SUGI 1988 Papers - Index

282 papers. Source: SAS Users Group International conference proceedings.

#	Title ×	Authors ×	Abstract ×	Keywords ×	Extras	Pages
1	A Backup System for SAS Data and Program Libraries Using SAS/AF Software	Kelly K. Antony, Donna M. Strauss	The backing up of files stored on disk on our IBM mainframe is done once the statistical analysis for the clinical or preclinical study is completed. This backup is necessary so that precious disk space can be freed up for ongoing studies. A menu-driven -backup system has been written on our IBM mainframe using `SAS/AF`. This `SAS/AF` application is used to copy SAS Data Libraries, SAS programs (stored as members of partitioned data sets) and, also data stored in flat sequential files, from disk to tape. This system also deletes files from the disk after the backup tapes have been created and generates listings of files stored on the backup tapes. The SAS System is used to generate the code and JCL for all programs in this Backup system. Programs are automatically submitted to the operating system through the `SAS/AF` program screens	proc step, proc contents	source	6
2	A Condensed Picture: Twenty-One Graphs on Two Pages	Madelyn A. Remsburg		proc step, proc greplay	source	3
3	A Data Base Interrogation System with the SAS System for Personal Computers	Stephen Mandel, J. Philip Miller		dataset, sas base	source	5
4	A Data Set Cross-Reference Report	William R. MacHose	Have you ever tried to find all the programs. job control language (J<Ll. and <LISTs that refer to a specific dataset? Programs. .lCL. and CLISTs are often held in parti tioned dataset libraries in an OS environment. This poster describes a SAS program that scans these libraries and prints a master index of dataset names and the library members that ref erence them	proc step, sample, sas library	source	3
5	A DataEase' to SAS Interface	Frank E. Harrell Jr., Lawrence H. Muhlbaier		variable, dataset, data dictionary, label statement, time variable, proc step, numeric variable, proc format, data governance, format statement, proc freq, date variable	source	5
6	A Documentation and Retrieval Strategy using the SAS System	John Podgurski		dataset, proc step, out=, data step, data dictionary, proc format, format statement, proc contents, proc print, label statement	source	6
7	A Flexible Disk Accounting and Reporting System under MVS Using the SAS System	Stephen W. Boyle	Efficient utilization of disk storage is a critical issue for many data processing operations. Physical size reductions and decreasing cost per megabyte of storage seldom keep pace wi th rapidly growing storage needs. Rarely used files are commonly kept for years because it is difficult to find out who is responsible for all of the files and whether they are necessary. Existing free space is frequently unused because those who need more disk are ignorant of the existing free space. A flexible disk accounting and reporting system for an IBM MVS/XA user was developed using SAS (1) to process data gathered by FDR COMPAKTOR (2). This system provides chargeback by disk volume and/or dataset name, free space accounting, dataset name convention assurance, and disk volume reorganization guidance	data step, management reporting	source	6
8	A Flow Chart for Users to Solve SAS Programming Problems	Joy Reel		sas log, numeric variable, error message, sample, proc step	source	3
9	A GPSS-like Language in the SAS System for Discrete-Event Simulation	Michael A. Greene, Gretchen K. Jones	This paper describes a program written in the SAS system which allows constructing a simulation using a s~bset of the GPSS language. Using this program, a dIscrete event model can be encoded as data (following the `CARDS` statement) and then run in the SAS environment. The advantage of this approach is the ability to use the statistical and graphics procedures in SAS for monitoring of the simulation during the run and for post simulation analysis	data step, data science, proc step	source	5
10	A Macro for All Seasons	Robert P. Beeman	The purpose of this brief overview is to provide the commercial applications programmer/analyst with both a general	error message, macro call, macro definition, data step	source	5
11	A Method for Hypotheses Testing in a Two-Way, Cross-classified Design with Missing Cells	Kenneth L. Koonce, Mattana Sanguanruang	Missing cells in a two-way. cross classified fixed effects model with interaction create problems in the estimation of treatment means and in the tests of hypotheses regarding main effects. This paper presents an alterna tive method for analyzing such data in two situations: (1) unobserved but existent treat ment combinations, and (2) nonexistent popula ti 005. For unobserved treatment combi nations. modified marginal means can be used to test hypotheses about mean parameters. For nonexis tent populations, linear relationships of means of all filled cells should be constructed, and the hypotheses tested according to the purposes of the researcher. This alternative method provides a rationale for selecting meaningful hypotheses when missing cells occur rather than depending upon the standard computer packages, since the result of this proposed	main effect, interaction effect, proc genmod, anova, fixed effects, hypothesis testing, sample, proc step	source	6
12	A Multivariate Approach to Social Inequality	Martin G. Ribe		logistic regression, estimate, sample, regression analysis, covariance	source	6
13	A SAS Application in a Large-Scale Foreign Epidemiological Study	Dwayne D. Oland, Sally K. Stansfield	The SAS system was used in a twelve year population registration-based epidemiological study conducted by the Hospital Albert Schweitzer in the rural Artibonite Valley of Haiti. The study incorporated vital statistics and health status data obtained during home visits by community health workers and mortality and morbidity records obtained from the hospital admission and outpatient reCOrds. The epidemiological analysis included the first detailed census of the area, disease morbidity and mortality rates, and the effect of socioeconomic, health program, and geographic factors on disease rates. Base SAS and `SAS/GRAPH8` provided the full spectrum of tools, procedures, and techniques required for effective data management and analysis in this study	sas base, proc step	source	4
14	A SAS Macro for Computing Pooled Intragroup Correlation Coefficients and Testing for Group Homogeneity	Norman A. Constantine, Michael S. Shing	Two or more separate groups are sometimes combined to compute a product-moment cor relation coefficient between a pair of variables based on all individuals in the combined group. The resulting correlation may be substantially different from the individual intragroup correlations, as well as from the intergroup correlation calculated on group aggregate statistics. Discrepancies are most. likely to occur when the groups differ strongly with respect to one or both of the correlated variables, or with respect to a third confounding variable. Which type of cor relation is appropriate for a particular application is a logical rather than statistical issue. There are times, for example in multisite clinical trials, when it is desirable to compute the weighted or unweighted average of the individual intragroup (site) correlations	correlation, sample, sample size	source	4
15	A SAS Macro for Estimating the Error Rates of Two Diagnostic Tests, Neither Being a Gold Standard	James J. Ashton, Melvin Moeschberger		likelihood function, sample	source	2
16	A SAS Macro System for Customized Report Writing	John Henry King	This paper describes a Macro system to aid users of the SAS system in report writing. Traditionally, customized reports are produced by writing PUT statements to place the heading text, body of the report, and footing text on the page. The disadvantages of this method are that it is tedious to program and difficult to change. The system described here changes the traditional coding technique with the use of a template and macro calls. The user enters the heading text, locations of variables, variable names and their formats, and the footing text into a text editor producing a template. This user defined template is processed by the macros to position the data on the page. The advantages of this system are 1) the user can see the format of the report before it is actually printed, 2) it is easier to layout the template than it is to write the PUT statements to achieve the same result, and 3) changes are more easily made	macro variable, call symput, put statement, macro put, call execute, macro conditional, macro loop, data step, macro definition, macro quoting function	source	6
17	A SAS Program for Selecting Subjects for a Case-Control Study	Gerald A. Kientzel, Terry L. Leet	The Case-Control Selection Program (CCSP) was developed to assist occupational epidemiologists in selecting subjects for case-control studies. This SAS program identifies the cases for the study and matches each case with R controls, based on certain specifications. The controls can be matched on sex, race, birth date, hire date, and/or length of employment, and can be selected with or without replacement from a pool of eligible controls. The controls can also be restricted to those employees who have not developed other exposure-related diseases. Demographic and disease-related information for the cases and controls selected for the study is stored in a permanent SAS data set for further analysis~ and appropriate printouts are generated	relative risk, dataset	source	2
18	A SAS/IML Software Macro for Testing Non-Nested Regression Models	Laura L. Bauer, Oral Capps Jr., Eric P. Smith		regression analysis, sample, proc reg, likelihood function, proc step	source	6
19	A Statistical Analysis of the Zip+4 Data Base	Peter Weiss		proc freq, data science, histogram	source	6
20	A Toolbox Approach to Networking SAS Experts or Reusing Instead of Reinventing the Wheel	Donald J. Henderson	Toolboxes are often used by data processing groups in the develop ment of computer systems. The concept of a toolbox is straight forward; generalized tools (e.g., software modules) are developed and packaged so that they may be easily used by an entire team or staff of people. `PROC PRINT` isan excellent example. SAS program mers do not have to write PUT and FILE statements each time they want to display a report of data; SAS Institute has developed a tool (`PROC PRINT`) and put it in a toolbox (the procedure library) available toaH SAS users. Macro tools presented at this and previous SUGIs provide other examples. Unfortunately, toolboxes have traditionally been used only for generic routines which can be taken "off the shelf" and used as is <e.g., the SAS procedure library and user developed libraries of procedures and Macros). Toolboxes have a much broader applica tion, however, especially in the educational and consultingenviron ment	sample, proc step	source	5
21	Administering SAS/SHARE' Software in an Information Center	Harriet Janes, Debbie Penland	Administering `SAS/SHARE` software in an information center is a matter of understanding the functions necessary for creating and maintaining the SAS server. Once the information center understands these functions. decisions need to be made on how these functjons will be delegated. This paper discusses how the information center at SAS Institute delegated these functions, the toofs that were developed to help in the administration, and the user support provided by the administrators. It is not the purpose of this paper to present a comprehensive discussion of the tech nical details of `SAS/SHARE` software	libname statement, proc step, dataset, autocall macro, batch job	source	7
22	Advanced SET and MERGE Processing	Robina G. Thornton, Merry G. Rabb		set statement, data step, do loop, conditional do loop, master data management, by group processing, %if-%then-%else, macro definition	source	10
23	Alternative Design for Corporate Information Systems Based on the SAS System	John G. Lord Jr.		data cleaning, sas base	source	3
24	An Algorithm to Solve Nonlinear Problems in General Equilibrium Theory and Operations Research	Kislaya Prasad	An algorithm to solve problems m Applied General Equilibrium Analysis and Operations Research is described. Typical problems m these areas are descibed along with an explanation of how the software can be used to solve them. The pwpose. in making these routines available from within SAS@ is for ease of use and wider accessibility		source	5
25	An Expert Sample Allocation Program	James W. Mergerson		sample, random sampling, sample size	source	2
26	An Expert System Interface	Julie A. Elston, David J. Purdon	In general, statistical software requires more analytical sophistication on the part of its users than most software. Users must be familiar with the syntax of the package, knowledgeable about statistical theory, and be able to converse with a variety of systems in order to use the software optimally. This places a tremendous burden on the amateur and occasional statistical software user. One answer to this common problem is to have a user-friendly front-end system to help users in being more productive with statistical software. This implies an interface which could access data from a variety of sources and run desired procedures. But more importantly, it also implies the ability to assist the user in applying basic statistical theory to the data analysis process. Expert system methodology presents a uniquely effective approach in developing such a system. We have built a demonstration prototype called THESA (The Expert Statistical ASSistant)	statistical software, regression analysis, correlation, autocorrelation, artificial intelligence, proc reg	source	5
27	An External File Interface for ISAM Files under VMS	Wayne Hester	A user-written external file interface (UFI) replaces the standard SAS I/O routines with another set of routines. These routines may allow a functionality not inherent to the SAS System. The inter face communicates with the SAS System through a data struc ture called the EXTIO communications vector. The ISAM driver provided by SAS Institute is an excellent example of a simple external file interface	key, proc sort	source	3
28	An Imputation System for a Large-Scale Survey	Eugene M. Burns, Doug Valcour		imputation, response variable, proc step, if-then-else, character variable, proc print, proc sort, sample, proc score, regression analysis, proc freq, macro definition	source	6
29	An Instrument for Evaluating the Quality of the Statistical Consulting Experience	David A. Ludwig, Judith M. Penny, James A. Penny		proc freq, proc step	source	6
30	An Integrated Clinical Data Management System at Allergan, Inc.	Jeffrey S. Kouba		data dictionary, clinical data, clinical research, dataset, audit trail, raw data, clinical trial	source	4
31	An Intelligent Interface to the SAS System for Use by Product Development Engineers	Debra D. Spencer		artificial intelligence, data science, regression analysis, dataset, proc step, sas programming, sample, anova, proc anova, histogram	source	6
32	An Interactive/Batch SAS Environment for VAX/VMS	Thomas J. Scott		sas log, sas batch	source	3
33	An On-Line Performance Tracking System Using MXG and SAS Software	Rodney L. Reish	An on-line performance tracking system has been developed at General Electric - Plastics Business Group using MXG and SAS. This system depicts important performance and capacity statistics for batch, IDMS/R [1], network, system and TSO processing. The system features a top-down reporting	batch processing, bar chart	source	6
34	Analyzing for Independence of Sequential Events with Markov Chain Models with the SAS System for Personal Computers	Paul F. Steblein		markov chain, sample size, sample, transition matrix, numeric functions, proc freq	source	5
35	Annotate with the GPLOT, GCHART, and GMAP Procedures: Using the Data System	Anthony L. Friebel	There are a number of ways within the Annotate Facility to describe the location, or display coordinates, that will be used by a function. The Annotate Facility accomplishes the task through reference systems. These	proc step, proc gchart, proc gplot, numeric variable, bar chart, if-then-else, proc gmap	source	10
36	Application of CHESMAP-An Environmental Data Mapping Program	Dan Jacobs, David G. Swartz	A computerized approach to displaying environmental data in maps is an important tool to those interested in analyzing the data. With the development of CHESMAP, customized maps are now available to anyone with access to SAS·. CHESMAP was written using the macro facility features of SAS and the graphics procedures of `SAS/GRAPH`·. The program creates data point, contour, overlay, time-plot or response surface multicolor maps in two- or three-dimensions. The map may be of the entire Chesapeake Bay region or a subregion based on longitude and latitude or by common areas (e.g. rivers, upper, lower, or mainstream Chesapeake Bay). This paper demonstrates the various options available with CHESMAP using some historical dissolved oxygen data from the Maryland portion of the Chesapeake Bay. While the environmental mapping program was developed using the Chesapeake Bay map coordinates, it can also produce maps of other geographic locations by using new map coordinates	macro variable, response variable, raw data, macro call	source	9
37	Application of the Analysis of Means	Peter R. Nelson		sample, anova, sample size, student t distribution	source	6
38	Application of the SAS Annotate Facility and Macro Language for Customized Graphics	Margaret F. Bryant, Peter P. Ruey, David L. Wahrenberger	Presenting results of c1101ca1 research often requires customized graphics not ;~;j~~~~~ ~~e~~~:~~rd T~:ti~~~ota~~ facility in SAS Version 5 is a powerful tool for use in meeting these needs. However, the required programming time is often unacceptable y and the modification of previous code is inefficient. At Ross Laboratories, our	scatter plot, macro definition, macro variable	source	5
39	Applied Research in Traffic Analysis: An Inductive Approach (Graphics and Moving Array Matrix Find Sample Expansion Factors)	Barbara C. Hertzog	Research by Washington State Department of Trans portation provides information about behavior of traffic for use in SUbsequent data collection and to improve predictions of vehicle volumes and truck weights used for highway design. Using previous obtained data, a system of SASI programs utilizes descriptive graphics interspersed with computation and inductive observations. Twenty-four hOUf traffic profile models with hourly mean volumes of 12 vehicle types were developed from actual data. Models were put through a moving array matrix producing all possible consecutive 4-hour samples to develop expansion factors for converting short samples to 24-hour volumes by time of sample. The entire process (which uses SAS means, plots, arrays, transpose, and macros) is broken down into a series of short programs with understandable steps not requiring complex mathematics or advanced SAS tech niques. In the moving matrix, the late p.m. hours are wrapped around to add to early a.m	sample, if-then-else	source	6
40	Automated Test Generator for Software Testing	K.L. Kanthan	A dynamic user-friendly IVP (Installa tion Verification program) system was designed to assure quality testing of online and batch software products. us ing several features of SAS. This menu system. driven by the user input. runs TPNS script simulations. generates vCL and Jobstreams for the objective of au tomated software testing. It facili tates browsing the results of the requested tests. Base SAS. `SAS/AF`. and the Display Manager software (OMS) made the development process easier. thus making the end product a flexible front end interface to testing of software products	macro definition, put statement, macro variable, sas base, macro facility	source	6
41	Automatic Footnoting of Specific Data in a Listing: An Example with Clinical Laboratory Data	James P. Young Jr.		proc print, sample	source	1
42	Automatic SAS Program Generation and Execution Using FORTRAN for a SAS/GRAPH Application	Chi-Kai King, Clara Yao	In a computer system that is based primarily on applications written in a high level language (eg FORTRAN), the integration of SRS Software may present some problems for both programmers and users. These problems include data accessibility, program and file maintenance, and user interface. This paper will present a FORTRRN program, running under uns, that addresses some of these problems. Using both user inputs and data retreived from specific data files, this program will generate a SJlS program to process and plot toxicology data such as animal body weights and food consumption. This FORTRAN program also creates a Del command procedure that invokes the SAS program. Control is passed to the command procedure at the end of the FOATAfUi prOgram. The SAS program uses POOC GAEPLRY to plot, display, group, delete, and save graphs without any user intervention	proc step, sample, dataset, proc gplot	source	4
43	Best Presentation of Data-Color	Linda W. Blazek, Douglas Scott		contour plot, proc step	source	7
44	Best Presentation of Data-Monochromatic	John G. Blodgett		proc step, if-then-else	source	7
45	Build as You Go: The Flexible Approach to Creating a Manufacturing Information System	Mary Felten, T. Godfrey Mackenzie	Growing worldwide competition has led manu facturing managers to focus on quality and pro ductivity improvements as key factors in reducing costs and increasing product quality. The Kohler Company's Cast Iron Division is uti lIzing several SAS* software products to build an integrated manufacturing information system to support these improvement efforts. The Cast Iron Division has maintained control and end user focus during construction of the system by adopting an incremental implementation strategy using engineering staff to write most of the code. This strategy along with the flexibility and integration of the various SAS products have af forded this build-as-you-go approach. This ap proach has contributed to acceptance of the new system and increased use of statistical process control techniques in managing the operation. Substantial cost reductions have resulted from the qualIty and productIvIty Improvement efforts durIng the past two years	dataset, proc step, sas base	source	4
46	Building a Window to a Data Archive, or How to Provide Access to Large Statistical Data Files	Tom Boggess	Researchers that use secondary data files, such as those produced by the Bureau of the Census, face the problems of obtaining information about the contents of such files and of formatting these diverse files for analytical evaluation. A data archive is charged with the responsibility to maintain and provide access to the data. An electronic informa tion and data retrieval system would seem an ideal interface between a data archive and its users. An information subsystem might allow users to obtain information about the databases by browsing or entering specific characteristics of interest, obtain ing information about which data bases contain such information, inform the user about the specifics of each field or variable in the database, and indicate any problems with merging disparate databases	phase i, proc step, dataset, phase ii, proc means	source	5
47	Building User-friendly Systems with Version 6 SAS/AF and SAS/FSP Software	Gail Kramer, K. Deva Kumar, Linda Wharton	A goal of end-user applications is to be user-friendly. It is impor tant that the applications programmer anticipate issues and questions that the user may have. The programmer must provide for a dialog between the user and the application so that deci sions can be made while the application is in use. In Version 6 `SAS/AF` and SASjFSP software, currently available in the PC environment, Screen Control Language {Sell is used to design applications that can prompt the user for all the information needed to complete a task. This tutorial expands on the ideas presented in the tutorial, "The Screen Control Language in Version 6 `SAS/AF` Software and `SAS/FSP` Software."	macro variable, proc print	source	4
48	CADAE: Computer-assisted Design and Analysis of Response Surface Experiments	Mark J. Levine	The CADAE computer program guides non-statisticians through experimental design, data entry/management and statistical analysis of a multi-objective searching scheme based upon Response Surface Methodology (RSM). The program consists of 2~ screens written usin~ the `SAS/AF` system on a MicroVAX II. The searching strategy is to fit a response surface model for each of several performance attributes (dependent variables) over a single factor space and to utilize contour plots and grid searches to identify the factor level combinations which meet	contour plot, sample size, sample, randomization, data science, anova	source	4
49	Changing Computing Environments in Midstream: You Can Lead a Horse to Water and Make It Drink	J.E. Diem, Alison Hartman		sas base, data wrangling	source	5
50	Color Hard Copy Technology-Plotters and Printers	Daniel Lee		data science, sample	source	5
51	Comment on Processing and Analysis of Supermarket Scanner Data	William M. Stanish		data step, proc step	source	1
52	Common Misunderstandings about the SAS Macro Facility	Earl R. Westerlund	The macro facility in base SAS software is a powerful tool which can increase the functionality and flexibility of SAS programs. The macro language is also very complex, and misunderstandings about how the language works can cause problems for beginning users. This tutorial will discuss a few of the common problems that beginners (and experienced users) have with the SAS macro facility. Intuitive, plain-English answers will be given to some questions that are frequently asked	macro variable, macro facility, macro definition, macro put, title statement, &variable, proc step	source	5
53	Comparison of Quantile Estimation Methods in the UNIVARIATE Procedure	Robert F. Carsel, Rudolph S. Parrish		sample, sample size, proc univariate, numeric functions, mse	source	6
54	Computer-aided NDA Review - A Microcomputer Approach	Gary N. Griffiths		dataset, proc download, decimal	source	4
55	Constructing Bar Charts with Customized Charting Symbols	Marje Martin, Bill Powers	By combining the use of `PROCEDURE OPTIONS` with the annotate facility in `SAS/GRAPH`, you have virtually unlimited control over picture production. This paper utilizes this combination to produce customized bar charts with bars constructed by stacking symbols either produced with `PROC GFONT` or selected from `SAS/GRAPH` fonts. `PROC GFONT` is used to construct the custom charting symbol, `PROC GPLOT` is used to produce axes scaled for the data set being charted, and the annotate facility is used to stack the appropriate number of symbols to produce each bar. I ntraduction Stylized bar charts constructed by stacking symbols which relate to the data being processed have become popular in presentations and th~ mass media. A bar chart depicting automobile sales across three years might be constructed by stacking the appropriate number o~ "automobile" symbols for each year's bar, with each symbol representing a fixed number of automobiles sold	call symput, proc step	source	5
56	Converting Existing Version 5 Applications to Version 6	Sue McGrath	Because of the many enhancements available with Version 6 `SAS/FSP`<!!I and `SAS/AF` software, all Version 5 `SAS/FSP` and `SAS/AF` applications need to be converted to Version 6 format in order to be run on a Version 6 system. This paper explains the steps of a conversion and its results, as weH as new Version 6 features for enhancing Version 5 applications. Also discussed are the CPORT, CIMPORT, and DOWNLOAD procedures available with base SAS software. Conversion of `SAS/AF` program screens from Version 5 format to Version 6 format is examined in detail, including the various conversion options	macro loop, macro variable, %if-%then-%else	source	8
57	Creating a SAS@ Language-Sensitive Editor	LeAnn P. Weaver	A unique full-screen editor was developed for the SAS System in the VMS operating environment. Digital Equipment Corporation created an editing pack age, the Langnage Sensitive Editor (LSE), for their V AX computers nsing the VMS operating system. LSE was nsed to create templates that represent SAS syntax. Also needed was the VAX Text Processing Utility (fPU). It was used to write procedures nsed to run SAS programs interactively. The LSE feature is an extra fea ture of VMS and mnst be licensed tonse it; however, VAX TPU is available with any version of VMS 45 or greater. This paper will be concerned with describing how the SAS LSE editor was created	character variable, proc step	source	4
58	Creating a Shipping Information System Using SAS Software and Barcodes	Sarah Ann Howey Dietrich	A unique shipping information system was designed using SAS- software and bar coded labels to get quick and accurate informa tion. Bar coding data saved time, reduced errors, and elimin'ated keying. SAS software was used to read and process the bar coded information, producing reports for UPS (United Parcel Service), RPS (Roadway Package System). other shippers, and SAS Institute Inc. management. A master shipping data base was created to track packages. The system allows the user to have immediate enhancements as the postal regulations and rates change and provides the best service at the least cost for the Institute and customer		source	4
59	Creating and Editing SYSTEM 2000 Data Bases Using the FSEDIT Procedure	Barbara A. Barrett	Applications with large SAS data sets may benefit by using SAS Institute's SYSTEM 2000 Data Management Software. Indexes are available to speed queries, and selection criteria can be applied to subset the data. Redundant values can be eliminated. Sorting and visual display of data relationships are provided. This paper examines new options available in Version 5 of `SAS/FSP` software's `FSEDIT` procedure on IBM computers for SYSTEM 2000 data bases	proc step, dataset, where statement, data governance	source	4
60	Creating Custom Symbols with the GFONT Procedure	Valerie Blettner	You can create a custom font of symbols by using the `GFONT` procedure	proc step	source	5
61	Creative Uses of Information Systems	Kirk Paul Lafler	This paper explores the creative uses of information systems using the SAS System. As the list of new features available on the SAS System for mainframes and for personal computers grows, the capabilities for developing sophisticated end-user applications become limitless. Numerous techniques and examples are presented illustrating the strengths found within the SAS System	artificial intelligence, data governance	source	5
62	Cubic Spline Smoothing: A Useful Tool for Curve Estimation	Mark S. DeHaan	Data analysis frequently involves fitting curves to data. Often the investigator has no idea what the underlying functional relationship is and ordinary functions or polynomials fit poor/yo This is especially common with experimental data that has a lot of noise and/or measurement error in it. In these cases, spline smoothing can be used to estimate and fit curves with excellent	regression analysis, weighted mean, polynomial regression	source	6
63	Customizing Graphics for Technical Reports Using SAS/GRAPH Software - Part I	Caroline M. Bobik, Arthur C. Singer	There is certain required information that must be displayed in graphs used for technical reports. One who is involved in presenting graphic information has a format which particularly appeals to that person. The format to which a given software package defaults may not be the desired format. Enhancement is possible by experimenting with available options for presentation until the desired format is obtained. This poster session presentation deals with application of `PROC GCHART` to data as an example of graphic format enhancement for technical reports	proc gchart, proc step	source	4
64	Customizing Graphics for Technical Reports Using SAS/GRAPH Software - Part II	Caroline M. Bobik, Arthur C. Singer	The volume of paper work generated in a research facility requires that reports be as concise as possible. FROC GREPLAY in SAS* software offers the ability to place several plots, notes and/or explanatory material on a single page	proc step, proc greplay, proc gplot, proc print	source	5
65	Cutting Costs in the DATA Step: Efficiency Techniques for Improving I/O and Processing Time	Neil H. Howard	Saving money is uppermost in the minds of DP managers, and the SAS System is certainly a "programrnercfficient" tool: minimal SAS code is required to perfonn traditionally code intensive tasks. Gratification and and results are immediate. Afamiliarquote suggests that the SAS System is easy to use, and just as easy to abuse. Because the SAS System is becoming increasingly popular as a development tool, it is essential to evaluate techniques for its effective use. This tutorial will define efficiency, suggest factors to be considered in measuring efficiency, and provide com parisons of techniques that improve efficiency. A primary area where the SAS System can be used more effectively isl/O. It should be noted that over 90% of the processing time in any given program involves reading and/or writing data (reference 1). This tutorial will illustrate how to avoid unnecessary I/O, minimize passes of the data, and control read-write loops	proc step, conditional do loop, raw data, proc format, dataset, put statement, input statement, proc append, set statement, sample, list input, cpu time	source	6
66	Data Entry in a Clinical Trial Using Windows in the SAS System for Personal Computers	MaryAnne DePesquo Hope	The SAS System for microcomputers was used to monitor study medications in a Veterans Administration Cooperative Studies Program clinical trial of vasodilators in congestive heart failure. This program was designed to facilitate data entry of discontinuation and restart of study medieations. The SAS software system using WINDOWS was utilized because of the requirements for end-user data entry, easily accessible files, and comprehensive data analysis	clinical trial, output	source	4
67	Data-based Water Quality Information System	Susan Campbell, Jeffery Finkeldey		sas log, in=	source	4
68	Dealing with Limitations of SAS/STAT Software on a PC	Thomas R. Birkett		proc genmod, proc reg, by group processing, proc step	source	4
69	Demystifying the FORMAT Procedure	Hallett German	(limit number of class variables)	proc step, proc format, format statement, proc freq, lookup	source	9
70	Designing a Drug Information System for the FDA Based on the SAS System	Nancy Bauer	By utilizing various SAS* software packages Merck & Company has designed and implemented an information system for the Food and Drug Administration (FDA) that can effectively be used for the review of clinical data, submitted as adjunct to the hard copy of a New Drug Application (NDA)	dataset, clinical data, proc freq	source	6
71	Detection of Influential Observations in Logistic Regression Using the SAS System	Charles S. Davis	Logistic regression is a widely-used technique for analyz ing data from epidemiologic and observational studies. Since the maximum likelihood fit of a logistic regression model is sen sitive to outlying responses and extreme points in the design space, the use of diagnostic tools similar to those developed for the standard linear model (see, e.g., Belsley, Kuh, and Welsch, _ 1980 and Cook and Weisberg, 1982) is recommended. While Pregibon (1981) and Cook and Weisberg (1982, Section 5.4) have developed diagnostic tools for logistic regression, these techniques are not readily-available in most statistical pack ages. In this paper, an easy-to-use SAS@ macro for calculat ing and displaying logistic regression diagnostics is described. The macro requires only the data set containing the values of the dependent and independent variables for each case and the output data set from PRDe LOGIST containing the pa. rameter estimates from the fitted model	logistic regression, regression analysis, logistic distribution, likelihood function, proc step, maximum likelihood estimation, proc reg, predictor variable, scatter plot, sample	source	6
72	Developing Full-Screen SAS/C' Applications on the IBM Mainframe	Pat Graham		sysrc, character variable	source	5
73	Digital Maps and the SAS System	James J. Knitis		proc step, proc gmap	source	6
74	Displaying Forest Research Data with SAS/GRAPH Software, PROC GPLOT, PROC GMAP, and PROC G3D	David C. Chojnacky, Mark E. Rubey, Carolyn K. Wraith	Graphic data display is an integral part of research at the Intermountain Research Station's Forest Survey Project of the Forest Service, U.S. Department of Agriculture. Graphics are used to explore data relationships and to produce customized graphs and maps of research results. Applications of `SAS/GRAPH` procedures `GPLOT`, GMAP, and G3D have been used to study tree growth relationships, map forest resources, and display response surface models. This paper discusses `GPLOT`, GMAP, and G3D as well as several graphics enhancements, the ANNOTATE feature, NOTE statement, AXIS statement, and LEGEND statement. Also mentioned are experiences running `SAS/GRAPH` software on a Data General MV8000 minicomputer and plotting results on Hewlett-Packard HP7475 and HP7550 pen plotters. Examples are given of graphs with oversized labels suitable for camera-ready slide production	proc step, proc gmap	source	3
75	Distribution System Design with SAS/OR Software and the Matrix Generator	Narendra Gangoli, Richard T. Jenkins	As part of its strategic planning process, Warner-Lambert regularly reviews its domestic distribution network. The objective is to determine the optimum number, location and size of its distribution centers which receive almost one million orders from approximately forty thousand customers. To fUrther complicate the problem, Warner-Lambert is comprised of three business groups with different service requirements. In the spring of 1987, Warner-Lambert formulated the problem as a mixed integer program with several hundred equations and almost two thousand variables. The difficulty was how to solve the problem in a short period of time. SAS-OR and the matrix generator developed by Marc-david Cohen l and `SAS/OR` were the answer. Within a three-month period, Warner-Lambert with the help of Mr	binary variable, data step	source	5
76	Don't Leave Your SAS Code in Dis-array	Robert Langill	Arrays are powerful, flexible and confusing. Using arrays in SAS code makes the code more compact and effective. By looking at how arrays work and why we use them, we can enjoy the advantages and avoid the angry bite of an abused array	array, numeric variable, data step, proc step	source	6
77	Dynamic Loading in the SAS/C Compiler, Theory and Practice	Stuart Newton	This paper explores some of the practical uses of the loadm and loadd functions of the SASIC compiler under CMS. Both benefits and drawbacks are discussed, along with tlgotcha'slf which accompany some uses of these functions	if-then-else, sample	source	6
78	Effective Use of Custom Report Writing Tools	Craig Dickstein	For SASR users needing capabilities over and above those provided by the reporting procedures this paper discusses the techniques for and the process, of generating periodic production reports. The design. construction, and implementation phases of the report writing project are discussed as necessary prerequisites for acquiring a superior end product. The FILE and PUT statements, system options, and `DATA STEP` functions are demonstrated to control the placement of variable and constant data on the printed report. Additionally, the Macro Facility is explored as a method for generating periodic reports. Since most of the material discussed is well explained, along with examples, in SAS System documentation, the emphasis is on a conceptual understanding of the report writing process	data step, put statement, sas programmer, macro facility, options statement, if-then-else, libname statement, macro variable, macro definition, raw data, dataset	source	6
79	Efficiency Techniques for SAS/AF Software Applications	Annette T. Harris		macro variable, proc step	source	3
80	Efficient Merge Operations for Historical Data	Leonard S. Lutomski	This paper briefly explores the relative efficiency of two ways of organizing historical data files in the SAS system -- as lists and as vectors of events - and presents the code for and a discussion of two useful variants of the SAS merge. These variants are (1) the date merge -- for the look-up of status information from historical data files and (2) the cross merge -- for generating the cross-products of BY-groups. The American Institutes for Research has found these two merge variants to be of substantial value. Their use has improved the efficiency of some extractions from large historical files by as much as 500%. Historical Data Files Historical files represent dated events. They typically have a subject identifier and a date as the two values minimally required to specify a unique event. The subject identifier specifies the individuals described by the file. The date specifies the time when some action affecting a relevant property of an individual occurred	data step, dataset	source	5
81	Efficient Programming with the SAS System for Personal Computers: Save Time, Save Space	Lisa Ann Horwitz, E. Tim Thompson	Users of the SAS System for Personal Computers are con fronted with lengthy processing time and storage limitations not usually encountered when running the SAS System under other operating environments. Therefore, efficient programming, while certainly an ideal in the mainframe and minicomputer worlds, becomes a necessity in the microcomputer world. On the other hand, users of the SAS System for PCs may have greater flexibil ity than their mainframe and minicomputer counterparts in the method they use to fun their SAS jobs and in how they configure their PC environment. Utilizing common sense programming techniques, as well as Release 6.03 syntax and configuration specifications, improves the efficiency, reduces the processing time of SAS programs, and cuts down on storage space for out put SAS data sets	dataset, where statement, proc step, if-then-else, subsetting if, set statement, cpu time, by group processing, proc format, put statement, pdv, keep statement, length statement, raw data, select	source	6
82	Efficient Use of a SAS Data Library as a Pseudo-relational Data Base	Vladimir Svirsky		set statement, data cleaning, proc step	source	6
83	Elements of Good Programming Style	Amy Caron	The itf{lOl"tance of good programnlng style cannot be under rated. Unfortunately. progranmers have a strong tendency to do SQ. We all I ike to believe that once a piece of code is put together. however haphazardly, the code wi II work properly from the point it is put into production until for ever after. Besides, why waste cleaning up the code if it is only going to be used once'? Writing code cleanly the first time not only increases the chance of getting it right buT also eases the task of fixing it in case it is not right. Furthennore. chances are likely that the program wi II be used more than onoe and it wi II probabl V be used by someone other than the author of the code. So in the interest of those writing code and those having to use code that has already been written. some elements of good pro gr-anming style wi II be discussed	if-then-else, character variable	source	3
84	Enhancements and Changes for Version 6 SAS/GRAPH Software	Mike Kalt	Users of Version 6 `SAS/GRAPH` software will find that programs written for Version 5 applications on other systems will run in Ver sion 6 with little or no modification. However, a number of new features have been implemented for Version 6 that give users new capabilities while retaining upward compatibility with Version 5. This paper discusses some of these new features	proc step, catalog entry	source	7
85	Estimating Logistic Regression Models When the Dependent Variable Has No Variance	Nicholas Scott Cardell, Dan Steinberg		sample, response variable, likelihood function, regression analysis, logistic regression, logistic distribution, proc step	source	2
86	Estimation of Row and Column Scores in the Linear-by-Linear Model for Two-Way Ordinal Contingency Tables	Charles S. Davis	Consider the situation when both the column and row variables of a. two-dimensional table are ordinal. A simple log linear model that -utilizes the orderings of the TOWS and the columns is the linear-by-linear association model. A disad vantage of this model is the necessity of assigning scores to the categories of the row and column variables. Although the integer scores are most commonly used in practice, there is no obvious choice of scores for many variables and the researcher may not wish to assume equal spacings. In this paper, we consider the alternative of treating the scores a.s parameters to be estimated from the data rather than as numbers to be supplied by the researcher. The resulting model was first discussed by Goodman, who referred to it as the "RC model." The RC model has the same general appearance as the linear~by-linear association model, except that scores for ordinal variables are treated as parameters	proc freq, correlation, proc cancorr, goodness of fit, null hypothesis, likelihood function, ordinal variable, proc step, dataset	source	6
87	Exact and Monte Carlo Methods for Computing Non-Parametric Significance Tests	Cyrus R. Mehta, Nitin R. Patel, Pralay Senchaudhuri		sample, monte carlo, confidence interval, proc step, proc freq, exact test, odds ratio, parametric test, wilcoxon test, wilcoxon rank-sum	source	6
88	Experience the Thrill of Label Creation: The Power of the FORMS Procedure	Rebecca Swartz		sas base, proc step	source	6
89	Extending the SAS System on Microcomputers with SAS/IML Software, the Interactive Matrix Language	Linda P. Atkinson		dataset, proc reg, proc iml, proc step, regression analysis, correlation	source	10
90	Extraction from Enormous Sequential Health Care Claims Data Files Using Multiple Format Tables	Nancy Fox, Craig K. Ray		dataset, proc format, macro definition, cpu time, proc step, insurance claims, lookup, macro variable	source	7
91	Extrapolation Sampling: Model-based Finite Population Sampling in Litigation Support	Timothy Wyant		sample, population	source	5
92	Forecasting as a Tool for Decision Making	Joel Fingerman		confidence interval, probability distribution, proc calis, sample	source	6
93	Forecasting the Economy Using the Composite Leading Index, Its Components, and a Rational Expectations Alternative	Lung-ho Lin, Steven C. Myers	The purpose of this paper is to evaluate the usefulness of the composite index of leading indicators and that of its component series and to propose a new approach to compile a more reliable and versatile leading index. The new	index, numeric functions, time series analysis	source	6
94	Future Direction of Biomedical Information Systems in the Pharmaceutical Industry Based on the SAS System	James F. Sattler		proc step, clinical trial, data review, relational database, clinical data, dataset, proc freq	source	4
95	Gaining a Competitive Advantage in Business Using SAS/QC Software	Matt D. Leet, Rodney L. Reish	A qual ity control chart system has been developed at General Electri~ - Silicone Products Business Division using `SAS/QC`. This system has saved the Business thousands of dollars by displaying out-of-spec trends during the manufacturing process. Additionally, the system provides a competitive weapon when marketing GE - Silicone products. The control chart system consists of three phases	quality control, sample, sas batch	source	3
96	Generalized Least Squares Regression with Partially Correlated Data	Andrew J. R. Gillespie	One assumption of the ordinary least squares (OLS) regression model is that the observations on the response variable Yare statistically uncorrelated. In data sets obtained from cluster sampling (usually a more cost-effective sampling method than simple random sampling), this . assumption is violated; observat1ons may be uncorrelated between clusters, but correlated within clusters. Application of OLS methods to such data can lead to serious underestimates of the precision of estimated regression coefficients. Many methods have been proposed for dealing with such intracluster. correlation, including genera11zed least squares (GLS), random coefficient regressions, and jackknife estimators of variance. This paper describes two algorithms written in the SAS Interactive Matrix Language (`SAS/IML`~) to fit GLS models based on work by scott and Holt (1984) and Fuller and Battese (1973). The algorithms pass through the data twice	correlation, sample, regression analysis, cluster sampling, covariance, response variable	source	4
97	Generating SAS INPUT Statements Using PROC COPYLIB	Sheryll R. Miller		proc step, input statement, dataset, proc copy	source	3
98	Generation of Serial Reports and Graphs in a Production Environment	Trisha Horman, Jeffrey Joseph	IN THE INDUSTRIAL SETTING, THERE IS AN EFFICIENT WAY TO GENERATE SERIAL CUSTOMIZED PRODUCTION GRAPHICS FROM MULTI-INPUT DATA. CURRENTLY, REPORTS ARE RECEIVED FROM A COMMERCIAL REPORTING AGENCY ON A PC DISKETTE. A REPORT IS RECEIVED ON A BI-WEEKlY BASIS PROVIDING TRACKING OF A PRODUCT'S MARKET SHARE IN ALL U.S. DISTRICTS FOR A 13 WEEK PERIOD. THE CURRENT ROUTINE TO HANDLE PRODUCTIQN OF THE REPORTS AND GRAPHICS IS LABOR INTENSIVE. THE REPORT OATA ON THE DISKETTE MUST BE REFORMATTED TO A FLAT FILE USING COBOL AND THE REPORTS AND GRAPHICS ARE PRODUCED BY USING SEVERAL SOFTWARE PACKAGES. EACH OF THE PLOTS WERE PREVIOUSLY GENERATED INDIVIDUALLY. A MORE EFfICIENT ROUTINE WAS DESIGNED USING SAS* IN A CMS OR TSO ENVIRONMENT TO pRODUCE THE REPORTS AND GRAPHICS FROM THE REPORT DATA GENERATED BY THE COMMERCIAL REPORTING AGENCY. THE REPORT DATA IS REfORMATTED AND A SAS DATASET PRODUCED USING SAS CODE	call symput, dataset, macro facility, proc step, macro definition	source	5
99	Getting Started with SAS/AF Software and Doing It Right	Howard Levine		proc step, macro variable	source	8
100	Graphical Presentation of Clinical Data Using SAS Macros	Pan-Yu Lai, Thomas E. Schneider	The adage, II a picture is worth a thousand words" is especi ally true when graphically displaying clinical safety data such as vital 5 igns, 1 aboratory assays, and reGs. Si nee clinical abnormalities generally depend on normal ranges and/or the magnitudes of changes from baseline values, a data plot which is able to clearly reveal these kinds of information and also label the "abnormal patients" by their identification numbers is very helpful to medical reviewers. This paper presents a SAS macro which performs this task on an x-v plot using the X-axis for baseline values and the V-axis for follow-up va1'les~ This macro autor:1atically extracts the necessary infannation from the input data set to produce the desired graphs. It also handles points with missing baseline values without excluding these points or distorting the graphs	outlier, proc step, macro variable, vital signs	source	4
101	Happy Haven Psychiatric Center: Dates in DO Loops	Ira A. Bader		do loop, dataset, sysdate9	source	5
102	Helping Naive Users: An Expert System to Prepare Initial SAS Analysis	David King, Shoko Tsuji, Laurie Zant	This paper examines how an expert system was built that prepares exploratory analyses with SAS. We focus on the knowledge used in the system and how it is represented. New. users of SAS often have difficulty utiliZing SAS effectively due to the tasks of preparing DATA steps, selecting appropriate PROCs for the type of analysis desired, as well as learning SAS's syntax. Expert systems can enable a new user to tap the power of SAS immediately by writing a SAS program based on a description of the user's data and requirements. At Teknowledge, we have written an expert system using our expert system shell, COPERNICUS TM, which produces a syntactically correct SAS program to analyze data using descriptive statistics. We assume the user of the expert system is familiar with statistics and has the data in hand. During a consultation, the user is asked questions about the nature of the variables used in the study	proc step, data analyst, data governance, data science	source	4
103	Horizontal Contour Lines Using the G3D Procedure	Arthur L. Carpenter	The relationship of _the vertical variable to the two horizontal variables in a three dimensional plot is easily shown by the surface produced by `PROC G3D`. It is, however, not always easy to determine the value of the vertical variable at some point on the surface. What is needed is the ability to plot contour lines or points of constant vertical value on the plot 1 s surface. A macro was written that can be inserted in the user's code prior to the `PROC G3D` step. The locations of_ the constant vertical points are calculated; those points that are not masked, hidden by the plot surface, are then placed on the plot as labels using the ANNOTATE facility. I ntraduction `PROC G3D` plots surfaces in three dimensions. The plot is formed by "stretching" an X-Y grid in the Z direction and the resulting plot takes ona three dimensional appearance as if the surface were being viewed from an angle (Figure 1). The viewer may TILT and ROTATE this plot to best advantage	proc step, macro call	source	5
104	Hosting a Mini SUGI at Your Site	John Chenier, Loretta Golby		proc step, proc sort	source	3
105	How Fast Is the SAS System for Personal Computers on Various PCs?	Nelson R. Pardee	How fast does PC/SAS run on one computer compared to another? One general comparison is the time required to run the TESTBASE installation program. This paper com pares times for a variety of PC's, mainframes. and mini computers. PC/8AS Is slow?! My first experience with PC/SAS was on a PC/XT with a hard disk that was slightly faster than IBM's. TESTBASE, the installation test job. took 22 :m1nutes. Worse. I watched the lines of of SAS code scroll down the we window with agonizing slowness. realizing that my jobs would scroll just as slowly. And it gets worse: when I got my own PC. Testbase took 27 minutes to run! Clearly. all PC's are not created equal when it comes to running SAS (and SAS Institute has clearly stated that it was designed for the AT even though it runs on the PC). What performance can be- expected? Complaining about how slowly SAS runs is not new. Users have been complaining for years that mainframe SAS is too slow	data step, cpu time	source	6
106	How SAS Software Makes It Easier to Use the IBM Response Time Monitor	Gregory Hawryschuk		bar chart, capacity planning	source	2
107	Implementation of a Hashing Routine in SAS Software	Craig K. Ray	This paper is intended to be an extension to the paper I A Comparison of Table Lookup Techniques (reference 1). In that paper, three techniques for table lookup were discussed and the applicability of each was compared	hash table, lookup, dataset, hashing function	source	7
108	Implementing Structured Design of Menu-Driven Systems with SAS/AF Software and the Macro Facility	John M. Rinehart	Structured system analysis and design techniques have come to be commonly used in "traditional" data processing environments, with the goal of producing information systems characterized by modularity, top-down structure, dear functionality, and maintainability. The increasing use of "fourth generation languages," with their reduced coding, simplified file management, and quick prototyping, has tempted some system designers to side~step these recently learned and proven techniques and implement systems "on the fly." For Shame. Structured analysis uses the Data Flow Diagram as a tool to describe the logical "what" of an information system. Several past SUGr presenters have noted the direct correspondence ofSAS· Software datasets and DATAIPROC job steps to elements of the DFD. Structured design goes on to specify "how" the information system will accomplish its appointed tasks	macro facility, proc step, macro call, macro variable, macro definition	source	5
109	Increasing Your Personal Productivity: A VAX/VMS Command Procedure for Submitting SAS Jobs as Background Processes	William A. Rosen	This coomand procedure wi II permit the user to submit $AS jobs for execution as a background process, thereby freeing the terminal for other activities		source	4
110	Information System Resource Management and Job Accounting in a Complex Network Environment	Keith A. Jones	The use of job accounting to provide for nwre cost-effective sys tem resource management is a traditional function within both large and small host processing environments. Chargeback and DP cost accounting have evolved bared upon the business princi ples of management accounting and variQUS methodologies cur rently exist to support a variety of specific management objectives. The application of such methodologies to network. processing environments, however, is sliD in the early stages of development. Network job accounting methodologies must rely equaOy upon statistical as well as traditional management ac counting techniques, and additional tools and frameworks need to be df!VeJoped and tested in order to provide effective resource management of complex integrated system environments. This consideration becomes especially important both as small PC-based workstation clusters or LANs access host networks	data wrangling, sas base	source	5
111	Information Systems in the Pharmaceutical Industry	Theresa J. Hart		pharmaceutical industry, phase i	source	3
112	Installing and Maintaining SAS Release 6.03 under PC DOS	David C. Brumitt		sas base, sample	source	5
113	Integration of ORACLE and the SAS System in the Generation of Adverse Experience Reports	Chih-Wu Chang, Edward F.C. Pun	This paper describes the combined usage of ORACLE and SAS to present adverse experience (AE) data for large clinical trials. Large-scale c~inical trials are characterized by the large volume of data which is collected from multiple centers with different study designs. The majority of the database usually consists of information related to safety'. Among all the safety data collected, adverse experience is some of the most important i nformat i on correspond-i ng to the safety of a patient. Adverse experience data collected on case report forms includes at least the following: date/time of the occurrence, severity, course, duration, relationship to the study drug and the description of a particular adverse experience. Adverse experiences can be presented in many. different ways. Most commonly, they are grouped by body systems, and the number of patients reporting these AEs are tabulated side by side for each treatment group	proc sql, clinical trial, crf, proc tabulate, oracle, proc step, dataset, adverse events, clinical database	source	5
114	Interacting with Your Host System from the SAS System under OS	Eric C. Brinsfield	During many SAS applications it is often more convenient and efficient to perform TSO or system functions while still in the SAS environment. This paper offers tips on running external programs or commands and demonstrates techniques for submitting batch jobs from within the SAS System. It also describes several meth `ODS` of testing return codes and explains the relationship of the SAS environment to the ISPF environment. The discussion focuses on tasks that can be completed using base SAS soft ware	sysrc, proc step, macro variable, batch job, macro facility, macro definition, %if-%then-%else	source	5
115	Interfacing IBM VS FORTRAN Version 2 with the SAS/C "' Compiler, Version 3.OOF	Stephen L. Maier	This paper addresses the interfacing obstacles and solutions identified while developing a library of 'C' programs to be used as a subroutine library for a VS FORTRAN program. The initial difficulty in interfacing program modules Ulritten in these two languages is in the calling conventions and the run-time environment. However~ the task does not stop there~ basic language differences such as data structures and format complicate the issue. The `SAS/C` Compiler addresses many of these issues through the use of a compiler option to generate independent code that can be called from any host language which conforms to the IBM 370 linkage conventions. To complicate the issu(-.! even more ~ IBM "s VS FORTRAN does not fully conform to the IBM 370 linkage convention. The result requires a sophisticated interface library of subroutines which 'bridge' the interface between the IC· functions and the FORTRAN main program. The intricacies and problL~ in developing the 'bridge' routines is presented	decimal, second	source	4
116	Interfacing SQLDS Data Bases and the SAS System: PROC SQLEXT	Paul M. Roesti	FRlC SQIElIT is a procedure in SlIS Institute's new SlIS/SQL-IE" pra:Iuct. '!be procedure nay be used to create a SlIS dataset ~ data extracted fi:an 1m's V1VSP relational database, 5OIy'I:S' * SQIlL'S naintains data as tables of rows and oolumns. SlIS datasets are also logically kept in this fashia> with ciJservations analagous to SQIy'I:S rows and SlIS variables to SQIlL'S oolUllU1S. similarities and differences between the SQIlL'S data manip.llatian l~, SQL, and SlIS `DATA STEP` statements are presented. When extractirq data fi:an several 5OIy'I:S tables, there is djso1ssian of whether or not to use a COIlPlex SQL S= statement or a sinpler 50L = follCJNed by SlIS `DATA STEP` transfonnations. '!be paper diso'sses use of FRlC 50= in rns noninteractive m:xle. SlIS programs that nul in this m:xle require 00 intervention fi:an terminal users. Teclmiques for prepar:in;l SlIS program statements to invoke FRlC SQIElIT fi:an = ro.rt:ines are presented	proc sql, data step	source	5
117	Interlanguage Communication Issues with the SAS/C "'"Compiler, Using FORTRAN as an Example	Ingrid Ammondson	A multilanguage program is composed of routines written in two or more high-level languages. Multilanguage programs are useful in the IBM 370 environment for several reasons. Many subrou tine libraries in various languages are available to perform stan dard or specialized functions. such as mathematical or statistical analysis. Additionally, it is often desirable to migrate gradually from one programming language to another, such as from PL/I to C. Thirdly, there are functions that are best performed in a par ticular programming language, such as producing a complex report with COBOL. A typical multilanguage application wlll freely combine routines in several languages. This mingling of lan guages often requires that the programmer have a detailed understanding of the varied environmental requirements of each language, even though the actual programmming task does not have such requirements. There are a number of issues involved in ease of communication between C and other languages	exception handling, sample	source	5
118	Jonckheere's Test Procedure for Dose-Response Trend	Dorothy Hu Huang, Armand Robles		sample, macro variable, anova, sample size, proc step, proc glm, macro definition	source	6
119	Logit Models for Ordered Categorical Response Data	Alan Agresti	Logit models are presented for repeated observations on an ordinal categorical response variable. The models describe how the marginal distribution of the response depends on values of explanatory variables. Primary attention is given to a cumulative logit model that generalizes models proposed by McCullagh (1977, 1980). Also diocussed are an a.ljruoent categories logit model that is equivalent to a loglinear model for marginal response probabilities, and a mean response model. The models are illustrated with an example in which a drug and placebo are compared with respect to treatment of insomnia problems. When the covanates are categorical. a weighted least squares fit for each type of model can be obtained using `PROC`	logistic regression, predictor variable, response variable, sample, longitudinal data, covariance, proc freq, repeated measures, proc step, missing value, odds ratio, treatment effect, regression analysis	source	9
120	Low Maintenance Programming Techniques	Robert Virgile	Elements of programming style and technique aff eet the amount of effort required to maintain programs. In many cases. the "necessity" to maintain programs is a direct result of the tools and techniques used when writing the original program. An ultimate goal would be zero maintenance: a program that can be submitted at any time, with no changes to the code, for any set of input data. In practice, this approach demands a broad range of skill and knowledge using SAS@ software. creativity. and a willingness to put in extra programming effort up front. The final program should require little or no attention on the part of the programmer. This paper investigates some of the techniques that reduce maintenance, gives examples of relevant syntax. and applies the techniques to sample programming problems. Where Does the Time Go? A few categories stand out as major time gluttons. For example: Figuring out what a program does	proc step, credit card, proc format, proc print, proc sort, dataset, if-then-else, sample, comment statement, macro facility	source	10
121	Macro and Macro-like Applications for Minicomputers	Virginia M. Dineley, Toby Trott	Version 5 of the SAS System for minicomputers has limited sup port of the macro facility. This paper demonstrates programming techniques that utilize the power of the `DATA STEP`. as well as the `%INCLUDE` statement, `%LET` statement and the macro functions SYMGET and CALL SYMPUT.ln particular, we demonstrate how several examples from the SAS Guide to Macro Processing, Ver sion 5 Edition, can be adapted for use in the minicomputer envi ronments	macro variable, proc step, macro facility, &variable, call symput, proc print, proc means, horizontal bar chart, bar chart, by group processing	source	8
122	Menu-driven Statistical Defect Analysis System	Ed Jurgensen, Tom Lorman, Tom Pineo		proc gplot, raw data, proc step, standard deviation, structured data	source	5
123	Merging Data Sets When Distinct BY Variables Are Not Available	Arthur L. Carpenter	Data sets are typically merged with at least one BY variable. Usually the values of each BY variable are unique in at least one of the data setS. The `MERGE` statement is designed to deal with data sets that contain distinct BY groups. When these BY groups are not distinct or do not exist, BY variables may need to be adjusted or added to one or more of the data sets before the merging process can take place. Consider two data sets, each of which contains observations collected at various locations deSignated by a coordinate pair (X, Y). The resultant merged data set should contain all possible pairs of observations from the two data sets, that are taken within a stated distance of each other. In this case BY variables, which will create the desired data set, do not exist. A set of BY variables must be created that will contain not only information on location but also distance. The problem is compounded when the data sets are large and multiple passes of the data are not feasible	array, sample, drop statement, by group processing, proc step	source	4
124	MEST (Material Estimating) System or How PROC FSEDIT Saved My Tail Again	Tracy A. Cermack		proc step, macro variable, dataset, proc tabulate	source	5
125	Micro-to-Host Link: Features and Future Directions	Cheryl A. Garner, David F. Kolb	This paper discusses the latest enhancements to the micro-to-host link: and some planned enhancements for future releases. Among the newest features implemented in the current PC and host ver sions are catalog download, remote processing of full-screen graph ics procedures, protocol c.onverter support, and batch processing with DM commands	batch processing	source	6
126	Most Creative Use of the Software-Color	Mark O'Brien, Stephen Shepard		macro definition, proc step	source	9
127	Most Creative Use of the Software-Monochromatic	Kernon M. Gibes		proc step, sample, proc genmod	source	5
128	Multiple Concurrent Updates of a SAS Software Data Set: An Alternative to SAS/SHARE' Software	Edward B. Dailey, Jude L. Naes Jr.		proc step, dataset, macro call, macro definition, macro variable	source	5
129	Multivariate Normal Plotting Using Ordered Mahalanobisn Distances	NamJun Kang		correlation, mahalanobis distance, sample, proc step	source	6
130	Multixy: A User-friendly Front-End to the GPLOT Procedure	Hallett German, Thomas J. Higgins	This paper will give an overview of MULTIXY - an ISPF-driven system that generates SAS@ code. The following topics are discussed: 1. The user environment at GTE Laboratories and why there was a need for MULTIXY. 2. A history of the development of MULTIXY. 3. An overview of MULTIXY and some ISPF techniques: system capabilities ISPF review * system design and a typical run of	proc step, sample	source	5
131	New Developments in Pyschometric and Market Research Procedures	Warren F. Kuhfeld, Forrest W. Young	Three new SAS procedures, PROe TRANSREG (regression analysis with nonlinear variable transformations), PROe PRINQUAL (principal components analysis with nonlinear vari able transformations). and PROe CORRESP (correspondence analysis) will be available in a later version of Release 6.03 of `SAS/STAT`"" software. These procedures were developed initially at the University of North Carolina at Chapel Hill under a grant from SAS Institute. This paper discusses the current design of these procedUres	regression analysis, proc corresp, proc step, missing value, response variable, correlation, proc cancorr, conjoint analysis, predictor variable, proc princomp, proc glm, proc freq, data science	source	5
132	New Features in the REG Procedure for Version 6	Sandra D. Schlotzhauer		proc step, proc reg, regression analysis, run, anova	source	6
133	New Measurement Data Sources for MVS/XA and Its Subsystems	H. W. Barry Merrill		cpu time, sample	source	10
134	New SAS Facilities for Design of Experiments	Randall D. Tobias	Recent developments in industrial quality control have resulted in a resurgence of interest in statistical design of experiments. Two new `SAS/QC` software procedures are available for con structing experimental designs. The F ACTEX procedure is a gen eral construction procedure for orthogonally confounded fractional factorial designs. Fractional factorial designs, or designs that are based on them, form a large part of the list of commonly used experimental designs. The `OPTEX` procedure searches for an optimal design from a set of candidate points. This facility can be useful for generating designs in nonstandard situations. Macros and `SAS/AF` menu systems have been devel oped that use these procedures, together with existing SAS facilities for manipulating data, to provide the user with a full com plement of design construction tools	main effect, sample, proc step	source	3
135	On-Line Access to the SAS Sample Library, Usage Notes, and Tutorials in VM/CMS	Gonzalo R. Garriga, William G. LeBlanc, Robert M. Syren	The programs in the SAS Sample Library were placed on a read-only system disk, together with usage notes, tutorials, and several locally developed sample programs. Sample programs, tutorials I and usage notes were stored as individual files. Starting at the level of the CMS system HELP menu, users are able to progress through a series of HELPTASK menus, through the various statistical packages available, down .to the level of a specific sample program or tutorial, or to the usage note file. Users are thus able to quickly and easily access the SAS sample programs, tutorials, or usage notes without needing an extensive knowledge of operating system commands. Any sample program can easily be copied to a write accessible disk, modified, and executed. Usage notes and tutorials can be easily browsed by using the OMS editor's search commands	sample, regression analysis	source	5
136	Overview of the Report Writing Capabilities in Version 6 SAS/FSP Software	Patricia L. Berryman	This paper concentrates on the expanded report writing capabili ties available with Version 6.03 of `SAS/FSP` software. Each SASjFSP procedure is discussed separately and in combination, with emphasis on ways to enhance report writing with the new available features. A discussion of the cut and paste facility and the SAS text editor (available with Base SAS software) and how they complement the features available within the `SAS/FSP` pro cedures is included. Also discussed are the `SAS/FSP` global com mands (for example. SPRINT, FORM NAME, and PRTFILE) that are useful in generating reports. A discussion of effective use of FORMS and ways to use your SASUSER.PROFILE catalog is included	proc step, dataset, sas base, sas log	source	8
137	Panel Discussion: C Programming in the IBM Mainframe Environment	Richard L. Gimarc, Ronald W. Higgin, Carl Kass, Larry Schumacher, Keith Watts	This panel discussion will focus on the use of the `SAS/CN` compiler in the IBM mainframe environment. The panelists are all experienced software developers and have used `SAS/C` in a wide range of applications. The topics to be discussed span the spedrum of software development; from high-level system architectural considerations to interfacing with system-level services provided by the operating system		source	2
138	Paper System Development with SAS Software-Menus, Salad Bars, and Toolboxes	James F. Moore		proc step, proc tabulate	source	4
139	Partial Proportional Odds Models and the LOGIST Procedure	Frank E. Harrell Jr., Bercedis L. Peterson	The LOGIST procedure fits the multiple logistic regression model of Walker and Duncan (1967) to ordinal response variables that take on k+l possible values. In this model, logits of cumulative probabilities are expressed as linear functions of the predictor variables. The resulting regression coefficients are identical for each of the k cumulative probabilities, except for the intercept terms, CXJI j=l, *** ,k, which are ordered to reflect the order of the cumulative probabilities. The assumption of identical regression coefficients, referred to by McCullagh (1980) as proportional odds, implies that the log odds ratio for the relationship between a dichotomized response variable (Y ~j vs. Y <j) and a predictor variable does not depend on j, the point of dichotomization. Since this assumption is frequently inappropriate, the newest version of LOGIST permits the testing and fitting of partial proportional odds models. models which do not require identical log odds ratios	odds ratio, predictor variable, goodness of fit, likelihood function, regression analysis, response variable, logistic regression, logistic distribution, proc step	source	5
140	PBS: An Intelligent Statistical System for Laboratory Scientists	Stan Altan, Thomas F. Devlin	Preclinical laboratory scientists often have a need for rapid statistical analysis of their data. Many of their experiments follow standard protocols and can be classified into a few basic statistical designs e.g. to compare pre- and post-treatment responses, or to compare the effects of two or more treatments. The laboratory scientist's need for rapid evaluation of his data requires him to perform some statistical analyses immediately following the experiment. In this process, he may not test the distributional assumptions of the parti cul ar stati sti cal test he chooses. Failure to meet these assumptions can invalidate the inferences. PBS, an intelligent menu-driven system, was developed at Ortho Pharmaceutical Corporation using `SAS/AF` and the MACRO facil ity to allow 1 aboratory sci enti sts to statistically analyze many of their routine studies	data science, macro facility	source	3
141	Plotting beyond Defaults	Chris Potter		proc step, proc gchart, proc gplot, proc freq, regression analysis	source	9
142	Portability of Large SAS Code across OS Batch, CMS, VMS, and the SAS System for Personal Computers ..	Larry McEvoy	The SASCompanion for the eMS Operating System (1986 Edition) states on page 1, "Generally speaking, SAS programs and their	data step, input statement, proc step, proc freq, proc print, cpu time, error message, array, macro facility, macro definition, options statement	source	6
143	Power Analysis in an Enhanced GLM Procedure: What It Might Look Like	Ralph G. O'Brien, S. Paul Wright		power analysis, proc genmod, sample, confidence interval, sample size, standard deviation, effect size, proc glm, degrees of freedom, proc step, f-test, anova, significance level	source	6
144	Predicting Equipment Failure Using Weibull Analysis and SAS Software	Bruce Kay, Douglas E. Price	Ball bearings, aircraft engines and missile guidance systems have one thing in common: they are all subject to unexpected mechanical failure. Practically every industry is concerned with increased cost~ and loss of productivity and customer satisfaction associated with equipment failure. Although breakdowns are inevitable, it is possible to predict when a failure is likely to occur using statistical techniques.' This paper will investigate a method for estimating failure parameters and how to use these parameters to predict equipment failure. In particular, it will examine Weibull analysis, a statistical technique of wide applicability. Using this technique in conjunction with historical data, we can predict failures more accurately and minimize their impact. Our discussion will center on predicting aircraft engine failures; however, this technique applies to a wide range of applications	weibull distribution, probability distribution, cumulative frequency, likelihood function, proc step, proc lifereg, proc reg	source	6
145	Presenting Capacity Planning Reports to Management Using SAS Software	Albert Noisier	This paper provides an overview of an Online System, "CAPLAN" designed at the Overseas Telecommunications Commission by the Capacity Planning group using SAS Software. This Online system provides Managers, Computer Support, Data Centre, Systems programmers, Information Centre and Capacity Planning with graphs and reports related to the performance of the system, as well as workload analysis and forecast. The whole system runs daily and is fully automated. It generates daily, weekly, monthly and historical graphs. MERRILL's Software collects data from the RHF/SMF and builds Performance Databases. Then using `SAS/BASEC`') `SAS/FSPC`') `SAS/AFC`') and `SAS/GRAPHC`') the information is retrieved and manipulated to create reports and graphs that can be accessed online by authorised users	capacity planning, macro facility, proc step	source	8
146	PROC DICHOT: A SAS Procedure for Creating Dichotomous Variables	Paul Lambert	A SAS Procedure (`PROC` DICHOT) is presented which (a) computes the median for a given set of numeric variables, (b) performs a median split on those variables, determines whether or not the numeric variable is above or below the median and creates a new variable (DICOTn) whose value is either 1 or 0, and (d) outputs the variable(s) to a SAS DATASET. `PROC` DICHOT provides an option which allows processing the data in either sorted or non sorted form as well as an option which allows users to determine the logistic values of scores which are equal to the median. These new dichotomous variables may then be used as input to procedures which perform analysis on categorical variables. Examples of using `PROC` DICHOT are presented along with data regarding time and memory utilization using input SAS datasets with varying numbers of observations	proc step, dataset, numeric variable, logistic distribution, missing value, binary variable, interval variable, out=	source	6
147	PROC TWOSAMPLE A SAS Procedure for the Two-Sample T and Rank-Sum Tests with Extensions	Jon L. Kosanke, Peter C. O'Brien, Kenneth P. Offord	The TWOSAMFLE procedure performs two sample ~ tests as in PRoe TTEST and rank sum tests as in NPARIWAY. It also provides extensions of these two test procedures when the assumption of a linear relationship between the log~dds of group membership and the variable of interest does not hold. Descriptive statistics, plots, tests of assumptions and an option for an output data set of results are available	sample, proc step, class variable, degrees of freedom, homoscedasticity, dataset, out=, heteroscedasticity, confidence interval, proc means	source	6
148	Processing and Analyzing Supermarket Scanner Data	Charles F. Hofacker		proc step, predictor variable, proc means, proc freq, logistic distribution	source	6
149	Processing Large Files: The SUMMARY Procedure and Other Efficiency Tips	Juliana M. Ma		dataset, proc step, class variable, production run, proc means, set statement, sample, calendar year, out=, select, format statement, proc freq	source	9
150	Producing More Efficient Applications	Daniel J. Squillace			source	7
151	Producing Technical Materials with the Aid of SAS Software	Bobbie Wagoner	This paper focuses on the development of course materials at SAS Institute and describes the features of the Education Division's Course Composition System, a system developed at the Institute for the production and maintenance of course materials. The discussion centers on system capabilities that allow SAS programs, output, screen images, and graphics to be included in text, and how these features are provided using IBM's lSPF product, base SAS, `SAS/OMI`, and `SAS/GRAPH` software. The structure of the system is discussed, along with benefits of its use from a time and cost perspective. Although the primary use of the Course Composition System at the Institute is in the development of course materials I the system could be equally useful for preparation of professional papers, presentations, or any type of technical materials	batch job, sas base	source	6
152	Programming in the Large with the SAS/C' Compiler	Keith Watts		proc sql, batch job	source	10
153	Project Schedule Monitoring Featuring SAS/FSP and SAS/OR Software	Robert P. Hardy	This paper discusses the aspects of schedule commitment monitoring in a Project Management environment, with the use of SAS products featuring the data management and display capabilities of the Full Screen (`SAS/FSP`) and Operations Research (`SAS/OR`) products. The concept of schedule commitment monitoring offers an effectlve alternative to the traditional, and often complex, critical path method of network presentation and maintenance	sample, proc step	source	5
154	Read Compressed VSAM Files with SAS Software, an Example	Judy C. Ambler	The development of an efficient and productive way of providing easy access to compressed files using a fourth generation language has always been in the forefront at Pizza Hut Inc. Most of our finan cial system files are quite large and are also com pressed making access impossible by languages such as SAS. The vendor for the financial systems at Pizza Hut is McCormack and Dodge. This paper addresses the Payroll Master and Personnel Master files of the Payroll System	sysrc, sample	source	6
155	Recent Developments in Statistical Graphics	Lawrence H. Cox		proc sgplot, bootstrap	source	2
156	Regional SAS Users Groups: The NESUG Experience	Hallett German		data wrangling, group by	source	4
157	Regression and the SAS System: An Overview	Rudolf J. Freund		proc step, regression analysis, proc reg, anova, estimate, proc glm, response variable, feature selection, proc genmod, predictor variable, polynomial regression, correlation	source	9
158	Repeated Measures Analysis with the SAS System	Ramon C. Littell, Jane Pendergast		repeated measures, anova, covariance, proc step, univariate analysis, proc anova, response variable, correlation, multivariate analysis, f-test, proc genmod, manova, covariance structure, main effect, p-value, degrees of freedom, repeated measures anova, sample	source	7
159	Robust Estimation Using SAS Software	Will A. Sullivan, Liangg Yuh	The M-estimate was first introduced by Huber (1964) and was shown to be a powerful robust method for location and regression parameters. With a specified weight function, a macro program was developed, utilizing FROC UNIVARIATE, `PROC` LAV, and `PROC NLIN`, to automatically obtain M-estimates for the simple linear regression morlel. The one sample location problem can easily be solved as a special case. The maero procedure is shown to be very flexible and easily mortified for a variety of W functions, as well as, more complicated models. In addition, for comparison, the traditional least squares estimate and the least absolute value estimate are listed in the output	proc step, sample, regression analysis, proc nlin, robust estimator, proc reg	source	6
160	Running SAS Jobs in Batch under VAX/VMS	Michael D. Rhoads		batch job, sas log, proc step, sas batch	source	6
161	SAS Functions-Simple but Powerful Techniques	Steve G. Holson, Jenny Putman		date variable, character functions, lookup, sample	source	6
162	SAS Programming - Utilities Using SAS/AF and SAS/FSP Software	W.D. Hutchison, Geoffrey Wheatley	`SAS/AF` and `SAS/FSP` provide the tools to build interactive programming utilities. This paper describes two such utilities. One assists in tt~e management of permanent in-house SAS formats. The other allows users to reposition variables in their SAS datasets	proc step, sas library, macro definition, proc print, dataset, format statement	source	6
163	SAS Software for the UNIX Environment	T. Randall Betancourt, Chip Kelly, John Carl Zeigler	The use of SAS!!" software as an application solution for the UNIXI6t environment is presented in three parts: an overview of UNIX operating systems, a discussion of the fundamentals of UNIX operating systems. and a tutorial on using SAS software in the UNIX environment	proc calis, error message	source	5
164	SAS User-generated Inquiry (SUGINQ)	Barbara E. Lee, Andrew S. Schlegel		macro variable, proc step, proc sort, dataset	source	4
165	SAS/ACCESS'""Interface to IDMS/R	Belinda T. Weinbrecht		dataset, where statement, proc step, variable	source	6
166	SAS/AF Slide System for Graphics	Jack Bulkley	This papex discusses the development of a system for producing word slides for pre sentations. The system uses SAS/ AF, SASjFSP and SASjGRAPH. Much of the discussion of graphics design and user in terface could he applied to other projects		source	3
167	SAS/AF Utility Macros	Karen Crandall	This poster presents three macros to be used with `SAS/AF`: CKVAR, CKDSN. and CKNBR. CKVAR checks that proper SAS nomenclature has been used for naming a variable. This macro is used in situations where the user specifies a variable range (for example, NUMBERI-NUMBER8). The NA¥E attribute finds an error because the variahle. exceeds a length of 8. CKVAR parses the string into two variable names and checks each one for proper syntax. CKDSN checks that proper SAS nomenclature has been used for naming an output dataset. Once verified, the output dataset name is compared to the datasets that exist on the user's A disk. If a dataset of the same name is found, the user must use another dataset name. CKDSN is analogous to the INPUT attribute except it is used to cheek OUTPUT datasets. CKNBR validates a number that has been parsed fwm a string of information. This is analogous to the DATATYP macIO function except it permits the use of signs and decimal points	macro put, %if-%then-%else, macro definition, out=, macro variable, in=	source	3
168	SAS/GRAPH Scatter Maps	Rob Dolan, Sheila Fitzgerald Evans	This paper describes how to use the Annotate facility in `SAS/GRAPH` software to illustrate graphically the density of some geographic distribution function. For example, you may use such an illustration to study a market for the location of a ware house or manufacturing plant. The types of maps represented here are known as the choropleth (Greek choros, meaning place, and plethas, meaning number or quantity) map (Schmid 1983). Choropleth maps display range graded values for geographic areas by shading, coloring, or plac ing symbols on those areas. The areas that may be represented in a scatter map are ZIP code boundaries, counties, precincts, census tracts, states, regions, and entire countries. Counties are the basic graphical area repre sented in this paper and application. However, with the use of the Nationwide 5-Dlgit ZIP Code Boundary File supplied by Geo graphic Data Technology, Inc	proc step, response variable	source	5
169	SAS/GRAPH Software and the Computer Graphics Meta file	David Muck			source	4
170	SAS/GRAPH Software and the IBM 3820 Page Printer	LeRoy Bessler	SASjGRAPH was intenaced with the IBM 3820 Page Printer, using the GDDMFAM4 driver from SAS Institute Inc .* Cary, North Carolina, USA, and the Plot3820 driver from Image Research Corp., Santa Barbara, California, USA. Tests included 112 illustrations from the `SAS/GRAPH` User's Guide. Driver features, reliability, quality, print speed, computer resource requirements, and ease of use are compared. Introductiou IBM 3820 Page Printer The IBM 3820 Page Printer is a black-and-white laser printer that uses standard 8.5-inch by ll-inch cut-sheet paper. It can handle tabular reports, text, graphics, fixed-width fonts, proportional-width fonts, typographic fonts, electronic fonns, scanned image, etc. It supports portrait and landscape formats, and simplex, duplex, and tumble duplex printing. This printer is part of IBM's Advanced Function Printing (AFP) concept, which includes a growing family of sophisticated printers, and a collection of user and system-support software	proc step, cpu time	source	6
171	SAS/IML Software versus IBM/APL as a Scientific Programming Language	Stephanie Breytenbach, Gordon Erens	The relative merits of `SAS/IML` and IBM/APL as scientific pro gramming languages are discussed and illustrated using a num ber of algorithms from current scientific literature. It is argued that `SAS/IML` is sufficiently powerful for most scientific programming problems. Moreover it is easier for both the SAS user and the uninitiated to acquire the necessary level of expertise than is the case with IBMjAPL	proc iml, network data	source	4
172	SAS/QC and SAS/GRAPH Software: An Efficient and Easy Means to Analyze Product Characteristics and Capabilities	Jim R. McHenry	This paper introduces the user to the basic code needed to utilize `SAS/QC` and `SAS/GRAPH` to generate histograms and bar charts to analyze product character istics and process capabilities. The min imal code provides the novice user a better understanding of product and pro ceSses involved in their manufacturing. The code is simple and provides. the basis for more complex- and sophisticated prog ramming and statistical analysis. It is the intent of this paper to -demonstrate with simple code and data the potential wealth of information `SAS/QC` and `SAS/GRAPH` software can provide	histogram, bar chart, proc step	source	6
173	SAS/SHARE' Software: Implementation and Experiences	Janet C. Lind	, With the availability of `SAS/SHARE` software in February 1987, new opportunities for developing information systems became a reality. Although the product initially addresses only the concurrent access and update of "data" type members, it significantly expanded the versatility and interactive potential of SAS-based information systems. This paper will first discuss what is SASjSHARE, what will SASjSHARE do and not do, and how to install/implement `SAS/SHARE`. Secondly it will discuss writing macros to control `SAS/SHARE` server assignment, programming considerations and conversion hints. WHAT IS `SAS/SHARE`? `SAS/SHARE` software provides concurrent acces& and update to data-type members within a SAS data library by means of server. The server is a program which synchronizes the input/output requests for access and update to the members of the SAS data library. There can be multiple servers to improve performance	proc step, macro variable, libname statement, macro definition, %if-%then-%else	source	5
174	Saving Resources with the TABULATE Procedure: The Conversion of Annual Tables from RAMS to the SAS System	Diane Muffett	For the past seven years the Division of Financial Aid at Purdue University has run a standardized set of annual tables to produce demographic information on financial aid recipients. All the change in financial aid policy brought about during the Reagan administration has created an increased interest in the impact of financial aid on the student population. Several schools within the University expressed interest in knowing the types and amount of aid their students received and how the figures compared to those of other schools. As a result, last year the annual tables were broken down by school for the first time. This proved to be CPU expensive. The Division of Financial Aid made a decision to continue to provide this service to the schools so a new approach was desired. The programs had been written in RAHIS	if-then-else, cpu time, proc tabulate	source	6
175	SCOGEN: A System for Generating Rote SAS Code	Mark P. Branagh	A SAS based data management system in a mUltisite clinical trial environment has benefited from the development of a SAS code generator implemented in the dBase 111** programming language. A code gener ator is a program that generates other programs or portions of other programs. It can play an integral role in the auto mation of a production system where a series of similar programs is needed. System input is extracted from existing ASCII word-processing files and a menu driven data entry system. The output of the system consists of a series of ASCII files that include SAS dataset INPUT specifications, variable labels, macro variable assignments, and repetitive SAS macro invocations. Benefits of the system include greatly reduced programmer time spent keypunching rate SAS code, increased data and p~ogram integrity, and more easily maintained programs. The system has also facilitated efforts to work around a SAS compiler capacity limitation	format statement, macro variable, proc format, proc step, data step, macro call, label statement	source	4
176	Sending SAS/GRAPH Output to a Hewlett-Packard LaserJet Printer	Dennis M. Case	This paper describes a prototype `SAS/GRAPH`* driver for the HP LaserJet that was developed at HP's corporate computing Center. The paper also discusses several important issues that were uncovered during that development, including data communications between IBM and the LaserJet, vector-to-raster conversion, and selecting a protocol converter. These issues are presented in a manner designed to assist the reader in creating his own LaserJet driver. Finally, a few commercially available solutions are presented as well		source	5
177	Setting Computer Chargeback Rates Using the SAS System, CICS, and Lotus 1-2-3	Jeanne Buse	The charter of the IS department in a Washington State County govern ment requires it to be entirely self-sup porting by charging its customers for the resources they use. It should show neither a profit nor a loss. Further, one customer may not subsidize another. To accomplish these goals, customer billing is done by a usage based, flat rate chargeback system. Moreover, for the billing system to perform within the extremely stringent criteria laid down in the charter, a rigorous process for setting. chargeback rates is required, and a Rate Determina tion System (RDS) was developed. In 1987, the system was renovated to up grade the technology and to reduce the cost of setting rates. This paper discusses the revised RDS which includes SAS, CICS and Lotus components. It describes the system out puts, inputs, functions and architecture as well as the interfaces between the different components. System Outputs The RDS has three basic outputs		source	5
178	Share and Share Alike	Michael R. Gibson		sas library, libname statement, proc step	source	3
179	So You Think You Know the SAS System? or The Interactive SAS Quiz	Alan B. Howard	The SAS System for Personal Computers implements several features not yet found in mainframe or minicomputer versions: the WINDOW and DISPLAY statements and the Stored Program Facility. An application of the SAS System for Personal Computers was designed and written to take advantage of these unique features. This interactive program uses the WINDOW and DISPLAY statements to quiz a user on his/her knowledge of SAS syntax. The compiled program was saved using the Stored Program Facility. By storing and subsequently loading compiled code, about two minutes of waiting time is avoided to start the application	sas syntax, dataset	source	2
180	Sort Efficiency Using the SAS System Under VM/CMS	Robert L. Mahan	SAS applications heavily consume system resources. Collection and analysis of large data files is CPU and I/O intensive. Using SyncSort CMS with SAS applica tions however, can substantially reduce CPU time, SIO and elapsed time. Further, with new SAS interfaces, SyncSort CMS can decrease resource consumption up to 80%. The new SAS interfaces are: `PROC` SYNCSORT replaces `PROC SORT` within SAS applications. SYNCSORT SAS, (a CMS command), can sort the SAS data prior to invoking the SAS system. These features of SyncSort CMS are here delineated focusing upon user performance benefits. 1Ypical SAS Application Following is a description of a typical SAS application, the sorts within it, and the consumed computer resources	proc step, proc sort	source	3
181	Sorting Massive Data Sets	David S. Rubin		proc step, proc sort, by group processing, sample	source	4
182	SOS: An Expert System Front-end for Analysis of Linear Models in the SAS System	Michael P. Meredith, Robert P. Yerex	This paper covers applications of expert systems to assist users of complex statistical analysis pack ages in the areas of model determination, model statement specification, and pre-diagnostic warnings. Methods and techniques involving knowledge representation, rule based schemata, and user interface are briefly reviewed. Emphasis is upon practical considerations of implementing a working system called Stage One for SAS (or S.O.S.). S.O.S. is an expert system designed to help both new and experienced users construct model and procedure statements for the analysis of linear models. S.O.S. guides users through a consultation oriented interactive system which runs on XT or AT class micro-computers. Necessary information concerning the variables and their use are elicited from the user. The system returns a set of `SAS/STA` T'" program statements suitable for performing an analysis using one of several proce dures (e.g., REG, ANOVA, GLM)	predictor variable, interaction effect	source	5
183	SPAM: Spatial Autocorrelation Modeling	Andrew J. R. Gillespie	Given (1) a set {xlI x21 ..... x n } of n observations on some spatially distributed variable of interest X and (2) a rule to determine which pair {(xi,Xj):i ~ j} are considered to be "neigflbors", then spatial autocorrelation is defined as a condition where the values of individual observations are dependent upon the values of their respective neighbors. An index of autocorrelation is thus dependent not only upon the observations on X, but also on the hypothesis of structure implied by the rule used to define neighbors. This permits researchers to develop and test hypotheses of relationships among spatially distributed data. Such tests should be made before assumptions of spatial autocorrelation are built into a model. A program called SP~, written in SAS Interactive Matrix Language (`SAS/IML`) '., facilitates the development and testing of such hypotheses	sample, confidence interval, randomization	source	5
184	Spatial Pattern Analysis	John M. LaBore	Spatial pattern analysis is used extensively in studies of plant ecosystems. In recent years, animal ethologists (behaviorists) have used spatial pattern analysis in studying distributions of animals over fields and pastures during grazing experiments. The analytical foundation for pattern studies is the application of the Poisson distribution, which provides a useful approximation of the binomial probability distribution. Other measurements, such as the chi-square goodness of fit statistic, variance, mean crowding, index of patchiness, and Morisita's (1959) measure of dispersion can provide .further insight regarding randomness of population distribution. This paper discusses advantages and disadvantages of the different measurements of dispersion, and describes SAS code which computes all of the relevant measurements from quadrat count data	poisson distribution, proc freq, goodness of fit, proc sort, binomial distribution, proc step	source	6
185	SQL and the SAS System Version 6 and Beyond	Lewis Church Jr., Paul Kent	This paper provides an overview of Structured Query Language (SOL) and its implementation in Version 6 of the SAS System. The History of SOL The Relational Data Model, proposed by Codd (1970), represents data in tables. Structured Query Language (Sal) is a language used for accessing data stored in tables. The SAS data set con cept blends very nicely with the concept of a table in the relational data model. Both have columns (variables) and rows (observa tions). SAS data sets are a little more liberal than true relational model tables because they allow duplicate rows and have an inherent ordering. Nevertheless, they are similar enough to make Sal a useful language for accessing SAS data sets. The terms data base table and SAS data set are interchangeable in the con text of this paper. There are many commercial products that support SOL Early Sal -based systems were written for mainframes and minicom puters	where statement, select, proc means, create view, order by, dataset, proc sql	source	4
186	Statistics for Ordered Categorical Data	Thomas O'Gorman, Robert F. Woolson	In this talk we describe a method for combining a measure of association from several 2xC contingency tables. The columns in the 2xC tables are assumed to be ordered and the method uses the weighted least squares methodology of Grizzle, Starmer and Koch (1969). The method uses a transformation of the gamma statistic, a statistic proposed for the analysis of ordered contingency tables by Goodman. The procedure gives an estimate of the overall gamma statistic and a test for association. The `CATMOD` procedure of SAS can be used to calculate the estimate and the test statistic. In this talk we describe an SAS macro which integrates the results of this method into the usual output from the `FREQ` procedure. An example from the health sciences is presented as an illustration	proc step, odds ratio, proc calis, confidence interval, normal distribution, proc freq, sample, proc means, macro definition	source	7
187	Strategies for Shrinking PC Data Sets and Finding More Free Bytes	James P. Young Jr.	Efficient data storage, desired but often ig nored on mainframe computers, has become a necessity for SAS microcomputer users. This paper examines how SAS data and ASCII files become larger than necessary, and provides numerous strategies for more efficient storage	character variable, dataset, proc step	source	6
188	STRATRAN: Stratified Random Selection of Controls for Matched Case-Control Data (A MCSTRAT Preprocessor)	Henry A. Anderson, Jerry Burns, Lawrence P. Hanrahan, Joseph Olson	Epidemiologic case - control studtes associate causal and risk factors to elucidate disease processes. In addition to modelling, matching cases and controls on confounding factors limits their impact on the e~posure - disease relationship. `PROC` MCSTRAT in SAS analyzes matched case - control datil using condi ti Dna 1 1 ogt sti c regress; on (CLR), preserving the matei'd n9 or strCltCl information. STRATRAN facilitates the construction of matched case~control sets for ClR analysi s. Using the FREO procedure. STRATRAN generates tables of cases and controls according to any user specified strata (eg., age sex county, year, marihl status, etc.). Users specify the case~ control matching ratio {1:1, 1:2, 1:3 etc.). STRATRAN then generates N 'WHEN" statements correspollding to the strata to select an e)(act within strah random sample of controls, obtaining the desired matching ratio for each case. The data set is then processed to obtain the input st~ndard for PROe MeSTRAT~	proc step, data step, sample, regression analysis	source	5
189	Structure and Content: Integrity in Relations of SAS Data Sets	Roger D. Aitchison, Robert A. McLaughlin	The compentency of SAS datasets is based on the datasets' structure, content and relationship to other datasets. For the structure of 2 SM data sets to he equivalent every element in the 2 sets must be the same. That is, each data set must be a subset of the other. Data editing for completeness and consistency can be described at 4 levels. Three levels exist within a SAS dataset, the fourth is between data sets. Internal integrity can be assessed with inter, intra variable checks as well as intra observation checks. Inter-dataset logical scans for structural consistency can be facilitated with the use of the contents pcocedure with an output out option. This four tiered approach to data editing is described with emphasis on structural integrity. Examples of the lack of structural integrity commonly encountered are presented. SAS code to perform structural integrity checks is also included	dataset, variable	source	4
190	Subgrouping and Subsetting before Invoking the SHEWHART Procedure	Mark A. Soboslai		proc step, if-then-else, dataset, sas base, sample, raw data, sysdate9, datetime, sas function	source	6
191	Subsetting Data into Groups for Complete Processing within Each Group	Janice D. Callahan, Arthur L. Carpenter	It is often difficult to follow the results for a particular BY group, when a series of SAS@procedures with a common set of BY variables is processed. The output from a series of SAS procedures and data steps are often used to "tell a story", and while the	by group processing, proc step, order by, macro loop	source	5
192	Supporting SAS Software in a VMS Environment	Virginia M. Dineley, Joe Slater	Supporting the SAS software products in a VMS"" environment requires some knowledge about how SAS software and the VMS operating system interact as well as a general kn9wledge of pro gram debugging. This paper will help you support SAS software by discussing the following four areas: the more important interfaces between SAS software and the VMS operating system a debugging guide for tracing problems to the appropriate system area information to gather before calling the SAS Institute T echnleal Support Department to speed problem resolution hints on conserving resources	filename statement, cpu time, sas library, dataset, error message, memory usage, problem solving	source	5
193	SYSTEM 2000 Data Management Software Applications at the National Bureau of Standards	Bob Lee	This paper describes the major SYSTEM 2000-based systems -that have been developed at the National Bureau of Standards (NBS). Three views are presented the Application Systems, SYSTEM 2000 Usage, and the Data Base Administration views	sample, order by	source	6
194	SYSTEM 2000 Data Management Software Update	Ron Hults, Jim Nelson	During the past year, the Institute has spent considerable effort to improve and expand SYSTEM 2000 Data Management Software across the three major hardware systems: IBM , Unisys , and CDC. This effort ranges from new releases to documentation and new procedures. This paper gives an overview of these current and future developments. SAS System Interfaces to SYSTEM 2000 Software The SAS System and SYSTEM 2000 Data Management Software are a great match. The marriage of these two products provides the end user and programmer with the features and benefits of both systems in one integrated package. The SYSTEM 2000 user gets the world's foremost data analysis system, graphics, statistical analysis, decision support, project planning, quality assurance, and' fourth generation application development system. The SAS user gets enhanced data management, including indexing, sophisticated storage techniques, comprehensive security and data recovery features	proc step, macro facility	source	8
195	Teaming SAS/FSP Software and the SAS Macro Language to Allow Data Base Maintenance and Report Writing by Nonprogrammers	Mary Jill Trumper		dataset, call symput	source	8
196	Teflon and Tang: Forging Generic Macros Tools	Peter Kretzman	The SAS programming language arose originally as an ad hoc tool for quick data base subsetting, manipulation, and statistical reporting. Reusability or generic application of code was not really an issue. Now, though, large SAS sys tems are being constructed with tens of thousands of lines of code. Unfortunately. SAS, unlike a language such as C, is not particularly designed for the development of reusable, generic small tools ("black boxes") which serve as system "building blocks." The SAS macro facility can, however, do something to fill this gap if appropriately used. This tutorial outlines a standard idiom and philosophy to use when constructing and documenting useful and usable macro tools. Using this idiom, you can identify and de velop the tools you need as you go about your ordinary programming. In other words, not all of your ad hoc work has to be disposable	macro parameter, macro variable, proc step, dataset, macro definition, macro facility, proc contents, proc print, flat file, macro call, &variable, %if-%then-%else	source	5
197	Test Scoring and Item Analysis Using SAS Software	Ronald P. Cody, Patrice Gregory		proc step, proc tabulate	source	3
198	The Agricultural Chemicals Computerized Information System (ACCIS) at Rohm and Haas Company	J. Koch, George W. Partelow, T.H. Pierce, P. D. Williams	The AeCIS computer system at Rahm and Haas Company is an integrated information system used by the Agricultural Chemicals Group. It consists of many computer programs, files, and databases spanning three computer environments designed to provide the user, either biologist, chemist, or manager, with information on the biological activity -of 'queried chemical compounds	proc step, libname statement	source	5
199	The Analysis of Complete Blocks Using Methods Based on Ranks	Ronald L. Iman		f-test, rank test, proc rank, treatment effect, degrees of freedom, friedman test, anova, residual, f distribution, outlier, main effect, null hypothesis, sample	source	9
200	The Analysis of Covariance for Split-Unit and Repeated-Measures Experiments Using the GLM Procedure (I)	W.T. Federer, Z.D. Feng, Michael P. Meredith, Nancy J. Miles-McDermott	The analysis of covariance for split-unit and repeated measures experiments is not always straightforward when using a statisti cal software package such as the `GLM` procedure in `SAS/STA` T"" software. In order to demonstrate correct analyses. several data sets are examined. and annotated SAS output is given. Hypo thetical data are analyzed first without and then with the covariate included. The whole plot units are arranged in a randomized com plete block design (RCBD), and the covariate is measured on the subplot units. A second example has whole plot units arranged in a completely randomized design (CRD) and the covariate mea sured only on the whole plot. This paper demonstrates the commonly employed approach requiring two separate procedural calls to `PROC GLM` to obtain the sums of squares necessary to construct the proper ANOVA table. Complete ANOVA tables for both examples may be com· puted in a single procedural call to `PROC GLM`	ancova, proc glm	source	6
201	The Analysis of Covariance for Split-Unit and Repeated-Measures Experiments Using the GLM Procedure (II)	W.T. Federer, Michael P. Meredith, Nancy J. Miles-McDermott	This is the second of two papers on the analysis of covariance in split-unit and repeated measures studies. The first paper (Miles-McDermott, et aI., 1988) documented analysis difficulties using SAS `PROC GLM` and illustrated a two-step approach to conduct correct analyses for several examples. A one-step approach is presented in this paper and is illustrated with the same examples. The first exam ple has whole plot experimental units (EUs) in a RCBD and a covariate measured on each subplot. The second example has whole plot EUs arranged in a CRD with a covariate measured on each whole plot. Both Type I and III sums-of-squares are necessary to construct the proper ANOVA table when using the one-step approach which requires a single procedural call to `PROC GLM`. Although the correct ANOVA table may be easily constructed in a single call to GLM, additional work is required for estimation	proc calis, anova, proc genmod, proc step, proc glm, proc anova, repeated measures, residual, covariance	source	6
202	The Application of the SAS Macro and Autocall Facilities to a Customized Exploratory Data Analyzing System	Marc Alary, William A. R. Orban	The system, utilizing macro, AUTOCALL facility and all Basic procedures under version 5.16 of the SAS product, produces customized outputs for statistical analysis. It was designed to do exploratory data analysis and generate reportable outputs from a number of large data sets. The system can be functionally categorized into the following modules: 1) Primer; 2) Storage; 3) Assembler; 4) Processor. The Primer module defines the key and current Symbolic macro variables utilizing the system editor. The Storage module contains and catalogues all DATA and `PROC STEP` information for a specific predefined output. . The Assembler constructs the data set required for processing in either a permanent or temporary form with either standard or arrayed variables. It automatically generates the data set for one of three levels: sample, subsample, and sub-subsample	macro variable, macro call, sample, proc step, class variable, macro definition, data science	source	3
203	The Best of All Possible Worlds: Using the SAS System under MVS, PRIMOS, and MS/DOS	Frederick E. Pratter		cpu time, data wrangling	source	6
204	The CMS SAS System and the IBM 9370 Processor Series	James C. Weathers Ili	This paper gives a brief overview of the IBM 9370 processor series and offers some observations about installation of the SAS System in the eMS environment and its maintenance and performance on these machines	elapsed time, sample	source	8
205	The Comparative Efficiency of Table Look-Up Methods	Leonard S. Lutomski, Henry A. Mullally	This paper provides performance data on the comparative efficiency of four methods in the SAS system for the use of enumeratively defined functions or table lookup procedures. The methods tested are (1) MERGE, (2) formats with PUT, (3) binary search, and (4) golden section search. The simulations are run with three configurations: (1) SAS Version 6.02 on a PC's Limited 80386, (2) SAS Version 6.02 on a Sperry IT 80286, and (3) SAS Version 5.15 on an Amdahl 470N8. Within the implementation constraints imposed by each configuration and	lookup, format statement, proc format, proc step	source	6
206	The Development of a Prototype SAS/AF Software Catalog for User Training	Paul H. Allsing	User training is a sensitive area for Information Center/IS Training person nel. How do you train a new SAS user quickly, yet minimize the frustration often felt by an adult learner? Can you motivate experienced SAS users to work with new technology when the old code still functions? Is there a way to introduce new coding methods without challenging "the way we've ALWAYS done it"? There is no single solution to the training dilemma, because each user community has different needs. But, no matter what mix of CBT, VBT, classroom or one-on-one training is employed, a tool can be developed to help both new and experienced SAS users. The proposed user training tool is a prototype `SAS/M` catalog. Not applica tion specific, the training catalog contains highly commented "generic" `SAS/AF` code. Copied into a user's library during training, it contains examples of functions native to `SAS/AF`, and interfaces to other SAS products	data cleaning, macro variable	source	4
207	The Diagnosis Query System	Gary L. Katsanis		audit trail, if-then-else	source	6
208	The Exact Friedman Test and Multiple Comparison Procedure in Macro Form	George L. Barg, Dale F. Kraemer	This paper discusses the development of a macro to perform the EXACT Friedman test and an associated multiple comparison procedure. Also provided is an overview of the Friedman test and the multiple comparison used. SAS code for the macros is provided	friedman test, proc step, null hypothesis, significance level	source	6
209	The Influence of Economic Variables on Seminary Participation: A SAS/ETS Application	Archie J. Calise, Joseph E. Earley		regression analysis, binary variable, time series analysis, response variable, proc step, residual, proc reg	source	5
210	The Information Center Interface (TICI)	Robert C. Hoering		dataset, batch job	source	6
211	The KID-CARE System: Doctor-Friendly Reports	Ira A. Bader		macro variable, proc step, if-then-else, macro call, subsetting if	source	5
212	The Macropedia: Macros for Listing and Summarizing Data from Clinical Trials	Diane Wold, Paul A. McSorley, Sue Allen, Janna D. Young	The Macropedia, written by members of the Burroughs Wellcome Company Clinical Statistics Department, is a group of macros which handle routine programming tasks. The first ofthree finished macros prints summary statistics, by treatment group, for any number of numeric or categorical variables. The others handle clinical laboratory data. One prints results and attaches flags to values which are above or below normal ranges. The second prints and plots summary statistics by time. The programs allow the user to specify labels and formats as well as variables for analysis, and produce output suitable for inclusion in final reports. This flexibility is achieved by having the user construct a data set which includes formats, labels, and other information about the variables involved in the analysis. The name of this special data set is a parameter of the macro, which transfers the information to macro variables	sample, proc means	source	6
213	The Nuts and Bolts of Course Development	Jean Ussery	Many experienced users are faced with the challenge of developing a course to train others. Although experi enced with the software, these users may be bewil dered by a course development project. This paper describes a structured approach to course develop ment, an approach similar to structured programming familiar to most programmers. Following a course design method,ology helps ensure that the training accomplishes desired outcomes. Courses written without a development strategy may (or may,not) provide the desired training. These courses are frustrating to develop ~ and frustrating to attend. The techniques described here can be applied to the development of any training: a one-hour lesson or a three-day course, training to use a software package or improve writing skil1s, or training developed in different formats such as instructor-led or video-based training	sample, dataset, sas log, raw data	source	7
214	The Organization and Management of a Large Process-Control Data Base Using the SAS System	Terry Mahoney	At Novacorj AGEe. three process-control computers are used to automate the production of a world·scale petrochemical manufacturing facility that consists of two ethylene plants and one polyethylene plant. They have to deal with readings on some 20,000 tags (variables) per hour. An hourly, running record of these measurements is very useful in day to day and longer term planning, engineering and management activities. In order to manage the storage of the large volumes of data involved, this process information is transmitted on an hourly basis, around the clock, to a remote mainframe running MVS and TSO. There it is collected and periodically moved into a set of SAS data libraries, using parameterized SAS macros. These data libraries are part of the Site Production Reporting (SPR) System and arc further organized into a monthly time series that can be readily de-archived	dataset, libname statement	source	6
215	The SAS Institute Zap Library Management System (ZLMS)	James O. Ward	ZLMS is a system developed at SAS Institute Inc. to collect, verify, maintain, and re-distribute "zaps" from a multitude of sources. The purpose of ZLMS is to assure the quality of zaps, to enhance the information regarding zaps, and to improve the process of SAS software maintenance	audit trail, sas base	source	2
216	The SAS Macro Facility	Maria Nicholson, Bob Pulgino	Oneofthe most common questions asked bySAS programmers who are just beginning their introduction to the macro facility is: "Why use macro ... what's it for?". This tutorial attempts to demonstrate the advantages that macro can offer and the motivations that a SAS programmer who uses macro will have. Starting with an example SAS program that uses no macro; features are added which demon strate the common uses of macro and the exceptional power it can give the SAS programmer. 1.2. What Is the Macro Facility? To understand how the macro facility works and how to best use it~ you first must completely understand what it is and how it fits into a conceptual model of the SAS System as a whole. Themacro facility can best be thought of as a separate languagethat works with the standard SASlanguage. Itcontainsmanyofthesame type of programming statements and functions as the SAS `DATA STEP` language	macro variable, macro facility, macro definition, proc step, proc freq, if-then-else, proc print, %if-%then-%else, macro parameter, macro loop	source	7
217	The SAS Supervisor	Donald J. Henderson, Merry G. Rabb		dataset, by group processing, set statement, merge statement, keep statement, drop statement, proc step, do loop, output, execution phase, sas programmer, automatic variable	source	7
218	The SAS System and the Universities	Rudolf J. Freund			source	3
219	The SAS System by Forms: Putting Statistical Power into the Hands of an End-User	Jonathan W. Lee	The SAS System provides a high-level program ming tool for data analysis. However, the flexibility of the SAS language and its numerous options can both intimidate and confuse a casual user. This report dem onstrates how the full power of the SAS System can be integrated, so that end-users with limited training can perform their own analyses without sacrificing flexibili ty. SAS software has been integrated into a menu driven system that can be adapted to an extensible set of SAS procedures in a dynamic data environment; it operates under VMS@ and uses the DIGITAL@ FMS Forms Management System	proc step, macro variable, dataset, macro definition, sas base, label statement, proc gplot, macro facility	source	6
220	The SAS System on the Apple Macintosh's II under DOS: The Best of Both Worlds	Alvin M. Best III		sas base, batch job	source	5
221	The Screen Control Language (SCL) in Version 6 SAS/AF and SAS/FSP Software	Chris Bailey, Yao Chen	Explanations and examples provide the basis of this tutorial that explains the new screen control language available in Version 6 SAS! AF and SASjFSP software. The screen control language provides the ability to develop interactive applications. The functions available in the Sel provide the basis for developing conversational systems for end users. Examples will be used to demonstrate how the language works. Differences in Sel usage in `SAS/FSP` applications and `SAS/AF` applications will also be demonstrated through examples. Similarities and differ ences between the SAS `DATA STEP` and the screen control lan guage will also be emphasized. Contents 1. Requirement analysis 2. Design 3. Comparison vs `DATA STEP` Language 4. Programming Constructs and features 5. Programming Concepts 6. SAS Data Set Interface 7. Examples 8. Future Requirements Analysis For `SAS/FSP` software, one item of more functionality requested on the SASWARE ballot is cross-field validation in `PROC` FSEO IT		source	6
222	The Selling of SAS Software: Toward a Unified Analysis Environment in Basic Research	Mark H. Bradshaw		data science, proc step, dataset	source	2
223	The TVA Waterborne Commerce Economic Simulation Model: Overview and Application	Larry G. Bray, Reba L. Copeland		regression analysis, confidence level	source	4
224	The Use of PROC SICHAID in Marketing Research	Jay Magidson, Patrick Rooney	The capabilities of the SAS system have recently been extended by the availability of the CHAID technique as a SAS procedure. CHAID (CHi Squared Automatic Interaction Detector) is a tree-based-classificatIon technique designed to select a set of categorical independent variables that best predict respqpse on some categorical dependent variable. CHAID is useful for segm~nting COnsumers in market research as well as in general exploratory data analysis. unlike its most related technique, (AID, or Automatic Interaction Detection), CHAID is based on Chi-Squared statistics, is not limited to binary splits and incorporates a Bonferroni correction for collapsing categories. CHAID is contrasted with a "single variable"	credit card, sample, proc step, predictor variable	source	5
225	Time-Series Investigation of Anomalous Thermocouple Responses in a Liquid-Metal-Cooled Reactor	Kenny C. Gross, H. Pete Planchon, J. Poloncsik	A study was undertaken using .sAS software to investigate the origin of anomalous tempera ture measurements recorded by thermocouples (Tes) in an instrumented fuel assembly in a liquid-metal-cooled nuclear reactor. SAS macros that implement univariate and bivariate spectral decomposition techniques were employed to analyze data recorded during 'a series of experi ments conducted at full reactor power. For each experiment, data from physical sensors in the test assembly were digitized at a sampling rate of 2/s and recorded on magnetic tapes for sub sequent interactive processing with CHS SAS	proc step, correlation, proc freq, time series analysis	source	4
226	To Pool or Not to Pool in a Multicenter Clinical Trial	Gregory T. Schwemer	In a two factor (treatment and investigator) clinical trial design, "pooling" usually refers to a model reduction from the full cross classification to a -main effects" model. The removal of the treatment by investigator interaction terms is typically decided by the size of the p-value for the test of no interaction. This decision process historically is due to Type I and Type II error considerations. In clinical trials, possihle bias incurred by model reduction is of prime concern. In this paper, parameterization .corresponding to SAS Type I, Type II, and Type III sums of squares is illustrated and the decision to pool is examined from a bias perspective. An inspection of mean square errors in the estimation of treatment differences suggests that when the F statistic for the interaction hypothesis test exceeds unity, pooling should not be done	treatment effect, clinical trial, sample, sample size, anova, interaction effect	source	4
227	Tools and Techniques to Facilitate End-User Computing in an OS/TSO/ISPF Environment	Dan M. Hess		proc step, macro definition, macro variable	source	6
228	Tools to Organize CMS Files and Beautify SAS Programs	Terry Jares	A file organization scheme for SAS files in CMS and MVS environments and an EXEC to make SAS programs more readable (written in REXX on an IBM/3090) are presented. Both these_ tools are intended to be of use in enhancing programmer productivity	proc step, sas batch	source	2
229	Toronto Board of Education Teacher Allocation System Using SAS Software, EXEC, XEDIT Macros, and COBOL	Avtar Dhanota	TBETAS is a menu-driven system developed un der the CMS environment using SAS software, EXEC, XEDIT macros, and COBOL. The system is used to allocate teachers, based upon various factors, to 163 of the Toronto Board of Edu cation elementary and secondary schools. About five years ago this system was ori ginally developed using IBM's ISPF dialog management facility, EXEC, and COBOL. The old system did not meet the user's require ments. The user wanted multiple access to the data base and a comprehensive on-line locate/search facility. This was possible to provide using SAS softwarei therefore, about two years ago the old system was converted from ISPF to SAS software. This system also uses COBOL to access centrally maintained VSAM tables (table of school related parame ters) to generate reports instead of stand alone SAS "format codes. In this application SAS software also proved very useful to write programs which in turn write other SAS pro grams		source	5
230	Tracking Information Center Activity	Theresa J. Hart, Bela Lahner, Janet S. Thieme		dataset, proc step	source	3
231	Tracking Retail Sales from Product Registration Cards: An Illustrative Example Using SAS/GRAPH Software	Bill Schneider	Estimates of a manufacturer's monthly retail sales can be derived through the tracking of product reg/stratton `CARDS`. The Sales And Inventory Tracking (S.A.I. T.) system is a method by which these estimates can 'be obtained. By Inserting a manufacture dale stamped registration card Inside each packaged product and monitoring both the number and rate of `CARDS` returned, a mathematical model Is developed to forecast future card receipts. These card projections yield card response rate est/mates which, In turn, provide retail sales estimates. The fntent of this paper Is to present an illustrative example of the S.A.l. T. methodology. The presentation materials were developed using SASIGRAPH software. This gives an effective means of highlighting Important aspects ot the S.A.I. T. process and greatly simplifies the task ot describing and explainIng technical concepts of the methodology to a non technIcal audience	distribution channel, proc step	source	5
232	Transferring a Mainframe Data-checking Application to the SAS System for Personal Computers: The Agony and the Ecstasy	Dorothea Collins, Joan Derrico	The emergence of the SAS* System for personal computers generated enthusiasm for setting up the data `PROC`'essing for a new Cooperative Studies Program clinical trial in a microcomputer environment. Part of this involved transferring a generalized VS FORTRAN datachecking routine from a large-scale mainframe to a microcomputer using the SAS System for personal computers under DOS. This program is part of a larger data management procedure used for cooperative clinical trials. It was immediately apparent that the mainframe routines would require major redesigning in order to operate under the SAS System for personal computers and that our approach would require different methods than we would have used for the mainframe SAS System. This paper will discuss the redesign that was necessary to handle the table-driven error checking. An evaluation of the microcomputer versus mainframe SAS System for such an application will be presented	clinical trial, outlier	source	3
233	Transitions From a Mainframe to a Network Environment with the SAS System for Personal Computers	Clyde Dent, Ray Palmer, Eric Yu I. Wang	The purpose of this paper is, to describe the transition from a mainframe SAS system environment to a networked 386 PC enviroment using the SAS~ system for - personal computers. We found that lar~e scale analysis can be performed efflciently and effectively within a 386 PC, IB~ Token-Ring network environment, utilizing various methods of system enhancement and data management and manipulation. Three basic issues are presented: (1) Our network hardware and software confi~rationl (2) accurate and cost efficlent data set transfer methods from mainframe to PC and (3) effective control file templates for handling several large merged and unmerged data sets simultaneously. 'A brief discussion about the type and volume of data that is routinely gathered, managed and analyzed at the Institute for Health Promotion and Disease Prevention Research (IPR) is given	proc step, proc contents, variable list, keep statement, dataset	source	6
234	Two ISPF Tricks	Kenneth R. Whitaker	TRICK 1: Many people have written systems interfacing SAS with ISPF. Most use ISPF to generate SAS code ~sed on user input, then invoke SAS to process the request, then have ISPF display the results. This process is usually repeated for each request. The first trick is a way to avoid invoking SAS for each request. SAS is invoked once for the entire session. It processes the first request, then waits for the next request. TRICK 2: Have you ever to your ISPF session? with the GO option. been in a SAS Display Manager session and wanted to swap back Most people exit SAS to return to ISPF, then invoke SAS again The second trick is a way to call ISPF from within SAS. You can call ISPF from within your SAS session, do what you want to do in ISPF, then exit ISPF to return to your SAS session	sas library, elapsed time	source	6
235	Type I Error Rate for Small Sample Sizes Using Various Options of the CATMOD Procedure	J. Paul Costiloe, Andrew J. Cucchiara, Susan J. Kenny	The caution extended by SAS Institute Inc. to users of their categorical `MODEL` procedure, CATMOD, does not clearly identify the risks involved when analyzing models with small sample size. A SAS program was written and executed to assess these risks by variation of three binomial parameter values under the null hypothesis (P 0=0.1, 0.3, 0.5) with constant sample sizes of 2 to 120. For sample sizes less than twenty, the 4 dimensional sample space for a 23 categorical model was formed and the coordinates of each point used to construct cell frequencies for a categorical table with two factors and one binomial response,. The significance of the interaction term in these tables 'was determined by use of CATMOD. For each table that yielded a X2 value of 3.84 or greater, the binomial probability function of SAS (PROBNML) was used to calculate the probability associated with point in the sample space	sample, sample size, type i error, significance level, null hypothesis, binomial distribution, proc freq, parameter, response variable	source	4
236	UNI-PRO: A System to Screen Large Health Care Data Sets Using the SAS System	William J. McDonald, J. Jon Veloski	Government health insurance programs and private insurance companies routinely screen masses of patient claims data to find clues to what appear to be inefficient use of hospital resources, or care of substandard quality. Using the capabilities of SAS, techniques were developed to analyze data from health care billing systems. The UNI-PRO system uses specifications embedded in SAS FORMAT libraries, tables stored in sequential files, as well as more complex hard-coded criteria. It is possible to locate unusual patterns of care based on patient demographics, disease/procedure (ICD-9-CM) coding and other combinations of data on hospitalizations that help to guide subsequent manual review of detailed medical records. Algorithms using RETAIN statements and `SET` statement pointers make it possible to identify the possibility of premature discharge from hospitals, unexpected readmissions to hospitals and inappropriate transfers among health care providers	sample, set statement	source	6
237	Upgrading User-written Procedures for Version 5 SAS Software: A User's Perspective	Richard L. Gimarc		proc step, sas log, standard deviation	source	8
238	Use of Macros for Error Checking in Research Data Management Systems	Henk Pechler	This paper describes a system of SAS* Macros for checking and editing errors in research data. Representative applica tions involving inter-item consistencies and skip patterns are illustrated. The Macros are contained in a SAS AUTOCALL library and constitute a key component of the File Update System (FUS), a SAS-based data management system developed for the Infant Health and Development program. Advantages of using the Macro Facility in this context are structured modular code resulting in efficient software mainte nance. Potential dangers are compiler overflow problems and SAS programs that are difficult to read	macro facility, dataset	source	6
239	Use of the SAS System in an Undergraduate Course on Statistical Packages at the University of Central Florida	Thian S. Kheoh, Dale F. Kraemer, Julie Wildman-Pepe	In this paper we present our experience in teaching statistical packages including SAS to undergraduates at the University of Central Florida. The goals of the course are teaching the students to perform statistical analyses using statistical packages and teaching the students to interpret the output they obtain correctly. Students are primarily majors in statistics and in computer science and have usually had at least two previous courses in statistics. We discuss the topics covered in our course and discuss some of the feedback received from the studen"ts. The Department of Statistics at the Universi-ty of Central Florida (UeF) teaches a course Computer Processing of Statistical Data in which students are taught to use statistical packages including SAS to perform statistical analyses of data sets	proc step, data science, regression analysis, sample, proc means, proc freq	source	3
240	Use of the SAS System, CMS REXX, and ISPF to Generate Standardized Vital Signs Reports for Clinical Studies	Rong Q. Pan	In the CMS environment. a user-friendly system written in REXX language allows an interface between the SAS~ system and the Interactive System Productivity Facility (ISPF) while creating panel Screens for users to enter information for generating standardized vital signs reports. Based on the type of study design~ users are able to enter specific page titles for-the reports. control input SAS data sets, enter period codes (with nested period codes if they are appropriate for the study). and specify the measurement unit for each vital sign parameter. This system will generate the properly formatted outputs based on the specific information entered by users. This system requires: (1) a SAS format library which contains investigator codes, treatment codes and sex codes, (2) a SAS data set which contains the patient randomization schedule, (3) a SAS data set which contains patients' demographic information, and (4) a SAS data set which contains the vital signs test results	vital signs, dataset, clinical trial, sas base, format statement, proc step	source	6
241	User Surveys: The Key to Customer Satisfaction and Happiness	Sandra Robinson, Sallie R. Teaf		data science, statistical software	source	5
242	User-written Function for Extracting VMS User Process Information	Shearin Bizzell	The SAS System for the VMS'" environment supports usef written functions written in Pl/I, MACRO, and C. This presenta tion provides an introduction to SAS user-written functions, with an example, written in the C programming language, that extracts Job Process Information (JPI) without exiting the SAS System. Jpt includes characteristics of a user process such as user name, account name, privileges, limits and quotas for system resources, and priority. This function serves as a sample to other users interested in writing user-written functions. It is not intended as a complete implementation of the GET JPI service	variable, sas function	source	3
243	User-written Graphics Procedures Using SAS/GRAPH GKS Subroutines	Mona L. Hobby	The Production Department of Exxon USA has moved towards the use of SAS software for solving geologic problems. Geologists' tools are highly graphic in nature and the high volume of data involved for these specialized displays make the use of standard SASjGRAPH procedures inappro priate. To help resolve this problem, SAS Institute has developed an experimental library of GKS-compatible graphics subroutines that interface with SASjGRAPH. These subroutines were used at Exxon to develop SAS procedures that produce geologic displays, including log displays, cross sections, base maps, and contour maps	cross-sectional data, cpu time	source	6
244	Using Base SAS Software to Process Data Residing in a DB2 Data Base	Hank Leddon	,This paper is for those SAS software users who wish to retrieve and process cL1.ta residing in an IBM Database 2 (DB2) data base, but who do not yet have the newly released SASjDB2 software. We will illustrate a method to use in the interim which enables a SAS user to retrieve data from a DB2 data base and then process that data using the SAS language. This paper may be of interest to other SAS users because the mechanics of this procedure can be applied to other situations which might arise where data reside in a data base that can not he easily accessed by base SAS soft ware_	proc sql, sas base, proc step, input statement, by group processing	source	4
245	Using Multivariate Statistical Techniques for Quality Improvement	Lonnie C. Vance		sample, discriminant analysis, covariance	source	5
246	Using SAS Color Graphics for Video Image Analysis	James Borek, Alan Huber		cv, standard deviation	source	3
247	Using SAS Software for Estimating the Probability of Being in Response	Brenda W Gillespie, Charles B. Scott		likelihood function, maximum likelihood estimation, if-then-else	source	5
248	Using SAS Software to Simplify the Display and the Interpretation of Laboratory Data	Kathleen H. Arthur	This paper discusses the SAS MACRO used to support the technique for data analYSis proposed by Sogliero-Gilbert J Mosher J and Zubkoff. Their paper "A Procedure for the Simplification and Assessment of Lab Parameters in Clinical Trials", Drug Information Journal, Vol. 20 (1986), introduced two concepts to simplify the display and the interpretation of lab parameter data: 1) a normalization of each lab parameter value so that it can be interpreted without reference to its normal range and 2) a multivariate scoring system that combines lab parameters into functionally related groups. The SAS code used to carry out the normalization and multivariate scoring will be provided upon request	bar chart, normalization, clinical trial	source	6
249	Using SAS/AF Software to Develop Pre-Clinical Information Systems	Kerry G. Bemis		proc step, dataset	source	5
250	Using SAS/DMI Software for Interactive Coding, Testing, and Debugging of Any SAS DATA Step	James Walker	Interactive development and testing techniques can significantly reduce the time and resources required to produce reliable applications. This tutorial will present concepts and	sysrc, proc step	source	8
251	Using SAS/GRAPH Software to Create Graphics for Data Analysis	Terry D. Allen, Glen Buckner		bar chart, standard deviation	source	6
252	Using SAS/GRAPH Software to Integrate and Analyze Environmental Spatial Data	Ronald W. Matheny, Jay Jacob Wind		proc step, macro facility	source	6
253	Using SAS/OR Project Management to Plan and Manage a Major Electric Generation Facility Outage	Wayne G. Maruska		batch job, data wrangling	source	4
254	Using SAS/SQL-DSO "Software in the Information Center Environment	Thomas C. Edds, Cliff Schell	`SAS/SOL`-OS'" software is a powerful new tool for end-user com puting in the information center environment. A wide range of possibilities is now available to integrate SQl/OS data into exist ing SAS applications. The `SAS/SOL`-OS interface speeds new application development and enables you to take full advantage of the powerful report writing and graphics capabilities of SAS software. This paper discusses the use of this new product, and an example is presented	data wrangling, macro definition, proc sql	source	4
255	Using SAS/STAT Software's REG Procedure to Develop Sales Tax Audit Selection Models	Kirk L. Johnson, Richard W. Kulp		regression analysis, multicollinearity, predictor variable, binary variable, proc reg, outlier, proc step, condition index, feature selection	source	6
256	Using the FSCALC Procedure to Write Reports	Patti Brideson, Ben Cochran	The SAS System provides a variety of methods to generate reports. You can use the `DATA STEP` with PUT statements. Alter natively, you can use the `PRINT` procedure, the MEANS proce dure, or the `TABULATE` procedure. This paper focuses on using the FSCALC procedure in `SAS/FSP` software to generate mean ingful, detailed reports with a minimum of effort	proc step, proc print	source	6
257	Using the GMAP Procedure to Display Geographical Distribution of Surgical Manpower and Population in Oklahoma	Ronald C. Elkins, Mary M. Lane, Cynthia K. Murray	default, the respective values for the variable number of surgeons would have been used. An example is presented which uses `PROC GMAP` to represent a map with two response variables. The first response variable is indicated_ by shading patterns while the second is indicated by block height. The resulting map simultaneously indicates the distribution of board-certified surgeons and population by county in Oklahoma. In addition, maps are produced using data from two points in time for comparison, and also for one surgical subspecialty, thoracic surgery	response variable, proc gmap	source	3
258	Using the IBM PC as a Host Graphics Device	Sarah Darden			source	6
259	Using the Micro-to-Host Link to Access the SAS System under VMS "	Susan M. O'Connor	The SAS micro-ta-host link provides the ability to transfer files between a microcomputer and a host computer and to use local or remote SAS processing. The SAS Guide to the Micro-la-Host Link, Version 6 Edition fully documents this feature of base SAS software for the enhanced 6.02 release of the SAS micro-ta-host link and the TSQ, eMS and VMS"" host operating systems. This paper discusses the SAS micro-ta-host link between a PC SAS Release 6.03 session on a microcomputer using MS-OOS and a SAS Release 5.16 session under the VAXTMjVMS'" operating system. Suggestions are made for integrating the SAS System under VMS and the SAS System for pes for use in SAS applica tion development, or for using the SAS System for PCs as a work station for VMS	dataset, sas log, sample, proc upload, proc step, sas output	source	6
260	Using the Micro-to-Host Link with SAS/GRAPH Software	Roger L Chenoweth		sas log, proc step, proc download	source	4
261	Using the New Features of ANSI C	Mark Gass	The C programming language has evolved since its original defi nition by Kernighan and Ritchie (1978). The proposed ANSI C Language Standard incorporates many new features and defines the language with greater precision. This paper describes and demonstrates how to use the most significant new features. It also explains ANSI terminology so that advanced programmers can study the proposed Standard	if-then-else, decimal	source	10
262	Using the NLIN Procedure to Fit Segmented Regression Models: A Tutorial	George A. Milliken		proc reg, regression analysis, proc nlin, confidence interval	source	10
263	Using the SAS Language-Sensitive Editor	LeAnn P. Weaver	A unique full-screen editing tool was custontized to the SAS System for the VMS operating environment. It was developed using the V AX Langnage Sensitive Editor (LSE) and the VAX Text Processing Utility (TPU). Four attractive features of the editor include a LSE initialization file, langnage templates that contain SAS syntax, the ability to run short SAS programs and look at output quickly, and access to on-line help. The last three features of SAS LSE editor can be used by striking a few pre-defined keys. The features of the SAS LSE editor make it an attractive development tool and an al ternative to the SAS Display Manager, particularly for the VMS operating environment	proc step, sas log	source	5
264	Using the SAS System for Analysis of Dissertation Data	Twila Faye Turpen	Historical research often involves the collection .. editing, manipulation .. and analysis of varied types of data. My dissertation research design was a complicated process that involved the collection, manipulation, and analysis of varied bits and pieces of information from many differing types of archival, census, newspaper, aggregated report, business record, and other historical sources. This design called for a software package that provided tools that could be used to accomplish all these tasks efficiently, easily, as well as inexpensively. The SAS System fulfilled all these requirements for me. Through the use of the `DATA STEP`, I was abte to build many different types of permanent SAse System data sets which could then be SORTed. MERGEd, FSEDITed. to produce final data sets that fit my custom requirements. These data sets then could be used for the analytical phase of the dissertation process. These analyses included `PROC FACTOR`, `PROC`	proc step, proc freq, proc standard, proc univariate, correlation	source	6
265	Using the SAS System for Analysis of Means	Donna O. Fulenwider	This tutorial is designed as a sequel to the presentation entitled 'Application of the Analysis of Means' given by Dr. Peter R. Nel~ son in the Econometrics, Operations Research, and Quality Con trol Section. It includes a very brief review of the concept of analysis of means but focuses primarily on how SAS software can be used to perform analysis of means	sample, sample size, proc step, standard deviation, degrees of freedom, call symput, significance level, macro variable	source	8
266	Using the SAS System to Assess Local Influence in Regression Analysis with Censored Data	Luis A. Escobar, William Q. Meeker Jr.	This paper describes a SAS" macro for the assessing local in fluence in regression analysis with censored data. The macro allows one to identify those observations that have the great est local influence on the estimates of the parameters or func tions of the parameters of one's model	regression analysis, likelihood function, influential observation, predictor variable, macro call, goodness of fit, censoring, macro variable, right censoring	source	6
267	Using the SAS System to Facilitate Clinical Trials Research and NDA Approval	Martin J. Rosenberg	The pharmaceutical indust-ry, in conjunc tion with the Food and Drug Administra tion, has been experimenting with ways of using computer technology to facilitate and accelerate the process of introducing new drugs to the market. Some of these experimental approaches include "electronic 1t NOAs (new drug applica tions), remote data entry, and computer assisted monitoring of clinical trials. The SAS system has long been the standard software used in the pharmaceutical industry for performing statistical anal yses on clinical trials data. However, the SAS system has capabilities that go far beyond those of a statistical pack age, which make it an excellent choice for these new applications. Since most pharmaceutical companies already have considerable in-house expertise, the use of the SAS system is a particularly attractive alternative. This paper dis cusses these newer, nontraditional uses of the SAS System in the clinic_al research and NDA approval processes	clinical data, clinical trial, dataset, crf, proc step, data science, data review, sample	source	7
268	Using the SAS System to Meet Business Processing Needs in a Mainframe Environment	Lynn M. Huckert, Mary G. Pulling, George P. Riner		dataset, data validation rules	source	6
269	Using the SAS/GRAPH Template Facility to Rotate Multiple Plots per Page in Batch Computing	Patricia L. Gerend, Richard E. Raftery	Combining different graphics output on one page can aid in. comparisons and comprehension. Similarly, display of graphics output in verti cal, or portrait, style can facilitate viewing and emphasize a point. These two concepts can be combined to generate multiple plots per BY variable on the same page through creative use of the `SAS/GRAPH` template facility in `PROC GREPLAY`. The template can be called numerous times to produce such graphics output via batch computing for all values of the BY variable	proc step, proc greplay, by group processing, proc gplot	source	3
270	Using the TABULATE Procedure	Holly Whittle		proc tabulate, proc step, class variable, format statement, proc freq, proc format, proc print	source	5
271	Utilizing SAS Software to Generate Command Language Programs: Using CA-TELLAGRAF as an Example	Sue Huang		macro variable, proc calis, proc freq	source	5
272	Version 6 Macro Features for the PC	Jeff McDermott, Bruce M. Tindall		macro variable, symbolgen, if-then-else, %if-%then-%else, macro definition, mprint, macro facility, macro function, proc step, sas log, sysrc, macro quoting function, proc compare, proc print, macro flow control, macro loop	source	6
273	Version 6 SAS Data Base System Architecture: Current and Future Features	Stephen M. Beatrous, William D. Clifford	This paper is divided into two parts. The first part is an overview of the Multiple Engine Architecture (MEA). The MEA allows the SAS System to process any file type as if it were a SAS data set. The MEA is described in terms of its inner workings, its advantages, and its end-user interface. The second part of this paper describes some advanced data base features of the portable base SAS software I/O engine. It focuses on the index processor by explaining the architecture of an index, special features of the index processor, and how the engine uses indexes. Not all of the features described in this paper are available in all Version 6 releases	dataset, where statement, sas base, flat file, index, proc step, missing value, proc copy, proc sql	source	10
274	Version 6 SAS/AF Software Enhancements for CBT	John C. Boling, Susan Strozier		if-then-else, proc step	source	6
275	Version 6 SAS/GRAPH Support for National Characters	Douglas Walker	Version 6 `SAS/GRAPH` software offers an intuitive and flexible approach to national character display. It has direct support for French, English, German, Italian, Spanish, Danish, Norwegian, Swedish, Finnish, Japanese. and other languages. This paper discusses the problems with and methods of producing national characters on hardcopy and CRT devices	proc step	source	4
276	We-R-Drugs Pharmaceuticals Corporation: SAS Programming Written by Macros	Ira A. Bader		sas base, macro variable, proc step	source	5
277	Weighted Statistical Measures for Grouped Data	Nancy Stevens		median, standard deviation, proc step	source	6
278	Whatever Happened to the Old SAS Data Library?	Erika Mares Stone		libname statement, dataset, proc step, catalog, proc contents	source	8
279	Windows without Pains	Ross Merlin		sample, character variable	source	4
280	Workload and Revenue Projection for Computer Center Cost Recovery Systems	Tim Carrico, Nada Harris	This paper discusses the financial modelling and rate-setting used at a large government computer center. The discussion includes how to build a workable financial model and the selection of data sources for the forecasting process. Several SASjETS* procedures were used to predict workload growth for a six months period, and the results were tracked against actual usage. Guidelines and recommendations for forecasters based on this study are included	batch job, time series analysis, proc reg, proc step, data wrangling, regression analysis	source	9
281	Writing Macros with the PARMBUFF Option	Anthony B. Kreamer	This paper illustrates the use of the PARt.4BUFF option in calculating the number of parameter values (N) supplied in a macro coil. Once N is calculated, a predefined global macro variable is defined for each respective value of N. The primary advantage of u!ling this macro with the PARMBUFF option is that 0 user need only supply the appropriate porameters. The macro Scanner calculates the value of N and subsequently invokes the user's macro via the 7.&VAR statement. ItmlOIJUCTION When invoking Q macro, parameter values ore supplied to the macro by use of positional or keyword parameters. For some tasks, the number of parameter values may vary and cannot be predicted prior to invocation. Moreover, certain tasks may need to be performed an unknown number of times (N). These problems can be alleviated by using an iterative 00 loop. However when a large number of parameter volues need to be supplied to 0 macro, iterative DO loops become tedious and inefficient	keyword parameter, %macro parameters, title statement, do loop, macro variable, macro parameter	source	2
282	Writing Portable C Code: The SAS Institute Experience	Richard D. Langston	This paper illustrates how SAS Institute Inc. went about converting a large software system from PL/I and IBM 370 assembly language into a new ver sion written almost entirely in C. The paper briefly discusses the history of so~tware development at the Institute, then explains the rationale behind the company's decision to convert to a new language. From there, the Institute's experience in writing portable software is discussed. In addition, the pa per touches on the development process the Institute used, including the adaptation of software coding standards. Let us begin by tracing the history of program ming language usage at the Institute. SAS soft ware was originally implemented for OS-type oper ating systems (MVT, MFT, MVS, VSl). The SAS System was implemented using a combination of PL/I and assembly language		source	6

Back to Conf go 1989 go 1987

This site serves as a convenient hosting location and does not work as a publisher. All copyrights belong to their respective owners.

The information on this site is provided for general reference purposes only. Whilst every effort is made to ensure that the site is up to date and accurate, no warrant, nor any responsibility or liability for, the accuracy or completeness of the content or for any loss which may arise from reliance on information contained in this site is provided. Links to other sites are provided for your convenience but no responsibility or liability for the content of those sites or of any external site which links to this site is accepted. References to any specific commercial product, process, service, or corporation are also for information only, and do not constitute endorsement, recommendation or favoring.

SAS is a registered trademark of SAS Institute Inc. SAS, Statistical Analysis System, and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration.