Sensible Systems, Inc -- A leader in
Automated Support for Responsible Occupational Analysis
For Historical Interest
ONLY!
For current information,
please see: <http://www.icodap.org>
In preparing materials to "give away" mainframe CODAP from the United States Air Force to State and Local governments, the following handout was generated (Weissmuller, 1979) with some modern elaboration [1997]:
CODAP: Three Definitions -
1. The Acronym - An acronym for the "Comprehensive Occupational Data Analysis Programs". A set of [100 to 400] computer programs used to automate, process, organize and report occupational data. The CODAP system was originally developed by the Air Force Human Resources Laboratory (AFHRL) in the mid-1960's and has been continuously updated and enhanced by AFHRL [now the Human Resources Directorate of Armstrong Laboratory, Brooks AFB, TX]. Differing versions of this system are available for use on UNIVAC, IBM, and CDC computer equipment [and now on IBM PC compatibles]. The DoD executive agent for CODAP (U.S. Navy) currently has a contract to redesign and rewrite the CODAP system to be more machine independent. [In 1983 and 1988 the Air Force version of CODAP (the Fieldata version) was upgraded to new FORTRAN compilers and renamed ASCII CODAP for Unisys (Sperry-Univac) versions and CODAP II for IBM mainframes and the IBM RISC platforms. The atCODAP system, programmed in Turbo Pascal, went into operational use in 1987 on the PC platform as a service bureau system.]
2. The Procedures - A set of procedures and guidelines used in conjunction with the CODAP computer programs. These procedures cover areas such as survey instrument construction, survey administration, data processing support, and analysis techniques covering both computer products and related hardcopy sources. The analysis guidelines provide personnel management specialists with quantified [,detailed] job information for selection, classification, training, evaluation and job design. Recently developed techniques aid curriculum designers in determining the content and emphasis for relevant training courses.
3. The Theoretic Approach - A theoretic approach in occupational analysis which stresses the quantification and empirical testing of occupational factors over a [custom-developed and carefully] defined inventory of items. These items (usually task statements), along with background questions, are used to form clusters or groups, define categories and produce prioritized lists meaningful to managers. Typical examples of reports include composite job descriptions, group difference descriptions, variable summaries or cross-tab reports, and ordered task factor summaries. To aid in producing these products, the system incorporates such sophisticated data analysis techniques as hierarchical clustering, interrater reliability measures and regression analysis. [The regression analysis capability was dropped in the 1983 conversion to ASCII CODAP and replaced with interfaces to external statistical packages.] As heavy emphasis is placed on empirical testing of both factors and procedures, it should be noted that over a decade [(now three decades as of 2 July 1997)] of operational service has shown the utility of survey techniques for various size samples using task inventories to collect basic data from both job incumbents and subject matter experts [SMEs] (supervisors). The factors which have proven most useful to date include "relative time spent" from incumbents and "task learning difficulty" from supervisors. Though other factors and scales are under investigation, with one exception, none have undergone sufficient testing to warrant widespread usage. The "recommended training emphasis" factor which is collected from supervisory personnel has just been accepted as an approved task factor. [Research in the late 1980's targeted the use of occupational data to "weight" test areas in promotion tests. This "Automated Test Outline" (ATO) research effort successfully tested and validated another task factor called "Testing Importance".] Because the CODAP technology is being widely dispersed and many agencies are using the software, new [experimental] factors are appearing which have not been validated in any documented research program. Extreme caution should be exercised in copying survey booklets, especially those containing factors not mentioned above. [Some factors which sound appealing such as "Frequency of Performance" or "Level of involvement" may have strong intuitive appeal, but operationally using those metrics tend to provide less satisfactory job classifications and group job descriptions.
CODAP Terminology Issues
Because CODAP is a scientific methodology, common, fuzzy terms (like "position" and "job") are given more precise technical meanings to ensure clear communications (Mitchell, 1977). [Note: all citations in this web site are found in the Master CODAP Bibliography available under the Research button.]
To CODAP, "a position" is a body of work for which exactly one individual is hired.
To describe the work of related positions, CODAP uses the term "job." In this sense, "a job" is the work accomplished by a group of people (one or more) and may be defined by one or more set of occupational descriptors (typically "task statements", "tools used", "knowledge, skill, ability, and other - KSAOs" or "competencies").. The terms "job", "job description", "group job description", "special job description" and "composite job description" are used interchangeably.
In evaluating existing classification structures the work done by traditionally defined groups of workers (like Carpenter Apprentice and Carpenter Journeyman) may not give a clear picture of the actual work differences. In these cases, hierarchical clustering is used to empirically identify units of work that really exist in the workplace. These empirical units are called job types and represent relatively focused efforts by subsets of the workforce.
When analyzing issues from an Americans with Disability Act (ADA) perspective, CODAP can be used to establish "essential elements" (using "significance to job" rather than traditional "time spent") and identify non-traditional categories of "jobs" which seem to offer the most rapid promise for expanded opportunities via reasonable accommodations.
CODAP is used to organize and document all jobs in a given job family. A job family is that series of positions spanning from entry-level, training positions to the most senior positions considered to be within the reasonable career path progression. CODAP was developed by the United States Air Force (USAF) to manage a sophisticated personnel classification system (250 job families) impacting a multi-billion dollar training program. With SSI's commercially available, PC-based version of CODAP (atCODAP technology), CODAP makes sense for any medium to large organization with 30 or more employees in a given job family - especially if the job family is operating at geographically dispersed locations or involve jobs which have evolved over long periods or during times of rapid growth.
Perhaps my favorite pair of papers were presented by the U.S. Navy in 1985 and 1987. The paper in 1985 was entitled "Data Base Management Systems, Statistical Packages and Report Writers: Is CODAP Really Necessary?" That paper demonstrated that there was no computation in CODAP which could not be duplicated by SPSS. After the paper was presented at the International Occupational Analysts Workshop, I was flooded by pleas to prove the Navy wrong. I replied that the Navy was totally correct. CODAP is the right approach, not because it does something other packages CAN'T DO, CODAP is right approach because it PACKAGES FEATURES to support operational occupational analysis programs. I estimated to "mimic" CODAP's products would require TEN TIMES the manpower on the part of the computer technicians and analysts to obtain necessary and comparable products. As one might expect, the Navy's 1987 paper was entitled "NODAC Recants - SPSS as a Supplement to CODAP -- not a Replacement." Again, I say, the Navy is totally right.
The "problem" with some job analysis systems is that they take the existing classification structure as a "given" and simply work within those constraints. In today's world (even more so than when CODAP was developed in the 1960's), the world is changing around organizations and the organizations react internally. Over time, a perfectly good classification structure may have been replaced by a pragmatic allocation of work that kept the organization going in spite of the paper tiger. CODAP does an unbeatable job in using hierarchical clustering to group respondents into a classification structure which is empirically sound (not simply a rubber stamp for the pre-existing structure). CODAP has an internal system for labeling and comparing and contrasting (at the task, tool, knowledge, skill, ability, competency, personality dimension level) empirical job types. If the old classification structure is seriously flawed, it can be replaced by one which makes sense to both the workers and the personnel department. When this happens, the entire occupational area needs to be reassessed for promotion practices, career pathing, training programs, etc. Once a workable classification structure is established, composite job descriptions can be produced for each class, group difference descriptions can compare and contrast adjacent grade level, and task factors can be used (after processing by interrater reliability software) to prioritize tasks in new (or old) jobs for training, inclusion in promotion programs and performance appraisal systems.
At present, the principals in Sensible Systems, Inc. include Johnny J. Weissmuller (San Antonio, TX), Michael R.Staley (Seattle, WA), and Terry D. Lewis (Dallas, TX). In the 1970's and early 1980's, the principals in Sensible Systems, Inc. (SSI) were military and civil service civilians whose job it was to maintain *** and give away for FREE*** the mainframe version of CODAP. We were told hundreds of reasons why state & local governments wouldn't take the "free" federal government software. Sensible Systems, Inc. was founded in 1982 to migrate the occupational analysis concepts capabilities developed by the U.S. military in the 1960's for training and classification requirements to the broader needs of the non-military public sector and private industry for the 1980's and beyond. In addition to the changing focus of occupational analysis, the technology explosion also meant a migration from mainframe computers beat into submission only by career programmers to personal computers serving as tools to human resource professionals and support staff. Sensible Systems, Inc.'s (SSI) first products (automated surveys, 1982) were developed to support the mainframe CODAP systems. Later services (the CODAP Service Bureau, 1987) provided access to CODAP to users without mainframe computers. Most recently, SSI's atCODAP ("anchored to" CODAP, but not limited by it, 1992) made available a desktop PC version of CODAP.
The two primary conferences for occupational analysis at which CODAP is a central issue include the Annual International Military Testing Association (IMTA) Conference and the biannual International Occupational Analyst Workshop (IOAW). Location of the IMTA varies each year depending on the sponsoring agency. This year it is in Sydney Australia in October and next year is scheduled for Pensacola, Florida.
The Analysts' Workshop (IOAW) is held every-other year (approximately) in San Antonio Texas and is primarily hosted by the USAF Occupational Measurement Squadron (Mr. Jay Tartell).
Because scheduling arrangements are "fluid", I suggest you visit the home page for the Institute for Job and Occupational Analysis (IJOA) who closely tracks these conferences. Charles Holt of IJOA takes photos at these conferences and posts them in their web site (for posterity or blackmail?). SSI is a strong backer of the IJOA and highly recommends its professional pursuits, its products, and its people.
Visit the IJOA web site to check out scheduling for upcoming conferences: <http://www.ijoa.org>
While Sensible Systems is able to provide names of consultants, we would prefer to discuss your project and situation in order provide you with the best match. Sometimes a project only needs an independent consultant and sometimes a project requires a wide-spectrum of talent.
CAUTIONARY NOTE: When dealing with consulting firms who show long lists of "projects completed", be sure to require an explicit link to who on the CURRENT STAFF did those projects. OA is a small community and we move around (or pass on in Dr. Christal's case). For example, one consulting company can claim to have employed ALL of the following individuals (though not necessarily at the same time). These killer credentials include:
A killer set of credentials, right. Who is this organization? Your choices are:
The early history (prior to 1974) is given by the FOREWORD to AFHRL-TR-74-85. CODAP: Programmer Notes for the Subroutine Library on the Univac 1108. (Weissmuller, Barton, & Rogers, 1974). Annotations from 1997 are set off by brackets where appropriate. The FOREWORD read as follows:
The Comprehensive Occupational Data Analysis Programs (CODAP) System has been under continuous development for the past fifteen years. In its current form [now denoted FIELDATA CODAP,] it is a joint product of the Computational Sciences Division and the Occupational Research Division of AFHRL [now called the Human Resources Directorate of the USAF Armstrong Laboratory]. In general, the Occupational Research Division has developed input-output specifications for programs and program modifications, while the Computational Sciences Division has provided programming services. However, even this distinction has not always been a clear-cut one, since suggestions for programs improvements have come from individuals in both Divisions.
Since its conception, development of the CODAP system has been under the general direction of Dr. Raymond E. Christal, Chief of the Occupational Research Division. Dr. Christal also personally designed many of the programs and program modifications. [Dr. Christal passed away in April 1995 and, with the kind permission of Dr. Christal's wife, IJOA has established the Dr. Raymond E. Christal Award for Occupational Analysis. (See IJOA Home Page for Details)]
The earliest CODAP programs were written by Mr. Daniel Rigney and Mr. Wayne E. Fisher. This version was later expanded and reprogrammed (under contract) for execution on an IBM 7040 by Mr. Irwin R. Oats and Mr. James R. Hills of the Computer Sciences Corporation and later of Oats-Hills Incorporated, Houston, Texas. (Mr. Oats and Mr. Hills have translated CODAP programs for execution on the IBM 360 and 370 series computers, under contract with the U.S. Marine Corps and the Department of Defense. [This IBM 360 version of CODAP is available from the National Technical Information Service, NTIS.])
Since 1970 all CODAP innovations and modifications have been programmed in-house by the Computational Sciences Division. During this period, many new programs were added to the system, and nearly all old programs were extensively modified. During the last two years [1972-1974] the CODAP system has been completely rewritten for execution on the Univac 1108. This reprogramming was first undertaken by MSgt William D. Stacey, with the assistance of Sgt Johnny J. Weissmuller. After MSgt Stacey's departure from AFHRL, the translation and reprogramming were completed and checked out by Sgt Weissmuller, assisted by A1C Bruce B. Barton. It is this version of the CODAP system which is documented in this technical report.
Many individuals have contributed to the CODAP system, and it would be difficult to specifically enumerate each contribution. However, the authors would be remiss not to mention those individuals who have participated in the mainstream effort. It is possible, with the passing of time and failing of memories, that the following history contains serious omissions and inaccuracies, but it is correct in terms of currently available information.
The FOREWORD goes on to list individual programs identifying both the initial design (occupational analyst) and the initial programmer. To use Dr. Joe Ward's favorite lead-in: "To make a long story short.." here is the list of "initial design" contributors: Dr. Raymond E. Christal, Dr. Joe H. Ward, Jr., Dr. Robert A. Bottenberg, Mr. William J. Phalen, Mr. Wayne Archer, Mr. Johnny J. Weissmuller, Mr. W. L. Wells, Mr. C. R. Rogers, Mr. Harry Lawrence, Mr. Edward Adams, Mr. Philip Aitken-Cade and Dr. Joe Morsh. Major suggestions for operational improvements came from Mr. J. Myer, Mr. Terry D. Lewis, Steve D. Poole, Bill Feltner, Jr., Robert L. Vance and Harry Kudrick. The "initial programming" list includes: Daniel Rigney, Wayne E. Fisher, I. R. Oats, J. R. Hills, D. W. Hartmann, Edward L. Adams, William D. Stacey, Janice Buchhorn, Kathleen Hall, Paul Aron, and Johnny J. Weissmuller.
Later CODAP developments include the following projects. In 1975, the Air Force released the Univac version to Dr. Sid Gael of A T & T. Acting as a consultant, Johnny Weissmuller remotely installed, customized, and tested the software operating on a Univac in Murray Hill, New Jersey. In 1976, the Air Force released the Univac 1108 to the Graduate School of Business at the University of Texas at Austin. As they had a CDC 6600 mainframe, acting as a consultant, Johnny Weissmuller spent a year of weekends rehosting the system. This rehosted version of CODAP was sold to Control Data Corporation to recoup conversion expenses. The CDC version of CODAP was eventually given back (free of charge) to the Air Force for use on CDC computers at Lowry AFB in Colorado.
By the late 1970's the aging IBM 360 and 370 CODAP versions were not only missing many new developments for the occupational analysis community, but newer operating systems from IBM were threatening the viability of the basic package. A contract was awarded to Texas A&M University to redesign CODAP for the 1980's and the new product was to be called "CODAP80". Although many long hours were spent in developing a system that worked, interface designs, while acceptable for an academic environment, simply precluded use in a high-volume, operational setting. This project was formally closed in 1985.
The Air Force developed the very first CODAP programs on an IBM 650 in the early 1960's. They worked on the IBM 7040 in the mid 1960's and migrated to a third generation computer (Univac 1108) in late 1973. By late 1982, projected workload demanded improved processing efficiency and hence a CODAP rewrite project was initiated. This 1983 to 1985 project analyzed the Univac 1108 system (written in FIELDATA FORTRAN) and redesigned key features to increase both flexibility AND processing speed (both in terms of Machine time and Analyst time). The replacement system (on the upgraded Univac -- now a Sperry Series 1100) was written in ASCII FORTRAN and permitted full upper and lowercase text to be used. Because of the programming languages, the new system was called ASCII CODAP (reflecting the expanded character set available in the American Standard Code for Information Interchange system). This work was accomplished under contract with the MAXIMA Corporation. The lead designer and systems analyst was Michael R. Staley with Johnny J. Weissmuller as Project Manager. Darryl K. Hand served as lead programmer.