Cover Image

PAPERBACK
$108.00



View/Hide Left Panel
Click for next page ( 215


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 214
9 Conclusions and Recommendations The charge to the Committee on Occupational Classification and Analysis is to review the need for continuing the occupational analysis program of the U.S. Employment Service and its principal product, the Dictionary of Occupational Titles. The committee was asked to consider in executing this charge both the requirements of Employment Service operations and those of other users, public and private, for the kind of information provided. The preceding chapters have presented the evidence on which we base our conclusions and recommendations. CONCLUSIONS In terms of the charge, our conclusions are the following: 1. There is a strong and continuing need both within and outside the U.S. Employment Service for the kind of information provided by the Dictionary of Occupational Titles and certain other products based on it. 2. Substantial improvements in the procedures and products of the occupational analysis program are required in order to meet the national need for occupational information. 214

OCR for page 214
Conclusions and Recommendations 215 Conclusion 1, the continuing need for a document that provides occupational information, takes into account three functions of the DOT: as a dictionary, as a classification system, and as a source of material on occupational characteristics. DICTIONARY The DOT is first and foremost a dictionary, which defines more than 12,000 occupations through descriptions of their work content and cross-refer- ences an additional 16,000 occupational titles to these 12,000 defined occupations. As such it provides a common understanding as to what is meant when a particular occupational title is used; it is by far the most comprehensive source of occupational definitions available in the United States. This aspect of the DOT is of very great importance to a wide variety of users, as chapter 4 details. We believe that there would be almost unanimous agreement that such a document, providing a standardized terminology and standardized definitions of that terminology, is essential. Is it, however, specifically essential to the Employment Service's goals- its placement and counseling operation? We believe that it is. Some proponents of the matching of jobs and applicants by computer have suggested that keywording obviates the necessity for defined titles, since descriptions of a particular job and of a particular worker's attributes can be entered directly into the computer matching system without the intervening mechanism of a title. Such a conclusion seems to us unrealistic because it fails to recognize the role that the occupational title plays in everyday language and in the labor market. The occupational title is shorthand (or, perhaps better, "short talky. An employer placing a job order for a Computer Programmer does not expect to describe what a programmer does but only the particular requirements, within the general category of programmer, for a particular job. An applicant with experience as a Lumber Scaler is certainly better served if the placement interviewer knows or can find in the dictionary what a lumber scaler does, because local terminology may vary and because the interviewer may then be able to suggest other occupations that make use of similar skills. For this reason, then, a document that defines terms is essential to the Employment Service's operation; some mechanism for constantly revising such a document must be maintained as new terminology comes into use and new activities arise.

OCR for page 214
216 CLASSIFICATION WORK, JOBS, AND OCCUPATIONS The basic purpose of the classification structure of the DOT iS to organize occupational titles and definitions in an order that facilitates the matching of job applicants and jobs, by grouping together jobs and occupations that are relatively interchangeable in terms of the requirements they make of a worker. In the terminology of the occupational analysis program an individual worker holds a position; the set of positions in which workers perform essentially the same activities within a particular establishment is called a job; and the set of jobs in which similar activities are performed across a number of establishments is called an occupation. Jobs are known by many names, and hence a procedure is needed to group together similar jobs with different titles. The 12,099 occupations defined in the fourth edition DOT constitute a classification of a much larger number of jobs those held by some 100 million workers in the American labor force. If all job applicants knew exactly what jobs they were qualified and willing to perform, the classification structure of the DOT could be restricted to grouping job titles into occupational categories. However, many workers are in fact able to do different kinds of work. To optimize their employment opportunities, a classification structure is needed that links together all the occupations in which a worker with particular skills and qualifications might reasonably be employed. To serve as an effective job placement tool then, the DOT must be organized in such a way. In addition, the DOT classification should also be compatible with other widely used classifications to facilitate the reporting and comparison of occupational statistics. OCCUPATIONAL CHARACTERISTICS Closely related to the classification system are the attributes of occupations and of workers that the Employment Service calls worker functions and worker traits. These attributes provide information on such items as training time, working conditions, physical effort, etc. As chapters 3 and 4 detail, this information is used for many purposes, including vocational guidance, job placement, rehabilitation counseling, and the determination of program eligibility for training funds. Moreover, it is clear that the worker functions and worker traits would be even more widely used if these data were more readily available and if additional characteristics were measured. In sum, the DOT serves as the major source of occupational data currently available and would be sorely missed if it were discontinued. The need for the kind of information that is contained in the DOT iS confirmed

OCR for page 214
Conclusions and Recommendations 217 by the extent of its distribution. Over its 13-year life (1965-1977), 148,145 copies of the third edition DOT were sold by the U.S. Government Printing Office, and in the first 21 months of availability (through September 1979), 1 15,1 15 copies of the fourth edition DOT have been sold. Evidence for conclusion 2 that substantial improvements are needed in the occupational analysis program is found throughout the report: chapter 4 identifies the kind of occupational information that is needed but not currently available; chapter 5 identifies various organizational difficulties in the program; chapters 6 and 7 evaluate the procedures used to collect the occupational information contained in the DOT as well as its quality; and chapter 8 assesses the classification structure of the DOT from the standpoint of its usefulness in matching workers and jobs. The material presented in these chapters leads the committee to conclude that data collection procedures are deficient in important respects, particularly in the way in which occupations are selected for observation and analysis and in the way in which worker trait and worker function ratings are measured. Furthermore, the current classification structure of the DOT does not appear to be optimal for the purpose of matching jobs and workers, nor does the proposed keyword system appear to be an adequate substitute. Our conclusions that there is a strong need both to continue and to improve the DOT lead us to 3 general recommendations intended to strengthen the occupational analysis program and to 19 specific recom- mendations intended to improve the quality of the DOT and, more generally, to facilitate the development of occupational information of high quality. GENERAL RECOMMENDATIONS 1. The occupational analysis program should concentrate its efforts on the fundamental activity of job analysis and on research and development strategies for improving procedures, monitoring changes in job content, and identifying new occupations that are associated with the production and continuous updating of the Dictionary of Occupational Titles. The program should discontinue the publication of career guides. In the judgment of the committee, too much of the energy and resources of occupational analysis staff, both in the national office and the field centers, has been diverted from the central mission of the occupational analysis program: the production of the Dictionary of Occupational Titles (see chapter 5~. Primary attention should be devoted to research designed to improve the quality of occupational data, the management and execution

OCR for page 214
218 WORK, JOBS, AND OCCUPATIONS of the very complex data collection effort, and the preparation of supplements to and new editions of the DOT. Other appropriate tasks- insofar as they do not distract from the main task include the preparation of special reports for other agencies based on DOT data and training and technical assistance on the use of the DOT. The production of career guides and brochures should not be continued as a function of the occupational analysis program. Such activities should be the responsibility of other agencies currently engaged in this type of information dissemination. At the national level the products of the occupational outlook program of the Bureau of Labor Statistics are widely used; at the state and local level the recently organized career information services program, with its links to vocational education and other relevant state systems, provides information to state residents on employment opportunities available in their own localities. Both organizations are dependent on data gathered by the occupational analysis program, and strong communication channels among these agencies are essential. The division of labor between data gatherers and those charged with disseminating information to the public is a rational one, however, which will lead to better use of the quite different specialized skills called for in each of these responsibilities. Similarly, the Job Search Branch of the Division of Occupational Analysis should be relocated. The Job Search Branch is an effective unit, but it relies on information furnished by local Employment Service offices and has no particular connection to the major activities of the occupation- al analysis program. Moreover, its presence in the Division of Occupation- al Analysis may distract resources from the occupational analysis activities that should be the primary concern of the division. 2. A permanent, professional research unit of high quality should be established to conduct technical studies designed to improve the quality of the Dictionary of Occupational Titles as well as basic research designed to improve understanding of the organization of work in the United States. A number of the recommendations below relate to research needed in specific areas in order to strengthen the occupational analysis program. In our judgment, however, the gravest difficulty lies not in specific areas but in the general lack of a research orientation. The early editions of the Dictionary of Occupational Titles were at the forefront of the occupational analysis of their time. For later editions this is no longer true: the program has been allowed to stagnate. It will not become a vital force again unless the importance of quality research, well integrated into the academic

OCR for page 214
Conclusions and Recommendations 219 disciplines providing the basic foundations for occupational analysis, is recognized.) While the committee is not prepared to make detailed recommendations regarding the location of a research units or the exact size of its staff, we have firm opinions regarding the considerations that should be kept in mind in developing such a unit. First, we envision a unit with a relatively large, high-level stab, of the order of 10 Ph.D.-level scientists (sociologists, economists, psychologists, and statisticians), perhaps an equal number of B.A.- or M.A.-level research assistants, and a sufficient number of support staff. We thus envision a research unit that is larger than the current Occupational Analysis Branch in the national office of the Division of Occupational Analysis. We recognize that this is a period of budgetary restraint, but we would be derelict in our responsibiity if we did not express our strong conviction regarding what is needed for a viable federal occupational analysis program simply because of current (and perhaps short-run) budgetary limitations. 3. An outside advisory committee to the occupational analysis program should be established. Its members should be appointed by the Assistant Secretary ofLaborfor Employment and Training. This outside advisory committee should include representatives of employ- ers and of unions familiar with the problems of occupational classification and placement, persons from relevant academic disciplines, and members of the public. It should meet periodically, perhaps twice a year, to receive and review reports on the work of the occupational analysis program and to make recommendations on future activities.3 Four considerations underlie this recommendation: iA good example of a successful research capability within an operating agency is to be found in the Bureau of the Census. High-quality technical studies are produced by the Census Bureau on a continuing basis; staff regard themselves as professional social scientists and statisticians, have close ties with their academic disciplines, regularly attend professional meetings, and are frequently drawn from or move to academic positions. 2The committee spent some time discussing alternative organizational arrangements, ranging from the establishment of a new unit within the Division of Occupational Analysis to the creation of an entirely independent occupational research institute within the federal government but outside the Department of Labor. In the end, however, we decided that we did not have the necessary organizational knowledge to advise on the optimal mechanism for creating an occupational research capability, although we are firm in our judgment as to its necessity. 3Again, the Census Bureau provides a good example. The Advisory Committee on Population Statistics, which meets twice a year, plays an active role in recommending and reviewing procedures. Its members are drawn from representatives from the Population Association of America and other interested groups.

OCR for page 214
220 WORK, JOBS, AND OCCUPATIONS We believe that the periodic reporting to an informed outside advisory committee would have a salutary effect on the planning and organizational efficiency of the national office of the Division of Occupational Analysis or any successor unit. Such an advisory committee would help to prevent the research that is essential to the program from becoming swamped by the exigencies of operational considerations in an agency (the U.S. Employment Service) whose primary focus is operational. It is our impression that the needs of the occupational analysis program for adequate staff, in particular in the national office, have not received sufficient attention in the past. An outside advisory committee would strengthen the position of the program by providing it with a constituency. In our view, the occupational analysis program has not been successful in communicating its goals or its problems to those groups standing to benefit most from its activities. A public advisor committee would provide some liaison to these groups and help to enlist their cooperation. Finally, all organizations, inside or outside the government, tend inevitably to develop procedures that acquire a sacrosanct status unless they are moderated by outside influences. An outside advisory group could raise questions that force the staff to consider the usefulness of established procedures. SPECIFIC RECOMMENDATIONS The remainder of this chapter presents a set of recommendations suggesting ways to improve the Dictionary of Occupational Titles and, more generally, to facilitate the development of high-quality occupational information. Recommendations 4-8 concern data collection procedures; recommendations 9 and 10 concern the worker function and worker trait scales; recommendations 11-13 concern the classification structure of the DOT and the keyword system; recommendations 14 and 15 propose needed areas of research; and recommendations 1~22 deal with various organiza- tional and administrative issues. DATA COLLECTION PROCEDURES 4. On-site observation of job performance by trained occupational analysts, including interviews with workers and supervisors, should continue as a major mode of data collection; experimentation with other data collection procedures, however, should also be undertaken.

OCR for page 214
Conclusions and Recommendations 221 In the judgment of the committee a major strength of the DOT iS that the descriptions it contains are based on the analysis of specific jobs rather than on abstract descriptions of occupational categories. We encourage the continuation of this mode of data collection. A number of considerations have led us to this recommendation. Chief among them is the need for standardization in the identification of significant tasks, the use of terminology, and the writing of descriptions. Standardization of procedures requires the services of analysts trained to observe in a larger context than an individual firm. As an increasing proportion of jobs are found in the service sector, where variations in activities are less constrained by the requirements of the machinery and equipment that dominate the production sector, the need for standardiza- tion will probably become even greater. We find additional support for our position in the requests of private firms and governmental units (cited in chapter 5) for assistance from field centers and national office analysts in developing classification systems for their employees. It may in some cases be possible, however, to collect equally useful data via a written instrument a questionnaire, checklist, or task inventory. Attention should be devoted to developing a repertoire of data collection techniques by exploring the conditions under which each is most effective and using the optimal technique for each situation. 5. Staffing schedules for establishments in which job analyses are performed should continue to be collected and should be used for research purposes. The recently discontinued tabulation by sex of the number of workers in each occupation should be reinstated. Staffing schedules, which outline the distribution of jobs within establish- ments, are currently used only to identify activities unique to an industry or establishment. In our judgment, however, they have value for at least three other purposes that would substantially improve the occupational analysis program. First, staffing schedules could be used as a check on the representativeness of establishments selected for job analysis by comparing staffing schedule data with the occupational structure of industries revealed through other sources (for example, the decennial census and the occupational employment survey of the Bureau of Labor Statistics). Second, they provide a tool that if properly used could alert occupational analysts to significant changes in occupational structure that may indicate concomitant changes in work content. Third, staffing schedule data are a potentially rich source of information on the differences in occupational opportunities for men and women. Recently, however (in November

OCR for page 214
222 WORK, JOBS, AND OCCUPATIONS 1978), the occupational distribution of workers on the staffing schedule ceased to be tabulated separately by sex. In our judgment this change is unfortunate, since it destroys the usefulness of staffing schedule data for an extremely important research purpose. We urge that separate tabulations by sex be reinstated. 6. The selection of establishments and work activities for which job analyses are performed should be made according to a general sampling plan designedfor the particular requirements of occupational analysis. The committee recognizes that the variation in the number of job analyses per defined occupation documented in chapter 7 is not prima facie evidence of maldistribution of effort. Some occupations are clearly homogeneous in work content regardless of their geographical or industrial setting, whereas the homogeneity or heterogeneity of other occupations can be determined only by comparative job analyses. We can find no evidence, however, of the use of systematic procedures in the selection of sites, in the selection of jobs to be analyzed, or even in the designation of industries to be included. The task of the national office is to assign industries to field centers according to geographic concentration or, for those industries that are widely dispersed, to obtain geographic representation. The task of the field center, once an industry has been assigned, is to select establishments that represent different size units (in terms of aggregate employment levels) and/or known technological variations. As chapter 7 shows, both goals are very generally stated, and no clear procedures are established for attaining them. An example of this lack of clarity in the procedures followed is the assignment of industries to field centers by the national office. Industry assignments vary widely in scope: an assignment may be as wide as "retail trade," a category covering establishments engaged in diverse activities, or as narrow as "button," covering establishments engaged in "manufactur- ing buttons, parts of buttons, button blanks, etc." Neither the basis for the national office's decision to make an assignment broad or narrow nor the procedure by which a field center decides among the possibilities in an industry of broad scope is clear. The procedures involved in the selection of jobs for analysis are also unclear. The identification of the types of organizations that have unique types of jobs should be an important goal, but current practice appears to be founded on the premise that an establishment's product is a major distinguishing characteristic of its jobs, a premise that reflects the long- standing emphasis of the DOT on manufacturing jobs and their close association with specialized equipment. One consequence of this emphasis i

OCR for page 214
Conclusions and Recommendations 223 is that jobs outside the production sector are generally assumed to require fewer job analyses than those concerned with fabrication. Job analysis is therefore often not undertaken throughout an establishment but is confined to those jobs assumed to be unique to it. Although this limitation is not unreasonable, it may result in the self-perpetuation of an assumption that is no longer accurate. The fine line between unnecessary duplication of job analyses and unsupported assumptions of homogeneity is not easy to draw, but the resolution of this problem must receive attention as the economy shifts increasingly from production to service activities. In our judgment the set of procedures involved in the selection of jobs for analysis should be thoroughly overhauled so that data can be collected that are truly representative of work content. The Employment Service should seek technical assistance in designing procedures that are both consistent with its needs and statistically sound. (This is a logical function for an occupational research unit, perhaps with the participation of outside consultants.) 7. Procedures should be designed to monitor changes ire the job content of the economy. Both new occupations and changes in existing occupations should be identified. As we have noted in chapters 6 and 7, the fourth edition DOT appears to provide better coverage of occupations in traditional sectors of the labor market than in rapidly expanding sectors. We suspect that this is due to the way jobs are selected for analysis. To correct this tendency, we believe that procedures should be developed to monitor explicitly changes in job content in the economy. We consider first the problem of identifying new occupations. There are several ways this might be done. A range of sources could be continuously or periodically monitored to identify occupational titles not already included in the DOT. Potential sources include the occupational employ- ment surveys of the Bureau of Labor Statistics, job orders received by local Employment Service offices (indeed, such job orders are already a major source in the form of occupational code requests), classified ads in major newspapers, and the Current Population Survey (cPs) conducted monthly for the Bureau of Labor Statistics by the Bureau of the Census. Because of its rich potential we urge exploration of ways to use the cPs to monitor the emergence of new occupations. It should be noted that three past cPs samples have been assigned DOT codes by occupational analysts at the field centers (those from April 1967, April 1971, and March 1978 the last still in preparation). Preliminary experimentation could be undertaken using these surveys.

OCR for page 214
224 WORK, JOBS, AND OCCUPATIONS Second is the problem of how to identify changes in the content of existing occupations. This is more difficult, since there is no good way to know in advance of analysis whether the content of an occupation has substantially changed. It may be possible, however, to develop an information network using industrial, trade, and professional associations, labor unions, etc. to keep abreast of rapidly changing occupations. Moreover, it is likely that in those sectors of the occupational structure in which many new occupational titles are emerging there is also rapid change in the content of existing occupations. The national office should develop a monitoring system for identifying sectors of the occupational structure in which there is rapid change, in order to target the occupations in such sectors for intensive analysis. 8. The Dictionary of Occupational Titles should be expanded to include definitions of all occupations in the economy, whether or not they are serviced by the Employment Service. As chapter 4 documents, the DOT iS widely used outside the Employment Service because it is the most comprehensive source of occupational information available anywhere. As such it should attempt to be complete in its coverage of the occupations practiced in the United States today. The fact is, however, that it is very uneven, covering some occupations in great detail and others not at all. Several sources can be used to identify occupations not currently included in the DOT: the Current Population Survey described above, the Census Bureau's Alphabetical Index of Occupations and Industries, classifications of military occupational specialties, the federal government's occupational coding schemes, and various specialized occupational glos- saries. These lists should be compared with the list of occupational titles in the DOT. Any title found in another list but not in the DOT would then become a candidate for an intensive job analysis. Procedures should be designed to locate suitable jobs for analysis once they are identified. MEASUREMENT OF OCCUPATIONAL CHARACTERISTICS 9. The worker trait and worker function scales should be reviewed and, where it is appropriate, replaced with carefully developed multiple-item scales that measure conceptually central aspects of occupational content. The committee has found substantial reason to question the adequacy of the worker trait and worker function scales. First, they do not appear in

OCR for page 214
Conclusions and Recommendations 225 the aggregate to adequately reflect conceptually central aspects of occupational content. They omit, for example, measures of such important features as the organizational setting in which jobs occur and the degree of responsibility entailed in jobs for decisions, materials, or supervision.4 At the same time they include measures of interests, aptitudes, and tempera- ments, which are better thought of as worker characteristics than as attributes of jobs. Second, the existing scales have not been developed or validated in accordance with current psychometric standards for scale construction, and some of them have been shown (see chapter 7 and Appendix E) to have rather low reliability. Moreover, they are very redundant. In chapter 7 we show that most of the variation among occupations can be described by three factors, and almost all the remaining variation by an additional three factors. Third, many of the scales have limited use, as chapters 3 and 4 document. In part, this is the result of the way they are published. Although scores on the worker function scales (DATA, PEOPLE, and THINGS) are available for each DOT occupation because they are included as part of the occupational classification code scores on worker trait scales~for each occupation in the third edition DOT are available only in supplements. Ranges of scale scores are also published for groups of occupations in volume 2 of the third edition DOT. Scores on the worker traits scales for the fourth edition had not been published as of January 1980, although they are available on computer tape.5 The development of a new set of scales of occupational characteristics is a research activity that should be undertaken prior to the publication of the next edition of the DOT and then continued as an ongoing activity of the research unit. The first step is to determine what occupational information is needed by major users of the DOT, including the Employment Service. Suitable scales to elicit this information should then be developed and validated using standard psychometric procedures. Responsibility and supervision are highly relevant for job placement and for other purposes as well, including the analysis of career ladders (identified by many respondents to the user survey as highly desirable information to add to the DOT) and equal employment opportunity 1SSUeS. 5A tape containing all of the worker traits for the fourth edition (known as the DOT master tape) may be obtained from the National Technical Information Service (Document No. PB 298 315/AS).

OCR for page 214
226 WORK, JOBS, AND OCCUPATIONS 10. A research activity of first priority should be review of the training time (GED and SIPS, physical demand, and working condition scales. Our review of DOT uses indicates that the training time, physical demand, and working condition scales are used widely for making key determina- tions in a variety of employment-related programs by government and other agencies. In some instances the worker function (DATA, PEOPLE, and THINGS) scales are inappropriately used as substitutes for training time scales. This may occur because of lack of knowledge of the worker trait scales, since worker functions are included in the basic occupational code, while worker traits are treated as separate dimensions and, in the third edition DOT, were published in supplementary volumes. We believe the need for and interest in these occupational characteristics are sufficient to warrant continuous effort and special publication by the occupational analysis program. CLASSIFICATION ISSUES 11. A major activity of the occupational analysis program should be investigation' of cross-occupational linkages that indicate possible transferability of skills or experience. Hitherto, the occupational analysis program has done comparative job analysis only to the extent necessary to fit jobs into occupational units within the established classification. The implicit assumption with respect to matching workers and jobs has been that the classification structure itself will reveal the range of possible matches. In our judgment this is too narrow a use of the occupational analyst's skills and too rigid a conception of what constitutes "similar" work. An informed glance through the detailed occupational classification of the DOT reveals a number of instances in which similar work performed in different work settings results in two codes that differ at the most aggregated, one-digit level. This experience can probably be repeated with any classification system yet devised. A number of procedures to aid in identifying occupations for which the required tasks are sufficiently alike to permit transfer of skills could be proposed. Two that appear to have special promise are (1) the comparative analysis of skill requirements via task analysis or other structured job analysis procedures and (2) the empirical identification of"interchange- able" occupations via the analysis of rates of naturally occurring occupational mobility. The basic idea in the latter proposal is that if people

OCR for page 214
Conclusions and Recommendations 227 who leave a particular occupation are especially likely to take up certain other occupations, then those occupations are probably similar in their requirements, and hence job applicants with experience at one occupation could work at the others. Job history data are currently collected routinely from job applicants by local Employment Service offices. These data could be used to estimate rates of movement between occupational categories that are specific to local labor markets, and the validity of the suggested interchangeability could be reviewed by trained analysts. If valid linkages emerged, those occupations with high interchangeability rates could be listed together in job banks, could be matched in the keyword or other automated systems, and could be listed for the use of job placement interviewers. It is important to note that classifications for placement purposes need not list each occupation only once. For example, a job opening could be included at several different places within a job bank to facilitate the job search process, in much the same way that books are cross-referenced in a library catalogue. We urge full exploration of these possibilities. 12. The development of an automated procedure for matching job applicants with job openings should continue, but the current keyword system should not be accepted as optimal. Appendix G presents an evaluation of the keyword system, the most widely used method of computerized job matching attempted by the Employment Service. The conclusions in Appendix G support the findings of critics who have called the system inadequate and inadequately tested prior to its implementation. We wish to emphasize, however, the need for continued research and experimentation in the use of automated data processing in both the job analysis and placement operation of the Employment Service. The exploratory work done by our staff (presented in Appendixes G and H) is suggestive of the potential inherent in this tool for assessing and developing classifications. Time and resources have limited the extent to which this exploration could be undertaken, but we are convinced of its long-term value. Experimental work in computerized job matching should continue in tandem with the development of an improved classification. In this the experience gained from the keywording operation should be carefully evaluated. For example, the "complementary terms" concept used in keywording may present an alternative to the very detailed and probably overly inflexible coding system now used in the DOT. A simplified set of

OCR for page 214
228 WORK, JOBS, AND OCCUPATIONS occupations associated with a range of complementary terms may serve the purposes of placement better than either the nine-digit DOT code or the occupational unit framework of the keyword system. In urging additional experimental work, we wish to caution against the precipitous, large-scale implementation of poorly or incompletely tested schemes. The optimal strategy would be to conduct a series of small-scale studies before adopting any particular scheme. The classification system developed for the next edition of the DOT should be compatible with the standard system implemented by the Office of Federal Statistical Policy and Standards or its successor coordinating federal agency. That is, explicit procedures should be developed to enable the translation of occupational codes so that information can be organized and reported using a standardized classification. The relationship between the classification system used by the Employ- ment Service, embodied in the DOT, and that used by other governmental agencies is a crucial issue. The committee believes that arguments for a standardized classification for reporting occupational data are so compel- ling as to leave no doubt of the importance of this goal. Within the context of this report, the need for the Employment Service's operating statistics to be part of a standardized system is clear. Therefore an essential task is to ensure that occupational information generated by the Employment Service can be translated to allow reporting in terms of a standardized occupational classification. Congress has established a National Occupational Information Coordi- nating Committee (Public Law 94-482; October 12, 1976), which has as one of its responsibilities the development and implementation of an occupational information system " . . . which system shall include data on occupational demand and supply based on uniform definitions, standardized estimating procedures, and standardized occupational clas- sifications.... " Beyond this legislative requirement the committee believes that an understanding of the Employment Service's role in the labor market is essential to its proper functioning and that for such an understanding, Employment Service operating statistics must be related to aggregate data for the labor force. Without a standardized classification system this connection is impossible to make. We believe that the occupational analysis program should take a lead role in providing the material and expertise required to keep the Standard Occupational Classification (soc) up to date-a role that is compatible with its activity in developing the soc.

OCR for page 214
Conclusions and Recommendations OTHER NEEDED RESEARCH 229 14. Research priority should be given to developing criteria for defining "occupations"-the aggregation problem. What is an occupation? It is a set of jobs that are similar in some way, in terms of tasks, duties, responsibilities, organizational or industrial setting, status, etc. Occupational classifications group occupations in terms of their similarity according to one or several of these criteria or still others. Classifications differ in two ways: first, in terms of the criteria of similarity, the grouping principle; and second, in terms of the level of aggregation, the number of distinctions that are made between elements, or occupations, in the classification. The 1970 Census classification, for example, contains 441 occupations, while the fourth edition DOT contains 12,099 occupa- tions. Obviously, the census occupations on the whole encompass a more heterogeneous set of jobs than do the DOT occupations. Despite the greater specificity of DOT occupations, however, there appears to be great variation from one occupation to another in their degree of heterogeneity. For example, there are 70 kinds of Sewing Machine Operator, Garment, with the same 6-digit code (786.682), while there are 6 kinds of Secretary with the same 6-digit code (201.362~. Moreover, inspection of the occupational definitions suggests more variability among the 6 secretarial occupations than among the 70 kinds of sewing occupations. There appears to be no conceptual basis for delineating boundary lines between occupations. Research is needed both on the conceptual basis for defining occupa- tions and on the consistency with which occupational boundaries are drawn in the fourth edition DOT, to provide a basis for revisions in the fifth edition. In undertaking a review of the existing occupational categories in the fourth edition DOT, attention should be paid to the possibiity that certain categories of occupations (e.g., clerical or service occupations) are insufficiently differentiated, or that certain categories (e.g., benchwork occupations) are overly differentiated. We urge exploration of strategies for reviewing the consistency of specificity of DOT occupations. 15. Basic research should be undertaken on the operation of labor markets to improve understanding of the processes by which workers acquire jobs. The Employment Service could do a great deal to improve its ability to place workers in jobs through research on the processes by which workers acquire jobs. In chapter 8 we proposed an empirical procedure for defining

OCR for page 214
230 WORK, JOBS, AND OCCUPATIONS interchangeable clusters of jobs on the basis of occupational mobility rates. The usefulness of such a procedure would be even greater with a better understanding of occupational mobility processes in general. What kinds of jobs tend to be open to workers with particular sorts of experience? Which jobs are filled by those who have previously worked elsewhere, which are filled by those just entering the labor force, and which are filled only by promotion from within an establishment? To what extent do sex, age, or race continue to be barriers to occupational opportunities, and are such barriers concentrated in particular sectors of the labor force? This research is likely to be most fruitful if it builds on institutional and segmented market approaches to labor market analysis, since these approaches focus on the very job and market structures that are at issue here. The Employment Service, in particular the occupational analysis program, is in a unique position to conduct research on such questions. Job history data currently collected routinely in the course of job placement interviews and establishment studies currently conducted on a regular basis for the purpose of job analysis are valuable sources of data that should be exploited in the interest of improving the ability to match workers and jobs. These data sources should also be exploited to improve understanding of career progressions, typical patterns of movement from job to job. When respondents to the survey of DOT users were asked how future editions could be improved to meet their needs better, the inclusion of career ladders was most often mentioned; the majority indicated that they would find such information helpful. While there undoubtedly is substantial variability in career progressions, some indication of typical sequences of jobs would be very useful for counseling purposes. Two existing data sources could be used to produce such information. First, the job analysis schedules used by occupational analysts include information on the relation of the job being analyzed to other jobs specifically, promotion lines, transfer lines, and lines of supervision. This information could be used to describe typical career ladders within enterprises. Second, the work history data collected routinely from job applicants in local employment service offices could be used to describe typical career ladders involving mobility among enterprises, in the manner discussed in chapter 8. We urge that these possibilities be explored.

OCR for page 214
Conclusions and Recommendations ORGANIZATIONAL AND ADMINISTRATIVE ISSUES 231 16. The leadership of the national office in the occupational analysis program should be strengthened; greater attention should be given to coordination of field center activities; and the lines of federal authority should be clearly established. In view of the intensive management study (Booz, Allen & Hamilton, Inc., 1979) commissioned by the Department of Labor in tandem with its request to the National Academy of Sciences for a study of long-range needs, the committee has concerned itself only with those aspects of organization that are directly related to the substantive content of the occupational analysis program. In this context we strongly endorse the Booz, Allen & Hamilton conclusion that strong leadership and increased coordination by the national office are essential. Throughout our report (notably in chapters 5 and 6) are specific instances of the costs that lack of leadership by the national office have produced in terms of quality. We particularly support the follow-up recommendation of the Office of Technical Support (U.S. Department of Labor, 1979b) that a written agreement between the Employment and Training Administration and the host state of each occupational analysis field center lay out clearly the rights and preroga- tives of the federal government in the control of field center activities. The committee is not persuaded, however, that the Booz, Allen & Hamilton recommendation that the number of field centers be halved is, in the long run, a wise one. Although in the short run such a reduction may be a useful way to eliminate those field centers whose contribution to the program has, for a variety of reasons, been below the desirable level, in the long run, geographical dispersion seems to us to be a strength, particularly in view of the new trends in population dispersion currently taking place in the United States. The problems of coordination by the national office may be reduced by a reduction in the number of field centers, but the problems of communication between the occupational analysis program and local office operating staff will certainly be increased. 17. The collection and dissemination of occupational information by the occupational analysis program should be a continuous process; activity should notfluctuate with the timing of new editions of the DOT. This recommendation follows from recommendation 1-that the program should concentrate its effort on job analysis. Chapters 6 and 7 present evidence of the costs, in terms of thoroughness and quality, of gearing the

OCR for page 214
232 WORK, JOBS, AND OCCUPATIONS program so closely to the publication of the new edition. Beyond this, however, we believe that to be most useful to the Employment Service's operating offices and to other users, occupational information should be kept current by closely monitoring the introduction of new jobs and changes in the content of existing jobs. 18. Procedures followed in collecting data and developing the DOT should be carefully documented and publicly described. The committee found that many procedural decisions appear to be made on an ad hoc basis and to be poorly documented. The lack of documentation, experimentation, and research on the efficacy of the procedures used seems to the committee to be one of the most serious deficiencies in the occupational analysis program. Although we recognize that the Employment Service is an operating agency whose purpose is to deliver service, such a service cannot be delivered for the highly complex and continuously changing world with which the Employment Service deals on the basis of ad hoc decisions that are never documented or systematically communicated to persons in operational roles. The lack of documentation makes the review and evaluation of Employment Service occupational information difficult for users, who should be supplied with this essential information. 19. The data produced for the DOT should be made publicly available. As well as being underdocumented, the DOT iS underpublished, in the sense that a great deal of material of great value to researchers is not made easily available. Public-use computer tapes and attendant documentation should be created for each of the data sets used in the preparation of new editions of the DOT and deposited in data archives such as the National Technical Information Service and the Inter-University Consortium for Political and Social Research at the University of Michigan. For example, the third-to-fourth edition map, and the 1966 and 1971 cPs tapes coded with DOT codes (and the March 1978 cPs tape when it becomes available) should all be made publicly available.6 Public access to data used in preparing the DOT can do nothing but improve the quality of the DOT. 6The DOT master tape containing all of the worker trait codes for the fourth edition is already available (see note 5). In addition, the committee deposited two magnetic tapes with the National Technical Information Service and the Inter-University Consortium for Political and Social Research: (1) the April 1971 Current Population Survey (N = 60,441), which includes third and fourth edition DOT codes, and (2) a summary tape of DOT occupational characteristics, which was created from the 1971 cPs and provides average DOT scores and factor-based scale scores for the expanded (N = 574) 1970 Census occupational classification (for details, see Appendix F).

OCR for page 214
Conclusions and Recommendations 233 Second, for the benefit of users without ready access to computers, data on the characteristics of each occupation (currently the worker traits and worker functions) should be published, with exact scores for each DOT occupation. 20. A tabulation program should be instituted immediately to aggregate monthly data from Employment Service operations to the revised Standard Occupational Classification unit groups used in the 1980 Census of Population and subsequent current population surveys. The current version of the Standard Occupational Classification has attempted to provide an interim solution to the problem of compatibility by allocating each of the 12,099 nine-digit codes of the fourth edition DOT to one of the approximately 600 four-digit unit groups of the soc. A similar crossover listing between the classifications, to be used in the 1980 Census of Population and the sac unit groups, has been developed by the Bureau of the Census. It therefore becomes possible, if computerized operating statistics are available at a nine-digit level, to rearrange these data into the census classification (or any other classification system providing such a crossover listing). Both recommendation 20 and recommendation 13 are closely related to the congressional instruction to the secretary of labor to institute a uniform reporting program, under the Comprehensive Employment and Training Act, using a detailed occupational or training code, a term defined as "any occupational or training code equivalent in detail to the Standard Occupational Classification at the four-digit level" (Public Law 95-524; October 27, 1978; Section 313 (g)~31~. 21. A systematic program should be instituted to communicate additions and revisions of occupational definitions and their classification promptly to all operating staff in the Employment Service as well as to other interested persons. It is crucial to the successful operation of the Employment Service and to other major users of the DOT as well that the occupational information provided by the DOT be up to date. It is in those sectors of the occupational structure that are most rapidly changing that the need for information is greatest. For this reason it is insufficient to rely on the periodic publication of new editions of the DOT. A mechanism should be established to transmit information continuously on new and changing occupations and on newly established linkages between occupations to all concerned persons. What we have in mind is a monthly news bulletin, issued by the occupational analysis program and circulated to all Employment Service personnel and

OCR for page 214
234 WORK, JOBS, AND OCCUPATIONS to other interested parties, and an annual supplement to the DOT incorporating all such information produced in the preceding year. 22. The next edition of the DOT should not be issued until substantial improvements in the occupational analysis program have been made, following the recommendations made here. There is no need to rush to a fifth edition of the DOT, especially if a program of continuous updating and dissemination of occupational information is developed as proposed above. Such a program would serve the needs of users for up-to-date occupational information by keeping the fourth edition current. This would permit time for a fifth edition to be fundamentally redesigned on the basis of the research proposed here-on the classification structure, the measurement of occupational characteris- tics, the definition of occupations, data collection procedures, and so on. We would expect such research to continue indefinitely and to serve as the basis for further modifications of subsequent editions of the DOT. Hence we are not proposing delay until completion of a single massive research effort, but rather delay until a permanent, ongoing research effort has been well begun and has borne fruit.