Current Occupational Analysis Systems
This appendix describes current occupational analysis systems in more detail than is provided in the main text. Category/enumerative systems are presented first followed by descriptions of six illustrative descriptive analytic systems.
The ISCO system uses two key concepts: job and skill. Job is defined as "a set of tasks and duties executed, or meant to be executed, by one person." Skill is defined as "the ability to carry out the tasks and duties of a given job." Operationally, four levels of skill are defined, entirely in terms of achieved education. The lowest level approximates primary school (about sixth grade in United States), the second approximates secondary school (about the twelfth grade in United States) but includes apprenticeships, the third approximates college education but not obtaining a degree, and the fourth includes undergraduate and graduate college education (International Labour Office, 1990:2–3).
The ISCO-88 structure is hierarchical, with 10 major groups at the top, 28 submajor groups, 116 minor groups, and 390 unit groups. Eight of the 10 major groups are categorized at one of the
four skill levels ("armed forces" and "legislators, senior officials and managers," were not so categorized). For example, all occupations in the major group "clerks" are categorized at the second skill level, and all those in "elementary occupations" are at the first level. All descriptions are verbal, and no quantitative data are provided. Each major group, submajor group, and minor group is described by a general duty description and a list of tasks (usually no more than a brief paragraph in length). The lowest level "unit groups" also include names of "example occupations" and related occupations, in addition to the general duty and task list descriptions.
This development of this structure was "carried out in line with the recommendations and decisions of the Thirteenth and Fourteenth International Conferences of Labour Statisticians, held at the International Labour Office, Geneva, in 1982 and 1987" (International Labour Office, 1990:1). The underlying source data consist of population censuses, statistical surveys, and administrative records maintained at the national level.
The Organization for Economic Cooperation and Development (1997) recently reviewed the use of ISCO-88 in Europe and elsewhere around the world. It reached a number of conclusions:
- ISCO-88 has superseded ISCO-68 and has become the model for new national classifications in many countries, even those with previously existing systems.
- Levels of reliability of classifying occupations into ISCO categories remain fairly low, at 75 percent for the most detailed levels of categorization (about 350 categories).
- Aggregating to higher levels of categorization improves the correspondence of across-nation coding (aggregation to about the "submajor" level in ISCO-88 terms).
- International comparability of ISCO-88 results is improved through technical assistance to participating countries in the use of the system.
The Australian Standard Classification of Occupations (ASCO) uses the same concepts of skill level and skill specializa-
tion as does ISCO-88 and has eight major groups, 52 minor groups, 282 unit groups, and 1,079 occupations. An occupation is defined as a set of jobs with similar sets of tasks. In 1993, changes likely to be made to ASCO were seen as including (Madden and Tam, 1993):
- Developing procedures for monitoring changes in industry, vocational education, and training to keep ASCO up to date,
- A movement toward the use of competencies (specific skills, knowledge, and training designed to meet industry standards) rather than educational qualifications and duration of training and experience as indicators of skill level,
- Increasing use of job tasks rather than job titles for classifying into occupations, because jobs are becoming broader and titles less reliable indicators of job content, and
- Modifying the major group structure of ASCO to meet user problems, including the need for career path analysis.
The Netherlands Standard Classification of Occupations 1992 (NSCO'92) also classifies occupations by skill level and specialization, but it differs primarily in its operational definition of those concepts. This system uses the "most adequate training program, that is the training program that best prepares for the tasks and duties in the job" (Bakker, 1993:273) as the method to identify skill level and specialization for each job. To do this, the Netherlands Standard Classification of Education is used as the basic information to conduct the coding. Skill specialization coding is made according to the major (and minor) educational sectors in the Netherlands, e.g., agriculture, mathematics and natural sciences, and language and culture. Skill level is coded with a five-point scale that combines formal education and length of on-the-job experience. Beyond these higher-level criteria for coding, they include two interesting concepts: main tasks and specific skills. If level and specialization are not adequate, then a list of 128 main tasks is used to differentiate the occupation. Examples of these 128 "task clusters" include "managing supervisors and decision-making general policy," "check, inspect, examine, verify,
test, sort," and "navigate a ship." If still further differentiation is needed, then a list of 11 specific skills are used (e.g., quantitative skills defined as activities in which it is important to perform calculations; serviceability defined as activities in which it is important to render service to other people). These concepts are similar to the "generalized work activities" and "basic and cross-functional skills" included in the Occupational Information Network (O*NET™), the Department of Labor's replacement for the Dictionary of Occupational Titles. In operation, the five skill levels crossed with the 13 major skill specializations produce 43 occupational "classes."1 With the invocation of minor skill specializations, 121 occupational groups are formed, and with the addition of the 128 "main tasks" criteria, 1,211 occupations are formed.
The United Kingdom replaced two earlier classification systems, the Classification of Occupations and Dictionary of Occupational Titles (CODOT) and the 1980 version of the Classification of Occupations (CO80) with its Standard Occupational Classification. This effort was coincident with the revision of ISCO-68, so an effort was made "to achieve the closest feasible harmonization" between the British SOC and ISCO-88 (White, 1993). Beginning with the 350 entities in the 1980 Classification of Occupations, modifications were made to fit with ISCO classification criteria. These new code groups were tried out against data from the 1981 census of population and a sample of job vacancies sent to job centers. The resulting structure had 9 major groups, 22 submajor groups, 77 minor groups, and 371 occupational unit groups.
Canada replaced its Canadian Classification and Dictionary
of Occupations (CCDO) in 1991 with "two classifications described as the National Occupational Classification and the Standard Occupational Classification" (Nijhowne and Silver, 1993). The two classifications share a common framework: a hierarchical structure with 514 "unit" groups and 139 "minor" groups in common. The minor groups are organized into 47 major groups in the Standard Occupational Classification (SOC 1991) and into 26 major groups in the National Occupational Classification (NOC 1991). Both have 10 broad occupational categories at the top of the hierarchy.
The SOC 1991 is primarily used for enumeration purposes (e.g., for the Canadian census coding of occupations). The 514 unit groups are described in terms of the principal tasks and duties of the jobs in the unit group. The NOC 1991 also contains other characteristics of the group, such as educational requirements, consistent with its purpose of classifying and describing occupations for labor market transaction. The titles shown as examples are generally the same for the two systems, "but some are unique to each classification" (p. 305). Regarding the military, the SOC 1991 includes just two groups, commissioned officers and other ranks. The NOC 1991 has these two groups, but they include only those military jobs that do not have a civilian counterpart. Military jobs with civilian counterparts are placed in the appropriate occupational unit group within the NOC 1991. The NOC 1991 is used to classify 25,000 job titles into 522 unit groups and does not contain the dictionary-like definitions of its predecessor. Rather, it "serves as a framework whose main function is to provide structure and meaning to the labour market as a whole (Roberts, 1993:320)." One component of NOC 1991 is a matrix, defined by skill level (four levels of type and length of education, training, or experience required for employment in an occupation) and skill type (broadly organized into 9 broad occupational categories, omitting the military category from the 10 broad categories). The 139 minor groups are displayed in this matrix.
Descriptive Analytic Systems
Position Analysis Questionnaire
The Position Analysis Questionnaire (PAQ) (McCormick et al., 1969) is a worker-oriented job analysis technique with a long history of research, development, and use with a variety of human resources applications. The PAQ consists of 187 items listing work behaviors and job elements at a level of abstraction that permits work to be described across a broad range of occupations. Completed by subject-matter experts (job incumbents, supervisors, or job analysts who are very familiar with job content), the PAQ reflects a simple model of work performance following an information input—processing—work output sequence. PAQ items are organized into six divisions: information input (e.g., use of written materials), mental processes (e.g., problem solving), work output (e.g., assembling), relationships with other persons (e.g., instructing), job context (e.g., high temperature), and other job characteristics (e.g., work schedule). Five-point response scales are used to assess importance, time spent, extent of use, possibility of occurrence, and applicability of job elements. "Does not apply" is also provided as an option for all items.
The developers of the PAQ provide computerized scoring services and a normative database that permits comparisons of jobs in one organization to similar jobs in other organizations or to all jobs in the database. In addition to summary statistics for job elements (PAQ items), factor analytically derived job dimension scores (e.g., visual input from devices/materials) and estimates of attributes required to perform the job (e.g., visual acuity) can be obtained. This information can be used to estimate the validity of tests of job attributes used for selection purposes. PAQ scores can also be used for job evaluation, to estimate pay rates based on normative wage data for similar jobs in the U.S. economy.
The PAQ has been widely used by human resource professionals and researchers. It has spawned a substantial body of research studies. Major advantages of the instrument include its broad applicability across occupations and availability of the normative database. Numerous criticisms have also been raised by
reviewers, including its advanced reading level (a college graduate reading level is required, according to Ash and Edgell, 1975), its content geared too heavily toward manufacturing occupations for an instrument that purports to apply to all jobs (DeNisi et al., 1987), and results that are too general in nature to specify the type of work actually done in a job.
In response to the criticism that the PAQ is too heavily weighted toward blue-collar occupations, in 1986 the PAQ's authors introduced a second worker-oriented instrument called the Professional and Managerial Position Questionnaire (PMPQ). Designed for analysis of managerial, scientific, technical, and staff jobs, the PMPQ consists of 98 items assessing 6 job functions (planning/scheduling, processing of information and ideas, exercising judgment, communicating, interpersonal activities and relationships, and technical activities), personal requirements (e.g., education and training required), and other information (e.g., personnel supervised). As with the PAQ, computerized scoring services and normative data are also available for the PMPQ.
Fleishman Job Analysis System
The Fleishman Job Analysis System (FJAS) is based on extensive experimental and factor analytic research on the nature of human abilities (Fleishman and Quaintance, 1984). Conducted over a 40-year period, this research program consisted of a wide variety of laboratory tasks designed to elicit performance from subjects drawing on one or more hypothesized underlying abilities. Task batteries were systematically varied to hone in on specific abilities and to delineate the boundaries of their application. Thus, the research linked task characteristics to ability requirements to produce the Fleishman Taxonomy of Human Abilities. The Fleishman taxonomy provides detailed descriptions of 52 abilities, including cognitive (e.g., oral comprehension, number facility), physical (e.g., explosive strength, arm-hand steadiness), psychomotor (e.g., rate control, reaction time), and sensory-perceptual (e.g., depth perception, speech recognition) domains. Nine social-interactive abilities (e.g., persuasion, persistence) and 13 job skills and knowledge (e.g., mechanical knowledge, driving) are the most recent additions (Fleishman, 1992).
A measurement system was also developed to evaluate jobs and tasks for their requisite abilities. The ability requirement scales (Fleishman, 1992) provide definitions, additional information to differentiate each ability from other similar abilities in the taxonomy, and 7-point behaviorally anchored rating scales to aid subject matter experts in estimating the amount of each ability needed to successfully perform a job or task. If tasks are rated, an ability profile for a job can be taken as an average (or weighted average, e.g., by task importance) of abilities required across tasks.
Reliability and interrater agreement are well established (see Fleishman and Mumford, 1988), as is the construct validity of the taxonomy and methods (Fleishman and Mumford, 1989). The FJAS has been especially useful in the development of valid tests linked to job requirements (Fleishman and Mumford, 1988,1989).
Occupational Analysis Inventory and the General Work Inventory
The Occupational Analysis Inventory (OAI) is designed to be more relevant to occupational education and guidance, rather than to applied problems in the work setting, which are the focus of systems like the position analysis questionnaire (Cunningham et al., 1983). The inventory includes 617 items, called "work elements," divided across the five categories of information received, mental activities, work behavior, work goals, and work context. Each item is rated on one of four scales: significance, extent, applicability, or a special scale for that element. The three nonspecific scales are relative ratings with adjectivally anchored scale points, e.g., "to a very small extent" at the lower end of the extent scale and "to a great extent" at the higher end. The OAI is characterized as a research tool and it is stated that it is "advisable for the OAI job rater to have college-level reading comprehension, plus some preparatory orientation and practice with the instrument" (Cunningham, 1988:981).
Empirical work has been completed to evaluate the reliability and validity of the OAI (Cunningham, 1988; Cunningham et al., 1983). A study of the reliability of OAI ratings was conducted using 12 job analysts and 21 trained psychology graduate students who rated 215 jobs using written task descriptions from the
U.S. Employment Service. Correlations were computed between two independent raters for each OAI work element. The mean correlation was .53 and the median was .56. Several studies aimed at evaluating the construct validity of the OAI have been conducted, including the comparisons of clusters of occupations obtained with the OAI on several tests and inventories (68 of the 92 measures showed statistically significant discrimination between the clusters), the prediction of mean occupational scores on the General Aptitude Test Battery using OAI factor scores, (median cross-validated multiple correlations were .60 for mental and .24 for motor abilities), bivariate correlations between OAI attribute-requirement estimates and mean scores of job incumbents (statistically significant correlations at the .05 level were found for 38 of 55 analyses), and analyses of variance to relate OAI need-requirement estimates to job satisfaction scores (12 of 15 analyses provided supporting evidence).
The OAI shows generally excellent measurement characteristics, when it is applied in the recommended manner—using college educated, trained analysts. Most of the reported empirical work has been conducted using "paper jobs," that is, written job descriptions from the U.S. Employment Service. It is not clear that it would work as well if used in the field by job incumbents, supervisors, or other occupational experts, many of whom would not be college-trained or be available for special training on the OAI.
A replacement for the OAI, the General Work Inventory (GWI), is shorter and written less technically and could be a more practical alternative for large-scale data collection. This instrument was developed for use by "any literate respondent who is familiar with the job to be analyzed" (Cunningham et al., 1990:34). It has 268 items organized into 8 sections and uses "part of the job" and "extent of occurrence" rating scales, both of which have 9 points and are adjectivally anchored. Research using this inventory in the military showed mean retest reliabilities (for single raters) of .62 across all items, and a mean correlation of profiles of ratings (again, for single raters) of .74, comparable with other similar studies. Ballentine et al., (1992) used the GWI to create a hierarchical structure of Air Force occupations that showed intuitive meaning and corresponded to existing Air Force classifica-
tions, although the comparison was somewhat influenced by artifactual correspondence between the two systems.
The stream of work represented by the OAI and GWI demonstrates well the utility of using a descriptive system designed to be applied to the general population of occupations but still retaining enough specificity to provide meaningful differentiations between occupations, to link to assessments of persons, and to form useful occupational structures based on the information obtained from the system.
Common Metric Questionnaire
The common metric questionnaire (CMQ) (The Psychological Corporation, 1993) was developed by Harvey as a "worker-oriented" job analysis instrument designed to have applicability to a broad range of exempt and nonexempt jobs. It is organized into five major sections (general background, contacts with people, making decisions, physical and mechanical activities, and work setting) with several subsections in each. In addition to general background items that ask about respondent and job characteristics (e.g., tenure in present job, work schedule), the CMQ consists of 242 behaviorally specific items (e.g., in order to perform your job, do you use desktop or personal computers?). A matrix response format is used, such that if an item is indicated as performed, the respondent is asked to provide ratings for up to four additional scales (e.g., frequency, criticality, consequence of error). Thus, amount of information provided and amount of time needed to complete the instrument varies according to job scope and complexity.
A major advantage of the CMQ, according to its author, is the possibility of comparing even very dissimilar jobs by virtue of the instrument's common metric of work descriptors. This may be useful for purposes of establishing job progression and compensation systems. Broad applicability of the instrument is further supported by its use of an eighth-grade reading level, so that most job incumbents can complete it without assistance, and absolute rather than relative rating scales, so that responses can be compared across jobs. The CMQ can be scored in terms of 80 factor analytically-derived work dimensions or at the item level, thus
supporting human resource applications requiring relatively abstract (e.g., job classification) or specific (e.g., job descriptions) information.
The CMQ is a recent product, and there does not yet exist a substantial professional literature concerning its usage. The goals that Harvey set for the CMQ, however, particularly concerning ease of use and comparability of data across disparate jobs, are laudable and potentially fill a gap among worker-oriented job analysis instruments that preceded it.
Multipurpose Occupational Systems Analysis Inventory-Closed Ended
Developed by the Office of Personnel Management, the purpose of the Multipurpose Occupational Systems Analysis Inventory-Closed Ended (MOSAIC) is to collect data on a number of occupational descriptors in a standardized manner across occupations within large occupational families, and then to provide that information in readily accessible electronic databases. MOSAIC has been described as follows: "This system uses an automated occupational analysis approach that eliminates costly redundancies in the collection of data and provides technically sound and legally defensible procedures and documentation to support human resource management (HRM) decisions" (Gregory and Park, 1992:ii).
The report by Gregory and Park illustrates the use of MOSAIC. The occupation focus of the research project was executives, managers, and supervisors. A standard questionnaire was developed and administered to a stratified random, sample of over 20,000 federal executives, managers, and supervisors. The questionnaire contained a diverse set of items, or job descriptors, including: 151 job tasks rated in terms of importance for effective job performance; 22 competencies (a human quality or characteristic associated with the performance of managerial tasks, e.g., knowledge, skill, ability, trait, motive, or self-concept) rated in terms of importance, and needed proficiency at entry; and personal and organizational styles. Data were presented showing the percentage of respondents of various occupational types indicating they performed tasks, found competencies needed at en-
try, or were important for success. No data on interrater agreement were presented, but a 49 percent return rate was obtained.
Work Profiling System
The Work Profiling System (WPS), a product of Saville and Holdsworth (1990), is a worker-oriented job analysis instrument supported by expert system computer technology. The WPS is organized into two parts: job tasks and job context. The job tasks section consists of 325 behavior description items (called "tasks") organized into 8 sections (managing tasks, managing people, receiving information, thinking creatively, working with information, communicating, administrating, physical activities) and 30 subsections (e.g., planning/implementing, working with equipment/machinery). Examples of items ("tasks") include: planning a course or route for a journey or voyage; looking after the needs of young children; driving a car, van, or light truck. Items are rated on scales of time spent, importance, and effect of poor performance. Part two, job context, addresses 28 topics, such as education, training, and experience levels needed to perform the job, responsibility for financial resources, types of interpersonal contact, and job-related travel.
Goals for the system include providing an integrated and user-friendly system for job analysis and providing a knowledge base that can serve as the basis for matching people to jobs. Worker attributes are inferred from task ratings using an expert system derived from ratings of attribute-task linkages provided by experienced occupational psychologists. In addition to person-job match, this information base is intended to support such human resource applications as job descriptions, job classification, performance appraisal criteria, job design, and human resource planning.
As is the case with the CMQ, the WPS is a recent product that does not yet have a substantial professional literature concerning its usage. Its objectives as stated by its developers are ambitious, providing a comprehensive methodology for building human resource systems.