SUGI 1982 Papers - Index

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

#	Title ×	Authors ×	Abstract ×	Keywords ×	Extras	Pages
1	"The Top Ten ": Frequently Encountered Errors of Beginning SAS Programmers	Debby Vivari		proc step, data step, proc freq, sas programmer	source	3
2	A Comparison of Procedures for the Classification of Multivariate Observations	Cathryn Gust, Rebecca Klemm	Two multivariate statistical models probit and discriminant analyses, are compared as tech niques for two-group classification. The stan dard classification technique, linear dis criminant analysis, assumes the complementary groups are distributed multivariate normal with equal variance-covariance matrices. However, probit analysis assumes nothing about the structures of the variance-covariance matrices ar.d requires only that a linear combination of the explanatory variables be normally distributed. Utilizing the Statistir..:al Analysis System, these two appraches are compared on the basis of their misclassifi cation rates utilizing various actual and simulated situations	discriminant analysis, sample, predictor variable, covariance, interest rate, homoscedasticity, regression analysis	source	4
3	A Computer Graphics Information System for Upper Management	Donald R. Chapman, C. Royce Claytor	A primary requirement of upper management is to be informed quickly and concisely. Executives at Virginia Electric and Power Company felt that this need could be partially satisfied by employing the recent color graphics technology. They requested a graphics system by which they could gain new perspectives from graphic presentation of data. In response to this request, we developed the Management Information Facility (MIF). MIF is an interactive tool that provides a method for integrating several graphical techniques and information from many departments into a single display system that can provide decision-making information to upper management. `SAS/GRAPH` was chosen as the primary software tool to be used by each department for composition of special ized graphs because it allowed access to existing computer files as well as flexibility in prel::lenting graphical data	management reporting, bar chart, proc step	source	6
4	A Data Mangement and Forecasting System for the Non-Data Processing User..	Janet S. Bridgers, Kenneth A. Sprinkle	This data management and forec~sting system ;~ a customized interactive financlal system WhlCh supports data management for approximately 3000 construction projects funded by the Veterans Administration. Because none of the users are data processing personnel, the syst~m is packa~ed and driven with menus and user_-def1ned selectlDn parameters coded as CMS EXECS.. :hese EXE~S communicate with the user in appllcatlon-speclflc terminology. The EXECS translate user requests to runnable SAS code by writing SAS macros. In some cases EXECS are bypassed, and SAS programs conduct a user dialogue directly. This system tracks actual project construction costs by month; it also provldes deta11ed ~ost forecasts by month for the 1 ife of each proJect using base prof; Ie curves~ ~djust.ing for ~ctu~l costs incurred. In add1t10n, 1f a proJect s planned costs are known to deviate ~rom the standard prof; 1 e, its forecast may be adJuste~ to properly reflect any anomalies	proc step, macro definition, sensitivity analysis, data cleaning	source	6
5	A General Linear Models Approach for Comparing the Response of Several Species in Acute Toxicity Tests	Karen L. Daniels, Michael P. Farrell, Jonathan C. Goyert, Rodney H. Strand	Acute to~icity tests (bioassays) estimate the concentration of a chemical required to produce a response (usually death) in fifty percent of a population <the lCSO). Simple conparisons of lC50 vaLues among several species are often inadequate because species can have identical LCSJ values while their overall response to a chemical may differ in either the threshoLd concentration (interce~t) or the rate of response (slope). A sequentiaL approach using a generaL linear model is presented for testing oifferences among species in their overall response to a chemical. This method tests for equality of slopes followed by a test for equality of regression lines. This procedure ~mploys the Statistical Analysis System's General Linear Models procedUre for conductirg a weighted least squares anal)sis ~ith a covariable. *Research sponsored by the Office of rlealth and EnvironmentaL Research, U.S. Department of Energy, under contract ~-74r5-eng-26 with tJnion Carbide Corporation. Publication No	regression analysis, proc glm, logistic distribution, data science, proc model, response variable	source	6
6	A General Purpose TSO Interface to SAS Using SPF	John L. Dickinson		data science, sas output	source
7	A Machine-Readable Master Index to the 1976-1981 SUGI Proceedings	Arnold W. Bragg Jr.			source	1
8	A Pharmacokinetic Analysis System (PAS)	Raymond P. Chen, Daniel M. Wildman		confidence interval, proc reg, regression analysis, data science, model training, estimate, pk parameter, adpp, output delivery system	source
9	A Proposal for the Contents of a Project Archive	Ronald W. Helms, James D. Hosking		data cleaning, dataset, sample, data wrangling, proc step	source	8
10	A SAS Based Management Information Data Base	Wayne B. Finley, William C. Pannell		data wrangling, dataset	source
11	A SAS Fundamentals Course for Novice Users	Joe Bellew, Roger B. Glaser, Clark M. Roberts	A prevalent challenge for many of today's industrial computer facilities in satisfying the support, service and expertise levels demanded by the user. The programming staff is often back logged and engaged in long-term projects, where as the user is typically confronted with short term computer applications. This paper addresses the successful approach taken by the San Diego Gas & Electric Company (SDG&E) in responding to the needs of its computer user community. SAS (~tatistical !nalysis ~stem) was adopted as one of the major SDG&E user languages, principally because of its breadth of applicability. A SAS fundamentals course, designed for non-programmers, has been offered to SDG&E employees in support of their growing needs for computer sophistication. The SDG&E SAS fundamentals course is inte grated with a prerequisite course in TSO and SPF. Our SAS course is comprised of ten lecture and workshop periods which cover a five-week interval	sample, proc print	source	8
12	A SAS MACRO for a Generalized Friedman Test	Mary S. Hirsch	The Friedman (1937) Test for a randomized complete block design has been generalized by K.M. Patel (975), along the lines of Gehan (965) and Breslow (970), for a randomized block design when observations are subject to arbitrary right censorship. The distribution of the censoring variable may vary among the blocks. Patel (1975) has shown the asymptotic null distribution of the test statistic to be chisquare. A SAS macro for this nonparametric test is presented. (Keywords: Macro, Friedman, Nonparametric, Censored)	macro definition, friedman test	source	2
13	A SAS Profitability Analysis System	Robert A. Jones		proc step, dataset	source
14	An Algorithm for Changing a Matrix in Correlation Form into a Positive Definite Matrix with a Certain Condition Number	George A. Milliken	When information from different sources is used to construct a matrix of correlation co efficients, the matrix may not be positive def inite. If it is not positive definite it is not a matrix in correlation form. A technique is described which will convert the matrix to correlation form by shrinking the correlation cDefficients. The method is such that the resulting matrix can be chosen so that it has a specified condition number. A weighting scheme is also described so that correlation coeffic ients which are known with more confid'ence will be shrunk less than those with less confidence	correlation, condition index, eigenvalue	source	5
15	An Information and Tracking System for Surveys	Carl B. Backman, Cathryn Gust		sample, sas base	source	3
16	An Integration of System 2000 and Statistical Analysis System (SAS)	Roger E. Booker	This paper describes two software systems. SYSTEM 2000 and SAS*. and the benefits that a "marriage ll between the two can bring to a research environ ment. The I'marriage l' in this case is provided by a SAS procedure, called FROG S2KSAS, that can be used to select data from any SYSTEM 2000 data base and transform it inlo a SAS data set. The research environment is the National Water Data Storage and Retrieval System (WATSTORE) that is maintained by the U.S. Geological Survey.lJ WATSTORE is a large (2.5 gigabyte) data base, representing the primary source of machine readable information about the quantity, quality, and availability of the Nation's water resources. It contains eight major subsystems, two of Which employ SYSTEM 2000 as the Data Base Management System (DBMS). The examples shown here were extracted from the Ground Water Site Inventory System (GWSI). a subsystem of WATSTORE	dataset, where statement, proc step, output, missing value, data standard, format statement, data science	source	7
17	Analysis of Effects on Variances Using GLM	Robert M. Hamer, Barry H. Schwab	This paper discusses a new technique for performing ADS, as well as a set of SAS `PROC` MATRIX statements which transforms data so that ADS can be performed using `PROC GLM` (or `PROC ANOVA`). This technique was developed by O'Brien (1979, 1981), and some of the brief literature review as well as the mathematics are taken directly from his work. Traditional tests for performing hypothesis testing on cell variances (e.9., F =Sl 2/S/ Bartlett's test, F max, etc.) have many diffi culties. Some are usuable only for testing two group hypotheses, some are overly sensitive (will reject Ho: when very small differences are present) and nearly all are sensitive to viola tions of the distributional assumptions. Thus, in cases where the distributional assumptions cannot be met, such tests may do more harm than good	anova, sample, proc step	source	5
18	Applications of SAS (FUNCAT) to the Analysis of Epidemiological/Survivorship Data	Saeid B. Amini, William R. Clarke, Robert F. Woolson	The statistical literature of recent years has included numerous advances which touch directly upon the analysis of epidemiological. e.g., case-control or follow-up, data. Many of the procedures are pertinent to categorical data involving linear, log-linear or more complicated functions of the various response categories. In this paper we describe the applicability of FUNCAT to a number of these problems. The paper is developed through three examples of case control data and one example of a follow-up study. The utility of FUNCAT is evident in each of these problems, and by logical exLension Lo a variety of others. Previously published data are utilized as illustrations	likelihood function, odds ratio, degrees of freedom, maximum likelihood estimation, estimate, chi-square test, regression analysis, logistic distribution, data science	source	9
19	Approaches to Estimating Covariance Matrices with Partial Multinomial Data	Mark W. Riggs, W.B. Smith		correlation, proc cancorr, sample, covariance, likelihood function, maximum likelihood estimation, missing value, normal distribution, anova, regression analysis, monte carlo, multivariate analysis	source	10
20	Approaches to Estimation and Testing in the Mixed Model	R. Kirk Steinhorst		anova, fixed effects, proc mixed, main effect, covariance	source
21	Architecture Alternatives for SAS/CPE Systems	Arturo G. Maria		capacity planning, if-then-else	source	4
22	Arrays	Peter L. Rikard		do over, end	source	3
23	Articulation Research System: A Tool for Community College Curriculum Evaluation	Russell Gilbert	Recognizing the need to evaluate the community college graduate's general education preparation for upper division course work in the State University Sys tem of Florida, Santa Fe Community 001- lege developed the Articulation Research System. The system 1s currently being used by seven of the 28 public community colleges in Florida. In determining the user requirements, it was decided that the system needed to (1) follow former community college stu dents through their bachelor degree pro grams at the State Universities, (2) ~e able to compare student performance 1D community college coursew6rk with relat ed upper division coursework, (3) be sufficiently generalized to be able to be used by other community colleges, (4) be flexible enough to accommodate the specific student record formats at each individual community college and (5) be friendly enough to be used by curriculum evaluation administrators at the commu nity colleges	dataset, batch processing	source	6
24	ASPP: Automated SAS Program Producer	Theresa J. Hart, David A. Pyne		dataset, data dictionary, raw data, sample	source	4
25	Basic Time Series Tools for Business Forecasting	Mary Sue Dickerson, Franklin S. Young		time series analysis, outlier, seasonal adjustment, regression analysis	source	6
26	BCAS - A USER FRIENDLY INFORMA'IION SYSTEft INCORPORATING SAS AND RAns	Arthur Bocian		proc step, dataset, correlation	source	5
27	Biplotsin SAS	Michael C. Tsianco		sample, data visualization, proc freq	source
28	BISARS Revisited	Erich Doubek, Edward T. Graham		dataset, batch job	source	6
29	CAI: Computer Assisted Instruction for New SAS Users '82	Kirk Paul Lafler, Clayton Stewart	Having the proper and adequate training for programmer/analysts is an essential and necessary objective every data pro cessing manager wishes to achieve. But we all know that it is not always possible to give that training due to project constraints, funding, etc. The end result is clear, the whole project suffers due to improperly trained programmers. Recognizing this need, we have developed a computer assisted instruction package (CAl) to train new SAS users in the syntax and structure of SAS. This paper discusses the method of design, module descriptions, and gives a course out- line used in developing the `SAS/CAI`	data wrangling	source	4
30	Can SAS Replace Traditional Programming Languages in Large Systems Development-A Manager's Perspective	J. Harvey Trimble Jr.		data transformation, structured data	source	3
31	Capacity Planning of Document Processing Equipment	Leslie Martin	The high cost and long lead times of acquiring specialh:ed documenL prucessing equipment in Lhe Cunsumer Financial Services DivisiDn of American Express demonstrated the need for a Capacity Planning System. The system was de signed to aid management in anticipating changes in demand for capacity and in planning for ad ditional needs before they arise. Forecasts of future patterns of equipment utilization have been developed by combining: o knowledge of hi~tori~~l trends, o forecasts of basic business indicators, and o notices of planned changes in operations	capacity planning, capacity utilization	source	3
32	Capacity Planning Using CapturelMVS., BEST. and SAS	Robert Ashton, Jeffrey Buzen, Leonard Lipner		capacity planning, data cleaning	source	4
33	Capacity Planning with a Single Workload Model	Henry Steinhauer III			source	8
34	Cartographic Evaluation of Environmental Management Strategies	Karen L. Daniels, Michael P. Farrell, Jonathan C. Goyert, Rodney H. Strand		output, proc gmap	source
35	Categorical Data Analysis Using the FUNCAT Procedure: An Example	Charles S. Davis	The Grizzle-Starmer-Koch (GSK) approach to the analysis of data from complex contingency tables Is readily accessible through the FUNCAT procedure. One of the examples of the use of this procedure provided In the SAS User's GuIde, 1979 Edition Involves the Ries-Smlth detergent preference data. This wei I-known data set was also analyzed using the GSK methodology by Johnson and Koch (1971). The relatively compl 1- eated analyses presented by these authors Incor porated design matrices and response functions which, In some cases, are not directly ava! lable In FUNCAT. The purpose of this tutorial 15 to demonstrate how the .Johnson-Koch analyses can be performed using SASe The practical detailS of this example I I lustrate both standard and non standard appl (cations of the FUNCAT procedure	proc step, degrees of freedom	source	7
36	Cluster Analysis by Least Means	Warren S. Sarle		cluster analysis, factor analysis	source	3
37	Comparison of Procedures for Exploratory Data Analysis	Randall S. Miller	The purpose of thIs paper is to document the SPLOl' Procedure and compare it with the SAS UNI VARIATE Procedure for the purpose of exploratory data analysis. As researchers, we may feel quite over whelmed when we first look at a group of data. Frequently, In our need to get an understanding of the data, we start with techniques such as mean, varian-ce and standard deviation. Unfor tunately, though, these are sUDlIllary techniques that apply to normal distributions. The prob lem with starting with these techniques is t:hat: they are sensitive to outliers. (It is for this reason that we use median age instead of mean age of a group; and median income instead of mean income of a group.) Thus, it 'is best to start with order sta tistics, for they assume no underlying normal distribution. Median, mode, interquartile range, and percentile ranks are order statistics. They are called such because they all require the data to be ordered, or ranked	proc univariate, box plot, proc step, proc freq	source	5
38	Computer Graphics in Selection of Screening Strategies	John K. Gohagan, Edward L. Spitznagel	Screening for a disease such as breast cancer with multiple test modali~ies (palpation, mammography, and thermography) will ordinarily have several distinct strategies optimal over the typical range of risk profiles and costs. Computer graphics haa played an important role in viewing the complexities of t~e problem and in making simplified screening recommendations	type ii error, proc freq	source	4
39	Computer Performance Reporting at Merck	Philip T. Newson		batch processing, proc step	source	6
40	Conditional Execution of Data and Proc Steps	Donald J. Henderson		proc step, by group processing	source
41	Constructing an Index from Hardcopy Documents Using SAS	Bob Sturgeon		proc sort, proc step	source
42	CPE's New Panacea: The SAS-based PMS	Shirley Linde, Sheryl Preston	Performance management systems (PMS) are an important tool in computer facility management and control. This paper presents a generalized approach to the design of a PMS and, as an example, the PMS designed by the MITRE Corporation for the Spacecraft Software D1v1s10n of NASA at Johnson Space Center	dataset, quality assurance	source	3
43	Curriculum Approach to SAS Training	Herbert J. Kirk		dataset, data wrangling, time series analysis, anova, regression analysis	source	3
44	Data Management in Clinical Trials Using VM/SP, CMS, and SAS	David I. Hopp, Robert W. Staley		data dictionary, proc format, array, clinical trial, proc step, data governance	source
45	Editing and Reformatting Character Data: Two New Functions for SAS	Stephen M. Beatrous		compress function, length statement	source	4
46	Effective Applications of SAS Procedures to Facilitate Transportation Planning	Howard McKamey, Robert E. Stammer Jr.		bar chart, horizontal bar chart, regression analysis, class variable, proc freq	source	5
47	Efficient Statistical Consulting in the Small University Environment	David M. Smith		time series analysis, sample	source	3
48	Enhancements in ALSCAL-82	Forrest W. Young	This paper presents the enhancements in ALSCAL-82, the 1982 version of the ALSCAL multi dimensional scaling procedure. The enhancements fall into three general categories: The range of scaling models has been extended by introducing two options to define dis tances according to the Generalized Euclidean Metric (GEM). The options are MODEL:GEMSCAL and DIRECTIONS=t. GEM is an individual differences model with many specific special cases. Specific GEM models include, among others: Carroll's IDIOSCAL model and Harshman's PARAFAC model, a model which assumes that each individual weights the dimensions of an orthogonal rotation of the stimulus space, each individual having a different rotation and weighting scheme. * Tucker's three-mode scaling model, a model which assumes that each individual weights the dimensions of the stimulus space, but that these dimensions are not orthogonal. The obliqueness of these dimensions varies across subjects	normalization, missing value, proc step	source	10
49	Environmetrics of Synfuels: IV. Project Results Tracking System (PRIS)	Karen L. Daniels, Jonathan C. Goyert, Rodney H. Strand, Michael P. Farrell	An ~ssessment of the environmental and 'Iealth effects of a low-BTU coal gasification facility on the University of Minnesota-Duluth campus began in 1978. The SAmpling program generated over 2.S million records characterizing, documenting, and estimating various environmental and process-related parameters. A Project Ppsults Tracking System (PRTS) was designed and implenented to account for and summarize the status of all s8"ples taken during the project. This paper presents the	sample, quality control, data science	source	9
50	Estimation of a Translog Cost Function Using SAS	George M. McCollister		estimate, price elasticity	source	3
51	Evaluating Linear Model Representations of Cubic Splines Using PROC REG	Thomas P. Capizzi, Hina Mehta		regression analysis, proc reg, proc step	source	6
52	Four Topics in Statistics	Rudolf J. Freund		proc step, outlier, anova, regression analysis, proc reg, proc genmod, proc glm, cook's distance, predictor variable, covariance, main effect	source	8
53	Full-Screen, Menu Directed Information Systems for Planning, Analysis and Control	Leonard J. Reingold, Jack Siler		data cleaning, proc step, dataset	source	6
54	Generating Combined Literature Indices	William M. Taylor, Barbara Taylor		index, dataset	source	5
55	GENSYS: A Generalized Batch Sequential Update System with SAS as a Foundation	Barry Masuda, Fred Yost		data dictionary, dataset, master data management	source	5
56	Geographic Display of a Multi-location Telecommunications Network	Gary C. Edgar	`SAS/GRAPH` offers `PROC GMAP` for producing choropleth (magnitude of re sponse within area) and surface (obi ique per spec tive view) maps. Provided that the unit area of interest is ava Hable. (Id. variable ; STATE. COUNTY or PROVINCF.), mapping a response data set to a SAS map is relatively straightforward. When the mapping of unique coordinates within a state or county is desired, you must create your own map data set. This paper discusses the creation of a telecolllUunicatioos network map data set, concatenation with a customized SAS STATES data set and projection. Illustrative examples of point-to-point mapping Bnd design methodology are presented	proc step, proc gmap, dataset	source	7
57	Graphic and Semi-graphic Displays in a Classroom Utilization Report: Rhetorical Considerations and Sample SAS Code	John D. Smith	A large univerSity's schedule of classes is described. A methodology proposed by Rockart was helpful in examining information needs regarding clas~room schedulIng of plannl:!rs, deans, and administrators at the University of Colorado and designing a classroom utilization report. Graphic and semi-graphic displays derived from Tukey's exploratory data analysis were used to involve information users in the design/pl:!rceptual proces~ as well as to provide more useful graphics. use a line printer and SAS code is included: The displays	box plot, proc step	source	6
58	Graphical Display of Scatter Data Using the Standard Deviation Ellipse	Charles G. Margolin, Charles E. Shipp	A new procedure based on classical equations is used to graphically characterize large volumes of coordinate points without drawing them. The construct, called the Standard Deviation Ellipse, is described and derived, and examples show how it is easier to recognize subtle differences in the data and how it makes presentation of the data more esthetically pleasing. A SAS procedure is described that can be used with SAS Proe PLOT or with `SAS/GRAPH` and two examples are given: one from CADAM statistics and one from professional sports showing height versus weight for various sports	standard deviation, proc step, correlation	source	6
59	Graphics Can Tell Your Story, Too	Richard A. Akemann		proc means, capacity planning	source	5
60	Graphics on the IBM 3800 and XEROX 9700 Printers	John Austin, Roy Hammond		batch job, data analyst	source
61	Graphics: An Important Tool for Communication in Epidemiology	Edward L. Frome, James R. Gray, Donna Hollis Hudson	The field of occupational ep_1demiology re quires communication between people from many varied disciplines and interests. Traditional	calendar year, poisson distribution, confidence interval	source	7
62	Helpful Hints for Harmonious Data Analysis	Philip Hanser		time series analysis, confidence interval	source	2
63	Hierarchical Data	Peter L. Rikard		dataset, drop statement	source	3
64	How to Avoid Typing and Proofreading PROC FREQ Tables	Vicki L. Cline, Carlos M. Puig	Two-way tables, such as those produced by `PROC FREQ`, have long been the principal means for summarizing survey and census results. Although direct copies of the PRoe FREQ output may be acceptable for internal use, tables are usually retyped for inclusion into final reports. For research groups that produce large volumes of tables, the typing and proofreading al:'e frequently costly, tedious, and error prone. When faced with the task of producing camera-ready copy for several hundred pages of two-way tables on a very tight schedule, we designed and developed a pL/I program (SASMOD) for completely automating the final report table production. SASMon accepts PRoe FREQ output written onto an as file. It selects counts, percentages, and totals in a user-specified combination. Then it removes extraneous lines and spaces, rounds figures to desired precision, eliminates lines consisting entirely of missing values, and repaginates the tables according to user directives	proc freq, proc step, missing value	source	6
65	Hypothesis Testing with Type IV Sums of Squares of the Computer Routine SAS GLM	G.F.S. Hudson, Shayle R. Searle		proc genmod, anova	source
66	Improving the Data Management Expertise of Experienced SAS Users	Joshua Sharlin	At many installations there are at least three levels of SAS users. At the top level are a few SAS experts. At the second level, are the majority of users, people who regularly use SAS, but often with brute force techniques. The third level includes beginning users. This paper pre sents ideas for a course to raise the level of data management expertise among second level SAS users. Such courses should emphasize understand ing rather than syntax, and be organized around a series of examples which address common data processing problems. The key idea is to present multiple methods for solving each problem and to explain the similarities and differences between the methods	retain, conditional do loop, set statement, proc step, cpu time, data governance	source	6
67	Influential Cases and Transformations	R. Dennis Cook, Sanford Weisberg		predictor variable, maximum likelihood estimation, regression analysis, likelihood function	source	6
68	Information-mapped SAS: Teaching SAS for Retention	P. L. Olympia		macro definition, title statement	source	6
69	Integration of IBM and Minicomputer Processors for SASApplications	Larry W. Claflin, Richard K. Keiser	Procedures, hardware, and software are available to minicomputer owners to access SAS cost-effectively. The processing needs of Ecological Analysts, Inc. (EA) have grown from the occasional use of SAS to analyze data sets exceeding the capacity of our PDP-l 1/70 to the regular use of SAS applications as a part of the corporate data processing system. The	batch processing, data wrangling	source	3
70	Interactive Execution of SAS Using TSO Clists and SPF Panels	Cary N. Prague	This paper traces the evolution Northeast Utilities has undergone to run SAS using TSO CUSTS and SPF panels. Throughout the last several years, we have made continuing progress toward executing SAS with more flexibility and speed, producing better results. The CLiSTS and panels that will be presented include. in order of increasing complexity. o o o o o A primitive CLiST that allocates sort space, asks file names and dataset names for infiles, and the SAS program dataset name. The SAS program is then executed and questions are asked to print the report and SAS program. An addition to the CLiST that allows specifying a save area, viewing of the	temporary dataset, sas log	source	6
71	Interactive Geocoding with SAS	John G. Blodgett		macro definition, dataset	source	6
72	Interactive Graphics for Clinical Trials	Judith A. Sromovsky		clinical trial, clinical data	source	6
73	Interim Error Handling Strategies in a SAS Based Research Data Management System	David H. Christiansen, James D. Hosking		interim analysis, missing value, exception handling	source	6
74	Interim Multipliers for Systems with Lagged Variables	Arbi Ben Abdallah, Robert P. Parks		proc step, sample	source	6
75	Iowa Survey of Dental Health: Stepwise Regression Results	J.D. Beck, J. Jakobsen, B. Townsend		sample, multiple regression, regression analysis, sample size	source	5
76	Junior Achievement Campaign Management System	Alen D. Baker, Sally W. Drew	Junimt:' Achievement (JA) is a non-profit organization which supports youth in the exploration of the free enterprise system. A national organization with local chapters forms a management shell with local companie~ providing the remainder of the volunteer man power. One company each year provides the leadership of a campaign drive to raise money for the entire year. The campaigns are only four weeks long and raise large sums very quickly when successful. Solid campaign management strategy is essential in raising over $100,000 in four weeks from over 2,000 contributors. This paper outlines the design and use of a SAS Campaign Management System, an excellent example of the flexible uses of SASe The power of solving business problems through the features of SAS have lead to possibili ties of broader user support from any type of non-profit organization	bar chart, management reporting	source	4
77	Linear Models with Dominant Variables	Peter S. Lufkin		estimate, correlation, regression analysis, predictor variable, proc calis, covariance, condition index	source	6
78	Macros for Matching Analysis	William M. Taylor		proc freq, proc step	source	3
79	Macros for Systematic Sample Selection and Variance Estimation from Ordered Frames	Josefina A. Lago		sample, systematic sampling, sample size	source
80	Maintaining Data Audit Trails in a SAS Based Research Data Management System	David H. Christiansen, James D. Hosking	One of the primary functions of any data manage ment system is the creation and maintenance of data audit trails, records tracking the flow of data through data entry, error detection and correction, and file updating. Unlike most com mercial data base management systems, SAS does not automatically provide such.3 facility. This paper describes the characteristics and uses of audit trails, and presents techniques for imple menting an audit/inventory subsystem as part of a SAS based research data management system	audit trail, data audit, data governance, sas base	source	6
81	Management of SAS Based Data Analysis	Wm. Ford Calhoun, James Holt	Mount Sinai School of Medicine, City University of New York A small group of data managers and programmers in the Research Computing Unit of the Biomathematical Sciences Department of the Mount Sinai School of Medicine of the City University of New York (CUNY) supports 5 statisticians on approximately 100 projects per year. The ]rojects range in size from large surveys of thousands of observations to small studies of a few pat1ents contain ing 7 or 8 variables. The Unit is able to maintain a high level of productivity by tightly managing and controlling its data analysis function. Components of the management system include computerized logs for: pro jects, programs and programmer activi ties; strict naming conventions for pro grams and files; standardized programming style, report generating programs and SAS related data management utilities	raw data, data governance	source	4
82	Managing "Not Small "Datasets in a Research Environment	Jeffrey S. Bass, Keith E. Muller, J.C. Smith	t~chniques in this paper work best with Many dixferent approaches work in fl~e5 ~ncluding l,OOn to 100,000 observa-.aoaging saall datasets. In contrast tions. datasets which are "not small" demanA Overview .In this paper three major top-efficient and sometimes specia1ized meth- ICS:are. disc~ssea . irst, .aDageme~t `ODS`. or resources are wasted. We suqgest strategles, lnolud~D9 planninq and att~- Strategies for avoidio9 some prohl.ems. tudes, arc empaasized. Secona, particl1-Beste icting the discuss~on to a research lac techuiq ues tal: manipllla t.ing Unot en.ironment helps constrain the task. sma~l" dataset$ dl:e discuss€d. Third, First. deL~ne "saall" dataset as any documentation, including meta-docu.enta-for which eoa.on fi~e and analysis tech- tion, is discussed	sample, data governance, output delivery system, dataset	source	6
83	Managing Large Survey Research Data Base	Steven Just, Francesca Seidita		proc step, data step, data cleaning, dataset	source	5
84	Minimum Sample Size Designs for Testing Interaction	James E. Dunn, Mary S. Hirsch		main effect, sample, proc genmod, sample size, degrees of freedom	source
85	Multidimensional Scaling by Least Squares	Cynthia H. Null, Warren S. Sarle		outlier, likelihood function, maximum likelihood estimation, proc nlin, regression analysis	source	8
86	New Time Series Features in SAS	David M. DeLong		covariance, moving average, likelihood function, time series analysis, correlation, proc step, regression analysis, proc arima	source	3
87	Optimal Stratified Sampling, with an Application to Auditing	Edward L. Spitznagel	The application of SAS to statistical sa~pling in auditing is presented, including the Dalenius-Hodges cum-square-root-of-F	sample, sample size, stratified sampling	source	3
88	Paired Comparison Estimation and Analysis: A SAS Macro	Ramon C. Littell, Walter W. Offen	Paired comparison experiments involve treatments compared pairwise by several subjects, with a binary response. For example, subjects may be asked to taste two samples of a particular fooo product, and choose the one with the better taste. The extension to the BradleY-Terry model presented by Rao and Kupper (1967) is used in this paper. This method allows for ties, where subjects may declare that no difference between the two samples can be discerned. This paper presents a SAS macro which performs an analysis of the results from a paired comparison experiment. First of all, maximum likelihood estimates of the Bradley-Terry treatment parameters and the tie parameter are found iteratively. From these, estimates of the pairwise preference probabilities and probabilities of ties are produced for each pair of treatments. Finally, the asymptotic variance-covariance matrix is estimated, and an asymptotic chi-square test for treatment differences is conducted	proc freq, covariance, likelihood function, maximum likelihood estimation, chi-square test, macro definition	source	5
89	Panel Discussion	Mason B. Nichols		proc step, error message	source	8
90	Parametric and Nonparametric Multiple Comparisons Using SAS	Chandu M. Patel	Many statistical computing packages including SAS provide limited methods of Multiple Comparison Procedures (MCP). There is a need for some convenient way to perform a variety of MCP. Many authors including Dunnett (1980a, 1980b) described several useful parametric MCP for three cases: and (III) nij,nj, ai2j,ai Rank analysis of variance by Kruskal and Wallis (1952) with correction for ties is described. Also, MCP based on ranks proposed by Dunn (1964) is described to compare treatments with a control and all possible comparisons. The purposes of this paper are to describe how SAS may be used to perform the computations for the above parametric and nonparametric MCP including the Kruskal-Wallis test with ties corrected, to provide macros using the MATRIX procedure for performing these calculations, and to illustrate them with a numerical example	sample, fixed effects, anova	source	6
91	Performance Analysis - Honeywell Level 6 Statistics	John L. Hall Jr.	GTE Data Services, Inc. has developed a number of computer performance routines using SAS (Statistical Analysis System). This paper will focus on one routine - the analysis of job accounting data from the Honeywell L6 mini computer. The Honeywell L6 Model 57 is the large scale model in the L6 series. The job accounting data recorded under the GECOS 600 operating system is less sophisticated than large mainframe computer data, such as IBM's S.M.F. (System Management Facility). However, this data represents the basic system perfor mance statistics. The analysis of these statistics involved a mixture of applications, which were time sharing, transaction and batch related. Elapsed and CPU timings were reviewed along with peripheral performance for disk, tape, printer, and card input/output devices. Memory swaps were also analyzed. The performance routines were verified by individual runs and validated during three comprehensive benchmark runs	proc means, capacity planning	source	6
92	Personnel and Program Management with SAS	William H. Seaton	SAS has been successful in overhaul ing and managing several important opera tions of a speech and hearing clinic and academic program in a University setting. The Clinical division serves several hundred clients each term. For each client and the students and supervisors who service them, a wide variety of data must be maintained for numerous purposes such as: s) scheduling of clients, stu dent clinicians, and their supervisors for therapy sessions: b) analyses of clinical operations and client character istics: d) student practicum hour reco,rd keeping: and, d) client record keeping. In the academic program, SAS has been used to assist in personnel management and financial analysis and planning. Several examples of SAS usage for mana ging program operations are given. Fi nally , the advantages provided by SAS are reviewed	proc step, clinical data	source	5
93	Pesticides and SAS: Data Processing for EPA Monitoring Networks	Dennis W. Herrin, David C. Moazed, David L. Myers, Jeanne R. Pagano, Deborah H. Whitehorne	Research Triangle Institute's computer Applications Center has provided data processing support for several EPA monitoring programs for the past two years. This support has included increasing user accessability to historical data and development of management information systems, as well as the processing of the monitoring data collected for the past several years. SAS has played an increasingly important role in carrying out these functions, in both interactive and batch modes. We have made use of SAS for docu mentation, merging data files, editing, and creation of management reports. Through the use of TSD CLISTS, routines have been developed to allow user submission of jobs without the need for the user to know SAS	sample, proc step, dataset, data wrangling, binary variable, management reporting, proc freq	source	6
94	Pharmacokinetic Analysis System: Structural Aspects	Raymond P. Chen, Daniel M. Wildman		dataset, proc step, sample	source	6
95	Practical Applications of PROC PRINTTO	Edward P. Barry, Amy Ko	By using Proe Printto, output from SAS procs can be manipu-lated to meet the needs and/or desires of the user. Simple programs are presented for creating concise summaries of analyses by reducing and enhancing output	proc step, proc genmod	source	6
96	Predicting Computer Needs for Banking	Robert L. Cofer		capacity planning, sample	source	2
97	Problems and Considertions in SAS/GRAPH Map Generation	Mitchell J. Tuller		proc step, proc gmap, sample	source	4
98	Problems in Spatial Data Analysis	Deanne Merrill		dataset, sample	source
99	PROC CRYPT: A Procedure to Encrypt and Decrypt SAS Data Sets	James D. Hosking, George Howard		proc step, dataset	source	6
100	PROC M204SAS: A SAS Interface with the Model 204 Data Base Management System	Christopher Rogers, Cary Toor	A SAS procedure, PROG M204SAS, was developed to enable SAS users to access data stored in a Model 204 DBMS file using the Inverted File Access Method. The procedure reads a Model 204 file and creates a SAS file using data from all or selected records and fields. Access to the Model 204 database is transparent to the SAS user and requires little knowledge of Model 204	numeric variable, variable, proc step, data dictionary	source	6
101	PROC PARCAT: A Procedure for Testing Average Partial Association	Ronald G. Fleming, J. Richard Landis, Douglas P. Schnautz	PRac PARCAT is a SAS procedure which provides tests of average partial association in three way cont.ingency tables within the framework of the multiple hyper geometric probability model. Primary attention is directed at the relation ship between two of the variables, controlling for the effects of a covariable. This	proc freq, proc step, correlation, degrees of freedom, sample, dataset, sample size, response variable	source	6
102	PROC TETCORR: Correlations between Categorical and Continuous Variables.	J. Philip Miller, Michael A. Province, Reimut Wette		correlation, likelihood function, proc step, maximum likelihood estimation, latent variable, binary variable, sample, proc calis, confidence interval	source	4
103	PROCs COMPARE and HEADER for Efficient Database Management in SAS	Martha McCrate, J. Philip Miller, Michael A. Province	This paper describes two SAS procedures that have little in common other than the fact that they are both straightforward solutions to problems that often occur in database management. The first problem is that of easily documenting and indexing changes that are made in SAS databases which are constantly being updated- that is, the prOblem of providing a complete audit trail for successive generations of SAS databases. While SAS automatically stores much of this information in the HISTORY records of a dataset, those records are not complete. In particular, the changes made by `PROC` EDITOR are not stored there. `PROC COMPARE` solves this problem by producing a machine-readable (SAS dataset) version of the difference between two datasets. The second problem is that of accessing the names, labels, formats, etc. of variables in a SAS program, for easy recoding	proc step, in=, proc compare, out=, dataset	source	6
104	PRONTO-A System for Maintenance of Duty Rosters of Shift Personnel	Anton A. Backer, Henk Esselman, Erik W. Tilanus		proc step, proc format	source
105	Repeated Measures Analysis for Program Evaluation	Garrett K. Mandeville		frequency analysis, proc freq	source
106	Repeated Measures ANOVA: The Multivariate Approach	S. Paul Wright	This paper covers the basic concepts underlying the multivariate approach to repeated measures ANOVA. There are two main reasons for using the multivariate rather than the univariate approach. One, the multivariate	anova, repeated measures, response variable, null hypothesis, sample, proc step, class variable, degrees of freedom, main effect, proc reg, proc genmod	source	4
107	Robustness of Discriminant Functions	Peter A. Lachenbruch	The purpose of this review paper is to describe methods of discriminant analysis, Lo outline the assumptions that are involved, to discuss the SAS and BMDP programs for discrimi nant analysis and to focus on the effects on the discriminant fum:tions when the assumptions are violat_ed. 1. Fundamentals of Discriminant Analysis The basic problem of discriminant analysis is to allocate a multivariate observation x to one of two populations. The al10catioo- is accomplished on the basis of the value of the observed random variable and was initially proposed by Fisher (1936) as the linear combina tion of observations which maximize the between group variance relative to the within-group variance: (1) where x, is the mean in the i th group and S is the co~on covariance matrix	discriminant analysis, sample, covariance, normal distribution, likelihood function, prior probability, bootstrap	source	7
108	SAS as a Convenient &Efficient Tool in the Management of Resources within an International Clinical Pharmacology Department	A. E. Edelmann, K.H. Jones, M.A.P. Meisinger		resource allocation, clinical research	source	4
109	SAS MACRO "REP2 "for Unweighted Means Analysis of Repeated Measures	Robert V. Zupkis	This paper presents a SAS MACRO "REP2," which will evaluate repeated measures data for the two-group case using unweighted means analy sis. "REP2" will provide the following test	repeated measures, covariance, sample, longitudinal data, macro definition, missing value, anova	source	8
110	SAS Panels - Using SAS with the New SPF's Dialog Manager	Larry Lucke, Patrick R. Timmins		macro definition, dataset, proc printto, proc step	source	6
111	SAS Procedures for ANOVA or Repeated Measures Designs	William G. Jackson Jr., Richard C. McNee, Carolyn J. Oakley		repeated measures, univariate analysis, anova, main effect, missing value	source	2
112	SAS Program for Ratio Estimates and Their Variances from Stratified Random Sampling	Wayne L. Cornelius		proc calis, sample	source	5
113	SAS Programs that Write Other Programs	Juan E. Ruz		proc step, character variable	source
114	SAS Report Generator: A SAS Program to Generate SAS Code	Robert L. Anderson, M.P Koranda	This paper presents a high level interface program that writes SAS code allowing users easy access to SAS's full capabilities and freeing the programmer from the chore of writing code to input data. Programming in a group that supports a large number of users, each accessing varying amounts of data and requiring different forms of data analysis, presents many problems. Some of the difficulties encountered are: interfacing with users ranging from nonprogrammers to highly experienced programmers, requests for one-time-only reports, the need for minor report modifications stemming from users wanting to look at the same data from different perspectives, and the problems associated with the duplication of code among users. A description of the high level interface and an explanation of the data tables required, along with examples of the SAS code generated, will be presented	proc step, sample, data science, proc print, dataset	source	7
115	SAS Video Training An Educational Perspective	John C. Boling		data wrangling, dataset	source	4
116	SAS, SAS/ETS, and CLIST: A User-Oriented Approach to Management Reporting	Michael P. Hudson, Roberta A. McBride		management reporting, control total	source	3
117	SAS/GRAPH and the Geometryof Statistical Distributions	Charles F. Hofacker, Paul Hoffman		degrees of freedom, normal distribution, probability distribution, gamma distribution, correlation, beta distribution, logistic distribution, covariance	source	8
118	SAS/GRAPH for the Non-SAS User	Cary N. Prague	Many users in Northeast Utilities would like to use computer graphics. For many who could not be convinced to learn SAS in depthl the use of `SAS/GRAPH` was not possible.· Through the development of fullscreen "menus", a non-SAS user may now produce `SAS/GRAPHS` with much the same capabilities as a SAS user. This presentation focuses on a fullscreen development which guides a non-SAS user through a series of "menus" where they can define their data, choose graph types, colors, labels, title.s or anything else they desire. The program builds the `SAS/GRAPH` statements and executes SAS. This package can speed the development of most `SAS/GRAPHS` even for the trained SAS user. Using a PL/I assembler interface to handle the fullscreen liD and to control PF keys, this product models itself after the IBM product- interactive chart utility. A user can quickly develop a `SAS/GRAPH` without any substantial training in SAS	proc step, bar chart, proc gplot	source	6
119	SAS/IMS-DLI Interface	John S. Wallace	Worldwide, a vast amount of data is maintained in DL/I data bases. Version 1 Release 1 of the `SAS/IMS`-DL/I product provided read and update access to this data from a SAS `DATA STEP` through an IMS/VS or CICS/DS/VS batch DL/I region. Version 1 Release 2 of the `SAS/IMS`-DL/I product expands this access to "online" data bases and IMS/vs message queues thl"ough a Batch Message Processing DL/I region. Th is Release also provides instaliation control of `SAS/IMS`-DL/I usage. In the near future an update to Release 2 will provide the facility to access CICS/OS/VS "online" data bases through a CICS Shared DL/I region. DL/I Description DL/I is a heirarchic data structure definition and access language. Data are stored in segments at the node points of the tree structure. Various DL/I access	input statement, put statement, infile statement, sample, variable	source	8
120	SCAN: An Interactive Graphics SAS Procedure for Specialized Correlation Analysis of Bivariate Data	David A. DeNardis, Robert N. Rodriguez	Continuous bivariate data are routinely ana lyzed with the aid of a scatter- plot and a corre lation. However, these tools have limitations: a scatter plot is not feasible if the data are weighted or grouped, and a correlation measures only the strength of ~ dependence. These limitations are overcome by a new procedure, referred to as SCAN, which was developed for the Statistical Analysis System (SAS). The SCAN Procedure implements a new graphical technique which displays detailed information about the joint distribution of a bivariate data set. The graphical display is referred to as a "correlation fingerprint", because it can be used to distinguish and identify correlation struc tures. For example, correlation fingerprints can reveal power transformation relationships and departures from bivariate normality. The interactive nature of the Procedure itself is also novel	correlation, scatter plot, pearson correlation, sample, data science, proc step	source	6
121	SD/SDM, a System Design and StatidDynamic Modeling Tool Set	Armen Gabrielian		proc freq, merge statement, structured data	source	6
122	Selecting Subset of Data	Donald J. Henderson		point= option, proc step, keep statement, proc print, dataset, lookup, put statement, set statement, proc sort	source
123	Set Merge and Update	Peter L. Rikard		by group processing, set statement, update statement, dataset, data merging, missing value	source	4
124	Simple Graphics	Cynthia Deitz		group by, proc freq	source	7
125	Some Aspects of the Rank Transform in Analysis of Variance Problems	Ronald L. Iman		sample, ancova, friedman test, rank test, covariance, proc glm, nonparametric test, normal distribution, f distribution, kruskal-wallis test, student t distribution	source
126	SPSS MANOVA and BMDP4V for SAS GLM Users	Kenneth N. Berk	With the help of `PROC` BMDP and `PROC` SPSS it is easy within SAS to use BMDP and SPSS. Both of these packages have new multivariate analysis of variance procedures which are especially convenient for repeated measures, univariate and multivariate. Although BMDP already had a fine univariate repeated measures program, BMDP2V, the new BMDP4V has the ability to do simultaneous univariate and multivariate analyses of repeated measures on one or several variables. SPSS MANOVA shares this and also the ability to generate orthogonal contrasts between subjects and within subjects. In some repeated measures situations~ SAS GLM may compute inappropriate ~um~ of squares. This paper discusses details of repeated measures features and others which supplement those of SAS GLM. The three programs are compared with respect to documentation~ control language. and output. A special consideration here is the choice of output - what is automatic, what is optional, and what can be turned off	proc genmod, repeated measures, anova, residual, proc step, degrees of freedom, main effect, regression analysis	source	5
127	Statistical Review of Inpatient Services: A Systematic Approach	Jennifer Anderson, Gregory Lenhart, Janice L. Weener		outlier, response variable	source	2
128	Strategic Analysis for Commodity Hedging	Ross Dennis, Nelson M. Fraiman, Murray Mohl		regression analysis, sample	source	9
129	Table Lookup Techniques	Donald J. Henderson		lookup, put statement, data step, if-then-else, proc step, format statement, proc freq	source	7
130	Teaching Programming and Documentation Techniques	William D. Phillips, Roger C. Race		proc print, hash table, proc step	source	6
131	Techniques for Creating Special TSO/CMS Commands that Generate and Execute SAS Programs	Dennis W. LaRue		proc step, macro definition	source	6
132	Techniques for Interrelating and Reducing Data Bases Using SAS	Robert U. Anderson, James W. Davis, Steven Flint		proc step, set statement, by group processing, data step, proc freq, proc means	source
133	The Analysis of Ecological Survey Data with SAS and EAP	Robert W. Smith		sample, cluster analysis, proc step, multiple regression, regression analysis	source
134	The Application of Risk Analysis to Marketing Information Support Systems	William R. Gjertsen		monte carlo, sas risk engine, proc freq, proc step	source
135	The Development of a SAS Software Package for Maintaining and Utilizing a Nutritional Analysis System	Penny T. Culley		cpu time, dataset	source
136	The Evaluation of Federal Programs in Education	Garrett K. Mandeville		repeated measures, covariance, treatment effect, factor analysis, anova	source
137	The Expected Value of Guttman's Coefficient of Reliability Based on the Multinomial Distribution	John A. McKillip, Janice L. Weener		numeric functions, proc calis	source
138	The Experimental Interactive Matrix Language	John P. Sall		proc step, macro facility	source	6
139	The Impact of Recent IBM Hardware and Software Announcements on the SAS Product Line	William H. Blair			source	4
140	The Rank Transformation in Discriminant Analysis	William Jay Conover, Ronald L. Iman	In order to decide which of two populations is the correct population to which an observed item belongs, the observed item is compared with samples of observations known to have come from the two populations. The sample that resembles most nearly the observed item determines the population to which the item is classified. Procedures for making these classifications fall in the category of discriminant analysis. The most popular discriminant analysis procedures are known as the LDF and QDF procedures. They are available, although not by those names, under `PROC DISCRIM`. These procedures rely on the assumption of normal distributions for their theoretical basis. If the populations are not normal, the effectiveness of these procedures is enhanced by using them on the ranks of the data, after `PROC RANK`, rather than on the data them selves. This result in a lower probability of misclassification in most cases	sample, covariance, sample size, discriminant analysis, proc means, normal distribution, parameter, proc step, proc discrim	source	10
141	The Role of Database Management Systems for Investigative Data	Gary Anderson		network optimization, data standard, proc sql	source
142	The Role of the Departmental Consultant in the Computer Center	Eileen Driscoll		data science, raw data	source	3
143	The TABULATE Procedure	Alan Eaton, Ken Howell		proc tabulate	source	11
144	Training Statistical Consultants in a University Biostatistics Consulting Center	Ronald G. Marks		sample, data science, correlation	source	4
145	Transposing Data	Donald J. Henderson		proc freq, proc step, proc transpose, proc print, data step, frequency analysis	source	7
146	Tuning the TSO Environment with SAS	Kenneth E. Levy		proc step, standard deviation	source
147	Two-stage Cyclic Queueing Model of Multiprogramming from the MICS Data Base	Gerald Stephen Hodge		cpu time, capacity planning, data cleaning	source	3
148	Unconventional Merges	Peter L. Rikard		set statement, proc step	source	3
149	Use of EXECs in File Management and Report Generation	Cynthia Deitz		_eof_, proc print	source	6
150	Use of SAS to Analyze Experiments Conducted over Time with Multivariate Responses and Autocorrelated Errors	Walter W. Stroup	An experimental procedure common in applied biological research involves the use of a simple experimental design to observe responses on mul tiple variables, often correlated, at several intervals over time. The purpose of such experi mentsis generally to evaluate treatment effects on changes in the vector of responses over time. In many such experiments, the vector of residuals exhibit autocorrelation which is strongest for observations contiguous in time and which weaken progressively with increasing lag times. Thus, the error structure can be described by a multi variate autoregressive model. While attention has been given to the use of SAS to analyze re peated measures experiments in the univariate easel there is no obvious method for using SAS to perform a multivariate analysis when such an error structure is present	proc step, proc glm, response variable, proc genmod, cv, correlation, time series analysis, anova	source	7
151	Using CMS Exec Files for Data Management and Report Generation	J.R. Barrett, G. Blum		raw data, input statement, dataset	source	4
152	Using PROC ALSCAL to Estimate Scale Values for Unidimensional Choice Models	Mark L. Davison, Phillip K. Wood		proc step, distance measure, sample	source	6
153	Using SAS and FORTRAN to Produce Mantel-Haenszel Statistics	Caroline A. Macera, Deborah A. Truesdale	The Mantel-Haenszel method is frequently used to obtain a SUll1l1ary odds ratio as an esti mate of relative risk for matched case-control studies. An interactive Fortran program calcu lates a Mantel-Haenszel summary odds ratio and chi-square statistics for any number of 2x2 tables, or subgroups. The information for each subgroup is entered by hand at the keyboard. When the number of subgroups is larger than ten, entry by hand becomes tedious. A SAS program was designed which passes crosstabulated infor mation produced by `PROC FREQ` to the Fortran program. The SAS program is generalized to allow easy manipulation of the variables of interest and handles missing data in such a way that the maximum amount of information is util ized. This paper describes the methodology used to create and pass i nformat ion from a SAS procedure to the Fortran program, as well as the prob 1 ems encountered indo I ng so	proc freq, odds ratio, proc step	source	4
154	Using SAS for Box-Cox Transformation in Regression Analysis	Ollie Frazier, Deepak A. Keshani		regression analysis, by group processing, proc step, likelihood function, sample	source	6
155	Using SAS for Estimation and Hypothesis Testing in an Unbalanced Repeated Measure Design	Kenneth L. Koonce, Charles J. Monlezun, David D. Morris		standard deviation, sample, repeated measures, sample size, main effect, proc calis	source	4
156	Using SAS in a (Non-statistics) Freshman Mathematics Course	Thomas A. Carnevale		proc step, transition matrix	source	6
157	Using SAS Interactively to Examine and Diagnose Psychiatric Patients	Ronald K. Elswick, Robert M. Hamer	This paper describes a SAS Program which in terviews a psychiatric patient and, based on the respDnses~ makes one or more psychiatric diag noses. The SAS Program runs interactively, and uses PUT and INPUT statements to communicate with the patient. A series of macros performs the diagnoses after the data are collected	input statement, dataset	source	5
158	Using SAS to Bridge the Gap between Computer Science and Quantitative Methods	David R. Dolkart		dataset, data wrangling	source	3
159	Using SAS to Introduce University Administrators to Computing	Andrew M. Smith	Clemson University is a 'large stgte-supported school located in northwestern South Carolina. Administrative personnel at Clemson are increas ingly in need of up-to-date information in order to manage their resources, but the administra tive programming group is already overbur dened with requests for computer-generated reports. As a result, colleges and departments must wait months or even years for the reports they need. Beginning in the spring of 1981, the Clemson University Computer Center began to offer a series of computing workshops tailored to admin istrators. These workshops taught general concepts of computing, the use of time-sharing terminals, and SAS. Because of these work shops, a number of administrative staff members became confident users of the computer, able to generate their own personalized reports using	data wrangling, data science	source	3
160	Using Spline Functions and the Bootstrap to Fit Differential Equation Models to Data	Thomas P. Capizzi, Robert D. Small		bootstrap, regression analysis, proc calis, proc reg, proc step	source	6
161	Variable Selection for Categorical Data Using SAS	Erik J. Bergstralh, Susan McGorray		feature selection, response variable, predictor variable, degrees of freedom, sample, chi-square test, sample size, nominal variable, likelihood function, proc freq	source	6
162	Weibull Regression and Hazard Estimation	Jan M. Hanley, William H. Rogers	Weibull regression is suitable for analyzing survival data in a regression-like format. This alternative to the Cox "proportional hazards" model offers several advantages: The analyst can estimate survival probabilities for individuals, together with confidence intervals. These help him interpret and describe results. A single parameter describes whether individuals have decreasing, stable, or increasing risk (hazard) functions. This helps test theoretical predictions about rising or falling risks. * The method is an M-estimate (from robustness theory), which makes available several practical results. We have written SAS procedures to carry out Weibull regression and calculate the auxiliary statistics required for a satisfying analysis. Several of our procedures lead to novel graphical presentations (e.g. estimates of the combined hazard function of the sample). Our paper will describe these methods and our SAS implementation	regression analysis, predictor variable, likelihood function, proc step, dataset, covariance, hazard function, sample, proc phreg, maximum likelihood estimation, outlier	source	11
163	Weighted Sequential Hot Deck Imputation Macros	Vincent G. Iannacchione		imputation, sample, macro definition, variable, dataset, sample weight	source	5
164	What is Cognitive Complexity? Alternative Scoring Procedures for the Modified Rep Test	Tom Trabin, Phillip K. Wood		standard deviation, correlation, eigenvalue, proc freq	source	6
165	Word Processing Using SAS	Ross Merlin	SAS provides many features that aid in the production of quality reports. There are times, however, wben graphs r charts, tables, even "custom-tailored" reports cannot properly convey some aspect of a report; explanations of aethods used and statements of	sample, data wrangling	source	2

Back to Conf go 1983 go 1981

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.