can be very costly to exit. Similarly, the ICPSR relies upon our long-term investment to continue doing what it does. This interdependency creates a hospitable environment for cooperation. This concluding section briefly explores the ways in which we can and should grow with the ICPSR.

    Public Relations and Policy

    As noted earlier, ORs bear the responsibility for membership maintenance. The previous section described methods in which for constructing and sustaining local partnerships. Viewed more broadly, part of you job as an OR is to publicize the ICPSR as a valuable resource available at your institution. In addition to the "data services" that we've discussed, ICPSR also sponsors a summer trainging program in quantitative methods and is home to the General Social Survey Data and Information Retrieval System. These both a incredible resources that are and should be publicized on your campus.

    In addition, you are also required to be aware of and adhere to ICPSR policies. The good news is that ICPSR is not overly laden with policies. However, you should be aware of the following:

    Basic Channels of Communication

    Building partnerships with ICPSR staff is undoubtedly the best way to make sure that you will get the most for your membership dues. a very good ICPSR policy is that you can contact individual staff directly (rather than going through a "help desk"). Get to know them and their specialties. Despite the incredible amount of work this small but dedicated staff does, they are always willing to help members. Knowing whom to call when you need assistance or have a question is very important. Much of this information can be obtained from the ICPSR web site. The ICPSR Web site provides a list of names, phone numbers, and email addresses for Administrative Staff. A second list of ICPSR Contacts (including area of specialty) is also available.

    Not to understate the importance of any ICPSR staff; but perhaps for a new O.R. the most important contacts will be:

    To tie this in with what Janet said earlier, the partnership with ICPSR is a two-way street. You must acquaint yourself with the staff there, but also make sure that you inform them of staff changes locally. You obtain services from them and at the same time may be called upon to provide services (either directly as in being asked to present at a biennial meeting or indirectly by informing them of problems in data, documentation, and the like). You should also take advantage of the fact that you can interact with ICPSR staff across all dimensions of data service. Like all of us, the staff are service providers and are open to suggestions on improving services. A common theme in my recent interactions with Janet, John Gray, and Mary Morris is that they do not hear from us very often and are not aware of bottlenecks that we might be encountering.

    In addition to direct interactions with ICPSR staff, you should also take advantage of the accummulated experience of the OR-community. OR-L is an an imformal discussion list where you can submit questions about particular datasets or problems that you are encountering. OR-L is also a venue for venting frustrations and voicing concerns about ICPSR practices and policies. For better or for worse, several members of the ICPSR staff are subscribed to to the list and have shown a great deal of responsiveness to our concerns.

    Membership in ICPSR can also be seen as a launching pad to other organizations that focus on data and data service issues. There are at least two groups that you might want to consider joining if you plan to be in data service over the long term. These are APDU (Association of Public Data Users), CESSDA and/or IFDO, and IASSIST (International Association for Social Science Service and Technology).

    The 21st Century and Beyond

    ICPSR members are incredibly diverse; including major research instutions and liberal arts colleges, individual members and federations of institutions. While all of us have primary responsibilities outside of the ICPSR, it is important we work together to keep the ICPSR-house stable and strong to meet the challenges that we all face.

    As we've tried to show in this session, "OR-ness" requires a certain amount of dedication to service and services. Certainly there are O.R.s (like faculty members) who've been assigned to the position but who are not able to provide even a basic level of service. In these cases it may be wise to try to locate someone in the campus library system (for example a reference librarian or a government documents librarian) who would be willing to help by taking on some of the responsibility. Although, granted, many librarians are overworked, it is likely that someone can be found to help -- if for no other reason that any librarian worth his or her salt will realize the value and importance of the membership.

    About DPLS Archive Catalog Internet Crossroads
    DPLS Home Page Publications Search DPLS 

    Email: disc@mailplus.wisc.edu
    Last updated 12 April, 2000.