A SURVEY OF ICPSR MEMBER INSTITUTIONS Ann Janda Vogelback Computing Center Northwestern University INTRODUCTION At the meeting of the ICPSR Official Representatives (ORs) held at Ann Arbor on November 11-14, 1983, a brief survey of member institutions was handed out to attending members. The survey was drawn up at Northwestern University by Lorraine Borman and Ann Janda (Vogelback Computing Center) in an effort to find out what other universities were doing in planning for a distributed computing environment. The questionnaire asked for information about what services are provided to access ICPSR data, mainframe and software usage, current extent of micro usage, and a final open-ended "future plans" question. Out of approximately 150 meeting participants, 38 responded to the questionnaire. This report discusses the results of the survey and provides two appendices: (A) an index to universities and their mainframes, and (B) a listing by university describing hardware, software, services, and future plans. (Appendix B available from Ann Janda, Vogelback Computing Center, Northwestern University, Evanston, Illinois 60201.) COMMENTARY ON SURVEY RESULTS BY QUESTION 1. Availability of ICPSR hardcopy codebooks: 18 at departmental site 15 at university library 9 at central computer site 6 at some other location The above results show that most institutions housed their codebooks at the department with the university library running a close second. The actual totals to this question ran higher because from the 38 responding institutions, 7 located their codebooks at more than one site, generally with a department/library combination. Research institutes largely fell into the "other" category, accounting for 4 of these 6 responses; one research institute identified with a department. 30 Computer programs such as database management systems to identify and access datasets? If yes, which? 13 Yes 25 No This was probably an ambiguous question. For those who answered "yes", the responses fell into two categories: (1) software that provided tape access to the data or (2) software that identified the studies by title or abstract. No two programs mentioned were alike. The software programs are listed in Appendix B. Personal assistance in identifying and accessing datasets? 23 at departmental site 12 at central computer site 8 at some other location 7 at university library According to the above results, by far the greatest assistance in identifying and accessing datasets is given in the department. As in question 1, more than one answer was checked, and indeed, 7 departments shared this function with the library and computing center. Most of the responses in the "other" category pertained to social science research laboratories or institutes. 4. Computer programs for graphic analysis of data? If yes, which? 24 Yes 11 No Of those who responded 'yes', most used SAS/GRAPH. SPSS PLOT was another popular package as was TEL-A-GRAF Display (spelled 3 different ways on the questionnaires). The graphics programs are also enumerated in Appendix B. 5. Consulting on statistical analysis of ICPSR data: 25 at central computer site 24 at departmental site 9 at some other location 1 at university library The above figures splitting the statistical consulting load between the department and the computing center accurately reflect the number of multiple responses falling into the combined department/computing center pattern. Slightly more than 31 half of the institutions provided this combination of statistical consulting. Most of the research labs ("other" location) shared this function with both the computing center and the department. How are your ICPSR data stored? 37 magnetic tape 20 system files, e.g., SPSS or SAS files \l raw data files disk packs Almost all of the member institutions store their data on magnetic tape—probably as they are originally received from the ICPSR. In addition to storage on magnetic tapes, more than half convert at least some of the data to SPSS or SAS system files. Disk packs also seem to enjoy considerable use. Do you offer subsetting services for large data files? 20 Yes 17 No The responses indicate that a little more than half of the institutions provide some form of subsetting for their users. Although the responses were straightforward, it is not clear whether the services provided only consulting and programs enabling the user to subset the large dataset—or actually "doing the subsetting" for . the user and delivering a smaller dataset based on user specifications. The latter option was intended. Is documentation available online? 17 Yes 19 No Although the responses to this question showed a fairly even split among ICPSR members in providing online documentation, some of the additional comments to this question indicated that this may have also been ambiguously phrased. While a few comments referred to availability of "abstracts", a few others referred to the use of machine-readable codebooks and dictionaries on tape. Obviously, online documentation could be construed either as study descriptions/abstracts—or as actual data documentation, e.g., codebooks and data dictionaries. While the online study descriptions were the target of this probe, the online codebooks have more interesting and powerful applications. What mainframe ICPSR data? 18 IBM 13 DEC 8 CDC computer(s) are used for storing and processing 3 PRIME 3 Others 2 UNIVAC The above figures indicate that leaders in hardware are IBM, DEC, and CDC— in that order. Within the "other" category are two AMDAHLs and a Hewlett-Packard 3000. A more detailed enumeration of specific models by university is indexed in Appendix A. 10. If more than one mainframe is used, how are data transferred? In all, there were 12 responses to this question, ranging from the use of DECNET, ETHERNET, tapes, to locally developed software and utilities. Specifics are found in Appendix B. 11. What mainframe analytical packages are used with ICPSR data? Mos tly Sometimes Seldom Never SPSS: 35 2 1 SAS: 10 11 1 14 BMDP: 1 13 15 7 OSIRIS: 8 10 18 SIR: 4 5 27 Other: 1 6 2 1 It is no surprise that SPSS and SAS are the most commonly used statistical packages for the mainframe. BMDP comes in as the next contender with high usage in the used "sometimes" and "seldom" categories. The comparatively low frequency of usage for SIR could most likely be explained by its relative newness in the market. Among packages named by respondents in the "other" category were MINITAB, ABC, TSP, SCSS, MIDAS, DATATEXT, and TROLL. From this group SCSS was noted as most commonly used by one institution; the remaining packages fell largely into the "sometimes/seldom" category with MINITAB usage noted at 5 institutions. 12. What use is made of microcomputers in analyzing ICPSR data? a great deal 8 some 30 none or virtually none Micros have not made much headway as consistently used tools for statistical analysis—yet. Although the number of responses in the "none or virtually none" category seems high, a few were 33 qualified by remarks referring to the early arrival of additional micros, pending changes to increase the use of micros, and current use by individual faculty and user groups. Also as more statistical packages become available, more use is anticipated. 13. If microcomputers are used for analyzing ICPSR data on your campus, please describe how the datasets are downloaded for use: There were a total of 9 responses to this question. Two institutions used KERMIT; the remainder variously responded with COMTTY, SMARTTERM PC, locally developed software, public domain software, an interactive mainframe program via SUPERWYLBUR—on through "the problem is currently being addressed." 14. If microcomputers are used, what are the most common machines and analysis programs? From 11 responses, the most frequently mentioned micro was the IBM PC followed by the APPLE HE. IBM-XTs appeared among the sprinkling of Victors, Zeniths, TERAKS, Radio Shacks (probably TRS-80s), and Tl Professionals. No clear pattern of software use emerged. Comments referred to uses other than statistical, e.g., as spreadsheets and as terminals hooked to mainframe statistical programs. It appears that presently micro stat packages are largely under consideration for use/purchase, e.g., one of the institutions is reviewing Statgraphics , Statpak, and other programs. 15. Are you planning to change your existing method of access and distribution in the next two-three years? If so, how? Twenty three respondents answered "yes" to this question. Their elaborations mainly pointed to the increased use of micros. For many this increase is linked with networking, workstations, shared storage, and subsetting datasets for micros. A few mentioned a move toward creating a social science laboratory. The complete references to future plans are printed in Appendix B. SUMMARY: WHAT HAVE WE LEARNED? We have reviewed information on hardware, software, and services provided to users of ICPSR data at 38 member institutions. For most institutions, the department is the most important site providing services to data users, e.g., codebooks and personal assistance in identifying datasets. The central computing site is next most important, and the library's role 34 is limited largely to storing codebooks . SPSS is by far the most common program for analyzing ICPSR data, and about half the institutions store data prepared for analysis in SPSS or SAS system files. ICPSR data are still analyzed primarily through computers at the central site, despite the popularity of microcomputers. Indeed, NO respondent reported extensive use of microcomputers, and 30 said that they made little or no use of microcomputers at present. Although most respondents reported future plans for analyzing ICPSR data with microcomputers, the revolution has not yet occurred . APPENDIX A: UNIVERSITIES AND MAINFRAMES INDEX ARIZONA STATE UNIVERSITY AUBURN UNIVERSITY AUSTRALIAN NATIONAL UNIVERSITY BARUCH COLLEGE, CUNY BOWDOIN COLLEGE BRYN MAWR COLLEGE CALIFORNIA STATE UNIVERSITY, FRESNO CARNEGIE-MELLON UNIVERSITY CENTRAL MICHIGAN UNIVERSITY COLBY COLLEGE CORNELL UNIVERSITY FLORIDA STATE UNIVERSITY HUNTER COLLEGE, CUNY INDIANA UNIVERSITY LOYOLA UNIVERSITY, CHICAGO NORTH TEXAS STATE UNIVERSITY NORTHWESTERN UNIVERSITY OHIO WESLEYAN PRINCETON UNIVERSITY RUTGERS UNIVERSITY SSRC DATA ARCHIVE SWARTHMORE COLLEGE TEMPLE UNIVERSITY TEXAS TECH UNIVERSITY TRINITY COLLEGE, YALE FEDERATION UNIVERSITY OF ALBERTA UNIVERSITY OF CALIFORNIA, LOS ANGELES UNIVERSITY OF FLORIDA UNIVERSITY OF IOWA UNIVERSITY OF OREGON UNIVERSITY OF PITTSBURGH UNIVERSITY OF SOUTHERN CALIFORNIA UNIVERSITY OF UTAH IBM, MODEL: CDC, MODEL: IBM, MODEL: DEC, MODEL: 1020 CDC, MODEL: 170 IBM, MODEL: 3081 IBM, MODEL: 3033 DEC * UNIVAC IBM DEC, MODEL: 10 OTHER: HP 3000 CDC, MODEL: 170 SERIES, 720 LOCALLY— 730 AND 760 CENTRALLY IN LOS ANGELES DEC, MODEL: 20 'S AND VAXS CDC * WE HOLD TAPES AND USE AS NEEDED DEC, MODEL: VAX 3081 * DEC, MODEL: : 760, 730 370 SYSTEM, 3081 ; VAX & PDP 11/44 /855 (MOST USE) * PRIME IBM, MODEL: 30335 IBM COMPATIBLE, MODEL: NAS 8040 DEC, MODEL: VAX 11/780 * CDC, MODEL: CYBER 170/730 DEC, MODEL: VAX 750 AND TWO 730' S IBM, MODEL: 3081 IBM COMPATIBLE, MODEL: NAS/ 9000-2 * DEC, MODEL: VAX 730, VAX 780 DEC, MODEL: 10 PRIME, MODEL: 750 CDC, MODEL: 172, 174 (SOON, 750) IBM, MODEL: 3033 IBM (AT YALE) OTHER: AMDAHL 5860 IBM, MODEL: 3033, AND 4341 IBM IBM, MODEL: 3033 * PRIME, MODEL: 750 IBM, MODEL: 4341 * DEC, MODEL: 1051 DEC, MODEL: 1099 IBM, MODEL: 360 UNIVAC 35 UNIVERSITY OF WASHINGTON UNIVERSITY OF WINDSOR UNIVERSITY OF WYOMING UTAH STATE UNIVERSITY WEST VIRGINIA UNIVERSITY CDC, MODEL: 750-150 IBM, MODEL: 3031 CDC IBM, MODEL: 4341 * DEC, MODEL: VAX 11/780 (5) OTHER: AMDAHL 420 BOOKS' POLITICAL TERRORISM A RESEARCH GUIDE TO CONCEPTS^ THEORIES^ DATA BASES AND LITERATURE by ALEX P. SCHMID, Centre for the Study of Social Conflicts, State University of Leiden, Leiden, The Netherlands 19 84 XIV +5 86 PAGES PRICEi US $40.50 ( IN USA & CANADA) DFL. 95.00 (REST OF WORLD) ISBN 0-444-85602-1 PAPEISACK APPROX. PUBLICATION DATE 12/84 This extensive handbook surveys contemporary social science thinking on political terrorism. As a reference work it provides the reader with the largest bibliography on the subject--a computer-based, partly annotated, 4000+ item, multi-disciplin- ary, multi-lingual literature survey covering aspects of theory and practice. Apart from regional and country entries, it carries subdivisions on such varied aspects as nuclear terrorism, hostage saving measures, state terrorism, etc. Divided into 21 major and 46 minor categories, this author-index biblio- graphy covers legal, psychological, sociological, military and ideological aspects of political terrorism. The handbook also includes a 130-page "World Directory of 'Terrorist' Organi- zations and Other Groups, Movements, and Parties involved in Political Vio- lence as Initiators or Targets of Armed Violence," which has been com- piled by A.J. Jongman. An 80-page survey of current thinking on the origins of terrorist violence in various contexts provides insight into the sociological, psychological, conspiratorial theories on the subject, and attention is given also to the theories of regime terrorism and those of the terrorists themselves. A further section in the volume discusses and evaluates available data bases for the study of terrorism such as those of the CIA, RAND, etc. The accessability and the reliability of data on terrorism are discussed and data requirements for social science research are indicated. Finally, the present state of the literature on terrorism is discussed, and research desiderata and strategies pointed out. CONTENTS: Forward by I.L. Horowitz, Introduction. Parts: I. Concepts. II. Theories. III. Data Bases. IV. Literature. For information or to order, write: Faxon Europe P.O. Box 197 1000 AD Amsterdam The Netherlands 36