12
THE 1971 CENSUS AND THE SOCIOECONOMICCLASSIFICATION OF OCCUPATIONS* PETER C. PINE0 IMcMaster University JOHN PORTER I Carleton University HUGH A. MC ROBERTS I Carleton University Apartirdurecensementde 1971, laclassificationdes professionsa subideschangements importants ; donc, les sociologues devraient developper des methodes nouvelles pour entreprendre I’analyse des resultats. Les auteurs discutent et evaluent la nouvelle classification et la (< Classification canadienne descriptive des professions )), et proposent une reorganisation des groupes de base a fin de formuler une classification socio-economique des professions plus utile que les grands groupes et sous- groupes du recensement. Radical changes in the occupation coding system for the 1971 Census of Canada mean that new strategies must be developed by sociologists to analyse census results. The structure of the new code, and of the Canadian Classification and Dictionary of Occupations, are discussed and evaluated. An alternative arrangement of the Unit Groups is proposed and demonstrated to be a more valuable socioeconomic classification of occupations than the Major and Minor Groups of the census. Researchers who have looked at the occupa- tional data of the 1971 Census of Canada will have noted the radical changes which have been made in the occupational codes and classifica- tions. Since the earlier occupational data were often criticized by sociologists who tried to use them (Pineo and Porter, 1967), it is important to determine whether or not the 1971 Census codes based on the Canadian Classification and Dictionary of Occupations (the CCDO codes as they are known) are any improvement. In the present paper we attempt to make such an as- sessment after an examination of how the codes were created. Our particular interest is the use of occupa- tional data for stratification research. We con- clude that the 1971 CCDO codes are but a slight improvement, and can only be used after a con- siderable reordering of occupational titles and their four-digit codes into a new classification. We have undertaken this reordering (it appears as an appendix to this article), and have tested the suitability of our new classification as a socioeconomic classification. While sophisti- cated scales of socioeconomic status, such as those developed by Blishen from the 1951. 1961, and 1971 censuses (1958; 1967; Blishen and McRoberts, 1976), play an extremely im- portant role, sociologists will, for a variety of reasons, continue to use classifications of occu- pations. We feel that the one we have con- structed can be used not only for census or other Statistics Canada materials, but also for organizing independently collected informa- tion. THE STRUCTURE OF THE NEW CENSUS CODE As in earlier years, the structure of the census occupational classification is described in a single compact volume (Canada, Statistics Can- ada, 1971). This publication, hereafter called the Coding Manual, is an indispensable tool for any quantitative sociologist who uses occu- pation as a variable in his work. In developing the 1971 system, however, Statistics Canada worked closely with the Department of Man- power and Immigration. For this reason, in ad- dition to the Coding Manual, study of two im- portant back-up documents which form the Canadian Classification and Dictionary of Oc- cupations, hereafter called the CCDO, is re- quired to fully understand the Census classifica- tion (Canada, Department of Manpower and Immigration, 1971).’ The CCDO volumes give detailed definitions and job descriptions and re- late these to the Census codes. The level of detail in the CCDO, which with seven digits of code allows for possible distinctions between * This classification was prepared as part of the Carleton-McMasterNational Occupational Mobility Study (CanadaCouncilNo. 2101-10). Other membersofthegroup are Monica Boyd, John Goyder, and Frank E. Jones. Frank Levin of Statistics Canada provided helpful comments. I In the us the two main coding systems are that of the Bureau of the Census and that of the Department of Labour (The Dictionary of Occupational Titles). What has happened is that the two Canadian equivalents are amalgamated into one code. Rev. canad. SOC. &Anth./Canad.Rev. SOC. &Anth. 14(1)1977

THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

Embed Size (px)

Citation preview

Page 1: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

PETER C. PINE0 IMcMaster University JOHN PORTER I Carleton University HUGH A . MC ROBERTS I Carleton University

Apartirdurecensementde 1971, laclassificationdes professionsa subideschangements importants ; donc, les sociologues devraient developper des methodes nouvelles pour entreprendre I’analyse des resultats. Les auteurs discutent et evaluent la nouvelle classification et la (< Classification canadienne descriptive des professions )), et proposent une reorganisation des groupes de base a fin de formuler une classification socio-economique des professions plus utile que les grands groupes et sous- groupes du recensement.

Radical changes in the occupation coding system for the 1971 Census of Canada mean that new strategies must be developed by sociologists to analyse census results. The structure of the new code, and of the Canadian Classification and Dictionary of Occupations, are discussed and evaluated. An alternative arrangement of the Unit Groups is proposed and demonstrated to be a more valuable socioeconomic classification of occupations than the Major and Minor Groups of the census.

Researchers who have looked at the occupa- tional data of the 1971 Census of Canada will have noted the radical changes which have been made in the occupational codes and classifica- tions. Since the earlier occupational data were often criticized by sociologists who tried to use them (Pineo and Porter, 1967), it is important to determine whether or not the 1971 Census codes based on the Canadian Classification and Dictionary of Occupations (the CCDO codes as they are known) are any improvement. In the present paper we attempt to make such an as- sessment after an examination of how the codes were created.

Our particular interest is the use of occupa- tional data for stratification research. We con- clude that the 1971 CCDO codes are but a slight improvement, and can only be used after a con- siderable reordering of occupational titles and their four-digit codes into a new classification. We have undertaken this reordering (it appears as an appendix to this article), and have tested the suitability of our new classification as a socioeconomic classification. While sophisti- cated scales of socioeconomic status, such as those developed by Blishen from the 1951. 1961, and 1971 censuses (1958; 1967; Blishen and McRoberts, 1976), play an extremely im- portant role, sociologists will, for a variety of reasons, continue to use classifications of occu-

pations. We feel that the one we have con- structed can be used not only for census or other Statistics Canada materials, but also for organizing independently collected informa- tion.

T H E S T R U C T U R E O F T H E N E W C E N S U S C O D E

As in earlier years, the structure of the census occupational classification is described in a single compact volume (Canada, Statistics Can- ada, 1971). This publication, hereafter called the Coding Manual, is an indispensable tool for any quantitative sociologist who uses occu- pation as a variable in his work. In developing the 1971 system, however, Statistics Canada worked closely with the Department of Man- power and Immigration. For this reason, in ad- dition to the Coding Manual, study of two im- portant back-up documents which form the Canadian Classification and Dictionary of Oc- cupations, hereafter called the CCDO, is re- quired to fully understand the Census classifica- tion (Canada, Department of Manpower and Immigration, 1971).’ The CCDO volumes give detailed definitions and job descriptions and re- late these to the Census codes. The level of detail in the CCDO, which with seven digits of code allows for possible distinctions between

* This classification was prepared as part of the Carleton-McMaster National Occupational Mobility Study (CanadaCouncilNo. 2101-10). Other membersofthegroup are Monica Boyd, John Goyder, and Frank E. Jones. Frank Levin of Statistics Canada provided helpful comments. I In the us the two main coding systems are that of the Bureau of the Census and that of the Department of Labour (The Dictionary of Occupational Titles). What has happened is that the two Canadian equivalents are amalgamated into one code.

Rev. canad. SOC. &Anth./Canad. Rev. SOC. &Anth. 14(1)1977

Page 2: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

92 / Peter C. Pineo, John Porter, and Hugh A. McRoberts

6670 different occupations, is in excess of that which would normally be used in most sociolog- ical research.

The Census code, based on the first four of the seven digits of the CCDO code, has four sections: an alphabetical list of occupations; Unit Groups into which the occupations are coded; Minor Groups; and Major Groups. These four sections, and the problems as- sociated with each, are described below, and illustrated by the occupation of ‘sociologist.’

I I Alphabeiical lisi This is a long list of occupational titles in which can be found for example, ‘Sociologist - Any hd-2313.’The ‘any ind’means ‘any industry.’ Sociologists are one of the few occupations that can be coded without knowing the industrial site of the work; for most occupations one must also know the industry and in research it should be asked whenever occupation is asked. The number, 2313, is the Unit Group Code into which sociologists are coded. Many rare and exotic occupations may be found in this list.

2 I The Unit Groups The Unit Group is the ‘fundamental particle’ of the code as far as all Statistics Canada tabula- tions are concerned and cannot be broken with- out reference to the CCDO. For each occupation listed, a four-digit number is given, which is the number indicating into which Unit Group the occupation belongs. Of course each listed oc- cupation cannot have its own unique code. The thousands of occupations in the alphabetical list are collapsed and pooled into some 486 Unit Groups. Sociologists are, as noted, put into Group 2313. The full title and a list of occupa- tions within the Unit Groups are given in a section of the Manual preceeding the alphabeti- cal list. One learns that sociologists are included with anthropologists as well as archaeologists, criminologists, ethnographers, penologists, and social ecologists. Thus in the Unit Group these seven somewhat different occupations are ir- reversibly pooled into a Unit Group called ‘Sociologists, Anthropologists and Related So- cial Scientists,’ to produce a category of sufficient size.

Demographers, on the other hand, appear in

the alphabetical list, but are coded with statisti- cians and actuaries (Unit Group 2181). Since forming the Unit Groups is an irreversible step this means that no tabulations of census data of sociologists and demographers, as such, is pos- sible.

A code cannot be better than the responses which go into it. Many sociologists in respond- ing to the census questions probably said ‘uni- versity professor’ and those who did would end up in an entirely different code - number 27 I I . Since there are only 430 cases coded as 2313 in the 1971 Census and since the membership of the CSAA was considerably larger at the time, a substantial number of sociologists and an- thropologists must have reported themselves as university professors.* One cannot, therefore, be at all sure that those coded as sociologists and anthropologists are representative of the whole professional group.

If the sociologist were a dean, and said so, he would go into code 1133 as some kind of man- ager. If he said he were a department chairman it is unclear what would happen; this occupa- tion does not appear in the alphabetical list. (It does appear, with a unique code, in the CCDO volumes, however.) And if he said he were a ‘research worker’ he would be banished to code 9919, ‘other occupations not elsewhere clas- sified,’ along with relief men, barrel washers, and oven dampers and strippers. Thus the Unit Groups irreversibly pool occupations into somewhat arbitrary, somewhat sensible large groupings. The arbitrary element is unavoida- ble; any coding scheme of the occupational world must pool to create some simplicity in what is a very complicated matter.

How good are the Unit Groups? Clearly if they are the irreversibly formed fundamental particle of the code the quality of everything which follows is dependent upon them. The principal criterion of classification used in form- ing them is said to be: ‘homogeneity in respect to work performed although, in some instances, other factors closely affecting the nature of the duties such as education, work environment, industrial process, or materials worked on are considered’ (Canada, Statistics Canada, 1974). To fully assess what this criterion means in practice it is necessary to consult the two vol-

z Simple status enhancement could be a reason for this. In the 1947 NORC study sociologists ranked twenty- seventh in occupational prestige while college professors ranked seventh (NORC, 1953:412-13). In 1965 in Canada ‘university professor’ ranked fourth out of 197, and generally considerably higher than the professions they teach, such as biologist, psychologist, etc. (Pine+Porter, 1967).

Page 3: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

The 1971 Census and the socioeconomic classification of occupations / 93

umes of the CCDO. There one finds, in many cases, that the occupations comprising the Unit Group are subdivided into three ‘clusters,’ A , B, and c, each representing different levels of General Educational Development (GED) and Specific Vocational Preparation (SVP). For example, Unit Group 8137 (Moulding, Core- making and Metal Casting Occupations) con- tains 6 occupations in Cluster A; 13 in Cluster B; and 31 in Cluster c.

For all the 50 occupations in Unit Group 8137 there is a very wide range of information, some of it scaled on many of those characteristics of occupations which theoretically are considered to be, and occasionally empirically demon- strated to be, related to status, such as edu- cation, difficulty or skill required, number of people controlled, freedom and independence, and so forth. Unfortunately all of this informa- tion is ‘lost’ beyond the four-digit level. It is clear that a large number of the Unit Groups are internally heterogeneous with respect to these ranking characteristics. It would be desirable to reaggregate by clusters so that cluster c’s for all occupations, for example, could be classified together and assigned a new code, but this is impossible.

Despite their status heterogeneity, the 1971 Unit Groups are probably no worse than those of earlier census codes. We have no way of revealing the latter’s weaknesses as we have here for the former. We show below that the socioeconomic classification of the Unit Groups, which resulted from a somewhat long and tedious consideration of ambiguous and un- familiar Unit Group titles in the CCDO, is a de- fensible one considering all the constraints.

One quick test suggests that however ques- tionable the Unit Groups may look, they are probably as good as those used previously. The Pineo-Porter list of occupations (1967) can be used as a baseline for comparison. Blishen found that some 88 Pineo-Porter titles had matches in the 1961 list of Unit Groups (196742). We can find almost 1 0 0 matches be- tween the Pine+Porter titles and the 1971 list of Unit Groups. Thus while very different from each other, the Unit Groups of both censuses are built of the kind of occupational titles which sociologists have typically found useful. The large number of Unit Groups somewhat guar- antees they will have value for sociological analysis. It would be almost impossible to create so many categories and not have sufficient raw material to produce a measure of

socioeconomic status. As well, in forming any code of occupations, regardless of their inten- tions, those designing it could scarcely have ignored social status, perhaps unconsciously, but also, too, because of using education as a criterion since it contributes so much to social status.

3 I The Minor Groups The Unit Groups are pooled, at the next level of abstraction, into Minor Groups. There are about six Unit Groups in each Minor Group - some 82 Minor Groups in all. Each Minor Group code is created by dropping the final digit of the four-digit Unit Group code. As the final digit of the code for sociologists, an- thropologists, and related social scientists is dropped, so that 2313 becomes 231, these occu- pations become pooled with all others having codes beginning with 231. Thus sociologists and anthropologists join with the political scien- tists and their related occupations (23 IS), the economists (231 I ) , and the psychologists (2315) to become a Minor Group called ‘occupations in the social sciences.’ Note that demographers are still not included. They head off in a differ- ent direction, ending in a Minor Group called ‘mathematics, statistics, systems analysis and related fields.’ So also, any sociologists coded as university professors, deans, or research workers have all headed off into different Minor Groups.

Little use can be made of the Minor Groups. Because there are 82 of them, they are too numerous to be used for analysis of contingency tables, and too heterogeneous to be of any other value.

4 I The Major Groups The final level of abstraction, or aggregation, is the forming of the Major Groups. There are 22

of these. The Major Groups are formed by dropping another digit from the Unit Group code. At this level sociologists and others in that Minor Group are pooled with several other Minor Groups, including such occupations as social workers (2331), lawyers (2343), librar- ians (2351). library technicians (2353), and vocational counsellors (2391). The whole then is Major Group 23. called ‘occupations in social science and related fields.’ Note that sociologists are not reunited with demog- raphers, who have joined a Major Group called ‘occupations in natural sciences, engineering and mathematics.’ Deans are in Major Group I I

Page 4: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

94 I Peter C. Pineo, John Porter, and Hugh A. McRoberts

TABLE I

OCCUPATIONAL PRESTIGE BY MAJOR GROUPS OF THE I971 CENSUS OCCUPATIONAL CODE*

Standard Name of Major Group No. of titles Mean score deviation

1 1 Managerial, administrative 19 66.9 13.9 21 Natural sciences, engineering,

mathematics 13 61.4 8.0 23 Social sciences and related fields 9 63.8 13.4 25 27 31 33 41 51 61 71

73

75 77

Religious occupations Teaching and related occupations Medicine and health Artistic, literary, recreational Clerical Sales Service Farming, horticultural, animal husbandry Fishing, hunting, trapping, and related occupations Forestry and logging occupations Mining, quarring, including oil and gas field occupations

2 4 8

15 18 16 20

6

2 3

4

70.3 63.0 66.4 54.2 38.7 40.5 34.8

35.0

23.5 27.9

33.5

3.5 17.8 14.5 8.1 7.4

12.5 15.3

10.0

0.1 11.2

8.3 81 /82 Processing 8 31.3 5 . 2 83 Machining and related occupations 6 37.9 8.9 85 Fabricating, assembling, repairing 9 40.9 , 8.4 87 Construction 11 36.3 9.4 91 Transport equipment operating 10 41.1 13.6 93 Materials handling 5 26.8 9.7 95 Other crafts and equipment operating 5 38.4 4.4

* Based on the 193 Pineo-Porter titles that were matched with Census Unit Groups.

‘managerial, administrative and related occupa- tions.’ University professors are in Group 27, ‘teaching and related occupations,’ where they are pooled with all manner of other teachers including, if one looks closely, driving instruc- tors. The moulding and coremaking groups are in Major Group 81, ‘processing occupations.’

Whether the Major Groups are a reasonable classification of occupations by socioeconomic level is important. All census tabulations give subtotals at the Major Group level. If they were reasonable, one would not have to go to the time-consuming labour of reorganizing and re-totalling all the Unit Group frequencies. Moreover, many census tabulations offer statis- tics only at the Major Group level. In so far as this is to be true of 1971 tabulations as well it is crucial to determine if any sociological sense can be made out of the Major Groups.

Value of the Major Groups as socioeconomic categories In Table I we have presented the mean prestige score for those occupations included in the Pineo-Porter prestige study which fall into each of the Census Major Groups. For example,

some 19 of the Pineo-Porter occupational titles fall into Major Group I I , Managerial and Ad- ministrative, and these 19 have prestige scores which average to 66.9 with a standard deviation

It was noted earlier that only some 1 0 0 of the Pineo-Porter titles had a close match with Cen- sus Unit Groups. Many of the rest can neverthe- less be quite easily put into a Major Group and with some pushing and shoving 193 of the 196 titles were placed in one Group or another. The fault here is not in the Census. Many of the Pineo-Porter titles are quite vague and many lack any reference to industrial site. The kind of stimuli which can be presented to a random sample for rating are not the same as are needed for accurate coding.

Table I parallels a table presented in the Pineo-Porter article as a test of the adequacy of the 1961 Census codes. There it was noted that ‘If we restrict the analysis to categories with at least five occupations in them and arbitrarily define homogeneity by a standard deviation of less than 10, only three of the nine possible census categories are homogeneous . . .’ (Pineo and Porter, 1967:35). Employing the same

Of 13.9.

Page 5: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

The 1971 Census and the socioeconomic classification of occupations / 95

TABLE I1

AGGREGATION OF CENSUS MAJOR GROUPS INTO CATEGORIES

Percentage of Number of workers labour force

Mean prestige Standard Category Male Female Male Female score* deviation

Upper white collar (Major Group 33 and less) 880,985 584,360 17.3 22.3 63.7 12.6

Clerical-sales-service (Major Groups41,51,61) 1,523,295 1,635,920 29.9 62.4 37.8 12.3

Crafts-trades and manual (including primary) (Major Group 71 or higher) 2,697,530 399,925 52.9 15.3 35.6 10.0

Total 5,101,955 2,620,205 100.0 100.0 ~~

NOTE: Some 904,900 workers are omittedOf these, 737,275 failed to give a codeable response; the balance a; in the highly heterogeneous Major Group 99, ‘Occupations Not Elsewhere Classified.’ SOURCE: Canada, Statistics Canada, ‘Advance Bulletin of the 1971 census of Canada,’ March 1974. * The mean prestige score is based on the 193 Pineo-Porter titles that were matched with Census Unit Groups.

criteria, it would appear that of 16 possible categories with at least 5 titles in the 1971 code, some 9 could be called homogeneous in socioeconomic standing, and 7 are not. There is some improvement over 1961, but it is still not a good code. And the 1971 Groups have de- veloped a difficulty not shared by those in 1961. Of the 16 Groups with at least five occupations in them, 7 have average prestige scores in the thirties. That is, there is a serious lack of dif- ferentiation between the 1971 Groups as well as high heterogeneity. Whenever possible it would appear desirable to ignore the Major Groups and reorganize the Unit Groups into some al- ternate poolings which more adequately mea- sure socioeconomic status.

The problem remains of what to do when reorganization is not possible. In Table I I we suggest one strategy. The top seven Major Groups, those numbered from I I to 33 and con- sisting mainly of professionals and managers, can be pooled to produce a category which could be called ‘upper white collar.’ While the standard deviation in prestige remains high, at 12.6, this group does have a mean prestige score which is appreciably higher than the balance: 63.7 as compared to around 36. About 20 per cent of the labour force falls into this category. Where tabulations are available at the Major Group level only, it would be possible to study separately the top one-fifth of the so- cioeconomic hierarchy. Unfortunately, no

equivalent differentiation further down the scale seems possible. The clerical-sales-service category can be distinguished from the crafts- trade (including farmers and primary workers). Table I I shows that both categories have fairly high standard deviations in prestige, and also the mean prestige scores are almost identical. Again, the problem is lack of differentiation. It would seem likely that even less can be learned from data organized in the 1971 Major Groups than in their equivalent, the ‘Divisions,’ of ear- lier c e n s ~ s e s . ~

Detailed inspection of the means and stan- dard deviations in Table I suggests which of the Major Groups create difficulties. The manage- rial group is heterogeneous but this seems in- evitable. No code seems able to handle this group. While the new group of natural scientists is quite homogeneous, the social scientists are a mixture -judges and playground directors are included. Teaching is a highly varied group. So also, the inclusion of some low status occupa- tions in the medicine and health group make it heterogeneous. Again there is evidence of the mixture of motives in the Census code, taking as it does socioeconomic status into account on occasion, if only by accident, and then at other times emphasizing common work site.

While the clerical group is quite homogene- ous, the service and sales groups cover a vast range by socioeconomic status. Farming is made somewhat heterogeneous by the inclusion

3 See The Verfical Mosaic (Porter, 1965) for an example of use of the 1951 and 1961 Divisions for stratification analysis. It is doubtful if as much could be achieved with the 1971 Major Groups.

Page 6: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

96 I Peter C. Pineo, John Porter, and Hugh A. McRoberts

of both farm owners and farm labourers in the same group. While forestry is a mixed group, mining, processing, machining, fabricating, and construction seem quite homogeneous. It is here that the problem of differentiation is clearest; while the last five Groups are quite homogeneous they also all have nearly identical mean prestige scores - ranging from 31.3 to 40.9.

Transport equipment operating, which in- cludes taxi drivers and airline pilots, is a varied category. So also the higher status of the power crane operator makes the materials handling category vary, although most jobs within this group are of very low status indeed, including garbage collector. A somewhat residual cate- gory of ‘other crafts and equipment operating’ including the printing trades, stationary engine operators, electronic equipment operators, and photographic processors, turns out to be un- usually homogeneous. The final Major Group, ‘occupations not elsewhere classified’ was omitted as obviously a purely residual category.

An alternative coding system Despite their deficiencies discussed earlier, the Unit Groups are the best building blocks for an alternative set of aggregating ~a tegor i e s .~ The code we propose is similar to that developed by Pineo-porter for the 1961 Census (1967:25); but there are also some adaptations and changes. The categories are given in Table 1 1 1 . One prin- ciple used in creating the categories was to group together, where possible, occupations which are recurrently said by sociologists to have special affinities even if they differ some- what in social standing. The category of ‘fore- men’ is the clearest example. Also, ‘profession- als’ can be meaningfully grouped. They may share some moderate consciousness of kind. They certainly share a common objective con- dition - that of prolonged formal training.

A second principle influencing the forming of the categories was the idea that the old clerical- sales-service category should now be broken up, and the occupations in it be allowed to dis- tribute through skill levels just as manual occu- pations do. Thus, rather new categories are

created such as: skilled clerical-sales-service, semiskilled clerical-sales-service, and unskilled clerical-sales-service. These groups could be reassembled and placed above foremen to re- trieve the older arrangement of lower white collar.

Other alterations from 1961 are minor. In forming the group of high-level managers only those holding top jobs in large firms were in- cluded. Heads of small concerns, when possi- ble, and minor functionairies in large concerns are put in with middle management, officials, and small business.

The category of ‘supervisors’ was suggested by the Census rather than designed in advance. Supervisors are white collar foremen - that is, they supervise clerical, sales, and service oper- ations. They could have been combined with foremen, but again to make the reassembling of the old lower white collar group possible they are distinguished. Farm owners and operators are distinguished from farm labourers and are placed around the level of skilled crafts and trades.

The category of self-employed professionals is very small as most professionals are now employees. It is, however, a distinctive group and making isolation of it possible seemed wise. As well, certain anachronisms must be permit- ted. An overly contemporaneous code could not be used to code the father’s occupation in an intergenerational study of occupational mobil- ity, for example.

With 16 categories, the code may be too de- tailed for some purposes. Collapsing in various ways is possible. The top three categories can be combined to form a managerial and profes- sional group. As noted earlier, skilled, semi- skilled, and unskilled clerical, sales and service can be recombined, with or without super- visors, to form the old lower white collar group. Or each could be combined with the manual labouring category at the same skill level.

Problems in assigning the Unit Groups to categories The job of assigning each Unit Group to one of the categories in the code was formidable. The

4 One attempt to organize the Census Unit Groups into a more sociologically meaningful code is already provided in the literature by Gower (1974). This code uses aggregating categories named: Managerial, Profes- sional, Technical, Crafts and Trades, Semiskilled, Low Skill. As a socioeconomic scale these sets seem promising, but the code, so far as one can tell from the write-up, was poorly executed. Lavish use has been made of the category ‘semiskilled’ as aresidual category. Some 58 percent of the labour force falls into the semiskilled category.

Page 7: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

The 1971 Census and the socioeconomic classification of occupations / 97

results are given in the Appendix. There each category is listed and for each the numbers of all Unit Groups which form it are given. Using the Appendix one can reorganize the data from census tabulations given at the Unit Group level into the categories of Table 1 1 1 . ~ Some of the difficulties in making the assignments derive from the census code itself. A large number of Unit Groups have the phrase ‘not elsewhere classified’ (or n.e.c.) in the title. The detailed description of each of these n.e.c. groups must be carefully inspected to get some idea of what sort of occupations are in the group and how it should be assigned.

Another frequent Unit Group title is ‘Inspect- ing, Testing, Grading and Sampling Occupa- tions, (such and such an) Industry.’ Estimates were made of the probable skill level required in each case that this code appears and they were distributed appropriately through the various skill levels of the manual labouring categories. Finally, it must be remembered that the formal title given to each Unit Group is an abbrevia- tion. In every case some search of the actual detailed contents of the group was necessary.

A team of four made the initial assignments, using whatever specialized knowledge of the work world each member possessed. Other sociologists were consulted on occasion. The most difficult occupations to classify were the largely ‘invisible’ ones that are located in fac- tories. A two-hour interview with the owner of a small machining plant proved extremely useful. To do the job perfectly a variety of such inter- views would be desirable and we do not pretend that the assignments we have made are perfect.

A word should be said about the implicit criterion underlying the classification. One could call the criterion an informed assessment of the social standing (or occupational prestige) of the occupation. Ideally we would have wished to have each occupation classified by an experienced sociologist who not only was famil- iar with the nature of the job but also aware of the typical attitudes of others in the society towards incumbents of the occupation. Such judgments should be strongly influenced by the educational level and income of the occupa- tion but other characteristics of the occupation would be taken into account when appropriate.

While census data on education and income

were not consulted in making the assignments, the number of workers was considered and some Unit Groups with very few,workers in them were assigned quickly with little careful thought (e.g., Mineral Ore Treating Occupa- tions, n.e.c.). I t may be these categories which are most frequently found to be misplaced. The consequences of such misplacements, how- ever, cannot be great. An attempt was also made in making the assignments, to avoid being influenced by the extent to which the job was typically seen as ‘women’s work’ or ‘men’s work.’

In assigning the occupations two problems recurred which have substantive significance. They involved the distinction, previously refer- red to, between unskilled and semiskilled work, on the one hand, and the distinction between clerical-sales-service and manual work on the other.

The old rule of thumb, that semiskilled work takes some six months on-the-job training and unskilled work virtually none, seemed less use- ful at this time than in earlier years, and seems to have been abandoned by the creators of the CCDO in favour of the refined General Educa- tional Development scale and the Special Voca- tional Preparation scale. It is clear from our example of moulders and core makers and its respective clusters that many of the Unit Groups contain a wide range of ‘skill,’ and the judgmental task was one of where it should be placed with respect to the skilled, semiskilled, and unskilled categories for both clerical- sales-service and manual work.

The decision not to put all the clerical, sales, and service occupations into a single category forced a reconsideration of the meaning of this category. What should lead to an occupation being classified as clerical-sales-service rather than crafts and trades? Or, put another way, what was the common element which led to the formation of the old ‘lower white collar’ group- ing? One axis of differentiation was between office work and factory or outside work. The idea of dealing with people rather than with machines and things was another component. The use of brains rather than brawn was a fac- tor, as was discomfort such as heat, or health hazards such as exposure to asbestos fibre. As well, there was the question of the use on the job

5 Itshouldbenoted thatUnitGroups 1151 to 1158andggz1 to9926,notincludedintheAppendix.areindustrial subdivisions of Unit Groups I 149 (Managers and Administrators not elsewhere classified) and 9918 (Labourers and other elemental work not elsewhere classified) respectively.

Page 8: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

9 / Peter C. Pineo, John Porter, and Hugh A. McRoberts

TABLE 111

MEAN PRESTIGE SCORES OF SOCIOECONOMIC CATEGORIES*

No. of marched Mean prestige Standard Socio-economic category titles score deviation

01 Self-employed professionals 4 78.6 8 .7 02 Employed professionals 23 68.0 8 . 2 03 High-level management 6 67.7 5 . 6 04 Semi-professionals 23 56.7 8 .6 05 Technicians 2 67.2 0 . 4 06 Middle management 14 64.8 16.7 07 Supervisors 8 46.3 8 . 4 08 Foremen 2 51 .O 0.1 09 Skilled clerical-sales-service 10 47.7 5.6 10 Skilled crafts and trades 26 4 0 . 3 7 . 0 12 Semiskilled clerical-sales-service 20 34.2 8.9 13 Semiskilled manual 25 32.4 7 . 0 14 Unskilled clerical-sales-service 6 29.7 8.9 15 Unskilled manual 18 24.7 6 . 0 16 Farm labourers 2 23.3 2 .6 1 1 Farmers 4 40.9 5.3

* The means were calculated for the 193 Pineo-Porter titles that were matched with Census Unit Groups.

of skills acquired from formal schooling. Along with some skill in handling the public, the clerical-sales-service division typically re- quired of workers some literacy and skill with numbers. When such skills were rare because only a minority proceeded far in school, those having them may clearly have been higher in social standing than those who did not. But as such skills have become generally distributed and as factory and outside work frequently have required them as well, the historical, invidious distinction may well be questioned. For exam- ple, the moderate skills in arithmetic required of a cashier can hardly be a source of significantly higher social standing when a maintenance man today is required to complete complicated, numeric job reports. Obviously much more em- pirical work needs to be done to determine how completely the gap is closed or the manual- non-manual dividing line has become blurred, but there still seems to be grounds for separat- ing the two groups, and dividing them by skill levels as we have.

At this point, after all occupations were clas- sified on the basis of the judgment of the team of sociologists, a thorough review of virtually all placements was made using the considerable resources of the two volumes of the CCDO. Studying the CCDO revealed once again how very heterogeneous certain Unit Groups were and how arbitrary their classification must be. But beyond this, the use of the CCDO was re-

markably reassuring. Of the 486 Unit Groups which had been classified, less than 10 per cent were changed to another category after study of the CCDO.

Testing the occupation code Despite the many problems involved in decid- ing upon categories and assigning Unit Groups it can be shown that our occupation code is superior as a measure of social standing to the Census Major Groups. Again the Pineo-Porter occupational prestige scores can be used to make the test. In Table 111 the 196 Pineo-Porter titles are distributed through the categories of the occupation codes and for each category an average prestige score and the standard devia- tion are given.

The categories of the new classification are much more homogeneous indicators of social standing than are the Census Major Groups. Concentrating on categories with at least five matched occupations in them, 10 out of I I of the occupational categories are homogeneous - that is, have standard deviations less than 10 - while only 9 out of 16 of the census codes were. It was noted that 7 of the 16 Census Groups had average prestige scores in the thirties; only 2 of the new categories do. As well, it is possible to organize the socioeconomic categories into a nearly perfect gradient in average prestige scores from self-employed professionals down to farm labourers. No equivalent arrangement

Page 9: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

The 1971 Census and the socioeconomic classification of occupations / 99

TABLE IV

DISTRIBUTION OF MALES AND FEMALES I N 1971 LABOUR FORCE BY SOCIOECONOMIC CATEGORY ,

Males Females -

Socioeconomic cafegory number perrentage number percentage Total ~~~

Self-employed professionals 52910 1 . O 4205 0 . 2 57115 0 . 7 Employed professionals 374160 7 .1 223495 8 . 5 597655 7 . 6 High-level management 102805 2 . 0 13350 0.5 116155 1 . 5

324765 4 . 1 Semi-professionals 149040 2 .8 175725 6 . 7 Technicians 92795 1 . 8 37180 1 . 4 129975 1 . 6 Middle management 168705 3 . 2 47115 1 . 8 215820 2 . 7 Supervisors 350720 6 . 7 110065 4 . 2 460785 5 . 8 Foremen 276645 5 .3 7995 0 . 3 284640 3 . 6 Skilled clerical-sales-service 209050 4 . 0 447925 17.0 656975 8 . 3 Skilled crafts and trades 894975 17.1 44440 1 .7 939415 11.9 Semiskilled clerical-sales

service 416535 7 .9 75 1550 28 .4 1168085 14.8 Semiskilled manual 665620 12.7 194860 7 . 4 860480 10.9 Unskilled clerical-sales-

service 152145 2 . 9 248965 9 . 4 401110 5 . 1 Unskilled manual 959405 18.3 229035 8 . 7 1188440 15 .1 Farm labourers 156015 3 . 0 98100 3 . 7 254115 3 . 2 Farmers 226335 4 . 3 7680 0 . 3 234015 3 . 0

Totals 5,247,860 100.0 2,641,685 100.0 7,889,545 100.0

SOURCE: Statistics Canada, Advance Bulletin of the 1971 Census of Canada, March 1974, Catalogue 94-788.

of the Census Major Groups seems possible. The category of middle management, offi-

cials, and small business remains heterogene- ous, with a standard deviation of 16.7. The more restrictive criteria for assigning an occu- pation to high-level management have pro- duced homogeneity, but now, in comparison to the 1967 grouping, the variability is found in middle-range management.

Self-employed professionals again head the list, although the number in the category is small. The small category of high-level man- agement is found to be almost identical in rank to employed professionals.

The supervisory category is homogeneous, but has a mean prestige slightly lower than might have been expected. There were only two foreman titles in the Pine&Porter list so that little can be said about that category. Farm owners and operators form a homogeneous category when farm labourers are made into a separate one. The differences in the social standing of these two groups is quite marked.

Standard deviations under 10 in each case suggest that the attempt to split the clerical- sales-service group into three skill levels was successful. The range in social standing is also quite great, from 29.7 for the unskilled clerical- sales-service up to 47.7 for the skilled. It will be noted that in all cases the clerical-sales-service

category has higher social standing than the equivalent skill group from among the manual labourers.

From the Census one can compute the actual number of male and female workers who fall into each category (Canada, Statistics Canada, 1974). The results are presented in Table IV along with the proportions which each category contains of the male and female labour forces and the total labour force. There are few self- employed professionals and men outnumber women in all the high level categories. Super- visors are quite numerous, and most are men, and among foremen, there are very few women.

In all three categories representing the clerical-sales-service occupations, women out- number men. Despite attempts to avoid taking gender into consideration in deciding where to code the jobs it is clear that the clerical-sales- service category, at any skill level, is predom- inantly ‘women’s work.’ However, ‘super- visors’ who work in clerical-sales-service are predominantly men, a fact which gives rise to interesting questions.

In contrast, the crafts and trades are pre- dominantly male, as are the lower level man- ual workers. The extent of sexual differenti- ation is not equivalent in the two areas. Of some 2,226,000 workers in clerical, sales, and ser- vice, only 770,000, or around 35 per cent, are

Page 10: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

IOO / Peter C. Pmeo, John Porter, and Hugh A. McRoberts

male. In contrast, among some 3,000,000 man- ual workers only 68,000 or about 16 per cent are female. Finally, few women are classified as farm owners and operators. But women are quite numerous in the category of farm labour- ers. Perhaps wives of farm owners have been classified as farm labourers.

The classification would seem to cover the male labour force well. Only two codes - the self-employed professionals and technicians - have fewer than 100,ooo workers in them. And even the largest code, that of unskilled manual workers, has only 18 per cent of the male labour force in it. The coverage for women is less adequate. As well as the self-employed profes- sional group there are seven other categories with fewer than 100,ooo workers. Collapsing and regrouping are required if the code is to serve well as a measure of the occupational distribution of the female labour force. At this time women in the labour force are so concen- trated in specific occupations that no socio- economic coding system could escape near blank categories for women.

Non-census uses of the classijication As well as providing a mode of reorganizing census data into categories which are of more use to sociologists than the Major or Minor Groups, the classification herein proposed could be used for non-census data. Data col- lected in an ordinary questionnaire or interview survey could be coded into these categories. Not only would a respectable measure of the standing of the occupations be produced, but it would also be possible to compare the resulting distribution with the census to test for possible bias in the sample.

The coding could be wholly mechanical. With the use of the Census Coding Manual the vast majority of responses to an occupation question could be coded into Unit Groups and, through the use of the appendix to this paper, into the occupational categories. Relatively un- skilled coders could do this job with little super- vision. With a more skilled coding crew one might wish to use more judgment. A purely mechanical use of the system means that any deficiency in the Census Unit Groups affects the results. A coding crew could be taught to follow the mechanical procedure when it ap- pears sensible, but to consult the CCDO and substitute their own judgment when the Unit Group appears so heterogeneous as to interfere with the proper coding of the occupation. Or,

thirdly, one might wish to ignore the Census Coding Manual or CCDO entirely and fit the oc- cupational responses to the code categories using only coder judgment. Although the Cod- ing Manual and CCDO were used extensively in the assignment of the Unit Groups to the codes there remained an element of judgment. Thus, with skilled coders the use of coder judgment would be reasonable. Making use of the Coding Manual when it proves to have value does seem sensible and expeditious, however.

Comparisons with the Blishen scores The code herein proposed is not meant to com- pete with, but rather to complement the Blishen scale. There are at least two instances in which the use of the Blishen scale is obviously to be preferred. First, in writing for an audience which would find fault with the procedures in the formation of our occupational code the Blishen scale might be preferable. Secondly, in using statistics which require interval-level data, many might feel more at ease using the Blishen scale than the occupational classifica- tion.

In contrast, for some audiences the abstract character of the Blishen scale may restrict com- prehension, and the presentation of results in terms of popularly understood occupational categories may be preferable. As well there are occasions when specific categories found in the occupation code - such as foremen or profes- sionals - are specified in the theory or hypothesis and in this case the code would be of immediate use. Finally, we have noted that in using the occupational code it is possible to train coders so that they ovemde the purely mechanical use of the Census Code in classify- ing occupations. It is not possible to override the Coding Manual in using the Blishen scale.

A P P E N D I X . C E N S U S U N I T G R O U P S F O R M I N G E A C H S O C I O E C O N O M I C C A T E G O R Y

Self employed professionals

Employed professionals 1171, 2111, 2112,2113, 2114, 2131, 2133, 2142,

2 1 4 I , 2 3 4 3 ~ 3 I I 1 , 3 1 1 3 , 3 1 1 5

21439 214% 2145, 2147, 2151, 2153, 2154, 2155, 21579 2159, 2160, 2161, 2181, 2311, 2313, 2315, 23317 23419 2350, 2351, 2391, 2511, 2711, 2731, 27339 2793,279593151,3152,6116 High-level management 1 1 1 3 9 1130, 1131, 1132, 1133, 1134, 1135. 1137, 1 I47

Page 11: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

The 1971 Census and the socioeconomic classification of occupations / IOI

Sem i-professionals 21639 21839 2189, 23197 2333, 2339, 2349, 2399. 2513, 2519, 27199 2739, 2791, 27929 27973 27999 31177 31199 3I309313I7 3137,31539 315473311> 33139 33147 33157 3319, 333073332,33339 33353 33529 33559 33597 3373*9I 11,9151 Technicians 2117, 2119, 2135, 2139, 2165, 2169, 2353, 3155, 3156,3157,8116,8146,8176,8336,8396,8526, 85359 873698796791 1 3 9 9153 Middle management 1111, 1116, 1119, 1136, 1141, 1142, 1143, 1145, 1149, 1174, 1175, 11769 1179, 3370, 337195191, 6131,6141,7131 Supervisors 1115, 4110, 4130. 4140, 4150.4160, 4170, 4190, ~1~o,~1~o,5190,6120,6130,6160,6190,~1o Foremen

8210,8230,8250,8260,8290,8310,8330, 8350, 8370, 8390,8510,8530,8540,8550,8570, 8580, 8590,8710, 8730,8780,91 10, 9130, 9170. 9190.

Skilled clerical-sales-service 23593 31337 31349 33377 33397 41119 41313 41439 4I92351317 51333 51713 5172,51739 517495177, 95539 9557 Skilled crafts and trades 6111,6112,6113,~~11.7516,7~17.8131, 8133, 8135. 8137,8141, 8143, 8151, 8155. 8161,8163, 8165,8167,8231,8311,8313,8316,8333,8335, 83377835~,8376,8391,8395,8399,8525, 8529, 8537,8541,8553,8555,8581,8582,8583,8584, 8585,8586,8588,8591,8592,8731,8733,8735,

9531 9 95397 9551 7 9555.95597 9916 Farmers 71 12 Semiskilled clerical-sales-service

4141. 4151, 4153, 4155, 4157. 4159,4161, 4169,

7180, 7311, 7510, 7710,8110,8130,8150,8160,

93I0995I0,95307 95509 9590

8781, 8782, 8784, 8785, 8791,8795,9119, 9131, 91337 93117 95117 9512,95139 95147 951599517,

31353 31397 31599 4113,41337 41359 41379 41399

41719 4175. 41917 41939 41957 51353 51377 51793 5193,5199,6123,6125,6143,6144,6145 Semiskilled crafts and trades 6117,6121, 7313, 7513. 7711,7713, 7715, 7717, 7719, 8111,8113,8115, 8149.8153.8156,8171, 8173. 8211, 8213, 8215,8223,8225,8226,8227, 8233,8235,8236,8251,8253,8256,8261,8263, 8265,8267,8271, 8273,8275,8276,8293,8295, 8296,8315,8319,8331,8334,8339,8353,8355. 8356,8357,8371,8373~8379,8393,8511~8513, 8515.8523,8527,8531,8533,8534,8536,8546, 8549,8551,8557,8561,8562,8563, 8 5 6 8569, 8571,8573,8575,8576,8587,8589,8595~ 8596,

8711,8713.8739,8783,8786,8787.8793,8799, 91359 91579 9I71*9I79*9I917 9193.9315- 9519, 9591 7 9599 Unskilled clerical-sales-service 417274173941779 41799 41947 41979 41999 51417 5143,5149, 6129,6139,6147,6149,6199 Unskilled labourers 3375, 3379,5145.6115,6119,6133,6135,6162. 6165,6169,6191,6193,6198,7315,7319,7518, 7519,7718,8118,8119, 8148, 8158,8159,8178, 8179,8217,8221,8228,8229,8238,8239, 8258, 8259,8278,8279,8298,8299,8359,8528,8538, 8539,8548,8568,8578,8579,8593,8598,8599, 8715,8718, 8719,8738,8798,9139,9155, 9159, 91737 91759 919999313? 93179 93189 93193 9518, 9918,9919

7182,7I95.7I973 7199 Farm labourers

NOTE: All Unit Groups ending with the digit ‘9’ are residual categories (n.e.c.) and might well be reclassified as ‘not stated’ if the sacrifice of a relatively small proportion of the cases is ac- ceptable.

R E F E R E N C E S

Blishen, B.R. 1958 ‘The construction and use of an occupa-

tional class scale.’ Canadian Journal of Economics and Political Science 24:519:3I

1967 ‘A socio-economic index for occupations in Canada.’ Canadian Review of Sociol- ogy and Anthropology 4z41-53

Blishen, B.R., and H.A. McRoberts 1976 ‘A revised socio-economic index for oc-

cupations in Canada.’ Canadian Review of Sociology and Anthropology I3(1):719

Canada, Department of Manpower and Immi- gration 1971 Canadian Classification and Dictionary of

Occupations, Vol. I , Classifications and DeJinitions ; Vol. 2, Occupational Qual- iJcation Requirements

Canada, Statistics Canada. 197 I Occupation Classification Manual Census

of Canada 197 I Based on Canadian Clas- sification and Dictionary of Occupations, Volume 2. Ottawa: Information Canada

1974 1971 Census ofCanada, Volume 111, Part 2, Occupations: Appendix, List of Occu- pationcodes, Bulletin3.2-13. Ottawa: In- formation Canada

Page 12: THE 1971 CENSUS AND THE SOCIOECONOMIC CLASSIFICATION OF OCCUPATIONS*

IOZ / Peter C. Pineo, John Porter, and Hugh A. McRoberts

Gower, David 1974 ‘The occupational composition ofjob

vacancies. ’ Canadian Statistical Review 48(4):4,118-23

National Opinion Research Center I 953 ‘Jobs and occupation: A popular evalua-

tion.’inR. BendixandS.M. Lipset(eds.) Class, Status and Power. Glencoe: The Free Press: 41 1-26

Pineo, Peter C., and John Porter 1967 ‘Occupational prestige in Canada.’ Cana-

dian Review of Sociology and Anthropol- OgY 4( I : 24-40

Porter, John 1965 The Vertical Mosaic: An analysis of So-

cial Class and Power in Canada. Toronto: University of Toronto Press